There is a blog post that goes into more detail about the bigger picture, and walks through all the required pieces to make this work. It is available here: https://devblogs.microsoft.com/directx/directx-heart-linux . The rest of this cover letter will focus on the Linux Kernel bits.
Overview ========
This is the first draft of the Microsoft Virtual GPU (vGPU) driver. The driver exposes a paravirtualized GPU to user mode applications running in a virtual machine on a Windows host. This enables hardware acceleration in environment such as WSL (Windows Subsystem for Linux) where the Linux virtual machine is able to share the GPU with the Windows host.
The projection is accomplished by exposing the WDDM (Windows Display Driver Model) interface as a set of IOCTL. This allows APIs and user mode driver written against the WDDM GPU abstraction on Windows to be ported to run within a Linux environment. This enables the port of the D3D12 and DirectML APIs as well as their associated user mode driver to Linux. This also enables third party APIs, such as the popular NVIDIA Cuda compute API, to be hardware accelerated within a WSL environment.
Only the rendering/compute aspect of the GPU are projected to the virtual machine, no display functionality is exposed. Further, at this time there are no presentation integration. So although the D3D12 API can be use to render graphics offscreen, there is no path (yet) for pixel to flow from the Linux environment back onto the Windows host desktop. This GPU stack is effectively side-by-side with the native Linux graphics stack.
The driver creates the /dev/dxg device, which can be opened by user mode application and handles their ioctls. The IOCTL interface to the driver is defined in dxgkmthk.h (Dxgkrnl Graphics Port Driver ioctl definitions). The interface matches the D3DKMT interface on Windows. Ioctls are implemented in ioctl.c.
When a VM starts, hyper-v on the host adds virtual GPU devices to the VM via the hyper-v driver. The host offers several VM bus channels to the VM: the global channel and one channel per virtual GPU, assigned to the VM.
The driver registers with the hyper-v driver (hv_driver) for the arrival of VM bus channels. dxg_probe_device recognizes the vGPU channels and creates the corresponding objects (dxgadapter for vGPUs and dxgglobal for the global channel).
The driver uses the hyper-V VM bus interface to communicate with the host. dxgvmbus.c implements the communication interface.
The global channel has 8GB of IO space assigned by the host. This space is managed by the host and used to give the guest direct CPU access to some allocations. Video memory is allocated on the host except in the case of existing_sysmem allocations. The Windows host allocates memory for the GPU on behalf of the guest. The Linux guest can access that memory by mapping GPU virtual address to allocations and then referencing those GPU virtual address from within GPU command buffers submitted to the GPU. For allocations which require CPU access, the allocation is mapped by the host into a location in the 8GB of IO space reserved in the guest for that purpose. The Windows host uses the nested CPU page table to ensure that this guest IO space always map to the correct location for the allocation as it may migrate between dedicated GPU memory (e.g. VRAM, firmware reserved DDR) and shared system memory (regular DDR) over its lifetime. The Linux guest maps a user mode CPU virtual address to an allocation IO space range for direct access by user mode APIs and drivers.
Implementation of LX_DXLOCK2 ioctl ==================================
We would appreciate your feedback on the implementation of the LX_DXLOCK2 ioctl.
This ioctl is used to get a CPU address to an allocation, which is resident in video/system memory on the host. The way it works:
1. The driver sends the Lock message to the host
2. The host allocates space in the VM IO space and maps it to the allocation memory
3. The host returns the address in IO space for the mapped allocation
4. The driver (in dxg_map_iospace) allocates a user mode virtual address range using vm_mmap and maps it to the IO space using io_remap_ofn_range)
5. The VA is returned to the application
Internal objects ================
The following objects are created by the driver (defined in dxgkrnl.h):
- dxgadapter - represents a virtual GPU
- dxgprocess - tracks per process state (handle table of created objects, list of objects, etc.)
- dxgdevice - a container for other objects (contexts, paging queues, allocations, GPU synchronization objects)
- dxgcontext - represents thread of GPU execution for packet scheduling.
- dxghwqueue - represents thread of GPU execution of hardware scheduling
- dxgallocation - represents a GPU accessible allocation
- dxgsyncobject - represents a GPU synchronization object
- dxgresource - collection of dxgalloction objects
- dxgsharedresource, dxgsharedsyncobj - helper objects to share objects between different dxgdevice objects, which can belong to different processes
Object handles ==============
All GPU objects, created by the driver, are accessible by a handle (d3dkmt_handle). Each process has its own handle table, which is implemented in hmgr.c. For each API visible object, created by the driver, there is an object, created on the host. For example, the is a dxgprocess object on the host for each dxgprocess object in the VM, etc. The object handles have the same value in the host and the VM, which is done to avoid translation from the guest handles to the host handles.
Signaling CPU events by the host ================================
The WDDM interface provides a way to signal CPU event objects when execution of a context reached certain point. The way it is implemented:
- application sends an event_fd via ioctl to the driver
- eventfd_ctx_get is used to get a pointer to the file object (eventfd_ctx)
- the pointer to sent the host via a VM bus message
- when GPU execution reaches a certain point, the host sends a message to the VM with the event pointer
- signal_guest_event() handles the messages and eventually eventfd_signal() is called.
Sasha Levin (4): gpu: dxgkrnl: core code gpu: dxgkrnl: hook up dxgkrnl Drivers: hv: vmbus: hook up dxgkrnl gpu: dxgkrnl: create a MAINTAINERS entry
MAINTAINERS | 7 + drivers/gpu/Makefile | 2 +- drivers/gpu/dxgkrnl/Kconfig | 10 + drivers/gpu/dxgkrnl/Makefile | 12 + drivers/gpu/dxgkrnl/d3dkmthk.h | 1635 +++++++++ drivers/gpu/dxgkrnl/dxgadapter.c | 1399 ++++++++ drivers/gpu/dxgkrnl/dxgkrnl.h | 913 ++++++ drivers/gpu/dxgkrnl/dxgmodule.c | 692 ++++ drivers/gpu/dxgkrnl/dxgprocess.c | 355 ++ drivers/gpu/dxgkrnl/dxgvmbus.c | 2955 +++++++++++++++++ drivers/gpu/dxgkrnl/dxgvmbus.h | 859 +++++ drivers/gpu/dxgkrnl/hmgr.c | 593 ++++ drivers/gpu/dxgkrnl/hmgr.h | 107 + drivers/gpu/dxgkrnl/ioctl.c | 5269 ++++++++++++++++++++++++++++++ drivers/gpu/dxgkrnl/misc.c | 280 ++ drivers/gpu/dxgkrnl/misc.h | 288 ++ drivers/video/Kconfig | 2 + include/linux/hyperv.h | 16 + 18 files changed, 15393 insertions(+), 1 deletion(-) create mode 100644 drivers/gpu/dxgkrnl/Kconfig create mode 100644 drivers/gpu/dxgkrnl/Makefile create mode 100644 drivers/gpu/dxgkrnl/d3dkmthk.h create mode 100644 drivers/gpu/dxgkrnl/dxgadapter.c create mode 100644 drivers/gpu/dxgkrnl/dxgkrnl.h create mode 100644 drivers/gpu/dxgkrnl/dxgmodule.c create mode 100644 drivers/gpu/dxgkrnl/dxgprocess.c create mode 100644 drivers/gpu/dxgkrnl/dxgvmbus.c create mode 100644 drivers/gpu/dxgkrnl/dxgvmbus.h create mode 100644 drivers/gpu/dxgkrnl/hmgr.c create mode 100644 drivers/gpu/dxgkrnl/hmgr.h create mode 100644 drivers/gpu/dxgkrnl/ioctl.c create mode 100644 drivers/gpu/dxgkrnl/misc.c create mode 100644 drivers/gpu/dxgkrnl/misc.h
Add support for a Hyper-V based vGPU implementation that exposes the DirectX API to Linux userspace.
Signed-off-by: Sasha Levin sashal@kernel.org --- drivers/gpu/dxgkrnl/Kconfig | 10 + drivers/gpu/dxgkrnl/Makefile | 12 + drivers/gpu/dxgkrnl/d3dkmthk.h | 1635 +++++++++ drivers/gpu/dxgkrnl/dxgadapter.c | 1399 ++++++++ drivers/gpu/dxgkrnl/dxgkrnl.h | 913 ++++++ drivers/gpu/dxgkrnl/dxgmodule.c | 692 ++++ drivers/gpu/dxgkrnl/dxgprocess.c | 355 ++ drivers/gpu/dxgkrnl/dxgvmbus.c | 2955 +++++++++++++++++ drivers/gpu/dxgkrnl/dxgvmbus.h | 859 +++++ drivers/gpu/dxgkrnl/hmgr.c | 593 ++++ drivers/gpu/dxgkrnl/hmgr.h | 107 + drivers/gpu/dxgkrnl/ioctl.c | 5269 ++++++++++++++++++++++++++++++ drivers/gpu/dxgkrnl/misc.c | 280 ++ drivers/gpu/dxgkrnl/misc.h | 288 ++ 14 files changed, 15367 insertions(+) create mode 100644 drivers/gpu/dxgkrnl/Kconfig create mode 100644 drivers/gpu/dxgkrnl/Makefile create mode 100644 drivers/gpu/dxgkrnl/d3dkmthk.h create mode 100644 drivers/gpu/dxgkrnl/dxgadapter.c create mode 100644 drivers/gpu/dxgkrnl/dxgkrnl.h create mode 100644 drivers/gpu/dxgkrnl/dxgmodule.c create mode 100644 drivers/gpu/dxgkrnl/dxgprocess.c create mode 100644 drivers/gpu/dxgkrnl/dxgvmbus.c create mode 100644 drivers/gpu/dxgkrnl/dxgvmbus.h create mode 100644 drivers/gpu/dxgkrnl/hmgr.c create mode 100644 drivers/gpu/dxgkrnl/hmgr.h create mode 100644 drivers/gpu/dxgkrnl/ioctl.c create mode 100644 drivers/gpu/dxgkrnl/misc.c create mode 100644 drivers/gpu/dxgkrnl/misc.h
diff --git a/drivers/gpu/dxgkrnl/Kconfig b/drivers/gpu/dxgkrnl/Kconfig new file mode 100644 index 000000000000..8bfec781b9a7 --- /dev/null +++ b/drivers/gpu/dxgkrnl/Kconfig @@ -0,0 +1,10 @@ +# +# dxgkrnl configuration +# + +config DXGKRNL + tristate "Microsoft virtual GPU support" + depends on HYPERV + help + This driver supports Microsoft virtual GPU. + diff --git a/drivers/gpu/dxgkrnl/Makefile b/drivers/gpu/dxgkrnl/Makefile new file mode 100644 index 000000000000..11505a153d9d --- /dev/null +++ b/drivers/gpu/dxgkrnl/Makefile @@ -0,0 +1,12 @@ +# SPDX-License-Identifier: GPL-2.0 +# Makefile for the Linux video drivers. +# 5 Aug 1999, James Simmons, mailto:jsimmons@users.sf.net +# Rewritten to use lists instead of if-statements. + +# Each configuration option enables a list of files. + +# Uncomment to enable printing debug messages by default +#ccflags-y := -DDEBUG + +obj-$(CONFIG_DXGKRNL) += dxgkrnl.o +dxgkrnl-y := dxgmodule.o hmgr.o misc.o dxgadapter.o ioctl.o dxgvmbus.o dxgprocess.o diff --git a/drivers/gpu/dxgkrnl/d3dkmthk.h b/drivers/gpu/dxgkrnl/d3dkmthk.h new file mode 100644 index 000000000000..ed9a72986e57 --- /dev/null +++ b/drivers/gpu/dxgkrnl/d3dkmthk.h @@ -0,0 +1,1635 @@ +// SPDX-License-Identifier: GPL-2.0 + +/* + * Copyright (c) 2019, Microsoft Corporation. + * + * Author: + * Iouri Tarassov iourit@microsoft.com + * + * Dxgkrnl Graphics Port Driver user mode interface + * + */ + +#ifndef _D3DKMTHK_H +#define _D3DKMTHK_H + +#include "misc.h" + +#define D3DDDI_MAX_WRITTEN_PRIMARIES 16 +#define D3DDDI_MAX_MPO_PRESENT_DIRTY_RECTS 0xFFF + +#define D3DKMT_CREATEALLOCATION_MAX 1024 +#define D3DKMT_ADAPTERS_MAX 64 +#define D3DDDI_MAX_BROADCAST_CONTEXT 64 +#define D3DDDI_MAX_OBJECT_WAITED_ON 32 +#define D3DDDI_MAX_OBJECT_SIGNALED 32 + +struct d3dkmt_adapterinfo { + d3dkmt_handle adapter_handle; + struct winluid adapter_luid; + uint num_sources; + winbool present_move_regions_preferred; +}; + +struct d3dkmt_enumadapters2 { + uint num_adapters; + struct d3dkmt_adapterinfo *adapters; +}; + +struct d3dkmt_closeadapter { + d3dkmt_handle adapter_handle; +}; + +struct d3dkmt_openadapterfromluid { + struct winluid adapter_luid; + d3dkmt_handle adapter_handle; +}; + +struct d3dddi_allocationlist { + d3dkmt_handle allocation; + union { + struct { + uint write_operation :1; + uint do_not_retire_instance :1; + uint offer_priority :3; + uint reserved :27; + }; + uint value; + }; +}; + +struct d3dddi_patchlocationlist { + uint allocation_index; + union { + struct { + uint slot_id:24; + uint reserved:8; + }; + uint value; + }; + uint driver_id; + uint allocation_offset; + uint patch_offset; + uint split_offset; +}; + +struct d3dkmt_createdeviceflags { + uint legacy_mode:1; + uint request_vSync:1; + uint disable_gpu_timeout:1; + uint reserved:29; +}; + +struct d3dkmt_createdevice { + union { + d3dkmt_handle adapter; + void *adapter_pointer; + }; + struct d3dkmt_createdeviceflags flags; + d3dkmt_handle device; + void *command_buffer; + uint command_buffer_size; + struct d3dddi_allocationlist *allocation_list; + uint allocation_list_size; + struct d3dddi_patchlocationlist *patch_location_list; + uint patch_location_list_size; +}; + +struct d3dkmt_destroydevice { + d3dkmt_handle device; +}; + +enum d3dkmt_clienthint { + D3DKMT_CLIENTHINT_UNKNOWN = 0, + D3DKMT_CLIENTHINT_OPENGL = 1, + D3DKMT_CLIENTHINT_CDD = 2, + D3DKMT_CLIENTHINT_DX7 = 7, + D3DKMT_CLIENTHINT_DX8 = 8, + D3DKMT_CLIENTHINT_DX9 = 9, + D3DKMT_CLIENTHINT_DX10 = 10, +}; + +struct d3dddi_createcontextflags { + union { + struct { + uint null_rendering:1; + uint initial_data:1; + uint disable_gpu_timeout:1; + uint synchronization_only:1; + uint hw_queue_supported:1; + uint reserved:27; + }; + uint value; + }; +}; + +struct d3dkmt_createcontext { + d3dkmt_handle device; + uint node_ordinal; + uint engine_affinity; + struct d3dddi_createcontextflags flags; + void *priv_drv_data; + uint priv_drv_data_size; + enum d3dkmt_clienthint client_hint; + d3dkmt_handle context; + void *command_buffer; + uint command_buffer_size; + struct d3dddi_allocationlist *allocation_list; + uint allocation_list_size; + struct d3dddi_patchlocationlist *patch_location_list; + uint patch_location_list_size; + d3dgpu_virtual_address obsolete; +}; + +struct d3dkmt_destroycontext { + d3dkmt_handle context; +}; + +struct d3dkmt_createcontextvirtual { + d3dkmt_handle device; + uint node_ordinal; + uint engine_affinity; + struct d3dddi_createcontextflags flags; + void *priv_drv_data; + uint priv_drv_data_size; + enum d3dkmt_clienthint client_hint; + d3dkmt_handle context; +}; + +struct d3dddi_createhwcontextflags { + union { + struct { + uint reserved:32; + }; + uint value; + }; +}; + +struct d3dddi_createhwqueueflags { + union { + struct { + uint disable_gpu_timeout:1; + uint reserved:31; + }; + uint value; + }; +}; + +enum d3dddi_pagingqueue_priority { + D3DDDI_PAGINGQUEUE_PRIORITY_BELOW_NORMAL = -1, + D3DDDI_PAGINGQUEUE_PRIORITY_NORMAL = 0, + D3DDDI_PAGINGQUEUE_PRIORITY_ABOVE_NORMAL = 1, +}; + +struct d3dkmt_createpagingqueue { + d3dkmt_handle device; + enum d3dddi_pagingqueue_priority priority; + d3dkmt_handle paging_queue; + d3dkmt_handle sync_object; + void *fence_cpu_virtual_address; + uint physical_adapter_index; +}; + +struct d3dddi_destroypagingqueue { + d3dkmt_handle paging_queue; +}; + +struct d3dkmt_renderflags { + uint resize_command_buffer:1; + uint resize_allocation_list:1; + uint resize_patch_location_list:1; + uint null_rendering:1; + uint present_redirected:1; + uint render_km:1; + uint render_km_readback:1; + uint reserved:25; +}; +struct d3dkmt_render { + union { + d3dkmt_handle device; + d3dkmt_handle context; + }; + uint command_offset; + uint command_length; + uint allocation_count; + uint patch_location_count; + void *new_command_buffer; + uint new_command_buffer_size; + struct d3dddi_allocationlist *new_allocation_list; + uint new_allocation_list_size; + struct d3dddi_patchlocationlist *new_patch_pocation_list; + uint new_patch_pocation_list_size; + struct d3dkmt_renderflags flags; + uint64_t present_history_token; + uint broadcast_context_count; + d3dkmt_handle broadcast_context[D3DDDI_MAX_BROADCAST_CONTEXT]; + uint queued_buffer_count; + d3dgpu_virtual_address obsolete; + void *priv_drv_data; + uint priv_drv_data_size; +}; + +enum d3dkmt_standardallocationtype { + D3DKMT_STANDARDALLOCATIONTYPE_EXISTINGHEAP = 1, +}; + +struct d3dkmt_standardallocation_existingheap { + uint64_t size; +}; + +struct d3dkmt_createstandardallocationflags { + union { + struct { + uint reserved:32; + }; + uint value; + }; +}; + +struct d3dkmt_createstandardallocation { + enum d3dkmt_standardallocationtype type; + struct d3dkmt_standardallocation_existingheap existing_heap_data; + struct d3dkmt_createstandardallocationflags flags; +}; + +struct d3dddi_allocationinfo2 { + d3dkmt_handle allocation; + union { + winhandle section; + const void *sysmem; + }; + void *priv_drv_data; + uint priv_drv_data_size; + uint vidpn_source_id; + union { + struct { + uint primary:1; + uint stereo:1; + uint override_priority:1; + uint reserved:29; + }; + uint value; + } flags; + d3dgpu_virtual_address gpu_virtual_address; + union { + uint priority; + uint64_t unused; + }; + uint64_t reserved[5]; +}; + +struct d3dkmt_createallocationflags { + union { + struct { + uint create_resource:1; + uint create_shared:1; + uint non_secure:1; + uint create_protected:1; + uint restrict_shared_access:1; + uint existing_sysmem:1; + uint nt_security_sharing:1; + uint read_only:1; + uint create_write_combined:1; + uint create_cached:1; + uint swap_chain_back_buffer:1; + uint cross_adapter:1; + uint open_cross_adapter:1; + uint partial_shared_creation:1; + uint zeroed:1; + uint write_watch:1; + uint standard_allocation:1; + uint existing_section:1; + uint reserved:14; + }; + uint value; + }; +}; + +struct d3dkmt_createallocation { + d3dkmt_handle device; + d3dkmt_handle resource; + d3dkmt_handle global_share; + const void *private_runtime_data; + uint private_runtime_data_size; + union { + struct d3dkmt_createstandardallocation *standard_allocation; + const void *priv_drv_data; + }; + uint priv_drv_data_size; + uint alloc_count; + struct d3dddi_allocationinfo2 *allocation_info; + struct d3dkmt_createallocationflags flags; + winhandle private_runtime_resource_handle; +}; + +struct d3dddicb_destroyallocation2flags { + union { + struct { + uint assume_not_in_use:1; + uint synchronous_destroy:1; + uint reserved:29; + uint system_use_only:1; + }; + uint value; + }; +}; + +struct d3dkmt_destroyallocation2 { + d3dkmt_handle device; + d3dkmt_handle resource; + const d3dkmt_handle *allocations; + uint alloc_count; + struct d3dddicb_destroyallocation2flags flags; +}; + +struct d3dddi_makeresident_flags { + union { + struct { + uint cant_trim_further:1; + uint must_succeed:1; + uint reserved:30; + }; + uint value; + }; +}; + +struct d3dddi_makeresident { + d3dkmt_handle paging_queue; + uint alloc_count; + const d3dkmt_handle *allocation_list; + const uint *priority_list; + struct d3dddi_makeresident_flags flags; + uint64_t paging_fence_value; + uint64_t num_bytes_to_trim; +}; + +struct d3dddi_evict_flags { + union { + struct { + uint evict_only_if_necessary:1; + uint not_written_to:1; + uint reserved:30; + }; + uint value; + }; +}; + +struct d3dkmt_evict { + d3dkmt_handle device; + uint alloc_count; + const d3dkmt_handle *allocations; + struct d3dddi_evict_flags flags; + uint64_t num_bytes_to_trim; +}; + +struct d3dddigpuva_protection_type { + union { + struct { + uint64_t write:1; + uint64_t execute:1; + uint64_t zero:1; + uint64_t no_access:1; + uint64_t system_use_only:1; + uint64_t reserved:59; + }; + uint64_t value; + }; +}; + +enum d3dddi_updategpuvirtualaddress_operation_type { + D3DDDI_UPDATEGPUVIRTUALADDRESS_MAP = 0, + D3DDDI_UPDATEGPUVIRTUALADDRESS_UNMAP = 1, + D3DDDI_UPDATEGPUVIRTUALADDRESS_COPY = 2, + D3DDDI_UPDATEGPUVIRTUALADDRESS_MAP_PROTECT = 3, +}; + +struct d3dddi_updategpuvirtualaddress_operation { + enum d3dddi_updategpuvirtualaddress_operation_type operation; + union { + struct { + d3dgpu_virtual_address base_address; + d3dgpu_size_t size; + d3dkmt_handle allocation; + d3dgpu_size_t allocation_offset; + d3dgpu_size_t allocation_size; + } map; + struct { + d3dgpu_virtual_address base_address; + d3dgpu_size_t size; + d3dkmt_handle allocation; + d3dgpu_size_t allocation_offset; + d3dgpu_size_t allocation_size; + struct d3dddigpuva_protection_type protection; + uint64_t driver_protection; + } map_protect; + struct { + d3dgpu_virtual_address base_address; + d3dgpu_size_t size; + struct d3dddigpuva_protection_type protection; + } unmap; + struct { + d3dgpu_virtual_address source_address; + d3dgpu_size_t size; + d3dgpu_virtual_address dest_address; + } copy; + }; +}; + +enum d3dddigpuva_reservation_type { + D3DDDIGPUVA_RESERVE_NO_ACCESS = 0, + D3DDDIGPUVA_RESERVE_ZERO = 1, + D3DDDIGPUVA_RESERVE_NO_COMMIT = 2 +}; + +struct d3dkmt_updategpuvirtualaddress { + d3dkmt_handle device; + d3dkmt_handle context; + d3dkmt_handle fence_object; + uint num_operations; + struct d3dddi_updategpuvirtualaddress_operation *operations; + uint reserved0; + uint64_t reserved1; + uint64_t fence_value; + union { + struct { + uint do_not_wait:1; + uint reserved:31; + }; + uint value; + } flags; +}; + +struct d3dddi_mapgpuvirtualaddress { + d3dkmt_handle paging_queue; + d3dgpu_virtual_address base_address; + d3dgpu_virtual_address minimum_address; + d3dgpu_virtual_address maximum_address; + d3dkmt_handle allocation; + d3dgpu_size_t offset_in_pages; + d3dgpu_size_t size_in_pages; + struct d3dddigpuva_protection_type protection; + uint64_t driver_protection; + uint reserved0; + uint64_t reserved1; + d3dgpu_virtual_address virtual_address; + uint64_t paging_fence_value; +}; + +struct d3dddi_reservegpuvirtualaddress { + d3dkmt_handle adapter; + d3dgpu_virtual_address base_address; + d3dgpu_virtual_address minimum_address; + d3dgpu_virtual_address maximum_address; + d3dgpu_size_t size; + enum d3dddigpuva_reservation_type reservation_type; + uint64_t driver_protection; + d3dgpu_virtual_address virtual_address; + uint64_t paging_fence_value; +}; + +struct d3dkmt_freegpuvirtualaddress { + d3dkmt_handle adapter; + d3dgpu_virtual_address base_address; + d3dgpu_size_t size; +}; + +enum d3dkmt_memory_segment_group { + D3DKMT_MEMORY_SEGMENT_GROUP_LOCAL = 0, + D3DKMT_MEMORY_SEGMENT_GROUP_NON_LOCAL = 1 +}; + +struct d3dkmt_queryvideomemoryinfo { + winhandle process; + d3dkmt_handle adapter; + enum d3dkmt_memory_segment_group memory_segment_group; + uint64_t budget; + uint64_t current_usage; + uint64_t current_reservation; + uint64_t available_for_reservation; + uint physical_adapter_index; +}; + +enum qai_driverversion { + KMT_DRIVERVERSION_WDDM_1_0 = 1000, + KMT_DRIVERVERSION_WDDM_1_1_PRERELEASE = 1102, + KMT_DRIVERVERSION_WDDM_1_1 = 1105, + KMT_DRIVERVERSION_WDDM_1_2 = 1200, + KMT_DRIVERVERSION_WDDM_1_3 = 1300, + KMT_DRIVERVERSION_WDDM_2_0 = 2000, + KMT_DRIVERVERSION_WDDM_2_1 = 2100, + KMT_DRIVERVERSION_WDDM_2_2 = 2200, + KMT_DRIVERVERSION_WDDM_2_3 = 2300, + KMT_DRIVERVERSION_WDDM_2_4 = 2400, + KMT_DRIVERVERSION_WDDM_2_5 = 2500, + KMT_DRIVERVERSION_WDDM_2_6 = 2600, + KMT_DRIVERVERSION_WDDM_2_7 = 2700 +}; + +struct d3dkmt_adaptertype { + union { + struct { + uint render_supported:1; + uint display_supported:1; + uint software_device:1; + uint post_device:1; + uint hybrid_discrete:1; + uint hybrid_integrated:1; + uint indirect_display_device:1; + uint paravirtualized:1; + uint acg_supported:1; + uint support_set_timings_from_vidpn:1; + uint detachable:1; + uint compute_only:1; + uint prototype:1; + uint reserved:19; + }; + uint value; + }; +}; + +enum kmtqueryadapterinfotype { + KMTQAITYPE_UMDRIVERPRIVATE = 0, + KMTQAITYPE_UMDRIVERNAME = 1, + KMTQAITYPE_UMOPENGLINFO = 2, + KMTQAITYPE_GETSEGMENTSIZE = 3, + KMTQAITYPE_ADAPTERGUID = 4, + KMTQAITYPE_FLIPQUEUEINFO = 5, + KMTQAITYPE_ADAPTERADDRESS = 6, + KMTQAITYPE_SETWORKINGSETINFO = 7, + KMTQAITYPE_ADAPTERREGISTRYINFO = 8, + KMTQAITYPE_CURRENTDISPLAYMODE = 9, + KMTQAITYPE_MODELIST = 10, + KMTQAITYPE_CHECKDRIVERUPDATESTATUS = 11, + KMTQAITYPE_VIRTUALADDRESSINFO = 12, + KMTQAITYPE_DRIVERVERSION = 13, + KMTQAITYPE_ADAPTERTYPE = 15, + KMTQAITYPE_OUTPUTDUPLCONTEXTSCOUNT = 16, + KMTQAITYPE_WDDM_1_2_CAPS = 17, + KMTQAITYPE_UMD_DRIVER_VERSION = 18, + KMTQAITYPE_DIRECTFLIP_SUPPORT = 19, + KMTQAITYPE_MULTIPLANEOVERLAY_SUPPORT = 20, + KMTQAITYPE_DLIST_DRIVER_NAME = 21, + KMTQAITYPE_WDDM_1_3_CAPS = 22, + KMTQAITYPE_MULTIPLANEOVERLAY_HUD_SUPPORT = 23, + KMTQAITYPE_WDDM_2_0_CAPS = 24, + KMTQAITYPE_NODEMETADATA = 25, + KMTQAITYPE_CPDRIVERNAME = 26, + KMTQAITYPE_XBOX = 27, + KMTQAITYPE_INDEPENDENTFLIP_SUPPORT = 28, + KMTQAITYPE_MIRACASTCOMPANIONDRIVERNAME = 29, + KMTQAITYPE_PHYSICALADAPTERCOUNT = 30, + KMTQAITYPE_PHYSICALADAPTERDEVICEIDS = 31, + KMTQAITYPE_DRIVERCAPS_EXT = 32, + KMTQAITYPE_QUERY_MIRACAST_DRIVER_TYPE = 33, + KMTQAITYPE_QUERY_GPUMMU_CAPS = 34, + KMTQAITYPE_QUERY_MULTIPLANEOVERLAY_DECODE_SUPPORT = 35, + KMTQAITYPE_QUERY_HW_PROTECTION_TEARDOWN_COUNT = 36, + KMTQAITYPE_QUERY_ISBADDRIVERFORHWPROTECTIONDISABLED = 37, + KMTQAITYPE_MULTIPLANEOVERLAY_SECONDARY_SUPPORT = 38, + KMTQAITYPE_INDEPENDENTFLIP_SECONDARY_SUPPORT = 39, + KMTQAITYPE_PANELFITTER_SUPPORT = 40, + KMTQAITYPE_PHYSICALADAPTERPNPKEY = 41, + KMTQAITYPE_GETSEGMENTGROUPSIZE = 42, + KMTQAITYPE_MPO3DDI_SUPPORT = 43, + KMTQAITYPE_HWDRM_SUPPORT = 44, + KMTQAITYPE_MPOKERNELCAPS_SUPPORT = 45, + KMTQAITYPE_MULTIPLANEOVERLAY_STRETCH_SUPPORT = 46, + KMTQAITYPE_GET_DEVICE_VIDPN_OWNERSHIP_INFO = 47, + KMTQAITYPE_QUERYREGISTRY = 48, + KMTQAITYPE_KMD_DRIVER_VERSION = 49, + KMTQAITYPE_BLOCKLIST_KERNEL = 50, + KMTQAITYPE_BLOCKLIST_RUNTIME = 51, + KMTQAITYPE_ADAPTERGUID_RENDER = 52, + KMTQAITYPE_ADAPTERADDRESS_RENDER = 53, + KMTQAITYPE_ADAPTERREGISTRYINFO_RENDER = 54, + KMTQAITYPE_CHECKDRIVERUPDATESTATUS_RENDER = 55, + KMTQAITYPE_DRIVERVERSION_RENDER = 56, + KMTQAITYPE_ADAPTERTYPE_RENDER = 57, + KMTQAITYPE_WDDM_1_2_CAPS_RENDER = 58, + KMTQAITYPE_WDDM_1_3_CAPS_RENDER = 59, + KMTQAITYPE_QUERY_ADAPTER_UNIQUE_GUID = 60, + KMTQAITYPE_NODEPERFDATA = 61, + KMTQAITYPE_ADAPTERPERFDATA = 62, + KMTQAITYPE_ADAPTERPERFDATA_CAPS = 63, + KMTQUITYPE_GPUVERSION = 64, + KMTQAITYPE_DRIVER_DESCRIPTION = 65, + KMTQAITYPE_DRIVER_DESCRIPTION_RENDER = 66, + KMTQAITYPE_SCANOUT_CAPS = 67, + KMTQAITYPE_PARAVIRTUALIZATION_RENDER = 68, +}; + +struct d3dkmt_queryadapterinfo { + d3dkmt_handle adapter; + enum kmtqueryadapterinfotype type; + void *private_data; + uint private_data_size; +}; + +enum d3dkmt_escapetype { + D3DKMT_ESCAPE_DRIVERPRIVATE = 0, + D3DKMT_ESCAPE_VIDMM = 1, + D3DKMT_ESCAPE_VIDSCH = 3, + D3DKMT_ESCAPE_DEVICE = 4, + D3DKMT_ESCAPE_DRT_TEST = 8, +}; + +enum d3dkmt_drt_test_command { + D3DKMT_DRT_TEST_COMMAND_HANDLETABLE = 39, +}; + +struct d3dkmt_drt_escape_head { + uint signature; + uint buffer_size; + enum d3dkmt_drt_test_command command; +}; + +enum d3dkmt_ht_command { + D3DKMT_HT_COMMAND_ALLOC, + D3DKMT_HT_COMMAND_FREE, + D3DKMT_HT_COMMAND_ASSIGN, + D3DKMT_HT_COMMAND_GET, + D3DKMT_HT_COMMAND_DESTROY, +}; + +struct d3dkmt_ht_desc { + struct d3dkmt_drt_escape_head head; + enum d3dkmt_ht_command command; + uint index; + d3dkmt_handle handle; + uint object_type; + void *object; +}; + +struct d3dddi_escapeflags { + union { + struct { + uint hardware_access:1; + uint device_status_query:1; + uint change_frame_latency:1; + uint no_adapter_synchronization:1; + uint reserved:1; + uint virtual_machine_data:1; + uint driver_known_escape:1; + uint driver_common_escape:1; + uint reserved2:24; + }; + uint value; + }; +}; + +struct d3dkmt_escape { + d3dkmt_handle adapter; + d3dkmt_handle device; + enum d3dkmt_escapetype type; + struct d3dddi_escapeflags flags; + void *priv_drv_data; + uint priv_drv_data_size; + d3dkmt_handle context; +}; + +enum dxgk_render_pipeline_stage { + DXGK_RENDER_PIPELINE_STAGE_UNKNOWN = 0, + DXGK_RENDER_PIPELINE_STAGE_INPUT_ASSEMBLER = 1, + DXGK_RENDER_PIPELINE_STAGE_VERTEX_SHADER = 2, + DXGK_RENDER_PIPELINE_STAGE_GEOMETRY_SHADER = 3, + DXGK_RENDER_PIPELINE_STAGE_STREAM_OUTPUT = 4, + DXGK_RENDER_PIPELINE_STAGE_RASTERIZER = 5, + DXGK_RENDER_PIPELINE_STAGE_PIXEL_SHADER = 6, + DXGK_RENDER_PIPELINE_STAGE_OUTPUT_MERGER = 7, +}; + +enum dxgk_page_fault_flags { + DXGK_PAGE_FAULT_WRITE = 0x1, + DXGK_PAGE_FAULT_FENCE_INVALID = 0x2, + DXGK_PAGE_FAULT_ADAPTER_RESET_REQUIRED = 0x4, + DXGK_PAGE_FAULT_ENGINE_RESET_REQUIRED = 0x8, + DXGK_PAGE_FAULT_FATAL_HARDWARE_ERROR = 0x10, + DXGK_PAGE_FAULT_IOMMU = 0x20, + DXGK_PAGE_FAULT_HW_CONTEXT_VALID = 0x40, + DXGK_PAGE_FAULT_PROCESS_HANDLE_VALID = 0x80, +}; + +enum dxgk_general_error_code { + DXGK_GENERAL_ERROR_PAGE_FAULT = 0, + DXGK_GENERAL_ERROR_INVALID_INSTRUCTION = 1, +}; + +struct dxgk_fault_error_code { + union { + struct { + uint is_device_specific_code:1; + enum dxgk_general_error_code general_error_code:31; + }; + struct { + uint is_device_specific_code_reserved_bit:1; + uint device_specific_code:31; + }; + }; +}; + +enum d3dkmt_deviceexecution_state { + D3DKMT_DEVICEEXECUTION_ACTIVE = 1, + D3DKMT_DEVICEEXECUTION_RESET = 2, + D3DKMT_DEVICEEXECUTION_HUNG = 3, + D3DKMT_DEVICEEXECUTION_STOPPED = 4, + D3DKMT_DEVICEEXECUTION_ERROR_OUTOFMEMORY = 5, + D3DKMT_DEVICEEXECUTION_ERROR_DMAFAULT = 6, + D3DKMT_DEVICEEXECUTION_ERROR_DMAPAGEFAULT = 7, +}; + +struct d3dkmt_devicereset_state { + union { + struct { + uint desktop_switched:1; + uint reserved:31; + }; + uint value; + }; +}; + +struct d3dkmt_present_stats { + uint present_count; + uint present_refresh_count; + uint sync_refresh_count; + uint64_t sync_qpc_time; + uint64_t sync_gpu_time; +}; + +struct d3dkmt_devicepresent_state { + uint vidpn_source_id; + struct d3dkmt_present_stats present_stats; +}; + +struct d3dkmt_present_stats_dwm { + uint present_count; + uint present_refresh_count; + uint64_t present_qpc_time; + uint sync_refresh_count; + uint64_t sync_qpc_time; + uint custom_present_duration; +}; + +struct d3dkmt_devicepagefault_state { + uint64_t faulted_primitive_api_sequence_number; + enum dxgk_render_pipeline_stage faulted_pipeline_stage; + uint faulted_bind_table_entry; + enum dxgk_page_fault_flags page_fault_flags; + struct dxgk_fault_error_code fault_error_code; + d3dgpu_virtual_address faulted_virtual_address; +}; + +struct d3dkmt_devicepresent_state_dwm { + uint vidpn_source_id; + struct d3dkmt_present_stats_dwm present_stats; +}; + +struct d3dkmt_devicepresent_queue_state { + uint vidpn_source_id; + bool bQueuedPresentLimitReached; +}; + +enum d3dkmt_devicestate_type { + D3DKMT_DEVICESTATE_EXECUTION = 1, + D3DKMT_DEVICESTATE_PRESENT = 2, + D3DKMT_DEVICESTATE_RESET = 3, + D3DKMT_DEVICESTATE_PRESENT_DWM = 4, + D3DKMT_DEVICESTATE_PAGE_FAULT = 5, + D3DKMT_DEVICESTATE_PRESENT_QUEUE = 6, +}; + +struct d3dkmt_getdevicestate { + d3dkmt_handle device; + enum d3dkmt_devicestate_type state_type; + union { + enum d3dkmt_deviceexecution_state execution_state; + struct d3dkmt_devicepresent_state present_state; + struct d3dkmt_devicereset_state reset_state; + struct d3dkmt_devicepresent_state_dwm present_state_dwm; + struct d3dkmt_devicepagefault_state page_fault_state; + struct d3dkmt_devicepresent_queue_state present_queue_state; + }; +}; + +enum d3dkmdt_gdisurfacetype { + D3DKMDT_GDISURFACE_INVALID = 0, + D3DKMDT_GDISURFACE_TEXTURE = 1, + D3DKMDT_GDISURFACE_STAGING_CPUVISIBLE = 2, + D3DKMDT_GDISURFACE_STAGING = 3, + D3DKMDT_GDISURFACE_LOOKUPTABLE = 4, + D3DKMDT_GDISURFACE_EXISTINGSYSMEM = 5, + D3DKMDT_GDISURFACE_TEXTURE_CPUVISIBLE = 6, + D3DKMDT_GDISURFACE_TEXTURE_CROSSADAPTER = 7, + D3DKMDT_GDISURFACE_TEXTURE_CPUVISIBLE_CROSSADAPTER = 8, +}; + +struct d3dddi_rational { + uint numerator; + uint denominator; +}; + +enum d3dddiformat { + D3DDDIFMT_UNKNOWN = 0, +}; + +struct d3dkmdt_gdisurfacedata { + uint width; + uint height; + uint format; + enum d3dkmdt_gdisurfacetype type; + uint flags; + uint pitch; +}; + +struct d3dkmtd_stagingsurfacedata { + uint width; + uint height; + uint pitch; +}; + +struct d3dkmdt_sharedprimarysurfacedata { + uint width; + uint height; + enum d3dddiformat format; + struct d3dddi_rational refresh_rate; + uint vidpn_source_id; +}; + +struct d3dkmdt_shadowsurfacedata { + uint width; + uint height; + enum d3dddiformat format; + uint pitch; +}; + +enum d3dkmdt_standardallocationtype { + D3DKMDT_STANDARDALLOCATION_SHAREDPRIMARYSURFACE = 1, + D3DKMDT_STANDARDALLOCATION_SHADOWSURFACE = 2, + D3DKMDT_STANDARDALLOCATION_STAGINGSURFACE = 3, + D3DKMDT_STANDARDALLOCATION_GDISURFACE = 4, +}; + +struct d3dddi_synchronizationobject_flags { + union { + struct { + uint shared:1; + uint nt_security_sharing:1; + uint cross_adapter:1; + uint top_of_pipeline:1; + uint no_signal:1; + uint no_wait:1; + uint no_signal_max_value_on_tdr:1; + uint no_gpu_access:1; + uint reserved:23; + }; + uint value; + }; +}; + +enum d3dddi_synchronizationobject_type { + D3DDDI_SYNCHRONIZATION_MUTEX = 1, + D3DDDI_SEMAPHORE = 2, + D3DDDI_FENCE = 3, + D3DDDI_CPU_NOTIFICATION = 4, + D3DDDI_MONITORED_FENCE = 5, + D3DDDI_PERIODIC_MONITORED_FENCE = 6, + D3DDDI_SYNCHRONIZATION_TYPE_LIMIT +}; + +struct d3dddi_synchronizationobjectinfo2 { + enum d3dddi_synchronizationobject_type type; + struct d3dddi_synchronizationobject_flags flags; + union { + struct { + winbool initial_state; + } synchronization_mutex; + + struct { + uint max_count; + uint initial_count; + } semaphore; + + struct { + uint64_t fence_value; + } fence; + + struct { + winhandle event; + } cpu_notification; + + struct { + uint64_t initial_fence_value; + void *fence_cpu_virtual_address; + d3dgpu_virtual_address fence_gpu_virtual_address; + uint engine_affinity; + } monitored_fence; + + struct periodic_monitored_fence_t { + d3dkmt_handle adapter; + uint vidpn_target_id; + uint64_t time; + void *fence_cpu_virtual_address; + d3dgpu_virtual_address fence_gpu_virtual_address; + uint engine_affinity; + } periodic_monitored_fence; + + struct { + uint64_t reserved[8]; + } reserved; + }; + d3dkmt_handle shared_handle; +}; + +struct d3dkmt_createsynchronizationobject2 { + d3dkmt_handle device; + struct d3dddi_synchronizationobjectinfo2 info; + d3dkmt_handle sync_object; +}; + +struct d3dkmt_waitforsynchronizationobject2 { + d3dkmt_handle context; + uint object_count; + d3dkmt_handle object_array[D3DDDI_MAX_OBJECT_WAITED_ON]; + union { + struct { + uint64_t fence_value; + } fence; + uint64_t reserved[8]; + }; +}; + +struct d3dddicb_signalflags { + union { + struct { + uint signal_at_submission:1; + uint enqueue_cpu_event:1; + uint allow_fence_rewind:1; + uint reserved:28; + uint DXGK_SIGNAL_FLAG_INTERNAL0:1; + }; + uint value; + }; +}; + +struct d3dkmt_signalsynchronizationobject2 { + d3dkmt_handle context; + uint object_count; + d3dkmt_handle object_array[D3DDDI_MAX_OBJECT_SIGNALED]; + struct d3dddicb_signalflags flags; + uint context_count; + d3dkmt_handle contexts[D3DDDI_MAX_BROADCAST_CONTEXT]; + union { + struct { + uint64_t fence_value; + } fence; + winhandle cpu_event_handle; + uint64_t reserved[8]; + }; +}; + +struct d3dddi_waitforsynchronizationobjectfromcpu_flags { + union { + struct { + uint wait_any:1; + uint reserved:31; + }; + uint value; + }; +}; + +struct d3dkmt_waitforsynchronizationobjectfromcpu { + d3dkmt_handle device; + uint object_count; + d3dkmt_handle *objects; + uint64_t *fence_values; + winhandle async_event; + struct d3dddi_waitforsynchronizationobjectfromcpu_flags flags; +}; + +struct d3dkmt_signalsynchronizationobjectfromcpu { + d3dkmt_handle device; + uint object_count; + d3dkmt_handle *objects; + uint64_t *fence_values; + struct d3dddicb_signalflags flags; +}; + +struct d3dkmt_waitforsynchronizationobjectfromgpu { + d3dkmt_handle context; + uint object_count; + d3dkmt_handle *objects; + union { + uint64_t *monitored_fence_values; + uint64_t fence_value; + uint64_t reserved[8]; + }; +}; + +struct d3dkmt_signalsynchronizationobjectfromgpu { + d3dkmt_handle context; + uint object_count; + d3dkmt_handle *objects; + union { + uint64_t *monitored_fence_values; + uint64_t reserved[8]; + }; +}; + +struct d3dkmt_signalsynchronizationobjectfromgpu2 { + uint object_count; + d3dkmt_handle *objects; + struct d3dddicb_signalflags flags; + uint context_count; + d3dkmt_handle *contexts; + union { + uint64_t fence_value; + winhandle cpu_event_handle; + uint64_t *monitored_fence_values; + uint64_t reserved[8]; + }; +}; + +struct d3dkmt_destroysynchronizationobject { + d3dkmt_handle sync_object; +}; + +struct d3dkmt_opensynchronizationobject { + d3dkmt_handle shared_handle; + d3dkmt_handle sync_object; + uint64_t reserved[8]; +}; + +struct d3dkmt_submitcommandflags { + uint null_rendering:1; + uint present_redirected:1; + uint reserved:30; +}; + +struct d3dkmt_submitcommand { + d3dgpu_virtual_address command_buffer; + uint command_length; + struct d3dkmt_submitcommandflags flags; + uint64_t present_history_token; + uint broadcast_context_count; + d3dkmt_handle broadcast_context[D3DDDI_MAX_BROADCAST_CONTEXT]; + void *priv_drv_data; + uint priv_drv_data_size; + uint num_primaries; + d3dkmt_handle written_primaries[D3DDDI_MAX_WRITTEN_PRIMARIES]; + uint num_history_buffers; + d3dkmt_handle *history_buffer_array; +}; + +struct d3dkmt_submitcommandtohwqueue { + d3dkmt_handle hwqueue; + uint64_t hwqueue_progress_fence_id; + d3dgpu_virtual_address command_buffer; + uint command_length; + uint priv_drv_data_size; + void *priv_drv_data; + uint num_primaries; + d3dkmt_handle *written_primaries; +}; + +struct d3dkmt_setcontextschedulingpriority { + d3dkmt_handle context; + int priority; +}; + +struct d3dkmt_setcontextinprocessschedulingpriority { + d3dkmt_handle context; + int priority; +}; + +struct d3dkmt_getcontextschedulingpriority { + d3dkmt_handle context; + int priority; +}; + +struct d3dkmt_getcontextinprocessschedulingpriority { + d3dkmt_handle context; + int priority; +}; + +struct d3dkmt_setallocationpriority { + d3dkmt_handle device; + d3dkmt_handle resource; + const d3dkmt_handle *allocation_list; + uint allocation_count; + const uint *priorities; +}; + +struct d3dkmt_getallocationpriority { + d3dkmt_handle device; + d3dkmt_handle resource; + const d3dkmt_handle *allocation_list; + uint allocation_count; + uint *priorities; +}; + +enum d3dkmt_allocationresidencystatus { + D3DKMT_ALLOCATIONRESIDENCYSTATUS_RESIDENTINGPUMEMORY = 1, + D3DKMT_ALLOCATIONRESIDENCYSTATUS_RESIDENTINSHAREDMEMORY = 2, + D3DKMT_ALLOCATIONRESIDENCYSTATUS_NOTRESIDENT = 3, +}; + +struct d3dkmt_queryallocationresidency { + d3dkmt_handle device; + d3dkmt_handle resource; + d3dkmt_handle *allocations; + uint allocation_count; + enum d3dkmt_allocationresidencystatus *residency_status; +}; + +struct D3DDDICB_LOCK2FLAGS { + union { + struct { + uint reserved:32; + }; + uint value; + }; +}; + +struct d3dkmt_lock2 { + d3dkmt_handle device; + d3dkmt_handle allocation; + struct D3DDDICB_LOCK2FLAGS flags; + void *data; +}; + +struct d3dkmt_unlock2 { + d3dkmt_handle device; + d3dkmt_handle allocation; +}; + +enum D3DKMT_DEVICE_ERROR_REASON { + D3DKMT_DEVICE_ERROR_REASON_GENERIC = 0x80000000, + D3DKMT_DEVICE_ERROR_REASON_DRIVER_ERROR = 0x80000006, +}; + +struct d3dkmt_markdeviceaserror { + d3dkmt_handle device; + enum D3DKMT_DEVICE_ERROR_REASON reason; +}; + +struct D3DDDI_UPDATEALLOCPROPERTY_FLAGS { + union { + struct { + uint accessed_physically:1; + uint reserved:31; + }; + uint value; + }; +}; + +struct D3DDDI_SEGMENTPREFERENCE { + union { + struct { + uint segment_id0:5; + uint direction0:1; + uint segment_id1:5; + uint direction1:1; + uint segment_id2:5; + uint direction2:1; + uint segment_id3:5; + uint direction3:1; + uint segment_id4:5; + uint direction4:1; + uint reserved:2; + }; + uint value; + }; +}; + +struct d3dddi_updateallocproperty { + d3dkmt_handle paging_queue; + d3dkmt_handle allocation; + uint supported_segment_set; + struct D3DDDI_SEGMENTPREFERENCE preferred_segment; + struct D3DDDI_UPDATEALLOCPROPERTY_FLAGS flags; + uint64_t paging_fence_value; + union { + struct { + uint set_accessed_physically:1; + uint set_supported_segmentSet:1; + uint set_preferred_segment:1; + uint reserved:29; + }; + uint property_mask_value; + }; +}; + +enum d3dkmt_offer_priority { + D3DKMT_OFFER_PRIORITY_LOW = 1, + D3DKMT_OFFER_PRIORITY_NORMAL = 2, + D3DKMT_OFFER_PRIORITY_HIGH = 3, + D3DKMT_OFFER_PRIORITY_AUTO = 4, +}; + +struct d3dkmt_offer_flags { + union { + struct { + uint offer_immediately:1; + uint allow_decommit:1; + uint reserved:30; + }; + uint value; + }; +}; + +struct d3dkmt_offerallocations { + d3dkmt_handle device; + d3dkmt_handle *resources; + const d3dkmt_handle *allocations; + uint allocation_count; + enum d3dkmt_offer_priority priority; + struct d3dkmt_offer_flags flags; +}; + +enum d3dddi_reclaim_result { + D3DDDI_RECLAIM_RESULT_OK = 0, + D3DDDI_RECLAIM_RESULT_DISCARDED = 1, + D3DDDI_RECLAIM_RESULT_NOT_COMMITTED = 2, +}; + +struct d3dkmt_reclaimallocations2 { + d3dkmt_handle paging_queue; + uint allocation_count; + d3dkmt_handle *resources; + d3dkmt_handle *allocations; + union { + winbool *discarded; + enum d3dddi_reclaim_result *results; + }; + uint64_t paging_fence_value; +}; + +struct d3dkmt_changevideomemoryreservation { + winhandle process; + d3dkmt_handle adapter; + enum d3dkmt_memory_segment_group memory_segment_group; + uint64_t reservation; + uint physical_adapter_index; +}; + +struct d3dkmt_createhwcontext { + d3dkmt_handle device; + uint node_ordinal; + uint engine_affinity; + struct d3dddi_createhwcontextflags flags; + uint priv_drv_data_size; + void *priv_drv_data; + d3dkmt_handle context; +}; + +struct d3dkmt_destroyhwcontext { + d3dkmt_handle context; +}; + +struct d3dkmt_createhwqueue { + d3dkmt_handle context; + struct d3dddi_createhwqueueflags flags; + uint priv_drv_data_size; + void *priv_drv_data; + d3dkmt_handle queue; + d3dkmt_handle queue_progress_fence; + void *queue_progress_fence_cpu_va; + d3dgpu_virtual_address queue_progress_fence_gpu_va; +}; + +struct d3dkmt_destroyhwqueue { + d3dkmt_handle queue; +}; + +struct d3dkmt_submitwaitforsyncobjectstohwqueue { + d3dkmt_handle hwqueue; + uint object_count; + d3dkmt_handle *objects; + uint64_t *fence_values; +}; + +struct d3dkmt_submitsignalsyncobjectstohwqueue { + struct d3dddicb_signalflags flags; + uint hwqueue_count; + d3dkmt_handle *hwqueues; + uint object_count; + d3dkmt_handle *objects; + uint64_t *fence_values; +}; + +struct dxgk_gpuclockdata_flags { + union { + struct { + uint context_management_processor:1; + uint reserved:31; + }; + uint value; + }; +}; + +struct dxgk_gpuclockdata { + uint64_t gpu_frequency; + uint64_t gpu_clock_counter; + uint64_t cpu_clock_counter; + struct dxgk_gpuclockdata_flags flags; +} __packed; + +struct d3dkmt_queryclockcalibration { + d3dkmt_handle adapter; + uint node_ordinal; + uint physical_adapter_index; + struct dxgk_gpuclockdata clock_data; +}; + +struct d3dkmt_flushheaptransitions { + d3dkmt_handle adapter; +}; + +struct d3dkmt_getsharedresourceadapterluid { + d3dkmt_handle global_share; + winhandle handle; + struct winluid adapter_luid; +}; + +struct d3dkmt_invalidatecache { + d3dkmt_handle device; + d3dkmt_handle allocation; + uint64_t offset; + uint64_t length; +}; + +struct d3dddi_openallocationinfo2 { + d3dkmt_handle allocation; + void *priv_drv_data; + uint priv_drv_data_size; + d3dgpu_virtual_address gpu_va; + uint64_t reserved[6]; +}; + +struct d3dkmt_opensyncobjectfromnthandle { + winhandle nt_handle; + d3dkmt_handle sync_object; +}; + +struct d3dkmt_opensyncobjectfromnthandle2 { + winhandle nt_handle; + d3dkmt_handle device; + struct d3dddi_synchronizationobject_flags flags; + d3dkmt_handle sync_object; + union { + struct { + void *fence_value_cpu_va; + d3dgpu_virtual_address fence_value_gpu_va; + uint engine_affinity; + } monitored_fence; + uint64_t reserved[8]; + }; +}; + +struct d3dkmt_openresource { + d3dkmt_handle device; + d3dkmt_handle global_share; + uint allocation_count; + struct d3dddi_openallocationinfo2 *open_alloc_info; + void *private_runtime_data; + int private_runtime_data_size; + void *resource_priv_drv_data; + uint resource_priv_drv_data_size; + void *total_priv_drv_data; + uint total_priv_drv_data_size; + d3dkmt_handle resource; +}; + +struct d3dkmt_openresourcefromnthandle { + d3dkmt_handle device; + winhandle nt_handle; + uint allocation_count; + struct d3dddi_openallocationinfo2 *open_alloc_info; + int private_runtime_data_size; + void *private_runtime_data; + uint resource_priv_drv_data_size; + void *resource_priv_drv_data; + uint total_priv_drv_data_size; + void *total_priv_drv_data; + d3dkmt_handle resource; + d3dkmt_handle keyed_mutex; + void *keyed_mutex_private_data; + uint keyed_mutex_private_data_size; + d3dkmt_handle sync_object; +}; + +struct d3dkmt_queryresourceinfofromnthandle { + d3dkmt_handle device; + winhandle nt_handle; + void *private_runtime_data; + uint private_runtime_data_size; + uint total_priv_drv_data_size; + uint resource_priv_drv_data_size; + uint allocation_count; +}; + +struct d3dkmt_queryresourceinfo { + d3dkmt_handle device; + d3dkmt_handle global_share; + void *private_runtime_data; + uint private_runtime_data_size; + uint total_priv_drv_data_size; + uint resource_priv_drv_data_size; + uint allocation_count; +}; + +struct d3dkmt_shareobjects { + uint object_count; + const d3dkmt_handle *objects; /* per-process DXG handle */ + void *object_attr; /* security attributes */ + uint desired_access; + winhandle *shared_handle; /* output file descriptor */ +}; + +union d3dkmt_enumadapters_filter { + struct { + uint64_t include_compute_only:1; + uint64_t include_display_only:1; + uint64_t reserved:62; + }; + uint64_t value; +}; + +struct d3dkmt_enumadapters3 { + union d3dkmt_enumadapters_filter filter; + uint adapter_count; + struct d3dkmt_adapterinfo *adapters; +}; + +/* + * Dxgkrnl Graphics Port Driver ioctl definitions + * + */ + +#define LX_IOCTL_DIR_WRITE 0x1 +#define LX_IOCTL_DIR_READ 0x2 + +#define LX_IOCTL_DIR(_ioctl) (((_ioctl) >> 30) & 0x3) +#define LX_IOCTL_SIZE(_ioctl) (((_ioctl) >> 16) & 0x3FFF) +#define LX_IOCTL_TYPE(_ioctl) (((_ioctl) >> 8) & 0xFF) +#define LX_IOCTL_CODE(_ioctl) (((_ioctl) >> 0) & 0xFF) + +#define LX_IOCTL(_dir, _size, _type, _code) ( \ + (((uint)(_dir) & 0x3) << 30) | \ + (((uint)(_size) & 0x3FFF) << 16) | \ + (((uint)(_type) & 0xFF) << 8) | \ + (((uint)(_code) & 0xFF) << 0)) + +#define LX_IO(_type, _code) LX_IOCTL(0, 0, (_type), (_code)) +#define LX_IOR(_type, _code, _size) \ + LX_IOCTL(LX_IOCTL_DIR_READ, (_size), (_type), (_code)) +#define LX_IOW(_type, _code, _size) \ + LX_IOCTL(LX_IOCTL_DIR_WRITE, (_size), (_type), (_code)) +#define LX_IOWR(_type, _code, _size) \ + LX_IOCTL(LX_IOCTL_DIR_WRITE | \ + LX_IOCTL_DIR_READ, (_size), (_type), (_code)) + +#define LX_DXOPENADAPTERFROMLUID \ + LX_IOWR(0x47, 0x01, sizeof(struct d3dkmt_openadapterfromluid)) +#define LX_DXCREATEDEVICE \ + LX_IOWR(0x47, 0x02, sizeof(struct d3dkmt_createdevice)) +#define LX_DXCREATECONTEXT \ + LX_IOWR(0x47, 0x03, sizeof(struct d3dkmt_createcontext)) +#define LX_DXCREATECONTEXTVIRTUAL \ + LX_IOWR(0x47, 0x04, sizeof(struct d3dkmt_createcontextvirtual)) +#define LX_DXDESTROYCONTEXT \ + LX_IOWR(0x47, 0x05, sizeof(struct d3dkmt_destroycontext)) +#define LX_DXCREATEALLOCATION \ + LX_IOWR(0x47, 0x06, sizeof(struct d3dkmt_createallocation)) +#define LX_DXCREATEPAGINGQUEUE \ + LX_IOWR(0x47, 0x07, sizeof(struct d3dkmt_createpagingqueue)) +#define LX_DXRESERVEGPUVIRTUALADDRESS \ + LX_IOWR(0x47, 0x08, sizeof(struct d3dddi_reservegpuvirtualaddress)) +#define LX_DXQUERYADAPTERINFO \ + LX_IOWR(0x47, 0x09, sizeof(struct d3dkmt_queryadapterinfo)) +#define LX_DXQUERYVIDEOMEMORYINFO \ + LX_IOWR(0x47, 0x0a, sizeof(struct d3dkmt_queryvideomemoryinfo)) +#define LX_DXMAKERESIDENT \ + LX_IOWR(0x47, 0x0b, sizeof(struct d3dddi_makeresident)) +#define LX_DXMAPGPUVIRTUALADDRESS \ + LX_IOWR(0x47, 0x0c, sizeof(struct d3dddi_mapgpuvirtualaddress)) +#define LX_DXESCAPE \ + LX_IOWR(0x47, 0x0d, sizeof(struct d3dkmt_escape)) +#define LX_DXGETDEVICESTATE \ + LX_IOWR(0x47, 0x0e, sizeof(struct d3dkmt_getdevicestate)) +#define LX_DXSUBMITCOMMAND \ + LX_IOWR(0x47, 0x0f, sizeof(struct d3dkmt_submitcommand)) +#define LX_DXCREATESYNCHRONIZATIONOBJECT \ + LX_IOWR(0x47, 0x10, sizeof(struct d3dkmt_createsynchronizationobject2)) +#define LX_DXSIGNALSYNCHRONIZATIONOBJECT \ + LX_IOWR(0x47, 0x11, sizeof(struct d3dkmt_signalsynchronizationobject2)) +#define LX_DXWAITFORSYNCHRONIZATIONOBJECT \ + LX_IOWR(0x47, 0x12, sizeof(struct d3dkmt_waitforsynchronizationobject2)) +#define LX_DXDESTROYALLOCATION2 \ + LX_IOWR(0x47, 0x13, sizeof(struct d3dkmt_destroyallocation2)) +#define LX_DXENUMADAPTERS2 \ + LX_IOWR(0x47, 0x14, sizeof(struct d3dkmt_enumadapters2)) +#define LX_DXCLOSEADAPTER \ + LX_IOWR(0x47, 0x15, sizeof(struct d3dkmt_closeadapter)) +#define LX_DXCHANGEVIDEOMEMORYRESERVATION \ + LX_IOWR(0x47, 0x16, sizeof(struct d3dkmt_changevideomemoryreservation)) +#define LX_DXCREATEHWCONTEXT \ + LX_IOWR(0x47, 0x17, sizeof(struct d3dkmt_createhwcontext)) +#define LX_DXCREATEHWQUEUE \ + LX_IOWR(0x47, 0x18, sizeof(struct d3dkmt_createhwqueue)) +#define LX_DXDESTROYDEVICE \ + LX_IOWR(0x47, 0x19, sizeof(struct d3dkmt_destroydevice)) +#define LX_DXDESTROYHWCONTEXT \ + LX_IOWR(0x47, 0x1a, sizeof(struct d3dkmt_destroyhwcontext)) +#define LX_DXDESTROYHWQUEUE \ + LX_IOWR(0x47, 0x1b, sizeof(struct d3dkmt_destroyhwqueue)) +#define LX_DXDESTROYPAGINGQUEUE \ + LX_IOWR(0x47, 0x1c, sizeof(struct d3dddi_destroypagingqueue)) +#define LX_DXDESTROYSYNCHRONIZATIONOBJECT \ + LX_IOWR(0x47, 0x1d, sizeof(struct d3dkmt_destroysynchronizationobject)) +#define LX_DXEVICT \ + LX_IOWR(0x47, 0x1e, sizeof(struct d3dkmt_evict)) +#define LX_DXFLUSHHEAPTRANSITIONS \ + LX_IOWR(0x47, 0x1f, sizeof(struct d3dkmt_flushheaptransitions)) +#define LX_DXFREEGPUVIRTUALADDRESS \ + LX_IOWR(0x47, 0x20, sizeof(struct d3dkmt_freegpuvirtualaddress)) +#define LX_DXGETCONTEXTINPROCESSSCHEDULINGPRIORITY \ + LX_IOWR(0x47, 0x21, \ + sizeof(struct d3dkmt_getcontextinprocessschedulingpriority)) +#define LX_DXGETCONTEXTSCHEDULINGPRIORITY \ + LX_IOWR(0x47, 0x22, sizeof(struct d3dkmt_getcontextschedulingpriority)) +#define LX_DXGETSHAREDRESOURCEADAPTERLUID \ + LX_IOWR(0x47, 0x23, sizeof(struct d3dkmt_getsharedresourceadapterluid)) +#define LX_DXINVALIDATECACHE \ + LX_IOWR(0x47, 0x24, sizeof(struct d3dkmt_invalidatecache)) +#define LX_DXLOCK2 \ + LX_IOWR(0x47, 0x25, sizeof(struct d3dkmt_lock2)) +#define LX_DXMARKDEVICEASERROR \ + LX_IOWR(0x47, 0x26, sizeof(struct d3dkmt_markdeviceaserror)) +#define LX_DXOFFERALLOCATIONS \ + LX_IOWR(0x47, 0x27, sizeof(struct d3dkmt_offerallocations)) +#define LX_DXOPENRESOURCE \ + LX_IOWR(0x47, 0x28, sizeof(struct d3dkmt_openresource)) +#define LX_DXOPENSYNCHRONIZATIONOBJECT \ + LX_IOWR(0x47, 0x29, sizeof(struct d3dkmt_opensynchronizationobject)) +#define LX_DXQUERYALLOCATIONRESIDENCY \ + LX_IOWR(0x47, 0x2a, sizeof(struct d3dkmt_queryallocationresidency)) +#define LX_DXQUERYRESOURCEINFO \ + LX_IOWR(0x47, 0x2b, sizeof(struct d3dkmt_queryresourceinfo)) +#define LX_DXRECLAIMALLOCATIONS2 \ + LX_IOWR(0x47, 0x2c, sizeof(struct d3dkmt_reclaimallocations2)) +#define LX_DXRENDER \ + LX_IOWR(0x47, 0x2d, sizeof(struct d3dkmt_render)) +#define LX_DXSETALLOCATIONPRIORITY \ + LX_IOWR(0x47, 0x2e, sizeof(struct d3dkmt_setallocationpriority)) +#define LX_DXSETCONTEXTINPROCESSSCHEDULINGPRIORITY \ + LX_IOWR(0x47, 0x2f, \ + sizeof(struct d3dkmt_setcontextinprocessschedulingpriority)) +#define LX_DXSETCONTEXTSCHEDULINGPRIORITY \ + LX_IOWR(0x47, 0x30, sizeof(struct d3dkmt_setcontextschedulingpriority)) +#define LX_DXSIGNALSYNCHRONIZATIONOBJECTFROMCPU \ + LX_IOWR(0x47, 0x31, \ + sizeof(struct d3dkmt_signalsynchronizationobjectfromcpu)) +#define LX_DXSIGNALSYNCHRONIZATIONOBJECTFROMGPU \ + LX_IOWR(0x47, 0x32, \ + sizeof(struct d3dkmt_signalsynchronizationobjectfromgpu)) +#define LX_DXSIGNALSYNCHRONIZATIONOBJECTFROMGPU2 \ + LX_IOWR(0x47, 0x33, \ + sizeof(struct d3dkmt_signalsynchronizationobjectfromgpu2)) +#define LX_DXSUBMITCOMMANDTOHWQUEUE \ + LX_IOWR(0x47, 0x34, sizeof(struct d3dkmt_submitcommandtohwqueue)) +#define LX_DXSUBMITSIGNALSYNCOBJECTSTOHWQUEUE \ + LX_IOWR(0x47, 0x35, \ + sizeof(struct d3dkmt_submitsignalsyncobjectstohwqueue)) +#define LX_DXSUBMITWAITFORSYNCOBJECTSTOHWQUEUE \ + LX_IOWR(0x47, 0x36, \ + sizeof(struct d3dkmt_submitwaitforsyncobjectstohwqueue)) +#define LX_DXUNLOCK2 \ + LX_IOWR(0x47, 0x37, sizeof(struct d3dkmt_unlock2)) +#define LX_DXUPDATEALLOCPROPERTY \ + LX_IOWR(0x47, 0x38, sizeof(struct d3dddi_updateallocproperty)) +#define LX_DXUPDATEGPUVIRTUALADDRESS \ + LX_IOWR(0x47, 0x39, sizeof(struct d3dkmt_updategpuvirtualaddress)) +#define LX_DXWAITFORSYNCHRONIZATIONOBJECTFROMCPU \ + LX_IOWR(0x47, 0x3a, \ + sizeof(struct d3dkmt_waitforsynchronizationobjectfromcpu)) +#define LX_DXWAITFORSYNCHRONIZATIONOBJECTFROMGPU \ + LX_IOWR(0x47, 0x3b, \ + sizeof(struct d3dkmt_waitforsynchronizationobjectfromgpu)) +#define LX_DXGETALLOCATIONPRIORITY \ + LX_IOWR(0x47, 0x3c, sizeof(struct d3dkmt_getallocationpriority)) +#define LX_DXQUERYCLOCKCALIBRATION \ + LX_IOWR(0x47, 0x3d, sizeof(struct d3dkmt_queryclockcalibration)) +#define LX_DXENUMADAPTERS3 \ + LX_IOWR(0x47, 0x3e, sizeof(struct d3dkmt_enumadapters3)) +#define LX_DXSHAREOBJECTS \ + LX_IOWR(0x47, 0x3f, sizeof(struct d3dkmt_shareobjects)) +#define LX_DXOPENSYNCOBJECTFROMNTHANDLE2 \ + LX_IOWR(0x47, 0x40, sizeof(struct d3dkmt_opensyncobjectfromnthandle2)) +#define LX_DXQUERYRESOURCEINFOFROMNTHANDLE \ + LX_IOWR(0x47, 0x41, sizeof(struct d3dkmt_queryresourceinfofromnthandle)) +#define LX_DXOPENRESOURCEFROMNTHANDLE \ + LX_IOWR(0x47, 0x42, sizeof(struct d3dkmt_openresourcefromnthandle)) + +#define LX_IO_MAX 0x42 + +#endif /* _D3DKMTHK_H */ diff --git a/drivers/gpu/dxgkrnl/dxgadapter.c b/drivers/gpu/dxgkrnl/dxgadapter.c new file mode 100644 index 000000000000..aa50aa5335b3 --- /dev/null +++ b/drivers/gpu/dxgkrnl/dxgadapter.c @@ -0,0 +1,1399 @@ +// SPDX-License-Identifier: GPL-2.0 + +/* + * Copyright (c) 2019, Microsoft Corporation. + * + * Author: + * Iouri Tarassov iourit@microsoft.com + * + * Dxgkrnl Graphics Port Driver + * dxgadapter implementation + * + */ + +#include <linux/module.h> +#include <linux/hyperv.h> +#include <linux/pagemap.h> +#include <linux/eventfd.h> + +#include "dxgkrnl.h" + +int dxgadapter_init(struct dxgadapter *adapter, struct hv_device *hdev) +{ + int ret = 0; + char s[80]; + + UNUSED(s); + guid_to_luid(&hdev->channel->offermsg.offer.if_instance, + &adapter->luid); + TRACE_DEBUG(1, "%s: %x:%x %p %pUb\n", + __func__, adapter->luid.b, adapter->luid.a, hdev->channel, + &hdev->channel->offermsg.offer.if_instance); + + adapter->adapter_state = DXGADAPTER_STATE_STOPPED; + refcount_set(&adapter->refcount, 1); + init_rwsem(&adapter->core_lock); + + INIT_LIST_HEAD(&adapter->adapter_process_list_head); + INIT_LIST_HEAD(&adapter->shared_resource_list_head); + INIT_LIST_HEAD(&adapter->adapter_shared_syncobj_list_head); + INIT_LIST_HEAD(&adapter->syncobj_list_head); + init_rwsem(&adapter->shared_resource_list_lock); + + ret = dxgvmbuschannel_init(&adapter->channel, hdev); + if (ret) + goto cleanup; + + adapter->channel.adapter = adapter; + + ret = dxgvmb_send_open_adapter(adapter); + if (ret) { + pr_err("dxgvmb_send_open_adapter failed: %d\n", ret); + goto cleanup; + } + + adapter->adapter_state = DXGADAPTER_STATE_ACTIVE; + + ret = dxgvmb_send_get_internal_adapter_info(adapter); + if (ret) { + pr_err("get_internal_adapter_info failed: %d\n", ret); + goto cleanup; + } + +cleanup: + + return ret; +} + +void dxgadapter_stop(struct dxgadapter *adapter) +{ + struct dxgprocess_adapter *entry; + + dxgglobal_acquire_process_adapter_lock(); + + list_for_each_entry(entry, &adapter->adapter_process_list_head, + adapter_process_list_entry) { + dxgprocess_adapter_stop(entry); + } + + dxgglobal_release_process_adapter_lock(); + + if (dxgadapter_acquire_lock_exclusive(adapter) == 0) { + dxgvmb_send_close_adapter(adapter); + dxgadapter_release_lock_exclusive(adapter); + } + dxgvmbuschannel_destroy(&adapter->channel); +} + +void dxgadapter_destroy(struct dxgadapter *adapter) +{ + TRACE_DEBUG(1, "%s %p\n", __func__, adapter); + dxgmem_free(NULL, DXGMEM_ADAPTER, adapter); +} + +bool dxgadapter_acquire_reference(struct dxgadapter *adapter) +{ + return refcount_inc_not_zero(&adapter->refcount); +} + +void dxgadapter_release_reference(struct dxgadapter *adapter) +{ + if (refcount_dec_and_test(&adapter->refcount)) + dxgadapter_destroy(adapter); +} + +bool dxgadapter_is_active(struct dxgadapter *adapter) +{ + return adapter->adapter_state == DXGADAPTER_STATE_ACTIVE; +} + +/* Protected by dxgglobal_acquire_process_adapter_lock */ +void dxgadapter_add_process(struct dxgadapter *adapter, + struct dxgprocess_adapter *process_info) +{ + TRACE_DEBUG(1, "%s %p %p", __func__, adapter, process_info); + list_add_tail(&process_info->adapter_process_list_entry, + &adapter->adapter_process_list_head); +} + +void dxgadapter_remove_process(struct dxgprocess_adapter *process_info) +{ + TRACE_DEBUG(1, "%s %p %p", __func__, + process_info->adapter, process_info); + list_del(&process_info->adapter_process_list_entry); + process_info->adapter_process_list_entry.next = NULL; + process_info->adapter_process_list_entry.prev = NULL; +} + +void dxgadapter_add_shared_resource(struct dxgadapter *adapter, + struct dxgsharedresource *object) +{ + /* Adapter lock is held */ + down_write(&adapter->shared_resource_list_lock); + list_add_tail(&object->shared_resource_list_entry, + &adapter->shared_resource_list_head); + up_write(&adapter->shared_resource_list_lock); +} + +void dxgadapter_remove_shared_resource(struct dxgadapter *adapter, + struct dxgsharedresource *object) +{ + down_write(&adapter->shared_resource_list_lock); + if (object->shared_resource_list_entry.next) { + list_del(&object->shared_resource_list_entry); + object->shared_resource_list_entry.next = NULL; + } + up_write(&adapter->shared_resource_list_lock); +} + +void dxgadapter_add_shared_syncobj(struct dxgadapter *adapter, + struct dxgsharedsyncobject *object) +{ + down_write(&adapter->shared_resource_list_lock); + list_add_tail(&object->adapter_shared_syncobj_list_entry, + &adapter->adapter_shared_syncobj_list_head); + up_write(&adapter->shared_resource_list_lock); +} + +void dxgadapter_remove_shared_syncobj(struct dxgadapter *adapter, + struct dxgsharedsyncobject *object) +{ + down_write(&adapter->shared_resource_list_lock); + if (object->adapter_shared_syncobj_list_entry.next) { + list_del(&object->adapter_shared_syncobj_list_entry); + object->adapter_shared_syncobj_list_entry.next = NULL; + } + up_write(&adapter->shared_resource_list_lock); +} + +void dxgadapter_add_syncobj(struct dxgadapter *adapter, + struct dxgsyncobject *object) +{ + down_write(&adapter->shared_resource_list_lock); + list_add_tail(&object->syncobj_list_entry, &adapter->syncobj_list_head); + up_write(&adapter->shared_resource_list_lock); +} + +void dxgadapter_remove_syncobj(struct dxgsyncobject *object) +{ + down_write(&object->adapter->shared_resource_list_lock); + if (object->syncobj_list_entry.next) { + list_del(&object->syncobj_list_entry); + object->syncobj_list_entry.next = NULL; + } + up_write(&object->adapter->shared_resource_list_lock); +} + +int dxgadapter_acquire_lock_exclusive(struct dxgadapter *adapter) +{ + TRACE_DEBUG(1, "%s", __func__); + dxglockorder_acquire(DXGLOCK_ADAPTER); + down_write(&adapter->core_lock); + if (adapter->adapter_state != DXGADAPTER_STATE_ACTIVE) { + dxgadapter_release_lock_exclusive(adapter); + return STATUS_DEVICE_REMOVED; + } + return 0; +} + +void dxgadapter_acquire_lock_forced(struct dxgadapter *adapter) +{ + dxglockorder_acquire(DXGLOCK_ADAPTER); + down_write(&adapter->core_lock); +} + +void dxgadapter_release_lock_exclusive(struct dxgadapter *adapter) +{ + TRACE_DEBUG(1, "%s", __func__); + up_write(&adapter->core_lock); + dxglockorder_release(DXGLOCK_ADAPTER); +} + +int dxgadapter_acquire_lock_shared(struct dxgadapter *adapter) +{ + TRACE_DEBUG(1, "%s", __func__); + dxglockorder_acquire(DXGLOCK_ADAPTER); + down_read(&adapter->core_lock); + if (adapter->adapter_state == DXGADAPTER_STATE_ACTIVE) + return 0; + dxgadapter_release_lock_shared(adapter); + return STATUS_DEVICE_REMOVED; +} + +void dxgadapter_release_lock_shared(struct dxgadapter *adapter) +{ + TRACE_DEBUG(1, "dxgadapter_release_lock\n"); + up_read(&adapter->core_lock); + dxglockorder_release(DXGLOCK_ADAPTER); +} + +struct dxgdevice *dxgdevice_create(struct dxgadapter *adapter, + struct dxgprocess *process) +{ + struct dxgdevice *device = dxgmem_alloc(process, DXGMEM_DEVICE, + sizeof(struct dxgdevice)); + if (device) { + refcount_set(&device->refcount, 1); + device->adapter = adapter; + device->process = process; + dxgadapter_acquire_reference(adapter); + INIT_LIST_HEAD(&device->context_list_head); + INIT_LIST_HEAD(&device->alloc_list_head); + INIT_LIST_HEAD(&device->resource_list_head); + init_rwsem(&device->device_lock); + init_rwsem(&device->context_list_lock); + init_rwsem(&device->alloc_list_lock); + INIT_LIST_HEAD(&device->pqueue_list_head); + INIT_LIST_HEAD(&device->syncobj_list_head); + device->object_state = DXGOBJECTSTATE_CREATED; + + dxgprocess_adapter_add_device(process, adapter, device); + } + return device; +} + +void dxgdevice_stop(struct dxgdevice *device) +{ + struct dxgallocation *alloc; + struct dxgpagingqueue *pqueue; + struct dxgsyncobject *syncobj; + + TRACE_DEBUG(1, "%s: DXGKDEBUG %p", __func__, device); + dxgdevice_acquire_alloc_list_lock(device); + list_for_each_entry(alloc, &device->alloc_list_head, alloc_list_entry) { + dxgallocation_stop(alloc); + } + dxgdevice_release_alloc_list_lock(device); + + hmgrtable_lock(&device->process->handle_table, DXGLOCK_EXCL); + list_for_each_entry(pqueue, &device->pqueue_list_head, + pqueue_list_entry) { + dxgpagingqueue_stop(pqueue); + } + list_for_each_entry(syncobj, &device->syncobj_list_head, + syncobj_list_entry) { + dxgsyncobject_stop(syncobj); + } + hmgrtable_unlock(&device->process->handle_table, DXGLOCK_EXCL); + TRACE_DEBUG(1, "%s: end %p\n", __func__, device); +} + +void dxgdevice_mark_destroyed(struct dxgdevice *device) +{ + down_write(&device->device_lock); + device->object_state = DXGOBJECTSTATE_DESTROYED; + up_write(&device->device_lock); +} + +void dxgdevice_destroy(struct dxgdevice *device) +{ + struct dxgprocess *process = device->process; + struct dxgadapter *adapter = device->adapter; + d3dkmt_handle device_handle = 0; + + TRACE_DEBUG(1, "%s: %p\n", __func__, device); + + down_write(&device->device_lock); + + if (device->object_state != DXGOBJECTSTATE_ACTIVE) + goto cleanup; + + device->object_state = DXGOBJECTSTATE_DESTROYED; + + dxgdevice_stop(device); + + dxgdevice_acquire_alloc_list_lock(device); + + while (!list_empty(&device->syncobj_list_head)) { + struct dxgsyncobject *syncobj = + list_first_entry(&device->syncobj_list_head, + struct dxgsyncobject, + syncobj_list_entry); + list_del(&syncobj->syncobj_list_entry); + syncobj->syncobj_list_entry.next = NULL; + dxgdevice_release_alloc_list_lock(device); + + dxgsyncobject_destroy(process, syncobj); + + dxgdevice_acquire_alloc_list_lock(device); + } + + { + struct dxgallocation *alloc; + struct dxgallocation *tmp; + + TRACE_DEBUG(1, "destroying allocations\n"); + list_for_each_entry_safe(alloc, tmp, &device->alloc_list_head, + alloc_list_entry) { + dxgallocation_destroy(alloc); + } + } + + { + struct dxgresource *resource; + struct dxgresource *tmp; + + TRACE_DEBUG(1, "destroying resources\n"); + list_for_each_entry_safe(resource, tmp, + &device->resource_list_head, + resource_list_entry) { + dxgresource_destroy(resource); + } + } + + dxgdevice_release_alloc_list_lock(device); + + { + struct dxgcontext *context; + struct dxgcontext *tmp; + + TRACE_DEBUG(1, "destroying contexts\n"); + dxgdevice_acquire_context_list_lock(device); + list_for_each_entry_safe(context, tmp, + &device->context_list_head, + context_list_entry) { + dxgcontext_destroy(process, context); + } + dxgdevice_release_context_list_lock(device); + } + + { + struct dxgpagingqueue *tmp; + struct dxgpagingqueue *pqueue; + + TRACE_DEBUG(1, "destroying paging queues\n"); + list_for_each_entry_safe(pqueue, tmp, &device->pqueue_list_head, + pqueue_list_entry) { + dxgpagingqueue_destroy(pqueue); + } + } + + /* Guest handles need to be released before the host handles */ + hmgrtable_lock(&process->handle_table, DXGLOCK_EXCL); + if (device->handle_valid) { + hmgrtable_free_handle(&process->handle_table, + HMGRENTRY_TYPE_DXGDEVICE, device->handle); + device_handle = device->handle; + device->handle_valid = 0; + } + hmgrtable_unlock(&process->handle_table, DXGLOCK_EXCL); + + if (device_handle) { + up_write(&device->device_lock); + if (dxgadapter_acquire_lock_shared(adapter) == 0) { + dxgvmb_send_destroy_device(adapter, process, + device_handle); + dxgadapter_release_lock_shared(adapter); + } + down_write(&device->device_lock); + } + +cleanup: + + if (device->adapter) { + dxgprocess_adapter_remove_device(device); + dxgadapter_release_reference(device->adapter); + } + + up_write(&device->device_lock); + + dxgdevice_release_reference(device); + TRACE_DEBUG(1, "dxgdevice_destroy_end\n"); +} + +int dxgdevice_acquire_lock_shared(struct dxgdevice *device) +{ + down_read(&device->device_lock); + if (!dxgdevice_is_active(device)) { + up_read(&device->device_lock); + return STATUS_DEVICE_REMOVED; + } + return 0; +} + +void dxgdevice_release_lock_shared(struct dxgdevice *device) +{ + up_read(&device->device_lock); +} + +bool dxgdevice_is_active(struct dxgdevice *device) +{ + return device->object_state == DXGOBJECTSTATE_ACTIVE; +} + +void dxgdevice_acquire_context_list_lock(struct dxgdevice *device) +{ + dxglockorder_acquire(DXGLOCK_DEVICE_CONTEXTLIST); + down_write(&device->context_list_lock); +} + +void dxgdevice_release_context_list_lock(struct dxgdevice *device) +{ + up_write(&device->context_list_lock); + dxglockorder_release(DXGLOCK_DEVICE_CONTEXTLIST); +} + +void dxgdevice_acquire_alloc_list_lock(struct dxgdevice *device) +{ + dxglockorder_acquire(DXGLOCK_DEVICE_ALLOCLIST); + down_write(&device->alloc_list_lock); +} + +void dxgdevice_release_alloc_list_lock(struct dxgdevice *device) +{ + dxglockorder_release(DXGLOCK_DEVICE_ALLOCLIST); + up_write(&device->alloc_list_lock); +} + +void dxgdevice_acquire_alloc_list_lock_shared(struct dxgdevice *device) +{ + dxglockorder_acquire(DXGLOCK_DEVICE_ALLOCLIST); + down_read(&device->alloc_list_lock); +} + +void dxgdevice_release_alloc_list_lock_shared(struct dxgdevice *device) +{ + dxglockorder_release(DXGLOCK_DEVICE_ALLOCLIST); + up_read(&device->alloc_list_lock); +} + +void dxgdevice_add_context(struct dxgdevice *device, struct dxgcontext *context) +{ + down_write(&device->context_list_lock); + list_add_tail(&context->context_list_entry, &device->context_list_head); + up_write(&device->context_list_lock); +} + +void dxgdevice_remove_context(struct dxgdevice *device, + struct dxgcontext *context) +{ + if (context->context_list_entry.next) { + list_del(&context->context_list_entry); + context->context_list_entry.next = NULL; + } +} + +void dxgdevice_add_alloc(struct dxgdevice *device, struct dxgallocation *alloc) +{ + dxgdevice_acquire_alloc_list_lock(device); + list_add_tail(&alloc->alloc_list_entry, &device->alloc_list_head); + dxgdevice_acquire_reference(device); + alloc->owner.device = device; + dxgdevice_release_alloc_list_lock(device); +} + +void dxgdevice_remove_alloc(struct dxgdevice *device, + struct dxgallocation *alloc) +{ + if (alloc->alloc_list_entry.next) { + list_del(&alloc->alloc_list_entry); + alloc->alloc_list_entry.next = NULL; + dxgdevice_release_reference(device); + } +} + +void dxgdevice_remove_alloc_safe(struct dxgdevice *device, + struct dxgallocation *alloc) +{ + dxgdevice_acquire_alloc_list_lock(device); + dxgdevice_remove_alloc(device, alloc); + dxgdevice_release_alloc_list_lock(device); +} + +void dxgdevice_add_resource(struct dxgdevice *device, struct dxgresource *res) +{ + dxgdevice_acquire_alloc_list_lock(device); + list_add_tail(&res->resource_list_entry, &device->resource_list_head); + dxgdevice_acquire_reference(device); + dxgdevice_release_alloc_list_lock(device); +} + +void dxgdevice_remove_resource(struct dxgdevice *device, + struct dxgresource *res) +{ + if (res->resource_list_entry.next) { + list_del(&res->resource_list_entry); + res->resource_list_entry.next = NULL; + dxgdevice_release_reference(device); + } +} + +struct dxgsharedresource *dxgsharedresource_create(struct dxgadapter *adapter) +{ + struct dxgsharedresource *resource = dxgmem_alloc(NULL, + DXGMEM_SHAREDRESOURCE, + sizeof(struct + dxgsharedresource)); + if (resource) { + INIT_LIST_HEAD(&resource->resource_list_head); + refcount_set(&resource->refcount, 1); + dxgmutex_init(&resource->fd_mutex, DXGLOCK_FDMUTEX); + resource->adapter = adapter; + } + return resource; +} + +bool dxgsharedresource_acquire_reference(struct dxgsharedresource *resource) +{ + return refcount_inc_not_zero(&resource->refcount); +} + +void dxgsharedresource_release_reference(struct dxgsharedresource *resource) +{ + if (!refcount_dec_and_test(&resource->refcount)) + return; + if (resource->global_handle) + hmgrtable_free_handle_safe(&dxgglobal->handle_table, + HMGRENTRY_TYPE_DXGSHAREDRESOURCE, + resource->global_handle); + if (resource->runtime_private_data) + dxgmem_free(NULL, DXGMEM_RUNTIMEPRIVATE, + resource->runtime_private_data); + if (resource->resource_private_data) + dxgmem_free(NULL, DXGMEM_RESOURCEPRIVATE, + resource->resource_private_data); + if (resource->alloc_private_data_sizes) + dxgmem_free(NULL, DXGMEM_ALLOCPRIVATE, + resource->alloc_private_data_sizes); + if (resource->alloc_private_data) + dxgmem_free(NULL, DXGMEM_ALLOCPRIVATE, + resource->alloc_private_data); + dxgmem_free(NULL, DXGMEM_SHAREDRESOURCE, resource); +} + +void dxgsharedresource_add_resource(struct dxgsharedresource *shared_resource, + struct dxgresource *resource) +{ + down_write(&shared_resource->adapter->shared_resource_list_lock); + TRACE_DEBUG(1, "%s: %p %p", __func__, shared_resource, resource); + list_add_tail(&resource->shared_resource_list_entry, + &shared_resource->resource_list_head); + dxgsharedresource_acquire_reference(shared_resource); + dxgresource_acquire_reference(resource); + resource->shared_owner = shared_resource; + up_write(&shared_resource->adapter->shared_resource_list_lock); +} + +void dxgsharedresource_remove_resource(struct dxgsharedresource + *shared_resource, + struct dxgresource *resource) +{ + down_write(&shared_resource->adapter->shared_resource_list_lock); + TRACE_DEBUG(1, "%s: %p %p", __func__, shared_resource, resource); + if (resource->shared_resource_list_entry.next) { + list_del(&resource->shared_resource_list_entry); + resource->shared_resource_list_entry.next = NULL; + dxgsharedresource_release_reference(shared_resource); + resource->shared_owner = NULL; + dxgresource_release_reference(resource); + } + up_write(&shared_resource->adapter->shared_resource_list_lock); +} + +struct dxgresource *dxgresource_create(struct dxgdevice *device) +{ + struct dxgresource *resource = dxgmem_alloc(device->process, + DXGMEM_RESOURCE, + sizeof(struct dxgresource)); + if (resource) { + refcount_set(&resource->refcount, 1); + resource->device = device; + resource->process = device->process; + resource->object_state = DXGOBJECTSTATE_ACTIVE; + dxgmutex_init(&resource->resource_mutex, DXGLOCK_RESOURCE); + INIT_LIST_HEAD(&resource->alloc_list_head); + dxgdevice_add_resource(device, resource); + } + return resource; +} + +void dxgresource_free_handle(struct dxgresource *resource) +{ + struct dxgallocation *alloc; + + if (resource->handle_valid) { + hmgrtable_free_handle_safe(&resource->device->process-> + handle_table, + HMGRENTRY_TYPE_DXGRESOURCE, + resource->handle); + resource->handle_valid = 0; + } + list_for_each_entry(alloc, &resource->alloc_list_head, alloc_list_entry) { + dxgallocation_free_handle(alloc); + } +} + +void dxgresource_destroy(struct dxgresource *resource) +{ + /* device->alloc_list_lock is held */ + struct dxgallocation *alloc; + struct dxgallocation *tmp; + struct d3dkmt_destroyallocation2 args = { }; + int destroyed = test_and_set_bit(0, &resource->flags); + struct dxgdevice *device = resource->device; + + if (!destroyed) { + dxgresource_free_handle(resource); + if (resource->handle) { + args.device = device->handle; + args.resource = resource->handle; + args.flags.assume_not_in_use = 1; + dxgvmb_send_destroy_allocation(device->process, + device, + &device->adapter-> + channel, &args, NULL); + resource->handle = 0; + } + list_for_each_entry_safe(alloc, tmp, &resource->alloc_list_head, + alloc_list_entry) { + dxgallocation_destroy(alloc); + } + dxgdevice_remove_resource(device, resource); + if (resource->shared_owner) { + dxgsharedresource_remove_resource(resource-> + shared_owner, + resource); + resource->shared_owner = NULL; + } + } + dxgresource_release_reference(resource); +} + +void dxgresource_acquire_reference(struct dxgresource *resource) +{ + refcount_inc_not_zero(&resource->refcount); +} + +void dxgresource_release_reference(struct dxgresource *resource) +{ + if (refcount_dec_and_test(&resource->refcount)) + dxgmem_free(resource->process, DXGMEM_RESOURCE, resource); +} + +bool dxgresource_is_active(struct dxgresource *resource) +{ + return resource->object_state == DXGOBJECTSTATE_ACTIVE; +} + +int dxgresource_add_alloc(struct dxgresource *resource, + struct dxgallocation *alloc) +{ + int ret = STATUS_DEVICE_REMOVED; + struct dxgdevice *device = resource->device; + + dxgdevice_acquire_alloc_list_lock(device); + if (dxgresource_is_active(resource)) { + list_add_tail(&alloc->alloc_list_entry, + &resource->alloc_list_head); + alloc->owner.resource = resource; + ret = 0; + } + alloc->resource_owner = 1; + dxgdevice_release_alloc_list_lock(device); + return ret; +} + +void dxgresource_remove_alloc(struct dxgresource *resource, + struct dxgallocation *alloc) +{ + if (alloc->alloc_list_entry.next) { + list_del(&alloc->alloc_list_entry); + alloc->alloc_list_entry.next = NULL; + } +} + +void dxgresource_remove_alloc_safe(struct dxgresource *resource, + struct dxgallocation *alloc) +{ + dxgdevice_acquire_alloc_list_lock(resource->device); + dxgresource_remove_alloc(resource, alloc); + dxgdevice_release_alloc_list_lock(resource->device); +} + +bool dxgdevice_acquire_reference(struct dxgdevice *device) +{ + return refcount_inc_not_zero(&device->refcount); +} + +void dxgdevice_release_reference(struct dxgdevice *device) +{ + if (refcount_dec_and_test(&device->refcount)) + dxgmem_free(device->process, DXGMEM_DEVICE, device); +} + +void dxgdevice_add_paging_queue(struct dxgdevice *device, + struct dxgpagingqueue *entry) +{ + dxgdevice_acquire_alloc_list_lock(device); + list_add_tail(&entry->pqueue_list_entry, &device->pqueue_list_head); + dxgdevice_release_alloc_list_lock(device); +} + +void dxgdevice_remove_paging_queue(struct dxgpagingqueue *pqueue) +{ + struct dxgdevice *device = pqueue->device; + + dxgdevice_acquire_alloc_list_lock(device); + if (pqueue->pqueue_list_entry.next) { + list_del(&pqueue->pqueue_list_entry); + pqueue->pqueue_list_entry.next = NULL; + } + dxgdevice_release_alloc_list_lock(device); +} + +void dxgdevice_add_syncobj(struct dxgdevice *device, + struct dxgsyncobject *syncobj) +{ + dxgdevice_acquire_alloc_list_lock(device); + list_add_tail(&syncobj->syncobj_list_entry, &device->syncobj_list_head); + dxgsyncobject_acquire_reference(syncobj); + dxgdevice_release_alloc_list_lock(device); +} + +void dxgdevice_remove_syncobj(struct dxgsyncobject *entry) +{ + struct dxgdevice *device = entry->device; + + dxgdevice_acquire_alloc_list_lock(device); + if (entry->syncobj_list_entry.next) { + list_del(&entry->syncobj_list_entry); + entry->syncobj_list_entry.next = NULL; + dxgsyncobject_release_reference(entry); + } + dxgdevice_release_alloc_list_lock(device); + dxgdevice_release_reference(device); + entry->device = NULL; +} + +struct dxgcontext *dxgcontext_create(struct dxgdevice *device) +{ + struct dxgcontext *context = dxgmem_alloc(device->process, + DXGMEM_CONTEXT, + sizeof(struct dxgcontext)); + if (context) { + refcount_set(&context->refcount, 1); + context->device = device; + context->process = device->process; + context->device_handle = device->handle; + dxgdevice_acquire_reference(device); + INIT_LIST_HEAD(&context->hwqueue_list_head); + init_rwsem(&context->hwqueue_list_lock); + dxgdevice_add_context(device, context); + context->object_state = DXGOBJECTSTATE_ACTIVE; + } + return context; +} + +/* + * Called when the device context list lock is held + */ +void dxgcontext_destroy(struct dxgprocess *process, struct dxgcontext *context) +{ + struct dxghwqueue *hwqueue; + struct dxghwqueue *tmp; + + TRACE_DEBUG(1, "%s %p\n", __func__, context); + context->object_state = DXGOBJECTSTATE_DESTROYED; + if (context->device) { + if (context->handle) { + hmgrtable_free_handle_safe(&context->process-> + handle_table, + HMGRENTRY_TYPE_DXGCONTEXT, + context->handle); + } + dxgdevice_remove_context(context->device, context); + dxgdevice_release_reference(context->device); + } + list_for_each_entry_safe(hwqueue, tmp, &context->hwqueue_list_head, + hwqueue_list_entry) { + dxghwqueue_destroy(process, hwqueue); + } + dxgcontext_release_reference(context); +} + +void dxgcontext_destroy_safe(struct dxgprocess *process, + struct dxgcontext *context) +{ + struct dxgdevice *device = context->device; + + dxgdevice_acquire_context_list_lock(device); + dxgcontext_destroy(process, context); + dxgdevice_release_context_list_lock(device); +} + +bool dxgcontext_is_active(struct dxgcontext *context) +{ + return context->object_state == DXGOBJECTSTATE_ACTIVE; +} + +bool dxgcontext_acquire_reference(struct dxgcontext *context) +{ + return refcount_inc_not_zero(&context->refcount); +} + +void dxgcontext_release_reference(struct dxgcontext *context) +{ + if (refcount_dec_and_test(&context->refcount)) + dxgmem_free(context->process, DXGMEM_CONTEXT, context); +} + +int dxgcontext_add_hwqueue(struct dxgcontext *context, + struct dxghwqueue *hwqueue) +{ + int ret = 0; + + down_write(&context->hwqueue_list_lock); + if (dxgcontext_is_active(context)) + list_add_tail(&hwqueue->hwqueue_list_entry, + &context->hwqueue_list_head); + else + ret = STATUS_DEVICE_REMOVED; + up_write(&context->hwqueue_list_lock); + return ret; +} + +void dxgcontext_remove_hwqueue(struct dxgcontext *context, + struct dxghwqueue *hwqueue) +{ + if (hwqueue->hwqueue_list_entry.next) { + list_del(&hwqueue->hwqueue_list_entry); + hwqueue->hwqueue_list_entry.next = NULL; + } +} + +void dxgcontext_remove_hwqueue_safe(struct dxgcontext *context, + struct dxghwqueue *hwqueue) +{ + down_write(&context->hwqueue_list_lock); + dxgcontext_remove_hwqueue(context, hwqueue); + up_write(&context->hwqueue_list_lock); +} + +struct dxgallocation *dxgallocation_create(struct dxgprocess *process) +{ + struct dxgallocation *alloc = dxgmem_alloc(process, DXGMEM_ALLOCATION, + sizeof(struct + dxgallocation)); + if (alloc) + alloc->process = process; + return alloc; +} + +void dxgallocation_stop(struct dxgallocation *alloc) +{ + if (alloc->pages) { + release_pages(alloc->pages, alloc->num_pages); + dxgmem_free(alloc->process, DXGMEM_ALLOCATION, alloc->pages); + alloc->pages = NULL; + } + dxgprocess_ht_lock_exclusive_down(alloc->process); + if (alloc->cpu_address_mapped) { + dxg_unmap_iospace(alloc->cpu_address, + alloc->num_pages << PAGE_SHIFT); + alloc->cpu_address_mapped = false; + alloc->cpu_address = NULL; + } + dxgprocess_ht_lock_exclusive_up(alloc->process); +} + +void dxgallocation_free_handle(struct dxgallocation *alloc) +{ + dxgprocess_ht_lock_exclusive_down(alloc->process); + if (alloc->handle_valid) { + hmgrtable_free_handle(&alloc->process->handle_table, + HMGRENTRY_TYPE_DXGALLOCATION, + alloc->alloc_handle); + alloc->handle_valid = 0; + } + dxgprocess_ht_lock_exclusive_up(alloc->process); +} + +void dxgallocation_destroy(struct dxgallocation *alloc) +{ + struct dxgprocess *process = alloc->process; + struct d3dkmt_destroyallocation2 args = { }; + + dxgallocation_stop(alloc); + if (alloc->resource_owner) + dxgresource_remove_alloc(alloc->owner.resource, alloc); + else if (alloc->owner.device) + dxgdevice_remove_alloc(alloc->owner.device, alloc); + dxgallocation_free_handle(alloc); + if (alloc->alloc_handle && !alloc->resource_owner) { + args.device = alloc->owner.device->handle; + args.alloc_count = 1; + args.flags.assume_not_in_use = 1; + dxgvmb_send_destroy_allocation(process, + alloc->owner.device, + &alloc->owner.device->adapter-> + channel, &args, + &alloc->alloc_handle); + } + if (alloc->gpadl) { + TRACE_DEBUG(1, "Teardown gpadl %d", alloc->gpadl); + vmbus_teardown_gpadl(dxgglobal_get_vmbus(), alloc->gpadl); + TRACE_DEBUG(1, "Teardown gpadl end"); + alloc->gpadl = 0; + } + if (alloc->priv_drv_data) + dxgmem_free(alloc->process, DXGMEM_ALLOCPRIVATE, + alloc->priv_drv_data); + if (alloc->cpu_address_mapped) + pr_err("Alloc IO space is mapped: %p", alloc); + dxgmem_free(alloc->process, DXGMEM_ALLOCATION, alloc); +} + +struct dxgpagingqueue *dxgpagingqueue_create(struct dxgdevice *device) +{ + struct dxgpagingqueue *pqueue; + + pqueue = dxgmem_alloc(device->process, DXGMEM_PQUEUE, sizeof(*pqueue)); + if (pqueue) { + pqueue->device = device; + pqueue->process = device->process; + pqueue->device_handle = device->handle; + dxgdevice_add_paging_queue(device, pqueue); + } + return pqueue; +} + +void dxgpagingqueue_stop(struct dxgpagingqueue *pqueue) +{ + if (pqueue->mapped_address) { + int ret = dxg_unmap_iospace(pqueue->mapped_address, PAGE_SIZE); + + UNUSED(ret); + TRACE_DEBUG(1, "fence is unmapped %d %p", + ret, pqueue->mapped_address); + pqueue->mapped_address = NULL; + } +} + +void dxgpagingqueue_destroy(struct dxgpagingqueue *pqueue) +{ + struct dxgprocess *process = pqueue->process; + + TRACE_DEBUG(1, "%s %p %x\n", __func__, pqueue, pqueue->handle); + + dxgpagingqueue_stop(pqueue); + + hmgrtable_lock(&process->handle_table, DXGLOCK_EXCL); + if (pqueue->handle) { + hmgrtable_free_handle(&process->handle_table, + HMGRENTRY_TYPE_DXGPAGINGQUEUE, + pqueue->handle); + pqueue->handle = 0; + } + if (pqueue->syncobj_handle) { + hmgrtable_free_handle(&process->handle_table, + HMGRENTRY_TYPE_MONITOREDFENCE, + pqueue->syncobj_handle); + pqueue->syncobj_handle = 0; + } + hmgrtable_unlock(&process->handle_table, DXGLOCK_EXCL); + if (pqueue->device) + dxgdevice_remove_paging_queue(pqueue); + dxgmem_free(process, DXGMEM_PQUEUE, pqueue); +} + +struct dxgprocess_adapter *dxgprocess_adapter_create(struct dxgprocess *process, + struct dxgadapter *adapter) +{ + struct dxgprocess_adapter *adapter_info = dxgmem_alloc(process, + DXGMEM_PROCESS_ADAPTER, + sizeof + (*adapter_info)); + if (adapter_info) { + if (!dxgadapter_acquire_reference(adapter)) { + pr_err("failed to acquire adapter reference"); + goto cleanup; + } + adapter_info->adapter = adapter; + adapter_info->process = process; + adapter_info->refcount = 1; + dxgmutex_init(&adapter_info->device_list_mutex, + DXGLOCK_PROCESSADAPTERDEVICELIST); + INIT_LIST_HEAD(&adapter_info->device_list_head); + list_add_tail(&adapter_info->process_adapter_list_entry, + &process->process_adapter_list_head); + dxgadapter_add_process(adapter, adapter_info); + } + return adapter_info; +cleanup: + if (adapter_info) + dxgmem_free(process, DXGMEM_PROCESS_ADAPTER, adapter_info); + return NULL; +} + +void dxgprocess_adapter_stop(struct dxgprocess_adapter *adapter_info) +{ + struct dxgdevice *device; + + dxgmutex_lock(&adapter_info->device_list_mutex); + list_for_each_entry(device, &adapter_info->device_list_head, + device_list_entry) { + dxgdevice_stop(device); + } + dxgmutex_unlock(&adapter_info->device_list_mutex); +} + +void dxgprocess_adapter_destroy(struct dxgprocess_adapter *adapter_info) +{ + struct dxgdevice *device; + + dxgmutex_lock(&adapter_info->device_list_mutex); + while (!list_empty(&adapter_info->device_list_head)) { + device = list_first_entry(&adapter_info->device_list_head, + struct dxgdevice, device_list_entry); + list_del(&device->device_list_entry); + device->device_list_entry.next = NULL; + dxgmutex_unlock(&adapter_info->device_list_mutex); + dxgdevice_destroy(device); + dxgmutex_lock(&adapter_info->device_list_mutex); + } + dxgmutex_unlock(&adapter_info->device_list_mutex); + + dxgadapter_remove_process(adapter_info); + dxgadapter_release_reference(adapter_info->adapter); + list_del(&adapter_info->process_adapter_list_entry); + dxgmem_free(adapter_info->process, DXGMEM_PROCESS_ADAPTER, + adapter_info); +} + +/* + * Must be called when dxgglobal::process_adapter_mutex is held + */ +void dxgprocess_adapter_release(struct dxgprocess_adapter *adapter_info) +{ + TRACE_DEBUG(1, "%s %p %d", + __func__, adapter_info, adapter_info->refcount); + adapter_info->refcount--; + if (adapter_info->refcount == 0) + dxgprocess_adapter_destroy(adapter_info); +} + +int dxgprocess_adapter_add_device(struct dxgprocess *process, + struct dxgadapter *adapter, + struct dxgdevice *device) +{ + struct dxgprocess_adapter *entry; + struct dxgprocess_adapter *adapter_info = NULL; + int ret = 0; + + dxgglobal_acquire_process_adapter_lock(); + + list_for_each_entry(entry, &process->process_adapter_list_head, + process_adapter_list_entry) { + if (entry->adapter == adapter) { + adapter_info = entry; + break; + } + } + if (adapter_info == NULL) { + pr_err("failed to find process adapter info\n"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + dxgmutex_lock(&adapter_info->device_list_mutex); + list_add_tail(&device->device_list_entry, + &adapter_info->device_list_head); + device->adapter_info = adapter_info; + dxgmutex_unlock(&adapter_info->device_list_mutex); + +cleanup: + + dxgglobal_release_process_adapter_lock(); + return ret; +} + +void dxgprocess_adapter_remove_device(struct dxgdevice *device) +{ + TRACE_DEBUG(1, "%s %p\n", __func__, device); + dxgmutex_lock(&device->adapter_info->device_list_mutex); + if (device->device_list_entry.next) { + list_del(&device->device_list_entry); + device->device_list_entry.next = NULL; + } + dxgmutex_unlock(&device->adapter_info->device_list_mutex); +} + +struct dxgsharedsyncobject *dxgsharedsyncobj_create(struct dxgadapter *adapter, + struct dxgsyncobject *so) +{ + struct dxgsharedsyncobject *syncobj; + + syncobj = dxgmem_alloc(NULL, DXGMEM_SHAREDSYNCOBJ, sizeof(*syncobj)); + if (syncobj) { + refcount_set(&syncobj->refcount, 1); + INIT_LIST_HEAD(&syncobj->shared_syncobj_list_head); + syncobj->adapter = adapter; + syncobj->type = so->type; + syncobj->monitored_fence = so->monitored_fence; + dxgadapter_add_shared_syncobj(adapter, syncobj); + dxgadapter_acquire_reference(adapter); + init_rwsem(&syncobj->syncobj_list_lock); + dxgmutex_init(&syncobj->fd_mutex, DXGLOCK_FDMUTEX); + } + return syncobj; +} + +bool dxgsharedsyncobj_acquire_reference(struct dxgsharedsyncobject *syncobj) +{ + TRACE_DEBUG(1, "%s 0x%p %d", __func__, syncobj, + refcount_read(&syncobj->refcount)); + return refcount_inc_not_zero(&syncobj->refcount); +} + +void dxgsharedsyncobj_release_reference(struct dxgsharedsyncobject *syncobj) +{ + TRACE_DEBUG(1, "%s 0x%p %d", __func__, syncobj, + refcount_read(&syncobj->refcount)); + if (refcount_dec_and_test(&syncobj->refcount)) { + TRACE_DEBUG(1, "Destroying"); + if (syncobj->global_shared_handle) { + hmgrtable_lock(&dxgglobal->handle_table, DXGLOCK_EXCL); + hmgrtable_free_handle(&dxgglobal->handle_table, + HMGRENTRY_TYPE_DXGSYNCOBJECT, + syncobj->global_shared_handle); + hmgrtable_unlock(&dxgglobal->handle_table, + DXGLOCK_EXCL); + } + if (syncobj->adapter) { + dxgadapter_remove_shared_syncobj(syncobj->adapter, + syncobj); + dxgadapter_release_reference(syncobj->adapter); + } + dxgmem_free(NULL, DXGMEM_SHAREDSYNCOBJ, syncobj); + } + TRACE_DEBUG(1, "%s end", __func__); +} + +void dxgsharedsyncobj_add_syncobj(struct dxgsharedsyncobject *shared, + struct dxgsyncobject *syncobj) +{ + TRACE_DEBUG(1, "%s 0x%p 0x%p", __func__, shared, syncobj); + dxgsharedsyncobj_acquire_reference(shared); + down_write(&shared->syncobj_list_lock); + list_add(&syncobj->shared_syncobj_list_entry, + &shared->shared_syncobj_list_head); + syncobj->shared_owner = shared; + up_write(&shared->syncobj_list_lock); +} + +void dxgsharedsyncobj_remove_syncobj(struct dxgsharedsyncobject *shared, + struct dxgsyncobject *syncobj) +{ + TRACE_DEBUG(1, "%s 0x%p", __func__, shared); + down_write(&shared->syncobj_list_lock); + list_del(&syncobj->shared_syncobj_list_entry); + up_write(&shared->syncobj_list_lock); +} + +struct dxgsyncobject *dxgsyncobject_create(struct dxgprocess *process, + struct dxgdevice *device, + struct dxgadapter *adapter, + enum + d3dddi_synchronizationobject_type + type, + struct + d3dddi_synchronizationobject_flags + flags) +{ + struct dxgsyncobject *syncobj; + + syncobj = dxgmem_alloc(process, DXGMEM_SYNCOBJ, sizeof(*syncobj)); + if (syncobj == NULL) + goto cleanup; + syncobj->type = type; + syncobj->process = process; + switch (type) { + case D3DDDI_MONITORED_FENCE: + case D3DDDI_PERIODIC_MONITORED_FENCE: + syncobj->monitored_fence = 1; + break; + case D3DDDI_CPU_NOTIFICATION: + syncobj->cpu_event = 1; + syncobj->host_event = dxgmem_alloc(process, DXGMEM_HOSTEVENT, + sizeof(struct dxghostevent)); + if (syncobj->host_event == NULL) + goto cleanup; + break; + default: + break; + } + if (flags.shared) { + syncobj->shared = 1; + if (flags.nt_security_sharing) + syncobj->shared_nt = 1; + } + + refcount_set(&syncobj->refcount, 1); + + if (syncobj->monitored_fence) { + syncobj->device = device; + syncobj->device_handle = device->handle; + dxgdevice_acquire_reference(device); + dxgdevice_add_syncobj(device, syncobj); + } else { + dxgadapter_add_syncobj(adapter, syncobj); + } + syncobj->adapter = adapter; + dxgadapter_acquire_reference(adapter); + + TRACE_DEBUG(1, "%s 0x%p\n", __func__, syncobj); + return syncobj; +cleanup: + if (syncobj->host_event) + dxgmem_free(process, DXGMEM_HOSTEVENT, syncobj->host_event); + if (syncobj) + dxgmem_free(process, DXGMEM_SYNCOBJ, syncobj); + return NULL; +} + +void dxgsyncobject_destroy(struct dxgprocess *process, + struct dxgsyncobject *syncobj) +{ + int destroyed; + + TRACE_DEBUG(1, "%s 0x%p", __func__, syncobj); + + dxgsyncobject_stop(syncobj); + + destroyed = test_and_set_bit(0, &syncobj->flags); + if (!destroyed) { + TRACE_DEBUG(1, "Deleting handle: %x", syncobj->handle); + hmgrtable_lock(&process->handle_table, DXGLOCK_EXCL); + if (syncobj->handle) { + hmgrtable_free_handle(&process->handle_table, + HMGRENTRY_TYPE_DXGSYNCOBJECT, + syncobj->handle); + syncobj->handle = 0; + dxgsyncobject_release_reference(syncobj); + } + hmgrtable_unlock(&process->handle_table, DXGLOCK_EXCL); + + if (syncobj->cpu_event) { + if (syncobj->host_event->cpu_event) { + eventfd_ctx_put(syncobj->host_event->cpu_event); + if (syncobj->host_event->event_id) { + dxgglobal_remove_host_event(syncobj-> + host_event); + } + syncobj->host_event->cpu_event = NULL; + } + } + if (syncobj->monitored_fence) + dxgdevice_remove_syncobj(syncobj); + else + dxgadapter_remove_syncobj(syncobj); + if (syncobj->adapter) { + dxgadapter_release_reference(syncobj->adapter); + syncobj->adapter = NULL; + } + } + dxgsyncobject_release_reference(syncobj); + + TRACE_DEBUG(1, "%s end", __func__); +} + +void dxgsyncobject_stop(struct dxgsyncobject *syncobj) +{ + int stopped = test_and_set_bit(1, &syncobj->flags); + + TRACE_DEBUG(1, "%s", __func__); + if (!stopped) { + TRACE_DEBUG(1, "stopping"); + if (syncobj->monitored_fence) { + if (syncobj->mapped_address) { + int ret = + dxg_unmap_iospace(syncobj->mapped_address, + PAGE_SIZE); + + (void)ret; + TRACE_DEBUG(1, "fence is unmapped %d %p\n", + ret, syncobj->mapped_address); + syncobj->mapped_address = NULL; + } + } + } + TRACE_DEBUG(1, "%s end", __func__); +} + +void dxgsyncobject_acquire_reference(struct dxgsyncobject *syncobj) +{ + TRACE_DEBUG(1, "%s 0x%p %d", + __func__, syncobj, refcount_read(&syncobj->refcount)); + refcount_inc_not_zero(&syncobj->refcount); +} + +void dxgsyncobject_release_reference(struct dxgsyncobject *syncobj) +{ + TRACE_DEBUG(1, "%s 0x%p %d", + __func__, syncobj, refcount_read(&syncobj->refcount)); + if (refcount_dec_and_test(&syncobj->refcount)) { + if (syncobj->shared_owner) { + dxgsharedsyncobj_remove_syncobj(syncobj->shared_owner, + syncobj); + dxgsharedsyncobj_release_reference(syncobj-> + shared_owner); + } + if (syncobj->host_event) + dxgmem_free(syncobj->process, DXGMEM_HOSTEVENT, + syncobj->host_event); + dxgmem_free(syncobj->process, DXGMEM_SYNCOBJ, syncobj); + } +} + +struct dxghwqueue *dxghwqueue_create(struct dxgcontext *context) +{ + struct dxgprocess *process = context->device->process; + struct dxghwqueue *hwqueue = + dxgmem_alloc(process, DXGMEM_HWQUEUE, sizeof(*hwqueue)); + if (hwqueue) { + refcount_set(&hwqueue->refcount, 1); + hwqueue->context = context; + hwqueue->process = process; + hwqueue->device_handle = context->device->handle; + if (dxgcontext_add_hwqueue(context, hwqueue)) { + dxghwqueue_release_reference(hwqueue); + hwqueue = NULL; + } else { + dxgcontext_acquire_reference(context); + } + } + return hwqueue; +} + +void dxghwqueue_destroy(struct dxgprocess *process, struct dxghwqueue *hwqueue) +{ + TRACE_DEBUG(1, "%s %p\n", __func__, hwqueue); + hmgrtable_lock(&process->handle_table, DXGLOCK_EXCL); + if (hwqueue->handle) { + hmgrtable_free_handle(&process->handle_table, + HMGRENTRY_TYPE_DXGHWQUEUE, + hwqueue->handle); + hwqueue->handle = 0; + } + hmgrtable_unlock(&process->handle_table, DXGLOCK_EXCL); + + if (hwqueue->progress_fence_mapped_address) { + dxg_unmap_iospace(hwqueue->progress_fence_mapped_address, + PAGE_SIZE); + hwqueue->progress_fence_mapped_address = NULL; + } + dxgcontext_remove_hwqueue_safe(hwqueue->context, hwqueue); + + dxgcontext_release_reference(hwqueue->context); + dxghwqueue_release_reference(hwqueue); +} + +bool dxghwqueue_acquire_reference(struct dxghwqueue *hwqueue) +{ + return refcount_inc_not_zero(&hwqueue->refcount); +} + +void dxghwqueue_release_reference(struct dxghwqueue *hwqueue) +{ + if (refcount_dec_and_test(&hwqueue->refcount)) + dxgmem_free(hwqueue->process, DXGMEM_HWQUEUE, hwqueue); +} diff --git a/drivers/gpu/dxgkrnl/dxgkrnl.h b/drivers/gpu/dxgkrnl/dxgkrnl.h new file mode 100644 index 000000000000..20be1b5053c7 --- /dev/null +++ b/drivers/gpu/dxgkrnl/dxgkrnl.h @@ -0,0 +1,913 @@ +// SPDX-License-Identifier: GPL-2.0 + +/* + * Copyright (c) 2019, Microsoft Corporation. + * + * Author: + * Iouri Tarassov iourit@microsoft.com + * + * Dxgkrnl Graphics Port Driver headers + * + */ + +#ifndef _DXGKRNL_H +#define _DXGKRNL_H + +#include <linux/uuid.h> +#include <linux/kernel.h> +#include <linux/mutex.h> +#include <linux/semaphore.h> +#include <linux/refcount.h> +#include <linux/rwsem.h> +#include <linux/atomic.h> +#include <linux/spinlock.h> +#include <linux/gfp.h> +#include <linux/cdev.h> + +struct dxgprocess; +struct dxgadapter; +struct dxgdevice; +struct dxgcontext; +struct dxgallocation; +struct dxgresource; +struct dxgsharedresource; +struct dxgsyncobject; +struct dxgsharedsyncobject; +struct dxghwqueue; + +#include "misc.h" +#include "hmgr.h" +#include "d3dkmthk.h" + +struct dxgk_device_types { + uint post_device:1; + uint post_device_certain:1; + uint software_device:1; + uint soft_gpu_device:1; + uint warp_device:1; + uint bdd_device:1; + uint support_miracast:1; + uint mismatched_lda:1; + uint indirect_display_device:1; + uint xbox_one_device:1; + uint child_id_support_dwm_clone:1; + uint child_id_support_dwm_clone2:1; + uint has_internal_panel:1; + uint rfx_vgpu_device:1; + uint virtual_render_device:1; + uint support_preserve_boot_display:1; + uint is_uefi_frame_buffer:1; + uint removable_device:1; + uint virtual_monitor_device:1; +}; + +enum dxgobjectstate { + DXGOBJECTSTATE_CREATED, + DXGOBJECTSTATE_ACTIVE, + DXGOBJECTSTATE_STOPPED, + DXGOBJECTSTATE_DESTROYED, +}; + +struct dxgvmbuschannel { + struct vmbus_channel *channel; + struct hv_device *hdev; + struct dxgadapter *adapter; + spinlock_t packet_list_mutex; + struct list_head packet_list_head; + struct kmem_cache *packet_cache; + atomic64_t packet_request_id; +}; + +int dxgvmbuschannel_init(struct dxgvmbuschannel *ch, struct hv_device *hdev); +void dxgvmbuschannel_destroy(struct dxgvmbuschannel *ch); +void dxgvmbuschannel_receive(void *ctx); + +struct dxgpagingqueue { + struct dxgdevice *device; + struct dxgprocess *process; + struct list_head pqueue_list_entry; + d3dkmt_handle device_handle; + d3dkmt_handle handle; + d3dkmt_handle syncobj_handle; + void *mapped_address; +}; + +/* + * The structure describes an event, which will be signaled by + * a message from host. + */ +struct dxghostevent { + struct list_head host_event_list_entry; + u64 event_id; + struct dxgprocess *process; + struct eventfd_ctx *cpu_event; + struct completion *completion_event; + bool destroy_after_signal; + bool remove_from_list; +}; + +struct dxgpagingqueue *dxgpagingqueue_create(struct dxgdevice *device); +void dxgpagingqueue_destroy(struct dxgpagingqueue *pqueue); +void dxgpagingqueue_stop(struct dxgpagingqueue *pqueue); + +/* + * This is GPU synchronization object, which is used to synchronize execution + * between GPU contextx/hardware queues or for tracking GPU execution progress. + * A dxgsyncobject is created when somebody creates a syncobject or opens a + * shared syncobject. + * A syncobject belongs to an adapter, unless it is a cross-adapter object. + * Cross adapter syncobjects are currently not implemented. + * + * D3DDDI_MONITORED_FENCE and D3DDDI_PERIODIC_MONITORED_FENCE are called + * "device" syncobject, because the belong to a device (dxgdevice). + * Device syncobjects are inserted to a list in dxgdevice. + * + * A syncobject can be "shared", meaning that it could be opened by many + * processes. + * + * Shared syncobjects are inserted to a list in its owner + * (dxgsharedsyncobject). + * A syncobject can be shared by using a global handle or by using + * "NT security handle". + * When global handle sharing is used, the handle is created durinig object + * creation. + * When "NT security" is used, the handle for sharing is create be calling + * dxgk_share_objects. On Linux "NT handle" is represented by a file + * descriptor. FD points to dxgsharedsyncobject. + */ +struct dxgsyncobject { + refcount_t refcount; + enum d3dddi_synchronizationobject_type type; + /* + * List entry in dxgdevice for device sync objects. + * List entry in dxgadapter for other objects + */ + struct list_head syncobj_list_entry; + /* List entry in the dxgsharedsyncobject object for shared synobjects */ + struct list_head shared_syncobj_list_entry; + /* Adapter, the syncobject belongs to. NULL for stopped sync obejcts. */ + struct dxgadapter *adapter; + /* + * Pointer to the device, which was used to create the object. + * This is NULL for non-device syncbjects + */ + struct dxgdevice *device; + struct dxgprocess *process; + /* Used by D3DDDI_CPU_NOTIFICATION objects */ + struct dxghostevent *host_event; + /* Owner object for shared syncobjects */ + struct dxgsharedsyncobject *shared_owner; + /* CPU virtual address of the fence value for "device" syncobjects */ + void *mapped_address; + /* Handle in the process handle table */ + d3dkmt_handle handle; + /* Cached handle of the device. Used to avoid device dereference. */ + d3dkmt_handle device_handle; + union { + struct { + /* Must be the first bit */ + uint destroyed:1; + /* Must be the second bit */ + uint stopped:1; + /* device syncobject */ + uint monitored_fence:1; + uint cpu_event:1; + uint shared:1; + /* shared using file descriptor */ + uint shared_nt:1; + uint reserved:26; + }; + long flags; + }; +}; + +/* + * The object is used as parent of all sync objects, created for a shared + * syncobject. When a shared syncobject is created without NT security, the + * handle in the global handle table will point to this object. + */ +struct dxgsharedsyncobject { + refcount_t refcount; + /* Referenced by file descriptors */ + int host_shared_handle_nt_reference; + /* + * Handle in the global handle table. It is zero for NT + * security syncobjects + */ + d3dkmt_handle global_shared_handle; + /* Corresponding handle in the host global handle table */ + d3dkmt_handle host_shared_handle; + /* + * When the sync object is shared by NT handle, this is the + * corresponding handle in the host + */ + d3dkmt_handle host_shared_handle_nt; + /* Protects access to host_shared_handle_nt */ + struct dxgmutex fd_mutex; + struct rw_semaphore syncobj_list_lock; + struct list_head shared_syncobj_list_head; + struct list_head adapter_shared_syncobj_list_entry; + struct dxgadapter *adapter; + enum d3dddi_synchronizationobject_type type; + uint monitored_fence:1; +}; + +struct dxgsharedsyncobject *dxgsharedsyncobj_create(struct dxgadapter *adapter, + struct dxgsyncobject + *syncobj); +bool dxgsharedsyncobj_acquire_reference(struct dxgsharedsyncobject *syncobj); +void dxgsharedsyncobj_release_reference(struct dxgsharedsyncobject *syncobj); +void dxgsharedsyncobj_add_syncobj(struct dxgsharedsyncobject *sharedsyncobj, + struct dxgsyncobject *syncobj); +void dxgsharedsyncobj_remove_syncobj(struct dxgsharedsyncobject *sharedsyncobj, + struct dxgsyncobject *syncobj); + +struct dxgsyncobject *dxgsyncobject_create(struct dxgprocess *process, + struct dxgdevice *device, + struct dxgadapter *adapter, + enum + d3dddi_synchronizationobject_type + type, + struct + d3dddi_synchronizationobject_flags + flags); +void dxgsyncobject_destroy(struct dxgprocess *process, + struct dxgsyncobject *syncobj); +void dxgsyncobject_stop(struct dxgsyncobject *syncobj); +void dxgsyncobject_acquire_reference(struct dxgsyncobject *syncobj); +void dxgsyncobject_release_reference(struct dxgsyncobject *syncobj); + +extern struct device *dxgglobaldev; + +struct dxgglobal { + struct dxgvmbuschannel channel; + struct delayed_work dwork; + struct hv_device *hdev; + u32 num_adapters; + struct resource *mem; + u64 mmiospace_base; + u64 mmiospace_size; + dev_t device_devt; + struct class *device_class; + struct device *device; + struct cdev device_cdev; + struct dxgmutex device_mutex; + + /* list of created processes */ + struct list_head plisthead; + struct dxgmutex plistmutex; + + /* list of created adapters */ + struct list_head adapter_list_head; + struct rw_semaphore adapter_list_lock; + + /* List of all current threads for lock order tracking. */ + struct mutex thread_info_mutex; + struct list_head thread_info_list_head; + + /* protects acces to the global VM bus channel */ + struct rw_semaphore channel_lock; + + /* protects the dxgprocess_adapter lists */ + struct dxgmutex process_adapter_mutex; + + /* list of events, waiting to be signaled by the host */ + struct list_head host_event_list_head; + spinlock_t host_event_list_mutex; + atomic64_t host_event_id; + + /* Handle table for shared objects */ + struct hmgrtable handle_table; + + bool cdev_initialized; + bool devt_initialized; + bool vmbus_registered; +}; + +extern struct dxgglobal *dxgglobal; + +void dxgglobal_acquire_adapter_list_lock(enum dxglockstate state); +void dxgglobal_release_adapter_list_lock(enum dxglockstate state); +struct vmbus_channel *dxgglobal_get_vmbus(void); +struct dxgvmbuschannel *dxgglobal_get_dxgvmbuschannel(void); +void dxgglobal_acquire_process_adapter_lock(void); +void dxgglobal_release_process_adapter_lock(void); +void dxgglobal_add_host_event(struct dxghostevent *hostevent); +void dxgglobal_remove_host_event(struct dxghostevent *hostevent); +u64 dxgglobal_new_host_event_id(void); +void dxgglobal_signal_host_event(u64 event_id); +struct dxghostevent *dxgglobal_get_host_event(u64 event_id); +int dxgglobal_acquire_channel_lock(void); +void dxgglobal_release_channel_lock(void); + +/* + * Describes adapter information for each process + */ +struct dxgprocess_adapter { + /* Entry in dxgadapter::adapter_process_list_head */ + struct list_head adapter_process_list_entry; + /* Entry in dxgprocess::process_adapter_list_head */ + struct list_head process_adapter_list_entry; + /* List of all dxgdevice objects created for the process on adapter */ + struct list_head device_list_head; + struct dxgmutex device_list_mutex; + struct dxgadapter *adapter; + struct dxgprocess *process; + int refcount; +}; + +struct dxgprocess_adapter *dxgprocess_adapter_create(struct dxgprocess *process, + struct dxgadapter + *adapter); +void dxgprocess_adapter_release(struct dxgprocess_adapter *adapter); +int dxgprocess_adapter_add_device(struct dxgprocess *process, + struct dxgadapter *adapter, + struct dxgdevice *device); +void dxgprocess_adapter_remove_device(struct dxgdevice *device); +void dxgprocess_adapter_stop(struct dxgprocess_adapter *adapter_info); +void dxgprocess_adapter_destroy(struct dxgprocess_adapter *adapter_info); + +struct dxgprocess { + /* + * Process list entry in dxgglobal. + * Protected by the dxgglobal->plistmutex. + */ + struct list_head plistentry; + struct task_struct *process; + pid_t pid; + pid_t tgid; + /* how many time the process was opened */ + int refcount; + /* + * This handle table is used for all objects except dxgadapter + * The handle table lock order is higher than the local_handle_table + * lock + */ + struct hmgrtable handle_table; + /* + * This handle table is used for dxgadapter objects. + * The handle table lock order is lowest. + */ + struct hmgrtable local_handle_table; + d3dkmt_handle host_handle; + + /* List of opened adapters (dxgprocess_adapter) */ + struct list_head process_adapter_list_head; + + struct hmgrtable *test_handle_table[2]; + atomic_t dxg_memory[DXGMEM_LAST]; + struct dxgmutex process_mutex; +}; + +struct dxgprocess *dxgprocess_create(void); +void dxgprocess_destroy(struct dxgprocess *process); +void dxgprocess_release_reference(struct dxgprocess *process); +int dxgprocess_open_adapter(struct dxgprocess *process, + struct dxgadapter *adapter, d3dkmt_handle *handle); +int dxgprocess_close_adapter(struct dxgprocess *process, d3dkmt_handle handle); +struct dxgadapter *dxgprocess_get_adapter(struct dxgprocess *process, + d3dkmt_handle handle); +struct dxgadapter *dxgprocess_adapter_by_handle(struct dxgprocess *process, + d3dkmt_handle handle); +struct dxgdevice *dxgprocess_device_by_handle(struct dxgprocess *process, + d3dkmt_handle handle); +struct dxgdevice *dxgprocess_device_by_object_handle(struct dxgprocess *process, + enum hmgrentry_type t, + d3dkmt_handle handle); +void dxgprocess_ht_lock_shared_down(struct dxgprocess *process); +void dxgprocess_ht_lock_shared_up(struct dxgprocess *process); +void dxgprocess_ht_lock_exclusive_down(struct dxgprocess *process); +void dxgprocess_ht_lock_exclusive_up(struct dxgprocess *process); +struct dxgprocess_adapter *dxgprocess_get_adapter_info(struct dxgprocess + *process, + struct dxgadapter + *adapter); + +enum dxgadapter_state { + DXGADAPTER_STATE_ACTIVE = 0, + DXGADAPTER_STATE_STOPPED = 1, +}; + +/* + * This object represents the grapchis adapter. + * Objects, which take reference on the adapter: + * - dxgglobal + * - dxgdevice + * - adapter handle (d3dkmt_handle) + */ +struct dxgadapter { + struct rw_semaphore core_lock; + struct rw_semaphore adapter_process_list_lock; + refcount_t refcount; + /* Entry in the list of adapters in dxgglobal */ + struct list_head adapter_list_entry; + /* The list of dxgprocess_adapter entries */ + struct list_head adapter_process_list_head; + /* List of all dxgsharedresource objects */ + struct list_head shared_resource_list_head; + /* List of all dxgsharedsyncobject objects */ + struct list_head adapter_shared_syncobj_list_head; + /* List of all non-device dxgsyncobject objects */ + struct list_head syncobj_list_head; + /* This lock protects shared resource and syncobject lists */ + struct rw_semaphore shared_resource_list_lock; + struct winluid luid; + struct dxgvmbuschannel channel; + d3dkmt_handle host_handle; + enum dxgadapter_state adapter_state; + struct winluid host_adapter_luid; + winwchar device_description[80]; + winwchar device_instance_id[W_MAX_PATH]; +}; + +int dxgadapter_init(struct dxgadapter *adapter, struct hv_device *hdev); +bool dxgadapter_is_active(struct dxgadapter *adapter); +void dxgadapter_stop(struct dxgadapter *adapter); +void dxgadapter_destroy(struct dxgadapter *adapter); +bool dxgadapter_acquire_reference(struct dxgadapter *adapter); +void dxgadapter_release_reference(struct dxgadapter *adapter); +int dxgadapter_acquire_lock_shared(struct dxgadapter *adapter); +void dxgadapter_release_lock_shared(struct dxgadapter *adapter); +int dxgadapter_acquire_lock_exclusive(struct dxgadapter *adapter); +void dxgadapter_acquire_lock_forced(struct dxgadapter *adapter); +void dxgadapter_release_lock_exclusive(struct dxgadapter *adapter); +void dxgadapter_add_shared_resource(struct dxgadapter *adapter, + struct dxgsharedresource *res); +void dxgadapter_remove_shared_resource(struct dxgadapter *adapter, + struct dxgsharedresource *res); +void dxgadapter_add_shared_syncobj(struct dxgadapter *adapter, + struct dxgsharedsyncobject *so); +void dxgadapter_remove_shared_syncobj(struct dxgadapter *adapter, + struct dxgsharedsyncobject *so); +void dxgadapter_add_syncobj(struct dxgadapter *adapter, + struct dxgsyncobject *so); +void dxgadapter_remove_syncobj(struct dxgsyncobject *so); +void dxgadapter_add_process(struct dxgadapter *adapter, + struct dxgprocess_adapter *process_info); +void dxgadapter_remove_process(struct dxgprocess_adapter *process_info); + +/* + * The object represent the device object. + * The following objects take reference on the device + * - dxgcontext + * - device handle (d3dkmt_handle) + */ +struct dxgdevice { + enum dxgobjectstate object_state; + /* Device takes reference on the adapter */ + struct dxgadapter *adapter; + struct dxgprocess_adapter *adapter_info; + struct dxgprocess *process; + /* Entry in the DGXPROCESS_ADAPTER device list */ + struct list_head device_list_entry; + refcount_t refcount; + /* Protects destcruction of the device object */ + struct rw_semaphore device_lock; + struct rw_semaphore context_list_lock; + struct list_head context_list_head; + /* List of device allocations */ + struct rw_semaphore alloc_list_lock; + struct list_head alloc_list_head; + struct list_head resource_list_head; + /* List of paging queues. Protected by process handle table lock. */ + struct list_head pqueue_list_head; + struct list_head syncobj_list_head; + d3dkmt_handle handle; + uint handle_valid; +}; + +struct dxgdevice *dxgdevice_create(struct dxgadapter *a, struct dxgprocess *p); +void dxgdevice_destroy(struct dxgdevice *device); +void dxgdevice_stop(struct dxgdevice *device); +int dxgdevice_acquire_lock_shared(struct dxgdevice *dev); +void dxgdevice_release_lock_shared(struct dxgdevice *dev); +bool dxgdevice_acquire_reference(struct dxgdevice *dev); +void dxgdevice_release_reference(struct dxgdevice *dev); +void dxgdevice_add_context(struct dxgdevice *dev, struct dxgcontext *ctx); +void dxgdevice_remove_context(struct dxgdevice *dev, struct dxgcontext *ctx); +void dxgdevice_add_alloc(struct dxgdevice *dev, struct dxgallocation *a); +void dxgdevice_remove_alloc(struct dxgdevice *dev, struct dxgallocation *a); +void dxgdevice_remove_alloc_safe(struct dxgdevice *dev, + struct dxgallocation *a); +void dxgdevice_add_resource(struct dxgdevice *dev, struct dxgresource *res); +void dxgdevice_remove_resource(struct dxgdevice *dev, struct dxgresource *res); +void dxgdevice_add_paging_queue(struct dxgdevice *dev, + struct dxgpagingqueue *pqueue); +void dxgdevice_remove_paging_queue(struct dxgpagingqueue *pqueue); +void dxgdevice_add_syncobj(struct dxgdevice *dev, struct dxgsyncobject *so); +void dxgdevice_remove_syncobj(struct dxgsyncobject *so); +bool dxgdevice_is_active(struct dxgdevice *dev); +void dxgdevice_acquire_context_list_lock(struct dxgdevice *dev); +void dxgdevice_release_context_list_lock(struct dxgdevice *dev); +void dxgdevice_acquire_alloc_list_lock(struct dxgdevice *dev); +void dxgdevice_release_alloc_list_lock(struct dxgdevice *dev); +void dxgdevice_acquire_alloc_list_lock_shared(struct dxgdevice *dev); +void dxgdevice_release_alloc_list_lock_shared(struct dxgdevice *dev); + +/* + * The object represent the execution context of a device. + */ +struct dxgcontext { + enum dxgobjectstate object_state; + struct dxgdevice *device; + struct dxgprocess *process; + /* entry in the device context list */ + struct list_head context_list_entry; + struct list_head hwqueue_list_head; + struct rw_semaphore hwqueue_list_lock; + refcount_t refcount; + d3dkmt_handle handle; + d3dkmt_handle device_handle; +}; + +struct dxgcontext *dxgcontext_create(struct dxgdevice *dev); +void dxgcontext_destroy(struct dxgprocess *pr, struct dxgcontext *ctx); +void dxgcontext_destroy_safe(struct dxgprocess *pr, struct dxgcontext *ctx); +bool dxgcontext_acquire_reference(struct dxgcontext *ctx); +void dxgcontext_release_reference(struct dxgcontext *ctx); +int dxgcontext_add_hwqueue(struct dxgcontext *ctx, struct dxghwqueue *hq); +void dxgcontext_remove_hwqueue(struct dxgcontext *ctx, struct dxghwqueue *hq); +void dxgcontext_remove_hwqueue_safe(struct dxgcontext *ctx, + struct dxghwqueue *hq); +bool dxgcontext_is_active(struct dxgcontext *ctx); + +/* + * The object represent the execution hardware queue of a device. + */ +struct dxghwqueue { + /* entry in the context hw queue list */ + struct list_head hwqueue_list_entry; + refcount_t refcount; + struct dxgcontext *context; + struct dxgprocess *process; + d3dkmt_handle progress_fence_sync_object; + d3dkmt_handle handle; + d3dkmt_handle device_handle; + void *progress_fence_mapped_address; +}; + +struct dxghwqueue *dxghwqueue_create(struct dxgcontext *ctx); +void dxghwqueue_destroy(struct dxgprocess *pr, struct dxghwqueue *hq); +bool dxghwqueue_acquire_reference(struct dxghwqueue *hq); +void dxghwqueue_release_reference(struct dxghwqueue *hq); + +/* + * A shared resource object is created to track the list of dxgresource objects, + * which are opened for the same underlying shared resource. + * There are two types of sharing resource objects: + * - sharing by using a global handle (nt_security is false). + * The global handle is a handle in the handle table of dxgglobal. It points + * to a dxgsharedresource object. dxgk_open_resource() creates a dxgresource + * object using dxgsharedresource. + * - shariing by using a file descriptor handle (nt_security is true). + * FD is created by calling dxgk_share_objects and providing shandle to + * dxgsharedresource. The FD points to a dxgresource object, which is created + * by calling dxgk_open_resource_nt. dxgresource object is referenced by the + * FD. + * + * The object is referenced by every dxgresource in its list. + * + */ +struct dxgsharedresource { + /* Every dxgresource object in the resource list takes a reference */ + refcount_t refcount; + struct dxgadapter *adapter; + /* List of dxgresource objects, opened for the shared resource. */ + /* Protected by dxgadapter::shared_resource_list_lock */ + struct list_head resource_list_head; + /* Entry in the list of dxgsharedresource in dxgadapter */ + /* Protected by dxgadapter::shared_resource_list_lock */ + struct list_head shared_resource_list_entry; + struct dxgmutex fd_mutex; + /* Referenced by file descriptors */ + int host_shared_handle_nt_reference; + /* Handle in the dxgglobal handle table, when nt_security is not used */ + d3dkmt_handle global_handle; + /* Corresponding global handle in the host */ + d3dkmt_handle host_shared_handle; + /* + * When the sync object is shared by NT handle, this is the + * corresponding handle in the host + */ + d3dkmt_handle host_shared_handle_nt; + /* Values below are computed when the resource is sealed */ + uint runtime_private_data_size; + uint alloc_private_data_size; + uint resource_private_data_size; + uint allocation_count; + union { + struct { + /* Referenced by file descriptor */ + uint nt_security:1; + /* Cannot add new allocations */ + uint sealed:1; + uint reserved:30; + }; + long flags; + }; + uint *alloc_private_data_sizes; + uint8_t *alloc_private_data; + uint8_t *runtime_private_data; + uint8_t *resource_private_data; +}; + +struct dxgsharedresource *dxgsharedresource_create(struct dxgadapter *adapter); +bool dxgsharedresource_acquire_reference(struct dxgsharedresource *res); +void dxgsharedresource_release_reference(struct dxgsharedresource *res); +void dxgsharedresource_add_resource(struct dxgsharedresource *sres, + struct dxgresource *res); +void dxgsharedresource_remove_resource(struct dxgsharedresource *sres, + struct dxgresource *res); + +struct dxgresource { + refcount_t refcount; + enum dxgobjectstate object_state; + d3dkmt_handle handle; + struct list_head alloc_list_head; + struct list_head resource_list_entry; + struct list_head shared_resource_list_entry; + struct dxgdevice *device; + struct dxgprocess *process; + /* Protects adding allocations to resource and resource destruction */ + struct dxgmutex resource_mutex; + winhandle private_runtime_handle; + union { + struct { + uint destroyed:1; /* Must be the first */ + uint handle_valid:1; + uint reserved:30; + }; + long flags; + }; + /* Owner of the shared resource */ + struct dxgsharedresource *shared_owner; +}; + +struct dxgresource *dxgresource_create(struct dxgdevice *dev); +void dxgresource_destroy(struct dxgresource *res); +void dxgresource_free_handle(struct dxgresource *res); +void dxgresource_acquire_reference(struct dxgresource *res); +void dxgresource_release_reference(struct dxgresource *res); +int dxgresource_add_alloc(struct dxgresource *res, struct dxgallocation *a); +void dxgresource_remove_alloc(struct dxgresource *res, struct dxgallocation *a); +void dxgresource_remove_alloc_safe(struct dxgresource *res, + struct dxgallocation *a); +bool dxgresource_is_active(struct dxgresource *res); + +struct privdata { + uint data_size; + uint8_t data[1]; +}; + +struct dxgallocation { + /* Entry in the device list or resource list (when resource exists) */ + struct list_head alloc_list_entry; + /* Allocation owner */ + union { + struct dxgdevice *device; + struct dxgresource *resource; + } owner; + struct dxgprocess *process; + /* Pointer to private driver data desc. Used for shared resources */ + struct privdata *priv_drv_data; + d3dkmt_handle alloc_handle; + /* Set to 1 when allocation belongs to resource. */ + uint resource_owner:1; + /* Set to 1 when 'cpu_address' is mapped to the IO space. */ + uint cpu_address_mapped:1; + /* Set to 1 when the allocatio is mapped as cached */ + uint cached:1; + uint handle_valid:1; + /* GPADL address list for existing sysmem allocations */ + uint gpadl; + /* Number of pages in the 'pages' array */ + uint num_pages; + /* + * How many times dxgk_lock2 is called to allocation, which is mapped + * to IO space. + */ + uint cpu_address_refcount; + /* + * CPU address from the existing sysmem allocation, or + * mapped to the CPU visible backing store in the IO space + */ + void *cpu_address; + /* Describes pages for the existing sysmem allocation */ + struct page **pages; +}; + +struct dxgallocation *dxgallocation_create(struct dxgprocess *process); +void dxgallocation_stop(struct dxgallocation *a); +void dxgallocation_destroy(struct dxgallocation *a); +void dxgallocation_free_handle(struct dxgallocation *a); + +void ioctl_desc_init(void); +long dxgk_compat_ioctl(struct file *f, unsigned int p1, unsigned long p2); +long dxgk_unlocked_ioctl(struct file *f, unsigned int p1, unsigned long p2); + +int dxg_unmap_iospace(void *va, uint size); +int dxg_copy_from_user(void *to, const void __user *from, unsigned long len); +int dxg_copy_to_user(void *to, const void __user *from, unsigned long len); +static inline void guid_to_luid(guid_t *guid, struct winluid *luid) +{ + *luid = *(struct winluid *)&guid->b[0]; +} + +/* + * VM bus interface + * + */ +int dxgvmb_send_set_iospace_region(u64 start, u64 len, u32 shared_mem_gpadl); +int dxgvmb_send_create_process(struct dxgprocess *process); +int dxgvmb_send_destroy_process(d3dkmt_handle process); +int dxgvmb_send_open_adapter(struct dxgadapter *adapter); +int dxgvmb_send_close_adapter(struct dxgadapter *adapter); +int dxgvmb_send_get_internal_adapter_info(struct dxgadapter *adapter); +d3dkmt_handle dxgvmb_send_create_device(struct dxgadapter *adapter, + struct dxgprocess *process, + struct d3dkmt_createdevice *args); +int dxgvmb_send_destroy_device(struct dxgadapter *adapter, + struct dxgprocess *process, d3dkmt_handle h); +d3dkmt_handle dxgvmb_send_create_context(struct dxgadapter *adapter, + struct dxgprocess *process, + struct d3dkmt_createcontextvirtual + *args); +int dxgvmb_send_destroy_context(struct dxgadapter *adapter, + struct dxgprocess *process, d3dkmt_handle h); +int dxgvmb_send_create_paging_queue(struct dxgprocess *pr, + struct dxgvmbuschannel *ch, + struct dxgdevice *dev, + struct d3dkmt_createpagingqueue *args, + struct dxgpagingqueue *pq); +int dxgvmb_send_destroy_paging_queue(struct dxgprocess *process, + struct dxgvmbuschannel *ch, + d3dkmt_handle h); +int dxgvmb_send_create_allocation(struct dxgprocess *pr, struct dxgdevice *dev, + struct d3dkmt_createallocation *args, + struct d3dkmt_createallocation *__user + input_args, struct dxgresource *res, + struct dxgallocation **allocs, + struct d3dddi_allocationinfo2 *alloc_info, + struct d3dkmt_createstandardallocation + *standard_alloc); +int dxgvmb_send_destroy_allocation(struct dxgprocess *pr, struct dxgdevice *dev, + struct dxgvmbuschannel *ch, + struct d3dkmt_destroyallocation2 *args, + d3dkmt_handle *alloc_handles); +int dxgvmb_send_make_resident(struct dxgprocess *pr, struct dxgdevice *dev, + struct dxgvmbuschannel *ch, + struct d3dddi_makeresident *args); +int dxgvmb_send_evict(struct dxgprocess *pr, struct dxgvmbuschannel *ch, + struct d3dkmt_evict *args); +int dxgvmb_send_submit_command(struct dxgprocess *pr, + struct dxgvmbuschannel *ch, + struct d3dkmt_submitcommand *args); +int dxgvmb_send_map_gpu_va(struct dxgprocess *pr, d3dkmt_handle h, + struct dxgvmbuschannel *ch, + struct d3dddi_mapgpuvirtualaddress *args); +int dxgvmb_send_reserve_gpu_va(struct dxgprocess *pr, + struct dxgvmbuschannel *ch, + struct d3dddi_reservegpuvirtualaddress *args); +int dxgvmb_send_free_gpu_va(struct dxgprocess *pr, struct dxgvmbuschannel *ch, + struct d3dkmt_freegpuvirtualaddress *args); +int dxgvmb_send_update_gpu_va(struct dxgprocess *pr, struct dxgvmbuschannel *ch, + struct d3dkmt_updategpuvirtualaddress *args); +int dxgvmb_send_create_sync_object(struct dxgprocess *pr, + struct dxgvmbuschannel *ch, + struct d3dkmt_createsynchronizationobject2 + *args, struct dxgsyncobject *so); +int dxgvmb_send_destroy_sync_object(struct dxgprocess *pr, d3dkmt_handle h); +int dxgvmb_send_signal_sync_object(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dddicb_signalflags flags, + uint64_t legacy_fence_value, + d3dkmt_handle context, + uint object_count, d3dkmt_handle *object, + uint context_count, d3dkmt_handle *contexts, + uint fence_count, uint64_t *fences, + struct eventfd_ctx *cpu_event, + d3dkmt_handle device); +int dxgvmb_send_wait_sync_object_gpu(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + d3dkmt_handle context, uint object_count, + d3dkmt_handle *objects, uint64_t *fences, + bool legacy_fence); +int dxgvmb_send_wait_sync_object_cpu(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct + d3dkmt_waitforsynchronizationobjectfromcpu + *args, u64 cpu_event); +int dxgvmb_send_lock2(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_lock2 *args, + struct d3dkmt_lock2 *__user outargs); +int dxgvmb_send_unlock2(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_unlock2 *args); +int dxgvmb_send_update_alloc_property(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dddi_updateallocproperty *args, + struct d3dddi_updateallocproperty *__user + inargs); +int dxgvmb_send_mark_device_as_error(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_markdeviceaserror *args); +int dxgvmb_send_set_allocation_priority(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_setallocationpriority + *args); +int dxgvmb_send_get_allocation_priority(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_getallocationpriority + *args); +int dxgvmb_send_set_context_scheduling_priority(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + d3dkmt_handle context, + int priority, bool in_process); +int dxgvmb_send_get_context_scheduling_priority(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + d3dkmt_handle context, + int *priority, bool in_process); +int dxgvmb_send_offer_allocations(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_offerallocations *args); +int dxgvmb_send_reclaim_allocations(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + d3dkmt_handle device, + struct d3dkmt_reclaimallocations2 *args, + uint64_t * __user paging_fence_value); +int dxgvmb_send_change_vidmem_reservation(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + d3dkmt_handle other_process, + struct + d3dkmt_changevideomemoryreservation + *args); +int dxgvmb_send_create_hwqueue(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_createhwqueue *args, + struct d3dkmt_createhwqueue *__user inargs, + struct dxghwqueue *hq); +int dxgvmb_send_destroy_hwqueue(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + d3dkmt_handle handle); +int dxgvmb_send_query_adapter_info(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_queryadapterinfo *args); +int dxgvmb_send_submit_command_to_hwqueue(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_submitcommandtohwqueue + *args); +int dxgvmb_send_query_clock_calibration(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_queryclockcalibration + *args, + struct d3dkmt_queryclockcalibration + *__user inargs); +int dxgvmb_send_flush_heap_transitions(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_flushheaptransitions + *args); +int dxgvmb_send_open_sync_object(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + d3dkmt_handle h, d3dkmt_handle *ph); +int dxgvmb_send_open_sync_object_nt(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_opensyncobjectfromnthandle2 + *args, struct dxgsyncobject *syncobj); +int dxgvmb_send_query_alloc_residency(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_queryallocationresidency + *args); +int dxgvmb_send_escape(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_escape *args); +int dxgvmb_send_query_vidmem_info(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_queryvideomemoryinfo *args, + struct d3dkmt_queryvideomemoryinfo *__user + inargs); +int dxgvmb_send_get_device_state(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_getdevicestate *args, + struct d3dkmt_getdevicestate *__user inargs); +int dxgvmb_send_create_nt_shared_object(struct dxgprocess *process, + d3dkmt_handle object, + d3dkmt_handle *shared_handle); +int dxgvmb_send_destroy_nt_shared_object(d3dkmt_handle shared_handle); +int dxgvmb_send_open_resource(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + d3dkmt_handle device, bool nt_security_sharing, + d3dkmt_handle global_share, uint allocation_count, + uint total_priv_drv_data_size, + d3dkmt_handle *resource_handle, + d3dkmt_handle *alloc_handles); +int dxgvmb_send_get_standard_alloc_priv_data(struct dxgdevice *device, + enum d3dkmdt_standardallocationtype + alloc_type, + struct d3dkmdt_gdisurfacedata + *alloc_data, + uint physical_adapter_index, + uint *alloc_priv_driver_size, + void *prive_alloc_data); + +#endif diff --git a/drivers/gpu/dxgkrnl/dxgmodule.c b/drivers/gpu/dxgkrnl/dxgmodule.c new file mode 100644 index 000000000000..cae13d22801a --- /dev/null +++ b/drivers/gpu/dxgkrnl/dxgmodule.c @@ -0,0 +1,692 @@ +// SPDX-License-Identifier: GPL-2.0 + +/* + * Copyright (c) 2019, Microsoft Corporation. + * + * Author: + * Iouri Tarassov iourit@microsoft.com + * + * Dxgkrnl Graphics Driver + * Interface with Linux kernel and the VM bus driver + * + */ + +#include <linux/module.h> +#include <linux/eventfd.h> +#include <linux/hyperv.h> + +#include "dxgkrnl.h" + +struct dxgglobal *dxgglobal; +struct device *dxgglobaldev; + +#define DXGKRNL_VERSION 0x0001 +#define PCI_VENDOR_ID_MICROSOFT 0x1414 +#define PCI_DEVICE_ID_VIRTUAL_RENDER 0x008E + +// +// Interface from dxgglobal +// + +struct vmbus_channel *dxgglobal_get_vmbus(void) +{ + return dxgglobal->channel.channel; +} + +struct dxgvmbuschannel *dxgglobal_get_dxgvmbuschannel(void) +{ + return &dxgglobal->channel; +} + +int dxgglobal_acquire_channel_lock(void) +{ + dxglockorder_acquire(DXGLOCK_GLOBAL_CHANNEL); + down_read(&dxgglobal->channel_lock); + if (dxgglobal->channel.channel == NULL) { + pr_err("Failed to acquire global channel lock"); + return STATUS_DEVICE_REMOVED; + } else { + return 0; + } +} + +void dxgglobal_release_channel_lock(void) +{ + up_read(&dxgglobal->channel_lock); + dxglockorder_release(DXGLOCK_GLOBAL_CHANNEL); +} + +void dxgglobal_acquire_adapter_list_lock(enum dxglockstate state) +{ + TRACE_DEBUG(1, "%s", __func__); + dxglockorder_acquire(DXGLOCK_GLOBAL_ADAPTERLIST); + if (state == DXGLOCK_EXCL) + down_write(&dxgglobal->adapter_list_lock); + else + down_read(&dxgglobal->adapter_list_lock); +} + +void dxgglobal_release_adapter_list_lock(enum dxglockstate state) +{ + TRACE_DEBUG(1, "%s", __func__); + if (state == DXGLOCK_EXCL) + up_write(&dxgglobal->adapter_list_lock); + else + up_read(&dxgglobal->adapter_list_lock); + dxglockorder_release(DXGLOCK_GLOBAL_ADAPTERLIST); +} + +void dxgglobal_add_host_event(struct dxghostevent *event) +{ + spin_lock_irq(&dxgglobal->host_event_list_mutex); + list_add_tail(&event->host_event_list_entry, + &dxgglobal->host_event_list_head); + spin_unlock_irq(&dxgglobal->host_event_list_mutex); +} + +void dxgglobal_remove_host_event(struct dxghostevent *event) +{ + spin_lock_irq(&dxgglobal->host_event_list_mutex); + if (event->host_event_list_entry.next != NULL) { + list_del(&event->host_event_list_entry); + event->host_event_list_entry.next = NULL; + } + spin_unlock_irq(&dxgglobal->host_event_list_mutex); +} + +void dxgglobal_signal_host_event(u64 event_id) +{ + struct dxghostevent *event; + unsigned long flags; + + TRACE_DEBUG(1, "%s %lld\n", __func__, event_id); + + spin_lock_irqsave(&dxgglobal->host_event_list_mutex, flags); + list_for_each_entry(event, &dxgglobal->host_event_list_head, + host_event_list_entry) { + if (event->event_id == event_id) { + TRACE_DEBUG(1, "found event to signal %lld\n", + event_id); + if (event->remove_from_list || + event->destroy_after_signal) { + list_del(&event->host_event_list_entry); + event->host_event_list_entry.next = NULL; + TRACE_DEBUG(1, "removing event from list\n"); + } + if (event->cpu_event) { + TRACE_DEBUG(1, "signal cpu event\n"); + eventfd_signal(event->cpu_event, 1); + if (event->destroy_after_signal) + eventfd_ctx_put(event->cpu_event); + } else { + TRACE_DEBUG(1, "signal completion\n"); + complete(event->completion_event); + } + if (event->destroy_after_signal) { + TRACE_DEBUG(1, "destroying event %p\n", event); + dxgmem_free(event->process, + DXGMEM_EVENT, event); + } + break; + } + } + spin_unlock_irqrestore(&dxgglobal->host_event_list_mutex, flags); + TRACE_DEBUG(1, "dxgglobal_signal_host_event_end %lld\n", event_id); +} + +struct dxghostevent *dxgglobal_get_host_event(u64 event_id) +{ + struct dxghostevent *entry; + struct dxghostevent *event = NULL; + + spin_lock_irq(&dxgglobal->host_event_list_mutex); + list_for_each_entry(entry, &dxgglobal->host_event_list_head, + host_event_list_entry) { + if (entry->event_id == event_id) { + list_del(&entry->host_event_list_entry); + entry->host_event_list_entry.next = NULL; + event = entry; + break; + } + } + spin_unlock_irq(&dxgglobal->host_event_list_mutex); + return event; +} + +u64 dxgglobal_new_host_event_id(void) +{ + return atomic64_inc_return(&dxgglobal->host_event_id); +} + +void dxgglobal_acquire_process_adapter_lock(void) +{ + dxgmutex_lock(&dxgglobal->process_adapter_mutex); +} + +void dxgglobal_release_process_adapter_lock(void) +{ + dxgmutex_unlock(&dxgglobal->process_adapter_mutex); +} + +/* + * File operations + */ + +static struct dxgprocess *dxgglobal_get_current_process(void) +{ + /* + * Find the DXG process for the current process. + * A new process is created if necessary. + */ + struct dxgprocess *process = NULL; + struct dxgprocess *entry = NULL; + + dxgmutex_lock(&dxgglobal->plistmutex); + list_for_each_entry(entry, &dxgglobal->plisthead, plistentry) { + /* All threads of a process have the same thread group ID */ + if (entry->process->tgid == current->tgid) { + entry->refcount++; + process = entry; + TRACE_DEBUG(1, "found dxgprocess entry\n"); + break; + } + } + dxgmutex_unlock(&dxgglobal->plistmutex); + + if (process == NULL) + process = dxgprocess_create(); + + return process; +} + +static int dxgk_open(struct inode *n, struct file *f) +{ + int ret = 0; + struct dxgprocess *process; + struct dxgthreadinfo *thread; + + TRACE_DEBUG2(1, 0, "%s %p %d %d", + __func__, f, current->pid, current->tgid); + + thread = dxglockorder_get_thread(); + + /* Find/create a dxgprocess structure for this process */ + process = dxgglobal_get_current_process(); + + if (process) { + f->private_data = process; + } else { + TRACE_DEBUG(1, "cannot create dxgprocess for open\n"); + ret = -EBADF; + } + + dxglockorder_put_thread(thread); + TRACE_DEBUG2(1, 0, "%s end %x", __func__, ret); + return ret; +} + +static int dxgk_release(struct inode *n, struct file *f) +{ + struct dxgthreadinfo *thread; + struct dxgprocess *process; + + process = (struct dxgprocess *)f->private_data; + TRACE_DEBUG2(1, 0, "%s %p, %p", __func__, f, process); + + if (process == NULL) + return -EINVAL; + + thread = dxglockorder_get_thread(); + + dxgprocess_release_reference(process); + + dxglockorder_check_empty(thread); + dxglockorder_put_thread(thread); + + f->private_data = NULL; + TRACE_DEBUG2(1, 0, "%s end", __func__); + return 0; +} + +static ssize_t dxgk_read(struct file *f, char __user *s, size_t len, + loff_t *o) +{ + TRACE_DEBUG(1, "file read\n"); + return 0; +} + +static ssize_t dxgk_write(struct file *f, const char __user *s, size_t len, + loff_t *o) +{ + TRACE_DEBUG(1, "file write\n"); + return len; +} + +const struct file_operations dxgk_fops = { + .owner = THIS_MODULE, + .open = dxgk_open, + .release = dxgk_release, + .compat_ioctl = dxgk_compat_ioctl, + .unlocked_ioctl = dxgk_unlocked_ioctl, + .write = dxgk_write, + .read = dxgk_read, +}; + +/* + * Interface with the VM bus driver + */ + +static int dxgglobal_getiospace(struct dxgglobal *dxgglobal) +{ + /* Get mmio space for the global channel */ + struct hv_device *hdev = dxgglobal->hdev; + struct vmbus_channel *channel = hdev->channel; + resource_size_t pot_start = 0; + resource_size_t pot_end = -1; + int ret; + + dxgglobal->mmiospace_size = channel->offermsg.offer.mmio_megabytes; + if (dxgglobal->mmiospace_size == 0) { + TRACE_DEBUG(1, "zero mmio space is offered\n"); + return STATUS_NO_MEMORY; + } + dxgglobal->mmiospace_size <<= 20; + TRACE_DEBUG(1, "mmio offered: %llx\n", dxgglobal->mmiospace_size); + + ret = vmbus_allocate_mmio(&dxgglobal->mem, hdev, pot_start, pot_end, + dxgglobal->mmiospace_size, 0x10000, false); + if (ret) { + pr_err("Unable to allocate mmio memory: %d\n", ret); + return ret; + } + dxgglobal->mmiospace_size = dxgglobal->mem->end - + dxgglobal->mem->start + 1; + dxgglobal->mmiospace_base = dxgglobal->mem->start; + TRACE_DEBUG(1, "mmio allocated %llx %llx %llx %llx\n", + dxgglobal->mmiospace_base, + dxgglobal->mmiospace_size, + dxgglobal->mem->start, dxgglobal->mem->end); + + return 0; +} + +static char *dxg_devnode(struct device *dev, umode_t *mode) +{ + /* Set read/write permissions to the DXG device */ + if (mode) + *mode = 0666; + return NULL; +} + +static int dxgglobal_init_global_channel(struct hv_device *hdev) +{ + int ret = 0; + + TRACE_DEBUG(1, "%s %x %x", __func__, hdev->vendor_id, hdev->device_id); + { + TRACE_DEBUG(1, "device type : %pUb\n", &hdev->dev_type); + TRACE_DEBUG(1, "device channel: %pUb %p primary: %p\n", + &hdev->channel->offermsg.offer.if_type, + hdev->channel, hdev->channel->primary_channel); + } + + if (dxgglobal->hdev) { + /* This device should appear only once */ + pr_err("dxgglobal already initialized\n"); + ret = -EBADE; + goto error; + } + + dxgglobal->hdev = hdev; + + ret = dxgvmbuschannel_init(&dxgglobal->channel, hdev); + if (ret) { + pr_err("dxgvmbuschannel_init failed: %d\n", ret); + goto error; + } + + ret = dxgglobal_getiospace(dxgglobal); + if (ret) { + pr_err("getiospace failed: %d\n", ret); + goto error; + } + + ret = dxgvmb_send_set_iospace_region(dxgglobal->mmiospace_base, + dxgglobal->mmiospace_size, 0); + if (ret) { + pr_err("send_set_iospace_region failed\n"); + goto error; + } + + hv_set_drvdata(hdev, dxgglobal); + + if (alloc_chrdev_region(&dxgglobal->device_devt, 0, 1, "dxgkrnl") < 0) { + pr_err("alloc_chrdev_region failed\n"); + ret = -ENODEV; + goto error; + } + dxgglobal->devt_initialized = true; + dxgglobal->device_class = class_create(THIS_MODULE, "dxgkdrv"); + if (dxgglobal->device_class == NULL) { + pr_err("class_create failed\n"); + ret = -ENODEV; + goto error; + } + dxgglobal->device_class->devnode = dxg_devnode; + dxgglobal->device = device_create(dxgglobal->device_class, NULL, + dxgglobal->device_devt, NULL, "dxg"); + if (dxgglobal->device == NULL) { + pr_err("device_create failed\n"); + ret = -ENODEV; + goto error; + } + dxgglobaldev = dxgglobal->device; + cdev_init(&dxgglobal->device_cdev, &dxgk_fops); + ret = cdev_add(&dxgglobal->device_cdev, dxgglobal->device_devt, 1); + if (ret < 0) { + pr_err("cdev_add failed: %d\n", ret); + goto error; + } + dxgglobal->cdev_initialized = true; + +error: + return ret; +} + +static void dxgglobal_destroy_global_channel(void) +{ + dxglockorder_acquire(DXGLOCK_GLOBAL_CHANNEL); + down_write(&dxgglobal->channel_lock); + + TRACE_DEBUG(1, "%s", __func__); + + if (dxgglobal->devt_initialized) { + unregister_chrdev_region(dxgglobal->device_devt, 1); + dxgglobal->devt_initialized = false; + } + if (dxgglobal->cdev_initialized) { + cdev_del(&dxgglobal->device_cdev); + dxgglobal->cdev_initialized = false; + } + if (dxgglobal->device) { + device_destroy(dxgglobal->device_class, dxgglobal->device_devt); + dxgglobal->device = NULL; + dxgglobaldev = NULL; + } + if (dxgglobal->device_class) { + class_destroy(dxgglobal->device_class); + dxgglobal->device_class = NULL; + } + if (dxgglobal->mem) { + vmbus_free_mmio(dxgglobal->mmiospace_base, + dxgglobal->mmiospace_size); + dxgglobal->mem = NULL; + } + + dxgvmbuschannel_destroy(&dxgglobal->channel); + + if (dxgglobal->hdev) { + hv_set_drvdata(dxgglobal->hdev, NULL); + dxgglobal->hdev = NULL; + } + + TRACE_DEBUG(1, "%s end\n", __func__); + + up_write(&dxgglobal->channel_lock); + dxglockorder_release(DXGLOCK_GLOBAL_CHANNEL); +} + +static int dxgglobal_create_adapter(struct hv_device *hdev) +{ + struct dxgadapter *adapter; + int ret; + + TRACE_DEBUG(1, "%s", __func__); + + adapter = dxgmem_alloc(NULL, DXGMEM_ADAPTER, sizeof(struct dxgadapter)); + if (adapter == NULL) { + pr_err("failed to allocated dxgadapter\n"); + return STATUS_NO_MEMORY; + } + + ret = dxgadapter_init(adapter, hdev); + if (ret) { + dxgadapter_stop(adapter); + dxgadapter_release_reference(adapter); + } else { + dxgglobal_acquire_adapter_list_lock(DXGLOCK_EXCL); + + TRACE_DEBUG(1, "new adapter added %p\n", adapter); + + list_add_tail(&adapter->adapter_list_entry, + &dxgglobal->adapter_list_head); + dxgglobal->num_adapters++; + + dxgglobal_release_adapter_list_lock(DXGLOCK_EXCL); + } + + TRACE_DEBUG(1, "%s end: %d\n", __func__, ret); + return ret; +} + +static void dxgglobal_stop_adapter(struct hv_device *hdev) +{ + struct dxgadapter *adapter = NULL; + struct dxgadapter *entry; + struct winluid luid; + + guid_to_luid(&hdev->channel->offermsg.offer.if_instance, &luid); + + TRACE_DEBUG(1, "%s: %x:%x\n", __func__, luid.b, luid.a); + + dxgglobal_acquire_adapter_list_lock(DXGLOCK_EXCL); + + list_for_each_entry(entry, &dxgglobal->adapter_list_head, + adapter_list_entry) { + if (*(u64 *) &luid == *(u64 *) &entry->luid) { + adapter = entry; + break; + } + } + + if (adapter) + list_del(&adapter->adapter_list_entry); + + dxgglobal_release_adapter_list_lock(DXGLOCK_EXCL); + + if (adapter) { + dxgadapter_stop(adapter); + dxgadapter_release_reference(adapter); + } else { + pr_err("Adapter is not found\n"); + } + + TRACE_DEBUG(1, "%s end", __func__); +} + +static const struct hv_vmbus_device_id id_table[] = { + /* Per GPU Device GUID */ + { HV_GPUP_DXGK_VGPU_GUID }, + /* Global Dxgkgnl channel for the virtual machine */ + { HV_GPUP_DXGK_GLOBAL_GUID }, + { } +}; + +static int dxg_probe_device(struct hv_device *hdev, + const struct hv_vmbus_device_id *dev_id) +{ + int ret = 0; + struct dxgthreadinfo *thread = dxglockorder_get_thread(); + + dxgmutex_lock(&dxgglobal->device_mutex); + + TRACE_DEBUG(1, "probe_device\n"); + + if (uuid_le_cmp(hdev->dev_type, id_table[0].guid) == 0) { + /* This is a new virtual GPU channel */ + ret = dxgglobal_create_adapter(hdev); + } else if (uuid_le_cmp(hdev->dev_type, id_table[1].guid) == 0) { + /* This is the global Dxgkgnl channel */ + ret = dxgglobal_init_global_channel(hdev); + if (ret) { + dxgglobal_destroy_global_channel(); + goto error; + } + } else { + /* Unknown device type */ + pr_err("probe: unknown device type\n"); + ret = -EBADE; + goto error; + } + +error: + + TRACE_DEBUG(1, "probe_device end\n"); + + dxgmutex_unlock(&dxgglobal->device_mutex); + + dxglockorder_put_thread(thread); + + return ret; +} + +static int dxg_remove_device(struct hv_device *hdev) +{ + int ret = 0; + struct dxgthreadinfo *thread; + + TRACE_DEBUG(1, "%s\n", __func__); + + thread = dxglockorder_get_thread(); + dxgmutex_lock(&dxgglobal->device_mutex); + + if (uuid_le_cmp(hdev->dev_type, id_table[0].guid) == 0) { + TRACE_DEBUG(1, "Remove virtual GPU\n"); + dxgglobal_stop_adapter(hdev); + } else if (uuid_le_cmp(hdev->dev_type, id_table[1].guid) == 0) { + TRACE_DEBUG(1, "Remove global channel device\n"); + dxgglobal_destroy_global_channel(); + } else { + /* Unknown device type */ + pr_err("remove: unknown device type\n"); + ret = -EBADE; + } + + dxgmutex_unlock(&dxgglobal->device_mutex); + dxglockorder_put_thread(thread); + return ret; +} + +MODULE_DEVICE_TABLE(vmbus, id_table); + +static struct hv_driver dxg_drv = { + .name = KBUILD_MODNAME, + .id_table = id_table, + .probe = dxg_probe_device, + .remove = dxg_remove_device, + .driver = { + .probe_type = PROBE_PREFER_ASYNCHRONOUS, + }, +}; + +/* + * Interface with Linux kernel + */ + +static int dxgglobal_create(void) +{ + int ret = 0; + + TRACE_DEBUG(1, "%s", __func__); + + dxgglobal = dxgmem_alloc(NULL, DXGMEM_GLOBAL, sizeof(struct dxgglobal)); + if (!dxgglobal) { + pr_err("no memory for dxgglobal\n"); + return STATUS_NO_MEMORY; + } + + INIT_LIST_HEAD(&dxgglobal->plisthead); + dxgmutex_init(&dxgglobal->plistmutex, DXGLOCK_PROCESSLIST); + dxgmutex_init(&dxgglobal->device_mutex, DXGLOCK_GLOBAL_DEVICE); + dxgmutex_init(&dxgglobal->process_adapter_mutex, + DXGLOCK_PROCESSADAPTER); + + INIT_LIST_HEAD(&dxgglobal->thread_info_list_head); + mutex_init(&dxgglobal->thread_info_mutex); + + INIT_LIST_HEAD(&dxgglobal->adapter_list_head); + init_rwsem(&dxgglobal->adapter_list_lock); + + init_rwsem(&dxgglobal->channel_lock); + + INIT_LIST_HEAD(&dxgglobal->host_event_list_head); + spin_lock_init(&dxgglobal->host_event_list_mutex); + atomic64_set(&dxgglobal->host_event_id, 1); + + hmgrtable_init(&dxgglobal->handle_table, NULL); + + TRACE_DEBUG(1, "dxgglobal_init end\n"); + return ret; +} + +static void dxgglobal_destroy(void) +{ + if (dxgglobal) { + TRACE_DEBUG(1, "%s\n", __func__); + + if (dxgglobal->vmbus_registered) + vmbus_driver_unregister(&dxg_drv); + + dxgglobal_destroy_global_channel(); + hmgrtable_destroy(&dxgglobal->handle_table); + + dxgmem_free(NULL, DXGMEM_GLOBAL, dxgglobal); + dxgglobal = NULL; + TRACE_DEBUG(1, "%s end\n", __func__); + } +} + +static int __init dxg_drv_init(void) +{ + int ret; + + pr_err("%s Version: %x", __func__, DXGKRNL_VERSION); + + ret = dxgglobal_create(); + if (ret) { + pr_err("dxgglobal_init failed"); + return STATUS_NO_MEMORY; + } + + ret = vmbus_driver_register(&dxg_drv); + if (ret) { + pr_err("vmbus_driver_register failed: %d", ret); + return ret; + } + dxgglobal->vmbus_registered = true; + + ioctl_desc_init(); + + return 0; +} + +static void __exit dxg_drv_exit(void) +{ + struct dxgthreadinfo *thread; + + TRACE_DEBUG(1, "%s\n", __func__); + + thread = dxglockorder_get_thread(); + dxgglobal_destroy(); + thread->lock_held = true; /* No need to acquire internal locks */ + dxglockorder_put_thread(thread); + dxgmem_check(NULL, DXGMEM_GLOBAL); + + TRACE_DEBUG(1, "%s end\n", __func__); +} + +module_init(dxg_drv_init); +module_exit(dxg_drv_exit); + +MODULE_LICENSE("GPL"); +MODULE_DESCRIPTION("Microsoft Dxgkrnl virtual GPU Driver"); diff --git a/drivers/gpu/dxgkrnl/dxgprocess.c b/drivers/gpu/dxgkrnl/dxgprocess.c new file mode 100644 index 000000000000..fae2076631fd --- /dev/null +++ b/drivers/gpu/dxgkrnl/dxgprocess.c @@ -0,0 +1,355 @@ +// SPDX-License-Identifier: GPL-2.0 + +/* + * Copyright (c) 2019, Microsoft Corporation. + * + * Author: + * Iouri Tarassov iourit@microsoft.com + * + * Dxgkrnl Graphics Port Driver + * DXGPROCSS implementation + * + */ + +#include "dxgkrnl.h" + +/* + * Creates a new dxgprocess object + * Must be called when dxgglobal->plistmutex is held + */ +struct dxgprocess *dxgprocess_create(void) +{ + struct dxgprocess *process; + int ret; + + TRACE_DEBUG(1, "%s", __func__); + + process = dxgmem_alloc(NULL, DXGMEM_PROCESS, sizeof(struct dxgprocess)); + if (process == NULL) { + pr_err("failed to allocate dxgprocess\n"); + } else { + TRACE_DEBUG(1, "new dxgprocess created\n"); + process->process = current; + process->pid = current->pid; + process->tgid = current->tgid; + dxgmutex_init(&process->process_mutex, DXGLOCK_PROCESSMUTEX); + ret = dxgvmb_send_create_process(process); + if (ret) { + TRACE_DEBUG(1, "dxgvmb_send_create_process failed\n"); + dxgmem_free(NULL, DXGMEM_PROCESS, process); + process = NULL; + } else { + INIT_LIST_HEAD(&process->plistentry); + process->refcount = 1; + + dxgmutex_lock(&dxgglobal->plistmutex); + list_add_tail(&process->plistentry, + &dxgglobal->plisthead); + dxgmutex_unlock(&dxgglobal->plistmutex); + + hmgrtable_init(&process->handle_table, process); + hmgrtable_init(&process->local_handle_table, process); + INIT_LIST_HEAD(&process->process_adapter_list_head); + } + } + return process; +} + +void dxgprocess_destroy(struct dxgprocess *process) +{ + int i; + enum hmgrentry_type t; + d3dkmt_handle h; + void *o; + struct dxgsyncobject *syncobj; + struct dxgprocess_adapter *entry; + struct dxgprocess_adapter *tmp; + struct dxgadapter *adapter; + + TRACE_DEBUG(1, "%s", __func__); + + /* Destroy all adapter state */ + dxgglobal_acquire_process_adapter_lock(); + list_for_each_entry_safe(entry, tmp, + &process->process_adapter_list_head, + process_adapter_list_entry) { + adapter = entry->adapter; + dxgprocess_adapter_destroy(entry); + } + dxgglobal_release_process_adapter_lock(); + + i = 0; + while (hmgrtable_next_entry(&process->local_handle_table, + &i, &t, &h, &o)) { + switch (t) { + case HMGRENTRY_TYPE_DXGADAPTER: + dxgprocess_close_adapter(process, h); + break; + default: + pr_err("invalid entry in local handle table %d", t); + break; + } + } + + i = 0; + while (hmgrtable_next_entry(&process->handle_table, &i, &t, &h, &o)) { + switch (t) { + case HMGRENTRY_TYPE_DXGSYNCOBJECT: + TRACE_DEBUG(1, "Destroy syncobj: %p %d", o, i); + syncobj = o; + syncobj->handle = 0; + dxgsyncobject_destroy(process, syncobj); + break; + default: + pr_err("invalid entry in handle table %d", t); + break; + } + } + + hmgrtable_destroy(&process->handle_table); + hmgrtable_destroy(&process->local_handle_table); + + for (i = 0; i < 2; i++) { + if (process->test_handle_table[i]) { + hmgrtable_destroy(process->test_handle_table[i]); + dxgmem_free(process, DXGMEM_HANDLE_TABLE, + process->test_handle_table[i]); + process->test_handle_table[i] = NULL; + } + } + + TRACE_DEBUG(1, "%s end", __func__); +} + +/* + * Release reference count on a process object + */ +void dxgprocess_release_reference(struct dxgprocess *process) +{ + TRACE_DEBUG(1, "%s %d", __func__, process->refcount); + dxgmutex_lock(&dxgglobal->plistmutex); + process->refcount--; + if (process->refcount == 0) { + list_del(&process->plistentry); + process->plistentry.next = NULL; + dxgmutex_unlock(&dxgglobal->plistmutex); + + dxgprocess_destroy(process); + + if (process->host_handle) + dxgvmb_send_destroy_process(process->host_handle); + dxgmem_check(process, DXGMEM_PROCESS); + dxgmem_free(NULL, DXGMEM_PROCESS, process); + } else { + dxgmutex_unlock(&dxgglobal->plistmutex); + } +} + +struct dxgprocess_adapter *dxgprocess_get_adapter_info(struct dxgprocess + *process, + struct dxgadapter + *adapter) +{ + struct dxgprocess_adapter *entry; + + list_for_each_entry(entry, &process->process_adapter_list_head, + process_adapter_list_entry) { + if (adapter == entry->adapter) { + TRACE_DEBUG(1, "found process adapter info %p", entry); + return entry; + } + } + return NULL; +} + +/* + * Dxgprocess takes references on dxgadapter and dxgprocess_adapter. + */ +int dxgprocess_open_adapter(struct dxgprocess *process, + struct dxgadapter *adapter, d3dkmt_handle *h) +{ + int ret = 0; + struct dxgprocess_adapter *adapter_info; + d3dkmt_handle handle; + + *h = 0; + adapter_info = dxgprocess_get_adapter_info(process, adapter); + if (adapter_info == NULL) { + TRACE_DEBUG(1, "creating new process adapter info\n"); + adapter_info = dxgprocess_adapter_create(process, adapter); + if (adapter_info == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + } else { + adapter_info->refcount++; + } + + handle = hmgrtable_alloc_handle_safe(&process->local_handle_table, + adapter, HMGRENTRY_TYPE_DXGADAPTER, + true); + if (handle) { + *h = handle; + } else { + pr_err("failed to create adapter handle\n"); + ret = STATUS_NO_MEMORY; + goto cleanup; + } + +cleanup: + + if (ret) { + if (adapter_info) { + dxgglobal_acquire_process_adapter_lock(); + dxgprocess_adapter_release(adapter_info); + dxgglobal_release_process_adapter_lock(); + } + } + + return ret; +} + +int dxgprocess_close_adapter(struct dxgprocess *process, d3dkmt_handle handle) +{ + struct dxgadapter *adapter; + struct dxgprocess_adapter *adapter_info; + int ret = 0; + + if (handle == 0) + return 0; + + hmgrtable_lock(&process->local_handle_table, DXGLOCK_EXCL); + adapter = dxgprocess_get_adapter(process, handle); + if (adapter) + hmgrtable_free_handle(&process->local_handle_table, + HMGRENTRY_TYPE_DXGADAPTER, handle); + hmgrtable_unlock(&process->local_handle_table, DXGLOCK_EXCL); + + if (adapter) { + adapter_info = dxgprocess_get_adapter_info(process, adapter); + if (adapter_info) { + dxgglobal_acquire_process_adapter_lock(); + dxgprocess_adapter_release(adapter_info); + dxgglobal_release_process_adapter_lock(); + } else { + ret = STATUS_INVALID_PARAMETER; + } + } else { + pr_err("%s failed %x", __func__, handle); + ret = STATUS_INVALID_PARAMETER; + } + + return ret; +} + +struct dxgadapter *dxgprocess_get_adapter(struct dxgprocess *process, + d3dkmt_handle handle) +{ + struct dxgadapter *adapter; + + adapter = hmgrtable_get_object_by_type(&process->local_handle_table, + HMGRENTRY_TYPE_DXGADAPTER, + handle); + if (adapter == NULL) + pr_err("%s failed %x\n", __func__, handle); + return adapter; +} + +/* + * Gets the adapter object from the process handle table. + * The adapter object is referenced. + * The function acquired the handle table lock shared. + */ +struct dxgadapter *dxgprocess_adapter_by_handle(struct dxgprocess *process, + d3dkmt_handle handle) +{ + struct dxgadapter *adapter; + + hmgrtable_lock(&process->local_handle_table, DXGLOCK_SHARED); + adapter = hmgrtable_get_object_by_type(&process->local_handle_table, + HMGRENTRY_TYPE_DXGADAPTER, + handle); + if (adapter == NULL) + pr_err("adapter_by_handle failed %x\n", handle); + else if (!dxgadapter_acquire_reference(adapter)) { + pr_err("failed to acquire adapter reference\n"); + adapter = NULL; + } + hmgrtable_unlock(&process->local_handle_table, DXGLOCK_SHARED); + return adapter; +} + +struct dxgdevice *dxgprocess_device_by_object_handle(struct dxgprocess *process, + enum hmgrentry_type t, + d3dkmt_handle handle) +{ + struct dxgdevice *device = NULL; + void *obj; + + hmgrtable_lock(&process->handle_table, DXGLOCK_SHARED); + obj = hmgrtable_get_object_by_type(&process->handle_table, t, handle); + if (obj) { + d3dkmt_handle device_handle = 0; + + switch (t) { + case HMGRENTRY_TYPE_DXGDEVICE: + device = obj; + break; + case HMGRENTRY_TYPE_DXGCONTEXT: + device_handle = + ((struct dxgcontext *)obj)->device_handle; + break; + case HMGRENTRY_TYPE_DXGPAGINGQUEUE: + device_handle = + ((struct dxgpagingqueue *)obj)->device_handle; + break; + case HMGRENTRY_TYPE_DXGHWQUEUE: + device_handle = + ((struct dxghwqueue *)obj)->device_handle; + break; + default: + pr_err("invalid handle type: %d\n", t); + break; + } + if (device == NULL) + device = hmgrtable_get_object_by_type( + &process->handle_table, + HMGRENTRY_TYPE_DXGDEVICE, + device_handle); + if (device) + if (!dxgdevice_acquire_reference(device)) + device = NULL; + } + if (device == NULL) + pr_err("device_by_handle failed: %d %x\n", t, handle); + hmgrtable_unlock(&process->handle_table, DXGLOCK_SHARED); + return device; +} + +struct dxgdevice *dxgprocess_device_by_handle(struct dxgprocess *process, + d3dkmt_handle handle) +{ + return dxgprocess_device_by_object_handle(process, + HMGRENTRY_TYPE_DXGDEVICE, + handle); +} + +void dxgprocess_ht_lock_shared_down(struct dxgprocess *process) +{ + hmgrtable_lock(&process->handle_table, DXGLOCK_SHARED); +} + +void dxgprocess_ht_lock_shared_up(struct dxgprocess *process) +{ + hmgrtable_unlock(&process->handle_table, DXGLOCK_SHARED); +} + +void dxgprocess_ht_lock_exclusive_down(struct dxgprocess *process) +{ + hmgrtable_lock(&process->handle_table, DXGLOCK_EXCL); +} + +void dxgprocess_ht_lock_exclusive_up(struct dxgprocess *process) +{ + hmgrtable_unlock(&process->handle_table, DXGLOCK_EXCL); +} diff --git a/drivers/gpu/dxgkrnl/dxgvmbus.c b/drivers/gpu/dxgkrnl/dxgvmbus.c new file mode 100644 index 000000000000..f903d659c237 --- /dev/null +++ b/drivers/gpu/dxgkrnl/dxgvmbus.c @@ -0,0 +1,2955 @@ +// SPDX-License-Identifier: GPL-2.0 + +/* + * Copyright (c) 2019, Microsoft Corporation. + * + * Author: + * Iouri Tarassov iourit@microsoft.com + * + * Dxgkrnl Graphics Port Driver + * VM bus interface implementation + * + */ + +#include <linux/kernel.h> +#include <linux/completion.h> +#include <linux/slab.h> +#include <linux/eventfd.h> +#include <linux/hyperv.h> +#include <linux/mman.h> + +#include "dxgkrnl.h" +#include "dxgvmbus.h" + +/* + * The interface version is used to ensure that the host and the guest use the + * same VM bus protocol. It needs to be incremented every time the VM bus + * interface changes. DXGK_VMBUS_LAST_COMPATIBLE_INTERFACE_VERSION is + * incremented each time the earlier versions of the interface are no longer + * compatible with the current version. + */ +const uint DXGK_VMBUS_INTERFACE_VERSION = 27; +const uint DXGK_VMBUS_LAST_COMPATIBLE_INTERFACE_VERSION = 16; + +#define RING_BUFSIZE (256 * 1024) + +struct dxgvmbuspacket { + struct list_head packet_list_entry; + u64 request_id; + struct completion wait; + void *buffer; + u32 buffer_length; + int status; +}; + +int dxgvmbuschannel_init(struct dxgvmbuschannel *ch, struct hv_device *hdev) +{ + int ret; + + ch->hdev = hdev; + spin_lock_init(&ch->packet_list_mutex); + INIT_LIST_HEAD(&ch->packet_list_head); + atomic64_set(&ch->packet_request_id, 0); + + ch->packet_cache = kmem_cache_create("DXGK packet cache", + sizeof(struct dxgvmbuspacket), 0, + 0, NULL); + if (ch->packet_cache == NULL) { + pr_err("packet_cache alloc failed"); + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + ret = vmbus_open(hdev->channel, RING_BUFSIZE, RING_BUFSIZE, + NULL, 0, dxgvmbuschannel_receive, ch); + if (ret) { + pr_err("vmbus_open failed: %d", ret); + goto cleanup; + } + + ch->channel = hdev->channel; + +cleanup: + + return ret; +} + +void dxgvmbuschannel_destroy(struct dxgvmbuschannel *ch) +{ + kmem_cache_destroy(ch->packet_cache); + ch->packet_cache = NULL; + + if (ch->channel) { + vmbus_close(ch->channel); + ch->channel = NULL; + } +} + +static inline void command_vm_to_host_init0(struct dxgkvmb_command_vm_to_host + *command) +{ + command->command_type = DXGK_VMBCOMMAND_INVALID_VM_TO_HOST; + command->process = 0; + command->command_id = 0; + command->channel_type = DXGKVMB_VM_TO_HOST; +} + +static inline void command_vm_to_host_init1(struct dxgkvmb_command_vm_to_host + *command, + enum dxgkvmb_commandtype_global + type) +{ + command->command_type = type; + command->process = 0; + command->command_id = 0; + command->channel_type = DXGKVMB_VM_TO_HOST; +} + +void signal_guest_event(struct dxgkvmb_command_host_to_vm *packet, + u32 packet_length) +{ + struct dxgkvmb_command_signalguestevent *command = (void *)packet; + + TRACE_DEBUG(1, "%s global packet", __func__); + + if (packet_length < sizeof(struct dxgkvmb_command_signalguestevent)) { + pr_err("invalid packet size"); + return; + } + if (command->event == 0) { + pr_err("invalid event pointer"); + return; + } + dxgglobal_signal_host_event(command->event); +} + +void process_inband_packet(struct dxgvmbuschannel *channel, + struct vmpacket_descriptor *desc) +{ + u32 packet_length = hv_pkt_datalen(desc); + + if (channel->adapter == NULL) { + if (packet_length < sizeof(struct dxgkvmb_command_host_to_vm)) { + pr_err("Invalid global packet"); + } else { + struct dxgkvmb_command_host_to_vm *packet = + hv_pkt_data(desc); + TRACE_DEBUG(1, "global packet %d", + packet->command_type); + switch (packet->command_type) { + case DXGK_VMBCOMMAND_SETGUESTDATA: + break; + case DXGK_VMBCOMMAND_SIGNALGUESTEVENT: + case DXGK_VMBCOMMAND_SIGNALGUESTEVENTPASSIVE: + signal_guest_event(packet, packet_length); + break; + case DXGK_VMBCOMMAND_SENDWNFNOTIFICATION: + break; + default: + pr_err("unexpected host message %d", + packet->command_type); + } + } + } else { + pr_err("Unexpected packet for adapter channel"); + } +} + +void process_completion_packet(struct dxgvmbuschannel *channel, + struct vmpacket_descriptor *desc) +{ + struct dxgvmbuspacket *packet = NULL; + struct dxgvmbuspacket *entry; + u32 packet_length = hv_pkt_datalen(desc); + unsigned long flags; + + spin_lock_irqsave(&channel->packet_list_mutex, flags); + list_for_each_entry(entry, &channel->packet_list_head, + packet_list_entry) { + if (desc->trans_id == entry->request_id) { + packet = entry; + list_del(&packet->packet_list_entry); + break; + } + } + spin_unlock_irqrestore(&channel->packet_list_mutex, flags); + + if (packet) { + if (packet->buffer_length) { + if (packet_length < packet->buffer_length) { + TRACE_DEBUG(1, "invalid size %d Expected:%d", + packet_length, + packet->buffer_length); + packet->status = STATUS_BUFFER_OVERFLOW; + } else { + memcpy(packet->buffer, hv_pkt_data(desc), + packet->buffer_length); + } + } + complete(&packet->wait); + } else { + pr_err("did not find packet to complete"); + } +} + +/* Receive callback for messages from the host */ +void dxgvmbuschannel_receive(void *ctx) +{ + struct dxgvmbuschannel *channel = ctx; + struct vmpacket_descriptor *desc; + + TRACE_DEBUG(1, "%s %p", __func__, channel->adapter); + foreach_vmbus_pkt(desc, channel->channel) { + TRACE_DEFINE(u32 packet_length = hv_pkt_datalen(desc); + ) + TRACE_DEBUG(1, "next packet (id, size, type): %llu %d %d", + desc->trans_id, packet_length, desc->type); + if (desc->type == VM_PKT_COMP) { + process_completion_packet(channel, desc); + } else { + if (desc->type != VM_PKT_DATA_INBAND) + pr_err("unexpected packet type"); + process_inband_packet(channel, desc); + } + } +} + +int dxgvmb_send_sync_msg(struct dxgvmbuschannel *channel, + void *command, + u32 cmd_size, void *result, u32 result_size) +{ + int ret = 0; + unsigned long t; + struct dxgvmbuspacket *packet = NULL; + + if (cmd_size > DXG_MAX_VM_BUS_PACKET_SIZE || + result_size > DXG_MAX_VM_BUS_PACKET_SIZE) { + pr_err("%s invalid data size", __func__); + return STATUS_INVALID_PARAMETER; + } + + packet = kmem_cache_alloc(channel->packet_cache, 0); + if (packet == NULL) { + pr_err("kmem_cache_alloc failed"); + return STATUS_NO_MEMORY; + } + + if (channel->adapter == NULL) { + TRACE_DEFINE(struct dxgkvmb_command_vm_to_host *cmd = command; + ) + TRACE_DEBUG(1, "send_sync_msg global: %d %p %d %d", + cmd->command_type, command, cmd_size, + result_size); + } else { + TRACE_DEFINE(struct dxgkvmb_command_vgpu_to_host *cmd = command; + ) + TRACE_DEBUG(1, "send_sync_msg adapter: %d %p %d %d", + cmd->command_type, command, cmd_size, + result_size); + } + + packet->request_id = atomic64_inc_return(&channel->packet_request_id); + init_completion(&packet->wait); + packet->buffer = result; + packet->buffer_length = result_size; + packet->status = 0; + spin_lock_irq(&channel->packet_list_mutex); + list_add_tail(&packet->packet_list_entry, &channel->packet_list_head); + spin_unlock_irq(&channel->packet_list_mutex); + + ret = vmbus_sendpacket(channel->channel, command, cmd_size, + packet->request_id, VM_PKT_DATA_INBAND, + VMBUS_DATA_PACKET_FLAG_COMPLETION_REQUESTED); + if (ret) { + pr_err("vmbus_sendpacket failed"); + goto cleanup; + } + + TRACE_DEBUG(1, "waiting for completion: %llu", packet->request_id); + t = wait_for_completion_timeout(&packet->wait, (1000 * HZ)); + if (!t) { + TRACE_DEBUG(1, "timeout waiting for completion"); + ret = STATUS_TIMEOUT; + } else { + TRACE_DEBUG(1, "completion done: %llu %x", + packet->request_id, packet->status); + if (!NT_SUCCESS(packet->status)) + ret = packet->status; + } + +cleanup: + + kmem_cache_free(channel->packet_cache, packet); + return ret; +} + +static ntstatus dxgvmb_send_sync_msg_ntstatus(struct dxgvmbuschannel *channel, + void *command, u32 cmd_size) +{ + ntstatus status = STATUS_SUCCESS; + int ret = dxgvmb_send_sync_msg(channel, command, cmd_size, + &status, sizeof(ntstatus)); + if (ret) + status = STATUS_UNSUCCESSFUL; + return status; +} + +static int check_iospace_address(unsigned long address, uint size) +{ + if (address < dxgglobal->mmiospace_base || + size > dxgglobal->mmiospace_size || + address >= (dxgglobal->mmiospace_base + + dxgglobal->mmiospace_size - size)) { + pr_err("invalid iospace address %lx", address); + return STATUS_INVALID_PARAMETER; + } + return 0; +} + +int dxg_unmap_iospace(void *va, uint size) +{ + int ret = 0; + + TRACE_DEBUG(1, "%s %p %x", __func__, va, size); + + /* + * When an app calls exit(), dxgkrnl is called to close the device + * with current->mm equal to NULL. + */ + if (current->mm) { + ret = vm_munmap((unsigned long)va, size); + if (ret) + pr_err("vm_munmap failed %d", ret); + } + return ret; +} + +static uint8_t *dxg_map_iospace(uint64_t iospace_address, uint size, + unsigned long protection, bool cached) +{ + struct vm_area_struct *vma; + unsigned long va; + int ret = 0; + + TRACE_DEBUG(1, "%s: %llx %x %lx", + __func__, iospace_address, size, protection); + if (check_iospace_address(iospace_address, size)) { + pr_err("%s: invalid address", __func__); + return NULL; + } + + va = vm_mmap(NULL, 0, size, protection, MAP_SHARED | MAP_ANONYMOUS, 0); + if ((long)va <= 0) { + pr_err("vm_mmap failed %lx %d", va, size); + return NULL; + } + + down_read(¤t->mm->mmap_sem); + vma = find_vma(current->mm, (unsigned long)va); + if (vma) { + pgprot_t prot = vma->vm_page_prot; + + if (!cached) + prot = pgprot_writecombine(prot); + TRACE_DEBUG(1, "vma: %lx %lx %lx", + vma->vm_start, vma->vm_end, va); + vma->vm_pgoff = iospace_address >> PAGE_SHIFT; + ret = io_remap_pfn_range(vma, vma->vm_start, vma->vm_pgoff, + size, prot); + if (ret) + pr_err("io_remap_pfn_range failed: %d", ret); + } else { + pr_err("failed to find vma: %p %lx", vma, va); + ret = STATUS_NO_MEMORY; + } + up_read(¤t->mm->mmap_sem); + + if (ret) { + dxg_unmap_iospace((void *)va, size); + return NULL; + } + TRACE_DEBUG(1, "%s end: %lx", __func__, va); + return (uint8_t *) va; +} + +/* + * Messages to the host + */ + +int dxgvmb_send_set_iospace_region(u64 start, u64 len, u32 shared_mem_gpadl) +{ + ntstatus status; + struct dxgkvmb_command_setiospaceregion command = { }; + int ret = dxgglobal_acquire_channel_lock(); + + if (ret) + goto cleanup; + + command_vm_to_host_init1(&command.hdr, + DXGK_VMBCOMMAND_SETIOSPACEREGION); + command.start = start; + command.length = len; + command.shared_page_gpadl = shared_mem_gpadl; + status = dxgvmb_send_sync_msg_ntstatus(&dxgglobal->channel, &command, + sizeof(command)); + if (!NT_SUCCESS(status)) { + pr_err("send_set_iospace_region failed %x", status); + ret = status; + } + + dxgglobal_release_channel_lock(); +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_create_process(struct dxgprocess *process) +{ + int ret = 0; + struct dxgkvmb_command_createprocess command = { 0 }; + struct dxgkvmb_command_createprocess_return result = { 0 }; + char s[W_MAX_PATH]; + int i; + + ret = dxgglobal_acquire_channel_lock(); + if (ret) + goto cleanup; + + TRACE_DEBUG(1, "%s", __func__); + command_vm_to_host_init1(&command.hdr, DXGK_VMBCOMMAND_CREATEPROCESS); + command.process = process; + command.process_id = process->process->pid; + command.linux_process = 1; + s[0] = 0; + __get_task_comm(s, W_MAX_PATH, process->process); + for (i = 0; i < W_MAX_PATH; i++) { + command.process_name[i] = s[i]; + if (s[i] == 0) + break; + } + + TRACE_DEBUG(1, "create_process msg %d %d", + command.hdr.command_type, (u32) sizeof(command)); + ret = dxgvmb_send_sync_msg(&dxgglobal->channel, &command, + sizeof(command), &result, sizeof(result)); + if (ret) { + pr_err("create_process failed %d", ret); + } else if (result.hprocess == 0) { + pr_err("create_process returned 0 handle"); + ret = STATUS_INTERNAL_ERROR; + } else { + process->host_handle = result.hprocess; + TRACE_DEBUG(1, "create_process returned %x", + process->host_handle); + } + + dxgglobal_release_channel_lock(); + +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_destroy_process(d3dkmt_handle process) +{ + ntstatus status; + struct dxgkvmb_command_destroyprocess command = { 0 }; + + status = dxgglobal_acquire_channel_lock(); + if (!NT_SUCCESS(status)) + goto cleanup; + command_vm_to_host_init2(&command.hdr, DXGK_VMBCOMMAND_DESTROYPROCESS, + process); + status = dxgvmb_send_sync_msg_ntstatus(&dxgglobal->channel, + &command, sizeof(command)); + dxgglobal_release_channel_lock(); + +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, status); + return status; +} + +int dxgvmb_send_open_adapter(struct dxgadapter *adapter) +{ + int ret; + struct dxgkvmb_command_openadapter command = { }; + struct dxgkvmb_command_openadapter_return result = { }; + + command_vgpu_to_host_init1(&command.hdr, DXGK_VMBCOMMAND_OPENADAPTER); + command.vmbus_interface_version = DXGK_VMBUS_INTERFACE_VERSION, + command.vmbus_last_compatible_interface_version = + DXGK_VMBUS_LAST_COMPATIBLE_INTERFACE_VERSION; + + ret = dxgvmb_send_sync_msg(&adapter->channel, &command, sizeof(command), + &result, sizeof(result)); + if (ret) + goto cleanup; + + ret = result.status; + adapter->host_handle = result.host_adapter_handle; + +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_close_adapter(struct dxgadapter *adapter) +{ + int ret = 0; + struct dxgkvmb_command_closeadapter command; + + command_vgpu_to_host_init1(&command.hdr, DXGK_VMBCOMMAND_CLOSEADAPTER); + command.host_handle = adapter->host_handle; + + ret = dxgvmb_send_sync_msg(&adapter->channel, &command, sizeof(command), + NULL, 0); + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_get_internal_adapter_info(struct dxgadapter *adapter) +{ + int ret = 0; + struct dxgkvmb_command_getinternaladapterinfo command = { }; + struct dxgkvmb_command_getinternaladapterinfo_return result = { }; + + command_vgpu_to_host_init1(&command.hdr, + DXGK_VMBCOMMAND_GETINTERNALADAPTERINFO); + + ret = dxgvmb_send_sync_msg(&adapter->channel, &command, sizeof(command), + &result, sizeof(result)); + if (NT_SUCCESS(ret)) { + adapter->host_adapter_luid = result.host_adapter_luid; + wcsncpy(adapter->device_description, result.device_description, + sizeof(adapter->device_description) / sizeof(winwchar)); + wcsncpy(adapter->device_instance_id, result.device_instance_id, + sizeof(adapter->device_instance_id) / sizeof(winwchar)); + } + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +d3dkmt_handle dxgvmb_send_create_device(struct dxgadapter *adapter, + struct dxgprocess *process, + struct d3dkmt_createdevice *args) +{ + int ret; + struct dxgkvmb_command_createdevice command = { }; + uint cmd_size = sizeof(command); + struct dxgkvmb_command_createdevice_return result = { }; + + command_vgpu_to_host_init2(&command.hdr, DXGK_VMBCOMMAND_CREATEDEVICE, + process->host_handle); + command.flags = args->flags; + ret = dxgvmb_send_sync_msg(&adapter->channel, &command, cmd_size, + &result, sizeof(result)); + if (ret) + result.device = 0; + TRACE_FUNC_EXIT_ERR(__func__, ret); + return result.device; +} + +int dxgvmb_send_destroy_device(struct dxgadapter *adapter, + struct dxgprocess *process, d3dkmt_handle h) +{ + ntstatus status; + struct dxgkvmb_command_destroydevice command = { }; + uint cmd_size = sizeof(command); + + command_vgpu_to_host_init2(&command.hdr, DXGK_VMBCOMMAND_DESTROYDEVICE, + process->host_handle); + command.device = h; + + status = dxgvmb_send_sync_msg_ntstatus(&adapter->channel, &command, + cmd_size); + TRACE_FUNC_EXIT_ERR(__func__, status); + return status; +} + +d3dkmt_handle dxgvmb_send_create_context(struct dxgadapter *adapter, + struct dxgprocess *process, + struct d3dkmt_createcontextvirtual + *args) +{ + struct dxgkvmb_command_createcontextvirtual *command = NULL; + uint cmd_size; + int ret; + d3dkmt_handle context = 0; + + if (args->priv_drv_data_size > DXG_MAX_VM_BUS_PACKET_SIZE) { + pr_err("PrivateDriverDataSize is invalid"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + cmd_size = sizeof(struct dxgkvmb_command_createcontextvirtual) + + args->priv_drv_data_size - 1; + command = dxgmem_alloc(process, DXGMEM_VMBUS, cmd_size); + if (command == NULL) { + pr_err("failed to allocate memory for command"); + ret = STATUS_NO_MEMORY; + goto cleanup; + } + command_vgpu_to_host_init2(&command->hdr, + DXGK_VMBCOMMAND_CREATECONTEXTVIRTUAL, + process->host_handle); + command->device = args->device; + command->node_ordinal = args->node_ordinal; + command->engine_affinity = args->engine_affinity; + command->flags = args->flags; + command->client_hint = args->client_hint; + command->priv_drv_data_size = args->priv_drv_data_size; + if (args->priv_drv_data_size) { + ret = dxg_copy_from_user(command->priv_drv_data, + args->priv_drv_data, + args->priv_drv_data_size); + if (ret) + goto cleanup; + } + /* Input command is returned back as output */ + ret = dxgvmb_send_sync_msg(&adapter->channel, command, cmd_size, + command, cmd_size); + if (ret) { + goto cleanup; + } else { + context = command->context; + if (args->priv_drv_data_size) { + ret = dxg_copy_to_user(args->priv_drv_data, + command->priv_drv_data, + args->priv_drv_data_size); + if (ret) { + dxgvmb_send_destroy_context(adapter, process, + context); + context = 0; + } + } + } + +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, ret); + if (command) + dxgmem_free(process, DXGMEM_VMBUS, command); + return context; +} + +int dxgvmb_send_destroy_context(struct dxgadapter *adapter, + struct dxgprocess *process, d3dkmt_handle h) +{ + ntstatus status; + struct dxgkvmb_command_destroycontext command = { }; + uint cmd_size = sizeof(command); + + command_vgpu_to_host_init2(&command.hdr, DXGK_VMBCOMMAND_DESTROYCONTEXT, + process->host_handle); + command.context = h; + + status = dxgvmb_send_sync_msg_ntstatus(&adapter->channel, &command, + cmd_size); + TRACE_FUNC_EXIT_ERR(__func__, status); + return status; +} + +int dxgvmb_send_create_paging_queue(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct dxgdevice *device, + struct d3dkmt_createpagingqueue *args, + struct dxgpagingqueue *pqueue) +{ + struct dxgkvmb_command_createpagingqueue_return result; + struct dxgkvmb_command_createpagingqueue command; + int ret; + + command_vgpu_to_host_init2(&command.hdr, + DXGK_VMBCOMMAND_CREATEPAGINGQUEUE, + process->host_handle); + command.args = *args; + args->paging_queue = 0; + + ret = dxgvmb_send_sync_msg(channel, &command, sizeof(command), &result, + sizeof(result)); + if (ret) { + pr_err("send_create_paging_queue failed %x", ret); + goto cleanup; + } + + args->paging_queue = result.paging_queue; + args->sync_object = result.sync_object; + args->fence_cpu_virtual_address = + dxg_map_iospace(result.fence_storage_physical_address, PAGE_SIZE, + PROT_READ | PROT_WRITE, true); + if (args->fence_cpu_virtual_address == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + pqueue->mapped_address = args->fence_cpu_virtual_address; + pqueue->handle = args->paging_queue; + +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_destroy_paging_queue(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + d3dkmt_handle h) +{ + int ret; + struct dxgkvmb_command_destroypagingqueue command; + uint cmd_size = sizeof(command); + + ret = dxgglobal_acquire_channel_lock(); + if (ret) + goto cleanup; + + command_vgpu_to_host_init2(&command.hdr, + DXGK_VMBCOMMAND_DESTROYPAGINGQUEUE, + process->host_handle); + command.paging_queue = h; + + ret = dxgvmb_send_sync_msg(channel, &command, cmd_size, NULL, 0); + + dxgglobal_release_channel_lock(); + +cleanup: + + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +static int copy_private_data(struct d3dkmt_createallocation *args, + struct dxgkvmb_command_createallocation *command, + struct d3dddi_allocationinfo2 *input_alloc_info, + struct d3dkmt_createstandardallocation + *standard_alloc) +{ + struct dxgkvmb_command_createallocation_allocinfo *alloc_info; + struct d3dddi_allocationinfo2 *input_alloc; + int ret = 0; + int i; + uint8_t *private_data_dest = (uint8_t *) &command[1] + + (args->alloc_count * + sizeof(struct dxgkvmb_command_createallocation_allocinfo)); + + if (args->private_runtime_data_size) { + ret = dxg_copy_from_user(private_data_dest, + args->private_runtime_data, + args->private_runtime_data_size); + if (ret) + goto cleanup; + private_data_dest += args->private_runtime_data_size; + } + + if (args->flags.standard_allocation) { + TRACE_DEBUG2(1, 1, "private data offset %d", + (uint) (private_data_dest - (uint8_t *) command)); + + args->priv_drv_data_size = sizeof(*args->standard_allocation); + memcpy(private_data_dest, standard_alloc, + sizeof(*standard_alloc)); + private_data_dest += args->priv_drv_data_size; + } else if (args->priv_drv_data_size) { + ret = dxg_copy_from_user(private_data_dest, + args->priv_drv_data, + args->priv_drv_data_size); + if (ret) + goto cleanup; + private_data_dest += args->priv_drv_data_size; + } + + alloc_info = (void *)&command[1]; + input_alloc = input_alloc_info; + if (input_alloc_info[0].sysmem) + command->flags.existing_sysmem = 1; + for (i = 0; i < args->alloc_count; i++) { + alloc_info->flags = input_alloc->flags.value; + alloc_info->vidpn_source_id = input_alloc->vidpn_source_id; + alloc_info->priv_drv_data_size = + input_alloc->priv_drv_data_size; + if (input_alloc->priv_drv_data_size) { + ret = dxg_copy_from_user(private_data_dest, + input_alloc->priv_drv_data, + input_alloc-> + priv_drv_data_size); + if (ret) + goto cleanup; + private_data_dest += input_alloc->priv_drv_data_size; + } + alloc_info++; + input_alloc++; + } + +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int create_existing_sysmem(struct dxgdevice *device, + struct dxgkvmb_command_allocinfo_return *host_alloc, + struct dxgallocation *dxgalloc, + bool read_only, + const void *sysmem) +{ + int ret = 0; + void *kmem = NULL; + struct dxgkvmb_command_setexistingsysmemstore set_store_command = { }; + u64 alloc_size = host_alloc->allocation_size; + uint npages = alloc_size >> PAGE_SHIFT; + struct dxgvmbuschannel *channel = &device->adapter->channel; + + /* + * Create a guest physical address list and set it as the allocation + * backing store in the host. This is done after creating the host + * allocation, because only now the allocation size is known. + */ + + TRACE_DEBUG(2, "alloc size: %lld", alloc_size); + + dxgalloc->cpu_address = (void *)sysmem; + dxgalloc->pages = dxgmem_alloc(dxgalloc->process, DXGMEM_ALLOCATION, + npages * sizeof(void *)); + if (dxgalloc->pages == NULL) { + pr_err("failed to allocate pages"); + ret = STATUS_NO_MEMORY; + goto cleanup; + } + ret = get_user_pages_fast((unsigned long)sysmem, npages, !read_only, + dxgalloc->pages); + if (ret != npages) { + pr_err("get_user_pages_fast failed: %d", ret); + ret = STATUS_NO_MEMORY; + goto cleanup; + } + kmem = vmap(dxgalloc->pages, npages, VM_MAP, PAGE_KERNEL); + if (kmem == NULL) { + pr_err("vmap failed"); + ret = STATUS_NO_MEMORY; + goto cleanup; + } + ret = vmbus_establish_gpadl(dxgglobal_get_vmbus(), kmem, + alloc_size, &dxgalloc->gpadl); + if (ret) { + pr_err("establish_gpadl failed: %d", ret); + ret = STATUS_NO_MEMORY; + goto cleanup; + } + TRACE_DEBUG(1, "New gpadl %d", dxgalloc->gpadl); + + command_vgpu_to_host_init2(&set_store_command.hdr, + DXGK_VMBCOMMAND_SETEXISTINGSYSMEMSTORE, + device->process->host_handle); + set_store_command.device = device->handle; + set_store_command.allocation = host_alloc->allocation; + set_store_command.gpadl = dxgalloc->gpadl; + ret = dxgvmb_send_sync_msg_ntstatus(channel, &set_store_command, + sizeof(set_store_command)); + if (!NT_SUCCESS(ret)) { + pr_err("failed to set existing store: %x", ret); + goto cleanup; + } + +cleanup: + if (kmem) + vunmap(kmem); + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +static int process_allocation_handles(struct dxgprocess *process, + struct dxgdevice *device, + struct d3dkmt_createallocation *args, + struct + dxgkvmb_command_createallocation_return + *result, struct dxgallocation **dxgalloc, + struct dxgresource *resource) +{ + int ret = 0, i; + + hmgrtable_lock(&process->handle_table, DXGLOCK_EXCL); + if (args->flags.create_resource) { + ret = hmgrtable_assign_handle(&process->handle_table, resource, + HMGRENTRY_TYPE_DXGRESOURCE, + result->resource); + if (ret) { + pr_err("failed to assign resource handle %x", + result->resource); + } else { + resource->handle = result->resource; + resource->handle_valid = 1; + } + } + for (i = 0; i < args->alloc_count; i++) { + struct dxgkvmb_command_allocinfo_return *host_alloc; + + host_alloc = &result->allocation_info[i]; + ret = hmgrtable_assign_handle(&process->handle_table, + dxgalloc[i], + HMGRENTRY_TYPE_DXGALLOCATION, + host_alloc->allocation); + if (ret) { + pr_err("failed to assign alloc handle %x %d %d", + host_alloc->allocation, + args->alloc_count, i); + break; + } + dxgalloc[i]->alloc_handle = host_alloc->allocation; + dxgalloc[i]->handle_valid = 1; + } + hmgrtable_unlock(&process->handle_table, DXGLOCK_EXCL); + + if (ret) + goto cleanup; + + if (args->flags.create_shared && !args->flags.nt_security_sharing) { + struct dxgsharedresource *shared_resource = + resource->shared_owner; + shared_resource->host_shared_handle = result->global_share; + shared_resource->global_handle = + hmgrtable_alloc_handle_safe(&dxgglobal->handle_table, + shared_resource, + HMGRENTRY_TYPE_DXGSHAREDRESOURCE, + true); + if (shared_resource->global_handle == 0) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + args->global_share = shared_resource->global_handle; + TRACE_DEBUG(1, "Shared resource global handles: %x %x", + shared_resource->global_handle, + shared_resource->host_shared_handle); + } + +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +static int create_local_allocations(struct dxgprocess *process, + struct dxgdevice *device, + struct d3dkmt_createallocation *args, + struct d3dkmt_createallocation *__user + input_args, + struct d3dddi_allocationinfo2 *alloc_info, + struct + dxgkvmb_command_createallocation_return + *result, struct dxgresource *resource, + struct dxgallocation **dxgalloc, + struct dxgkvmb_command_destroyallocation + *destroy_buffer, uint destroy_buffer_size) +{ + int i; + int alloc_count = args->alloc_count; + uint8_t *alloc_private_data = NULL; + int ret = 0; + ntstatus status = STATUS_SUCCESS; + struct dxgvmbuschannel *channel = &device->adapter->channel; + + /* Prepare the command to destroy allocation in case of failure */ + command_vgpu_to_host_init2(&destroy_buffer->hdr, + DXGK_VMBCOMMAND_DESTROYALLOCATION, + process->host_handle); + destroy_buffer->device = args->device; + destroy_buffer->resource = args->resource; + destroy_buffer->alloc_count = alloc_count; + destroy_buffer->flags.assume_not_in_use = 1; + for (i = 0; i < alloc_count; i++) { + TRACE_DEBUG2(1, 1, "host allocation: %d %x", + i, result->allocation_info[i].allocation); + destroy_buffer->allocations[i] = + result->allocation_info[i].allocation; + } + + if (args->flags.create_resource) { + TRACE_DEBUG(1, "created resource: %x", result->resource); + ret = dxg_copy_to_user(&input_args->resource, &result->resource, + sizeof(d3dkmt_handle)); + if (ret) + goto cleanup; + } + + alloc_private_data = (uint8_t *) result + + sizeof(struct dxgkvmb_command_createallocation_return) + + sizeof(struct dxgkvmb_command_allocinfo_return) * (alloc_count - 1); + + for (i = 0; i < alloc_count; i++) { + struct dxgkvmb_command_allocinfo_return *host_alloc; + struct d3dddi_allocationinfo2 *user_alloc; + + host_alloc = &result->allocation_info[i]; + user_alloc = &alloc_info[i]; + if (alloc_info->sysmem) { + ret = create_existing_sysmem(device, host_alloc, + dxgalloc[i], + args->flags.read_only != 0, + alloc_info->sysmem); + if (ret) + goto cleanup; + } + dxgalloc[i]->num_pages = + host_alloc->allocation_size >> PAGE_SHIFT; + dxgalloc[i]->cached = host_alloc->allocation_flags.cached; + if (host_alloc->priv_drv_data_size) { + ret = dxg_copy_to_user(user_alloc->priv_drv_data, + alloc_private_data, + host_alloc->priv_drv_data_size); + if (ret) + goto cleanup; + alloc_private_data += host_alloc->priv_drv_data_size; + } + ret = dxg_copy_to_user(&args->allocation_info[i].allocation, + &host_alloc->allocation, + sizeof(d3dkmt_handle)); + if (ret) + goto cleanup; + } + + ret = process_allocation_handles(process, device, args, result, + dxgalloc, resource); + if (ret) + goto cleanup; + + ret = dxg_copy_to_user(&input_args->global_share, &args->global_share, + sizeof(d3dkmt_handle)); + +cleanup: + + if (ret) { + /* Free local handles before freeing the handles in the host */ + dxgdevice_acquire_alloc_list_lock(device); + if (dxgalloc) + for (i = 0; i < alloc_count; i++) + if (dxgalloc[i]) + dxgallocation_free_handle(dxgalloc[i]); + if (resource && args->flags.create_resource) + dxgresource_free_handle(resource); + dxgdevice_release_alloc_list_lock(device); + + /* Destroy allocations in the host to unmap gpadls */ + status = dxgvmb_send_sync_msg_ntstatus(channel, destroy_buffer, + destroy_buffer_size); + if (!NT_SUCCESS(status)) + pr_err("failed to destroy allocations: %x", status); + + dxgdevice_acquire_alloc_list_lock(device); + if (dxgalloc) { + for (i = 0; i < alloc_count; i++) { + if (dxgalloc[i]) { + dxgalloc[i]->alloc_handle = 0; + dxgallocation_destroy(dxgalloc[i]); + dxgalloc[i] = NULL; + } + } + } + if (resource && args->flags.create_resource) { + /* + * Prevent the resource memory from freeing. + * It will be freed in the top level function. + */ + dxgresource_acquire_reference(resource); + dxgresource_destroy(resource); + } + dxgdevice_release_alloc_list_lock(device); + } + + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_create_allocation(struct dxgprocess *process, + struct dxgdevice *device, + struct d3dkmt_createallocation *args, + struct d3dkmt_createallocation *__user + input_args, struct dxgresource *resource, + struct dxgallocation **dxgalloc, + struct d3dddi_allocationinfo2 *alloc_info, + struct d3dkmt_createstandardallocation + *standard_alloc) +{ + struct dxgkvmb_command_createallocation *command = NULL; + struct dxgkvmb_command_destroyallocation *destroy_buffer = NULL; + struct dxgkvmb_command_createallocation_return *result = NULL; + int ret; + int i; + uint result_size = 0; + uint cmd_size = 0; + uint destroy_buffer_size = 0; + uint priv_drv_data_size; + struct dxgvmbuschannel *channel = &device->adapter->channel; + + if (args->private_runtime_data_size >= DXG_MAX_VM_BUS_PACKET_SIZE || + args->priv_drv_data_size >= DXG_MAX_VM_BUS_PACKET_SIZE) { + ret = STATUS_BUFFER_OVERFLOW; + goto cleanup; + } + + /* + * Preallocate the buffer, which will be used for destruction in case + * of a failure + */ + destroy_buffer_size = sizeof(struct dxgkvmb_command_destroyallocation) + + args->alloc_count * sizeof(d3dkmt_handle); + destroy_buffer = dxgmem_alloc(process, DXGMEM_TMP, destroy_buffer_size); + if (destroy_buffer == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + /* Compute the total private driver size */ + + priv_drv_data_size = 0; + + for (i = 0; i < args->alloc_count; i++) { + if (alloc_info[i].priv_drv_data_size >= + DXG_MAX_VM_BUS_PACKET_SIZE) { + ret = STATUS_BUFFER_OVERFLOW; + goto cleanup; + } else { + priv_drv_data_size += alloc_info[i].priv_drv_data_size; + } + if (priv_drv_data_size >= DXG_MAX_VM_BUS_PACKET_SIZE) { + ret = STATUS_BUFFER_OVERFLOW; + goto cleanup; + } + } + + /* + * Private driver data for the result includes only per allocation + * private data + */ + result_size = sizeof(struct dxgkvmb_command_createallocation_return) + + (args->alloc_count - 1) * + sizeof(struct dxgkvmb_command_allocinfo_return) + + priv_drv_data_size; + result = dxgmem_alloc(process, DXGMEM_VMBUS, result_size); + if (result == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + /* Private drv data for the command includes the global private data */ + priv_drv_data_size += args->priv_drv_data_size; + + cmd_size = sizeof(struct dxgkvmb_command_createallocation) + + args->alloc_count * + sizeof(struct dxgkvmb_command_createallocation_allocinfo) + + args->private_runtime_data_size + priv_drv_data_size; + if (cmd_size > DXG_MAX_VM_BUS_PACKET_SIZE) { + ret = STATUS_BUFFER_OVERFLOW; + goto cleanup; + } + + TRACE_DEBUG(1, "command size, driver_data_size %d %d %ld %ld", + cmd_size, priv_drv_data_size, + sizeof(struct dxgkvmb_command_createallocation), + sizeof(struct dxgkvmb_command_createallocation_allocinfo)); + + command = dxgmem_alloc(process, DXGMEM_VMBUS, cmd_size); + if (command == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + command_vgpu_to_host_init2(&command->hdr, + DXGK_VMBCOMMAND_CREATEALLOCATION, + process->host_handle); + command->device = args->device; + command->flags = args->flags; + command->resource = args->resource; + command->private_runtime_resource_handle = + args->private_runtime_resource_handle; + command->alloc_count = args->alloc_count; + command->private_runtime_data_size = args->private_runtime_data_size; + command->priv_drv_data_size = args->priv_drv_data_size; + if (args->flags.standard_allocation) { + /* + * Flags.ExistingSysMem cannot be set from user mode, so it + * needs to be set it here. + */ + command->flags.existing_sysmem = 1; + } + + ret = copy_private_data(args, command, alloc_info, standard_alloc); + if (ret) + goto cleanup; + + ret = dxgvmb_send_sync_msg(channel, command, cmd_size, + result, result_size); + if (ret) { + pr_err("send_create_allocation failed %x", ret); + goto cleanup; + } + + ret = create_local_allocations(process, device, args, input_args, + alloc_info, result, resource, dxgalloc, + destroy_buffer, destroy_buffer_size); + if (ret) + goto cleanup; + +cleanup: + + if (destroy_buffer) + dxgmem_free(process, DXGMEM_TMP, destroy_buffer); + if (command) + dxgmem_free(process, DXGMEM_VMBUS, command); + if (result) + dxgmem_free(process, DXGMEM_VMBUS, result); + + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_destroy_allocation(struct dxgprocess *process, + struct dxgdevice *device, + struct dxgvmbuschannel *channel, + struct d3dkmt_destroyallocation2 *args, + d3dkmt_handle *alloc_handles) +{ + struct dxgkvmb_command_destroyallocation *destroy_buffer = NULL; + uint destroy_buffer_size = 0; + int ret = 0; + int allocations_size = args->alloc_count * sizeof(d3dkmt_handle); + + destroy_buffer_size = sizeof(struct dxgkvmb_command_destroyallocation) + + allocations_size; + destroy_buffer = dxgmem_alloc(process, DXGMEM_TMP, destroy_buffer_size); + if (destroy_buffer == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + command_vgpu_to_host_init2(&destroy_buffer->hdr, + DXGK_VMBCOMMAND_DESTROYALLOCATION, + process->host_handle); + destroy_buffer->device = args->device; + destroy_buffer->resource = args->resource; + destroy_buffer->alloc_count = args->alloc_count; + destroy_buffer->flags = args->flags; + if (allocations_size) + memcpy(destroy_buffer->allocations, alloc_handles, + allocations_size); + + ret = dxgvmb_send_sync_msg_ntstatus(channel, destroy_buffer, + destroy_buffer_size); + +cleanup: + + if (destroy_buffer) + dxgmem_free(process, DXGMEM_TMP, destroy_buffer); + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_make_resident(struct dxgprocess *process, + struct dxgdevice *device, + struct dxgvmbuschannel *channel, + struct d3dddi_makeresident *args) +{ + int ret = 0; + uint cmd_size; + struct dxgkvmb_command_makeresident_return result = { }; + struct dxgkvmb_command_makeresident *command = NULL; + + cmd_size = (args->alloc_count - 1) * sizeof(d3dkmt_handle) + + sizeof(struct dxgkvmb_command_makeresident); + command = dxgmem_alloc(process, DXGMEM_VMBUS, cmd_size); + if (command == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + ret = dxg_copy_from_user(command->allocations, args->allocation_list, + args->alloc_count * sizeof(d3dkmt_handle)); + if (ret) + goto cleanup; + command_vgpu_to_host_init2(&command->hdr, + DXGK_VMBCOMMAND_MAKERESIDENT, + process->host_handle); + command->alloc_count = args->alloc_count; + command->paging_queue = args->paging_queue; + if (device) + command->device = device->handle; + command->flags = args->flags; + + ret = dxgvmb_send_sync_msg(channel, command, cmd_size, + &result, sizeof(result)); + if (ret) { + pr_err("send_make_resident failed %x", ret); + goto cleanup; + } + + args->paging_fence_value = result.paging_fence_value; + args->num_bytes_to_trim = result.num_bytes_to_trim; + ret = result.status; + +cleanup: + + if (command) + dxgmem_free(process, DXGMEM_VMBUS, command); + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_evict(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_evict *args) +{ + int ret = 0; + uint cmd_size; + struct dxgkvmb_command_evict_return result = { }; + struct dxgkvmb_command_evict *command = NULL; + + cmd_size = (args->alloc_count - 1) * sizeof(d3dkmt_handle) + + sizeof(struct dxgkvmb_command_evict); + command = dxgmem_alloc(process, DXGMEM_VMBUS, cmd_size); + if (command == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + ret = dxg_copy_from_user(command->allocations, args->allocations, + args->alloc_count * sizeof(d3dkmt_handle)); + if (ret) + goto cleanup; + command_vgpu_to_host_init2(&command->hdr, + DXGK_VMBCOMMAND_EVICT, process->host_handle); + command->alloc_count = args->alloc_count; + command->device = args->device; + command->flags = args->flags; + + ret = dxgvmb_send_sync_msg(channel, command, cmd_size, + &result, sizeof(result)); + if (ret) { + pr_err("send_evict failed %x", ret); + goto cleanup; + } + args->num_bytes_to_trim = result.num_bytes_to_trim; + +cleanup: + + if (command) + dxgmem_free(process, DXGMEM_VMBUS, command); + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_submit_command(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_submitcommand *args) +{ + int ret = 0; + uint cmd_size; + struct dxgkvmb_command_submitcommand *command = NULL; + uint hbufsize = args->num_history_buffers * sizeof(d3dkmt_handle); + + cmd_size = sizeof(struct dxgkvmb_command_submitcommand) + + hbufsize + args->priv_drv_data_size; + command = dxgmem_alloc(process, DXGMEM_VMBUS, cmd_size); + if (command == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + ret = dxg_copy_from_user(&command[1], args->history_buffer_array, + hbufsize); + if (ret) + goto cleanup; + ret = dxg_copy_from_user((uint8_t *) &command[1] + hbufsize, + args->priv_drv_data, args->priv_drv_data_size); + if (ret) + goto cleanup; + + command_vgpu_to_host_init2(&command->hdr, + DXGK_VMBCOMMAND_SUBMITCOMMAND, + process->host_handle); + command->args = *args; + + ret = dxgvmb_send_sync_msg_ntstatus(channel, command, cmd_size); + +cleanup: + + if (command) + dxgmem_free(process, DXGMEM_VMBUS, command); + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_map_gpu_va(struct dxgprocess *process, + d3dkmt_handle device, + struct dxgvmbuschannel *channel, + struct d3dddi_mapgpuvirtualaddress *args) +{ + struct dxgkvmb_command_mapgpuvirtualaddress command; + struct dxgkvmb_command_mapgpuvirtualaddress_return result; + int ret = 0; + + command_vgpu_to_host_init2(&command.hdr, + DXGK_VMBCOMMAND_MAPGPUVIRTUALADDRESS, + process->host_handle); + command.args = *args; + command.device = device; + + ret = dxgvmb_send_sync_msg(channel, &command, sizeof(command), &result, + sizeof(result)); + if (ret) { + pr_err("%s failed %x", __func__, ret); + goto cleanup; + } + args->virtual_address = result.virtual_address; + args->paging_fence_value = result.paging_fence_value; + ret = result.status; + +cleanup: + + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_reserve_gpu_va(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dddi_reservegpuvirtualaddress *args) +{ + struct dxgkvmb_command_reservegpuvirtualaddress command; + struct dxgkvmb_command_reservegpuvirtualaddress_return result; + int ret = 0; + + command_vgpu_to_host_init2(&command.hdr, + DXGK_VMBCOMMAND_RESERVEGPUVIRTUALADDRESS, + process->host_handle); + command.args = *args; + + ret = dxgvmb_send_sync_msg(channel, &command, sizeof(command), &result, + sizeof(result)); + args->virtual_address = result.virtual_address; + + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_free_gpu_va(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_freegpuvirtualaddress *args) +{ + struct dxgkvmb_command_freegpuvirtualaddress command; + ntstatus status; + + command_vgpu_to_host_init2(&command.hdr, + DXGK_VMBCOMMAND_FREEGPUVIRTUALADDRESS, + process->host_handle); + command.args = *args; + + status = dxgvmb_send_sync_msg_ntstatus(channel, &command, + sizeof(command)); + TRACE_FUNC_EXIT_ERR(__func__, status); + return status; +} + +int dxgvmb_send_update_gpu_va(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_updategpuvirtualaddress *args) +{ + struct dxgkvmb_command_updategpuvirtualaddress *command = NULL; + uint cmd_size; + uint op_size; + int ret = 0; + + if (args->num_operations == 0 || + (DXG_MAX_VM_BUS_PACKET_SIZE / + sizeof(struct d3dddi_updategpuvirtualaddress_operation)) < + args->num_operations) { + ret = STATUS_INVALID_PARAMETER; + pr_err("Invalid number of operations: %d", + args->num_operations); + goto cleanup; + } + + op_size = args->num_operations * + sizeof(struct d3dddi_updategpuvirtualaddress_operation); + cmd_size = sizeof(struct dxgkvmb_command_updategpuvirtualaddress) + + op_size - sizeof(args->operations[0]); + command = dxgmem_alloc(process, DXGMEM_VMBUS, cmd_size); + if (command == NULL) { + ret = STATUS_NO_MEMORY; + pr_err("Failed to allocate command"); + goto cleanup; + } + + command_vgpu_to_host_init2(&command->hdr, + DXGK_VMBCOMMAND_UPDATEGPUVIRTUALADDRESS, + process->host_handle); + command->fence_value = args->fence_value; + command->device = args->device; + command->context = args->context; + command->fence_object = args->fence_object; + command->num_operations = args->num_operations; + command->flags = args->flags.value; + ret = dxg_copy_from_user(command->operations, args->operations, + op_size); + if (ret) + goto cleanup; + + ret = dxgvmb_send_sync_msg_ntstatus(channel, command, cmd_size); + +cleanup: + + if (command) + dxgmem_free(process, DXGMEM_VMBUS, command); + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +static void set_result(struct d3dkmt_createsynchronizationobject2 *args, + d3dgpu_virtual_address fence_gpu_va, uint8_t *va) +{ + args->info.periodic_monitored_fence.fence_gpu_virtual_address = + fence_gpu_va; + args->info.periodic_monitored_fence.fence_cpu_virtual_address = va; +} + +int dxgvmb_send_create_sync_object(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_createsynchronizationobject2 + *args, struct dxgsyncobject *syncobj) +{ + struct dxgkvmb_command_createsyncobject_return result = { }; + struct dxgkvmb_command_createsyncobject command = { }; + int ret = 0; + uint8_t *va = 0; + + command_vgpu_to_host_init2(&command.hdr, + DXGK_VMBCOMMAND_CREATESYNCOBJECT, + process->host_handle); + command.args = *args; + command.client_hint = 1; /* CLIENTHINT_UMD */ + + ret = dxgvmb_send_sync_msg(channel, &command, sizeof(command), &result, + sizeof(result)); + if (ret) { + pr_err("%s failed %d", __func__, ret); + goto cleanup; + } + args->sync_object = result.sync_object; + if (syncobj->shared) { + if (result.global_sync_object == 0) { + pr_err("shared handle is 0"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + args->info.shared_handle = result.global_sync_object; + } + + if (syncobj->monitored_fence) { + va = dxg_map_iospace(result.fence_storage_address, PAGE_SIZE, + PROT_READ | PROT_WRITE, true); + if (va == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + if (args->info.type == D3DDDI_MONITORED_FENCE) { + args->info.monitored_fence.fence_gpu_virtual_address = + result.fence_gpu_va; + args->info.monitored_fence.fence_cpu_virtual_address = + va; + { + unsigned long value; + + TRACE_DEBUG(1, "fence cpu address: %p", va); + ret = dxg_copy_from_user(&value, va, + sizeof(uint64_t)); + if (ret) + pr_err("failed to read fence"); + else + TRACE_DEBUG(1, "fence value: %lx", + value); + } + } else { + set_result(args, result.fence_gpu_va, va); + } + syncobj->mapped_address = va; + } + +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_destroy_sync_object(struct dxgprocess *process, + d3dkmt_handle sync_object) +{ + struct dxgkvmb_command_destroysyncobject command = { }; + ntstatus status; + + status = dxgglobal_acquire_channel_lock(); + if (status) + goto cleanup; + + command_vm_to_host_init2(&command.hdr, + DXGK_VMBCOMMAND_DESTROYSYNCOBJECT, + process->host_handle); + command.sync_object = sync_object; + + status = dxgvmb_send_sync_msg_ntstatus(dxgglobal_get_dxgvmbuschannel(), + &command, sizeof(command)); + + dxgglobal_release_channel_lock(); + +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, status); + return status; +} + +int dxgvmb_send_signal_sync_object(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dddicb_signalflags flags, + uint64_t legacy_fence_value, + d3dkmt_handle context, + uint object_count, + d3dkmt_handle __user *objects, + uint context_count, + d3dkmt_handle __user *contexts, + uint fence_count, + uint64_t __user *fences, + struct eventfd_ctx *cpu_event_handle, + d3dkmt_handle device) +{ + int ret = 0; + struct dxgkvmb_command_signalsyncobject *command = NULL; + uint object_size = object_count * sizeof(d3dkmt_handle); + uint context_size = context_count * sizeof(d3dkmt_handle); + uint fence_size = fences ? fence_count * sizeof(uint64_t) : 0; + uint8_t *current_pos; + uint cmd_size = sizeof(struct dxgkvmb_command_signalsyncobject) + + object_size + context_size + fence_size; + + if (context) + cmd_size += sizeof(d3dkmt_handle); + + command = dxgmem_alloc(process, DXGMEM_VMBUS, cmd_size); + if (command == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + command_vgpu_to_host_init2(&command->hdr, + DXGK_VMBCOMMAND_SIGNALSYNCOBJECT, + process->host_handle); + + if (flags.enqueue_cpu_event) + command->cpu_event_handle = (winhandle) cpu_event_handle; + else + command->device = device; + command->flags = flags; + command->fence_value = legacy_fence_value; + command->object_count = object_count; + command->context_count = context_count; + current_pos = (uint8_t *) &command[1]; + ret = dxg_copy_from_user(current_pos, objects, object_size); + if (ret) { + pr_err("Failed to read objects %p %d", + objects, object_size); + goto cleanup; + } + current_pos += object_size; + if (context) { + command->context_count++; + *(d3dkmt_handle *) current_pos = context; + current_pos += sizeof(d3dkmt_handle); + } + if (context_size) { + ret = dxg_copy_from_user(current_pos, contexts, context_size); + if (ret) { + pr_err("Failed to read contexts %p %d", + contexts, context_size); + goto cleanup; + } + current_pos += context_size; + } + if (fence_size) { + ret = dxg_copy_from_user(current_pos, fences, fence_size); + if (ret) { + pr_err("Failed to read fences %p %d", + fences, fence_size); + goto cleanup; + } + } + + ret = dxgvmb_send_sync_msg_ntstatus(channel, command, cmd_size); + +cleanup: + if (command) + dxgmem_free(process, DXGMEM_VMBUS, command); + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_wait_sync_object_cpu(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct + d3dkmt_waitforsynchronizationobjectfromcpu + *args, u64 cpu_event) +{ + int ret = 0; + struct dxgkvmb_command_waitforsyncobjectfromcpu *command = NULL; + uint object_size = args->object_count * sizeof(d3dkmt_handle); + uint fence_size = args->object_count * sizeof(uint64_t); + uint8_t *current_pos; + uint cmd_size = sizeof(*command) + object_size + fence_size; + + command = dxgmem_alloc(process, DXGMEM_VMBUS, cmd_size); + if (command == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + command_vgpu_to_host_init2(&command->hdr, + DXGK_VMBCOMMAND_WAITFORSYNCOBJECTFROMCPU, + process->host_handle); + command->device = args->device; + command->flags = args->flags; + command->object_count = args->object_count; + command->guest_event_pointer = (uint64_t) cpu_event; + current_pos = (uint8_t *) &command[1]; + ret = dxg_copy_from_user(current_pos, args->objects, object_size); + if (ret) + goto cleanup; + current_pos += object_size; + ret = dxg_copy_from_user(current_pos, args->fence_values, fence_size); + if (ret) + goto cleanup; + + ret = dxgvmb_send_sync_msg_ntstatus(channel, command, cmd_size); + +cleanup: + if (command) + dxgmem_free(process, DXGMEM_VMBUS, command); + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_wait_sync_object_gpu(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + d3dkmt_handle context, uint object_count, + d3dkmt_handle *objects, uint64_t *fences, + bool legacy_fence) +{ + ntstatus status; + struct dxgkvmb_command_waitforsyncobjectfromgpu *command = NULL; + uint fence_size = object_count * sizeof(uint64_t); + uint object_size = object_count * sizeof(d3dkmt_handle); + uint8_t *current_pos; + uint cmd_size = object_size + fence_size - sizeof(uint64_t) + + sizeof(struct dxgkvmb_command_waitforsyncobjectfromgpu); + + if (object_count == 0 || object_count > D3DDDI_MAX_OBJECT_WAITED_ON) { + status = STATUS_INVALID_PARAMETER; + goto cleanup; + } + command = dxgmem_alloc(process, DXGMEM_VMBUS, cmd_size); + if (command == NULL) { + status = STATUS_NO_MEMORY; + goto cleanup; + } + command_vgpu_to_host_init2(&command->hdr, + DXGK_VMBCOMMAND_WAITFORSYNCOBJECTFROMGPU, + process->host_handle); + command->context = context; + command->object_count = object_count; + command->legacy_fence_object = legacy_fence; + current_pos = (uint8_t *) command->fence_values; + memcpy(current_pos, fences, fence_size); + current_pos += fence_size; + memcpy(current_pos, objects, object_size); + + status = dxgvmb_send_sync_msg_ntstatus(channel, command, cmd_size); + +cleanup: + if (command) + dxgmem_free(process, DXGMEM_VMBUS, command); + TRACE_FUNC_EXIT_ERR(__func__, status); + return status; +} + +int dxgvmb_send_lock2(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_lock2 *args, + struct d3dkmt_lock2 *__user outargs) +{ + int ret = 0; + struct dxgkvmb_command_lock2 command = { }; + struct dxgkvmb_command_lock2_return result = { }; + struct dxgallocation *alloc = NULL; + + command_vgpu_to_host_init2(&command.hdr, + DXGK_VMBCOMMAND_LOCK2, process->host_handle); + command.args = *args; + + ret = dxgvmb_send_sync_msg(channel, &command, sizeof(command), + &result, sizeof(result)); + if (ret) + goto cleanup; + if (!NT_SUCCESS(result.status)) { + ret = result.status; + goto cleanup; + } + + hmgrtable_lock(&process->handle_table, DXGLOCK_SHARED); + alloc = hmgrtable_get_object_by_type(&process->handle_table, + HMGRENTRY_TYPE_DXGALLOCATION, + args->allocation); + if (alloc == NULL || alloc->cpu_address || alloc->cpu_address_refcount + || alloc->cpu_address_mapped) { + pr_err("%s invalid alloc", __func__); + ret = STATUS_INVALID_PARAMETER; + } else { + args->data = dxg_map_iospace((uint64_t) result. + cpu_visible_buffer_offset, + alloc->num_pages << PAGE_SHIFT, + PROT_READ | PROT_WRITE, + alloc->cached); + if (args->data == NULL) { + ret = STATUS_NO_MEMORY; + } else { + ret = dxg_copy_to_user(&outargs->data, &args->data, + sizeof(args->data)); + if (!ret) { + alloc->cpu_address = args->data; + alloc->cpu_address_mapped = true; + alloc->cpu_address_refcount = 1; + } else { + dxg_unmap_iospace(alloc->cpu_address, + alloc-> + num_pages << PAGE_SHIFT); + } + } + } + hmgrtable_unlock(&process->handle_table, DXGLOCK_SHARED); + +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_unlock2(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_unlock2 *args) +{ + ntstatus status; + struct dxgkvmb_command_unlock2 command = { }; + + command_vgpu_to_host_init2(&command.hdr, + DXGK_VMBCOMMAND_UNLOCK2, + process->host_handle); + command.args = *args; + + status = dxgvmb_send_sync_msg_ntstatus(channel, &command, + sizeof(command)); + TRACE_FUNC_EXIT_ERR(__func__, status); + return status; +} + +int dxgvmb_send_update_alloc_property(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dddi_updateallocproperty *args, + struct d3dddi_updateallocproperty *__user + inargs) +{ + int ret; + ntstatus status; + struct dxgkvmb_command_updateallocationproperty command = { }; + struct dxgkvmb_command_updateallocationproperty_return result = { }; + + command_vgpu_to_host_init2(&command.hdr, + DXGK_VMBCOMMAND_UPDATEALLOCATIONPROPERTY, + process->host_handle); + command.args = *args; + + status = dxgvmb_send_sync_msg(channel, &command, sizeof(command), + &result, sizeof(result)); + if (status == STATUS_PENDING) { + ret = dxg_copy_to_user(&inargs->paging_fence_value, + &result.paging_fence_value, + sizeof(uint64_t)); + if (ret) + status = ret; + } + TRACE_FUNC_EXIT_ERR(__func__, status); + return status; +} + +int dxgvmb_send_mark_device_as_error(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_markdeviceaserror *args) +{ + struct dxgkvmb_command_markdeviceaserror command = { }; + ntstatus status; + + command_vgpu_to_host_init2(&command.hdr, + DXGK_VMBCOMMAND_MARKDEVICEASERROR, + process->host_handle); + command.args = *args; + status = dxgvmb_send_sync_msg_ntstatus(channel, &command, + sizeof(command)); + TRACE_FUNC_EXIT_ERR(__func__, status); + return status; +} + +int dxgvmb_send_set_allocation_priority(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_setallocationpriority + *args) +{ + uint cmd_size = sizeof(struct dxgkvmb_command_setallocationpriority); + uint alloc_size = 0; + uint priority_size = 0; + struct dxgkvmb_command_setallocationpriority *command = NULL; + int ret = 0; + d3dkmt_handle *allocations; + + if (args->allocation_count > DXG_MAX_VM_BUS_PACKET_SIZE) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + if (args->resource) { + priority_size = sizeof(uint); + if (args->allocation_count != 0) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + } else { + if (args->allocation_count == 0) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + alloc_size = args->allocation_count * sizeof(d3dkmt_handle); + cmd_size += alloc_size; + priority_size = sizeof(uint) * args->allocation_count; + } + cmd_size += priority_size; + command = dxgmem_alloc(process, DXGMEM_VMBUS, cmd_size); + if (command == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + command_vgpu_to_host_init2(&command->hdr, + DXGK_VMBCOMMAND_SETALLOCATIONPRIORITY, + process->host_handle); + command->device = args->device; + command->allocation_count = args->allocation_count; + command->resource = args->resource; + allocations = (d3dkmt_handle *) &command[1]; + ret = dxg_copy_from_user(allocations, args->allocation_list, + alloc_size); + if (ret) + goto cleanup; + ret = dxg_copy_from_user((uint8_t *) allocations + alloc_size, + args->priorities, priority_size); + if (ret) + goto cleanup; + + ret = dxgvmb_send_sync_msg_ntstatus(channel, command, cmd_size); + +cleanup: + if (command) + dxgmem_free(process, DXGMEM_VMBUS, command); + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_get_allocation_priority(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_getallocationpriority + *args) +{ + uint cmd_size = sizeof(struct dxgkvmb_command_getallocationpriority); + uint result_size; + uint alloc_size = 0; + uint priority_size = 0; + struct dxgkvmb_command_getallocationpriority *command = NULL; + struct dxgkvmb_command_getallocationpriority_return *result; + int ret = 0; + d3dkmt_handle *allocations; + + if (args->allocation_count > DXG_MAX_VM_BUS_PACKET_SIZE) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + if (args->resource) { + priority_size = sizeof(uint); + if (args->allocation_count != 0) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + } else { + if (args->allocation_count == 0) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + alloc_size = args->allocation_count * sizeof(d3dkmt_handle); + cmd_size += alloc_size; + priority_size = sizeof(uint) * args->allocation_count; + } + result_size = + sizeof(struct dxgkvmb_command_getallocationpriority_return) + + priority_size; + cmd_size += result_size; + command = dxgmem_alloc(process, DXGMEM_VMBUS, cmd_size); + if (command == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + command_vgpu_to_host_init2(&command->hdr, + DXGK_VMBCOMMAND_GETALLOCATIONPRIORITY, + process->host_handle); + command->device = args->device; + command->allocation_count = args->allocation_count; + command->resource = args->resource; + allocations = (d3dkmt_handle *) &command[1]; + ret = dxg_copy_from_user(allocations, args->allocation_list, + alloc_size); + if (ret) + goto cleanup; + + result = (void *)((uint8_t *) &command[1] + alloc_size); + + ret = dxgvmb_send_sync_msg(channel, command, cmd_size, + result, result_size); + if (ret) + goto cleanup; + if (!NT_SUCCESS(result->status)) { + ret = result->status; + goto cleanup; + } + + ret = dxg_copy_to_user(args->priorities, (uint8_t *) result + + sizeof(struct + dxgkvmb_command_getallocationpriority_return), + priority_size); + +cleanup: + if (command) + dxgmem_free(process, DXGMEM_VMBUS, command); + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_set_context_scheduling_priority(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + d3dkmt_handle context, + int priority, bool in_process) +{ + struct dxgkvmb_command_setcontextschedulingpriority2 command = { }; + ntstatus status; + + command_vgpu_to_host_init2(&command.hdr, + DXGK_VMBCOMMAND_SETCONTEXTSCHEDULINGPRIORITY, + process->host_handle); + command.context = context; + command.priority = priority; + command.in_process = in_process; + status = dxgvmb_send_sync_msg_ntstatus(channel, &command, + sizeof(command)); + TRACE_FUNC_EXIT_ERR(__func__, status); + return status; +} + +int dxgvmb_send_get_context_scheduling_priority(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + d3dkmt_handle context, + int *priority, bool in_process) +{ + struct dxgkvmb_command_getcontextschedulingpriority command = { }; + struct dxgkvmb_command_getcontextschedulingpriority_return result = { }; + int ret; + + command_vgpu_to_host_init2(&command.hdr, + DXGK_VMBCOMMAND_GETCONTEXTSCHEDULINGPRIORITY, + process->host_handle); + command.context = context; + command.in_process = in_process; + ret = dxgvmb_send_sync_msg(channel, &command, sizeof(command), + &result, sizeof(result)); + if (!ret) { + ret = result.status; + *priority = result.priority; + } + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_offer_allocations(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_offerallocations *args) +{ + struct dxgkvmb_command_offerallocations *command; + int ret = 0; + uint alloc_size = sizeof(d3dkmt_handle) * args->allocation_count; + uint cmd_size = sizeof(struct dxgkvmb_command_offerallocations) + + alloc_size - sizeof(d3dkmt_handle); + + command = dxgmem_alloc(process, DXGMEM_VMBUS, cmd_size); + if (command == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + command_vgpu_to_host_init2(&command->hdr, + DXGK_VMBCOMMAND_OFFERALLOCATIONS, + process->host_handle); + command->flags = args->flags; + command->priority = args->priority; + command->device = args->device; + command->allocation_count = args->allocation_count; + if (args->resources) { + command->resources = true; + ret = dxg_copy_from_user(command->allocations, args->resources, + alloc_size); + } else { + ret = dxg_copy_from_user(command->allocations, + args->allocations, alloc_size); + } + if (ret) + goto cleanup; + + ret = dxgvmb_send_sync_msg_ntstatus(channel, command, cmd_size); + +cleanup: + if (command) + dxgmem_free(process, DXGMEM_VMBUS, command); + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_reclaim_allocations(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + d3dkmt_handle device, + struct d3dkmt_reclaimallocations2 *args, + uint64_t * __user paging_fence_value) +{ + struct dxgkvmb_command_reclaimallocations *command = NULL; + struct dxgkvmb_command_reclaimallocations_return *result = NULL; + int ret = 0; + uint alloc_size = sizeof(d3dkmt_handle) * args->allocation_count; + uint cmd_size = sizeof(struct dxgkvmb_command_reclaimallocations) + + alloc_size - sizeof(d3dkmt_handle); + uint result_size = sizeof(*result); + + command = dxgmem_alloc(process, DXGMEM_VMBUS, cmd_size); + if (command == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + command_vgpu_to_host_init2(&command->hdr, + DXGK_VMBCOMMAND_RECLAIMALLOCATIONS, + process->host_handle); + command->device = device; + command->paging_queue = args->paging_queue; + command->allocation_count = args->allocation_count; + command->write_results = args->results != NULL; + if (args->resources) { + command->resources = true; + ret = dxg_copy_from_user(command->allocations, args->resources, + alloc_size); + } else { + ret = dxg_copy_from_user(command->allocations, + args->allocations, alloc_size); + } + if (ret) + goto cleanup; + + if (command->write_results) + result_size += (args->allocation_count - 1) * + sizeof(enum d3dddi_reclaim_result); + result = dxgmem_alloc(process, DXGMEM_VMBUS, result_size); + if (result == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + ret = dxgvmb_send_sync_msg(channel, command, cmd_size, + result, result_size); + if (ret) + goto cleanup; + ret = dxg_copy_to_user(paging_fence_value, + &result->paging_fence_value, sizeof(uint64_t)); + if (ret) + goto cleanup; + + ret = result->status; + if (NT_SUCCESS(result->status) && args->results) + ret = dxg_copy_to_user(args->results, result->discarded, + sizeof(result->discarded[0]) * + args->allocation_count); + +cleanup: + if (command) + dxgmem_free(process, DXGMEM_VMBUS, command); + if (result) + dxgmem_free(process, DXGMEM_VMBUS, result); + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_change_vidmem_reservation(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + d3dkmt_handle other_process, + struct + d3dkmt_changevideomemoryreservation + *args) +{ + struct dxgkvmb_command_changevideomemoryreservation command = { }; + ntstatus status; + + command_vgpu_to_host_init2(&command.hdr, + DXGK_VMBCOMMAND_CHANGEVIDEOMEMORYRESERVATION, + process->host_handle); + command.args = *args; + command.args.process = other_process; + + status = dxgvmb_send_sync_msg_ntstatus(channel, &command, + sizeof(command)); + TRACE_FUNC_EXIT_ERR(__func__, status); + return status; +} + +int dxgvmb_send_create_hwqueue(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_createhwqueue *args, + struct d3dkmt_createhwqueue *__user inargs, + struct dxghwqueue *hwqueue) +{ + struct dxgkvmb_command_createhwqueue command_on_stack = { }; + struct dxgkvmb_command_createhwqueue *command = &command_on_stack; + uint cmd_size = sizeof(struct dxgkvmb_command_createhwqueue); + int ret = 0; + bool command_allocated = false; + + if (args->priv_drv_data_size > DXG_MAX_VM_BUS_PACKET_SIZE) { + pr_err("invalid private driver data size"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + if (args->priv_drv_data_size) + cmd_size += args->priv_drv_data_size - 1; + + /* Input command is returned back as output */ + cmd_size = DXGK_DECL_VMBUS_ALIGN_FOR_OUTPUT(cmd_size); + + if (args->priv_drv_data_size) { + command = dxgmem_alloc(process, DXGMEM_VMBUS, cmd_size); + if (command == NULL) { + pr_err("failed to allocate memory"); + ret = STATUS_NO_MEMORY; + goto cleanup; + } + command_allocated = true; + } + + command_vgpu_to_host_init2(&command->hdr, + DXGK_VMBCOMMAND_CREATEHWQUEUE, + process->host_handle); + command->context = args->context; + command->flags = args->flags; + command->priv_drv_data_size = args->priv_drv_data_size; + if (args->priv_drv_data_size) { + ret = dxg_copy_from_user(command->priv_drv_data, + args->priv_drv_data, + args->priv_drv_data_size); + if (ret) + goto cleanup; + } + + ret = dxgvmb_send_sync_msg(channel, command, cmd_size, + command, cmd_size); + if (ret) + goto cleanup; + + if (!NT_SUCCESS(command->status)) { + ret = command->status; + goto cleanup; + } + + ret = hmgrtable_assign_handle_safe(&process->handle_table, hwqueue, + HMGRENTRY_TYPE_DXGHWQUEUE, + command->hwqueue); + if (ret) + goto cleanup; + + hwqueue->handle = command->hwqueue; + + hwqueue->progress_fence_mapped_address = + dxg_map_iospace((unsigned long)command-> + hwqueue_progress_fence_cpuva, PAGE_SIZE, + PROT_READ | PROT_WRITE, true); + if (hwqueue->progress_fence_mapped_address == NULL) + goto cleanup; + + hwqueue->progress_fence_sync_object = command->hwqueue_progress_fence; + + ret = dxg_copy_to_user(&inargs->queue, &command->hwqueue, + sizeof(d3dkmt_handle)); + ret |= dxg_copy_to_user(&inargs->queue_progress_fence, + &command->hwqueue_progress_fence, + sizeof(d3dkmt_handle)); + ret |= + dxg_copy_to_user(&inargs->queue_progress_fence_cpu_va, + &hwqueue->progress_fence_mapped_address, + sizeof(inargs->queue_progress_fence_cpu_va)); + ret |= + dxg_copy_to_user(&inargs->queue_progress_fence_gpu_va, + &command->hwqueue_progress_fence_gpuva, + sizeof(uint64_t)); + if (args->priv_drv_data_size) + ret |= dxg_copy_to_user(args->priv_drv_data, + command->priv_drv_data, + args->priv_drv_data_size); + +cleanup: + if (ret) { + pr_err("%s failed %x", __func__, ret); + if (hwqueue->handle) { + hmgrtable_free_handle_safe(&process->handle_table, + HMGRENTRY_TYPE_DXGHWQUEUE, + hwqueue->handle); + hwqueue->handle = 0; + } + if (command->hwqueue) + dxgvmb_send_destroy_hwqueue(process, channel, + command->hwqueue); + } + if (command_allocated) + dxgmem_free(process, DXGMEM_VMBUS, command); + return ret; +} + +int dxgvmb_send_destroy_hwqueue(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + d3dkmt_handle handle) +{ + ntstatus status; + struct dxgkvmb_command_destroyhwqueue command = { }; + + command_vgpu_to_host_init2(&command.hdr, DXGK_VMBCOMMAND_DESTROYHWQUEUE, + process->host_handle); + command.hwqueue = handle; + + status = dxgvmb_send_sync_msg_ntstatus(channel, &command, + sizeof(command)); + TRACE_FUNC_EXIT_ERR(__func__, status); + return status; +} + +int dxgvmb_send_query_adapter_info(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_queryadapterinfo *args) +{ + struct dxgkvmb_command_queryadapterinfo *command = NULL; + uint cmd_size; + int ret = 0; + + cmd_size = sizeof(*command) + args->private_data_size - 1; + command = dxgmem_alloc(process, DXGMEM_VMBUS, cmd_size); + if (command == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + ret = dxg_copy_from_user(command->private_data, + args->private_data, args->private_data_size); + if (ret) + goto cleanup; + + command_vgpu_to_host_init2(&command->hdr, + DXGK_VMBCOMMAND_QUERYADAPTERINFO, + process->host_handle); + command->private_data_size = args->private_data_size; + command->query_type = args->type; + + ret = dxgvmb_send_sync_msg(channel, command, cmd_size, + command->private_data, + command->private_data_size); + if (ret) + goto cleanup; + switch (args->type) { + case KMTQAITYPE_ADAPTERTYPE: + case KMTQAITYPE_ADAPTERTYPE_RENDER: + { + struct d3dkmt_adaptertype *adapter_type = + (void *)command->private_data; + adapter_type->paravirtualized = 1; + adapter_type->display_supported = 0; + adapter_type->post_device = 0; + adapter_type->indirect_display_device = 0; + adapter_type->acg_supported = 0; + adapter_type->support_set_timings_from_vidpn = 0; + break; + } + default: + break; + } + ret = dxg_copy_to_user(args->private_data, command->private_data, + args->private_data_size); + +cleanup: + if (command) + dxgmem_free(process, DXGMEM_VMBUS, command); + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_submit_command_to_hwqueue(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_submitcommandtohwqueue + *args) +{ + int ret = 0; + uint cmd_size; + struct dxgkvmb_command_submitcommandtohwqueue *command = NULL; + uint primaries_size = args->num_primaries * sizeof(d3dkmt_handle); + + cmd_size = sizeof(*command) + args->priv_drv_data_size + primaries_size; + command = dxgmem_alloc(process, DXGMEM_VMBUS, cmd_size); + if (command == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + if (primaries_size) { + ret = dxg_copy_from_user(&command[1], args->written_primaries, + primaries_size); + if (ret) + goto cleanup; + } + if (args->priv_drv_data_size) { + ret = dxg_copy_from_user((char *)&command[1] + primaries_size, + args->priv_drv_data, + args->priv_drv_data_size); + if (ret) + goto cleanup; + } + + command_vgpu_to_host_init2(&command->hdr, + DXGK_VMBCOMMAND_SUBMITCOMMANDTOHWQUEUE, + process->host_handle); + command->args = *args; + + ret = dxgvmb_send_sync_msg_ntstatus(channel, command, cmd_size); + +cleanup: + if (command) + dxgmem_free(process, DXGMEM_VMBUS, command); + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_query_clock_calibration(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_queryclockcalibration + *args, + struct d3dkmt_queryclockcalibration + *__user inargs) +{ + struct dxgkvmb_command_queryclockcalibration command; + struct dxgkvmb_command_queryclockcalibration_return result; + int ret; + + command_vgpu_to_host_init2(&command.hdr, + DXGK_VMBCOMMAND_QUERYCLOCKCALIBRATION, + process->host_handle); + command.args = *args; + + ret = dxgvmb_send_sync_msg(channel, &command, sizeof(command), + &result, sizeof(result)); + if (ret) + goto cleanup; + ret = dxg_copy_to_user(&inargs->clock_data, &result.clock_data, + sizeof(result.clock_data)); + if (ret) + goto cleanup; + ret = result.status; + +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_flush_heap_transitions(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_flushheaptransitions *args) +{ + struct dxgkvmb_command_flushheaptransitions command; + ntstatus status; + + command_vgpu_to_host_init2(&command.hdr, + DXGK_VMBCOMMAND_FLUSHHEAPTRANSITIONS, + process->host_handle); + status = + dxgvmb_send_sync_msg_ntstatus(channel, &command, sizeof(command)); + TRACE_FUNC_EXIT_ERR(__func__, status); + return status; +} + +int dxgvmb_send_query_alloc_residency(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_queryallocationresidency + *args) +{ + int ret = 0; + struct dxgkvmb_command_queryallocationresidency *command = NULL; + uint cmd_size = sizeof(*command); + uint alloc_size = 0; + uint result_allocation_size = 0; + struct dxgkvmb_command_queryallocationresidency_return *result = NULL; + uint result_size = sizeof(*result); + + if (args->allocation_count > DXG_MAX_VM_BUS_PACKET_SIZE) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + if (args->allocation_count) { + alloc_size = args->allocation_count * sizeof(d3dkmt_handle); + cmd_size += alloc_size; + result_allocation_size = args->allocation_count * + sizeof(args->residency_status[0]); + } else { + result_allocation_size = sizeof(args->residency_status[0]); + } + result_size += result_allocation_size; + + command = dxgmem_alloc(process, DXGMEM_VMBUS, cmd_size); + if (command == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + command_vgpu_to_host_init2(&command->hdr, + DXGK_VMBCOMMAND_QUERYALLOCATIONRESIDENCY, + process->host_handle); + command->args = *args; + if (alloc_size) { + ret = dxg_copy_from_user(&command[1], args->allocations, + alloc_size); + if (ret) + goto cleanup; + } + + result = dxgmem_alloc(process, DXGMEM_VMBUS, result_size); + if (result == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + ret = dxgvmb_send_sync_msg(channel, command, cmd_size, + result, result_size); + if (ret) + goto cleanup; + if (!NT_SUCCESS(result->status)) { + ret = result->status; + goto cleanup; + } + ret = dxg_copy_to_user(args->residency_status, &result[1], + result_allocation_size); + +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, ret); + if (command) + dxgmem_free(process, DXGMEM_VMBUS, command); + if (result) + dxgmem_free(process, DXGMEM_VMBUS, result); + return ret; +} + +int dxgvmb_send_escape(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_escape *args) +{ + int ret = 0; + struct dxgkvmb_command_escape *command = NULL; + uint cmd_size = sizeof(*command); + + if (args->priv_drv_data_size > DXG_MAX_VM_BUS_PACKET_SIZE) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + cmd_size = cmd_size - sizeof(args->priv_drv_data[0]) + + args->priv_drv_data_size; + + command = dxgmem_alloc(process, DXGMEM_VMBUS, cmd_size); + if (command == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + command_vgpu_to_host_init2(&command->hdr, + DXGK_VMBCOMMAND_ESCAPE, + process->host_handle); + command->adapter = args->adapter; + command->device = args->device; + command->type = args->type; + command->flags = args->flags; + command->priv_drv_data_size = args->priv_drv_data_size; + command->context = args->context; + if (args->priv_drv_data_size) + ret = dxg_copy_from_user(command->priv_drv_data, + args->priv_drv_data, + args->priv_drv_data_size); + if (ret) + goto cleanup; + + ret = dxgvmb_send_sync_msg(channel, command, cmd_size, + command->priv_drv_data, + args->priv_drv_data_size); + if (ret) + goto cleanup; + + if (args->priv_drv_data_size) + ret = dxg_copy_to_user(args->priv_drv_data, + command->priv_drv_data, + args->priv_drv_data_size); + +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, ret); + if (command) + dxgmem_free(process, DXGMEM_VMBUS, command); + return ret; +} + +int dxgvmb_send_query_vidmem_info(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_queryvideomemoryinfo *args, + struct d3dkmt_queryvideomemoryinfo *__user + output) +{ + int ret = 0; + struct dxgkvmb_command_queryvideomemoryinfo command = { }; + struct dxgkvmb_command_queryvideomemoryinfo_return result = { }; + + command_vgpu_to_host_init2(&command.hdr, + dxgk_vmbcommand_queryvideomemoryinfo, + process->host_handle); + command.adapter = args->adapter; + command.memory_segment_group = args->memory_segment_group; + command.physical_adapter_index = args->physical_adapter_index; + + ret = dxgvmb_send_sync_msg(channel, &command, sizeof(command), + &result, sizeof(result)); + if (ret) + goto cleanup; + + ret = dxg_copy_to_user(&output->budget, &result.budget, + sizeof(output->budget)); + if (ret) + goto cleanup; + ret = dxg_copy_to_user(&output->current_usage, &result.current_usage, + sizeof(output->current_usage)); + if (ret) + goto cleanup; + ret = dxg_copy_to_user(&output->current_reservation, + &result.current_reservation, + sizeof(output->current_reservation)); + if (ret) + goto cleanup; + ret = dxg_copy_to_user(&output->available_for_reservation, + &result.available_for_reservation, + sizeof(output->available_for_reservation)); + +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_get_device_state(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_getdevicestate *args, + struct d3dkmt_getdevicestate *__user output) +{ + int ret; + struct dxgkvmb_command_getdevicestate command = { }; + struct dxgkvmb_command_getdevicestate_return result = { }; + + command_vgpu_to_host_init2(&command.hdr, + dxgk_vmbcommand_getdevicestate, + process->host_handle); + command.args = *args; + + ret = dxgvmb_send_sync_msg(channel, &command, sizeof(command), + &result, sizeof(result)); + if (ret) + goto cleanup; + + if (!NT_SUCCESS(result.status)) { + ret = result.status; + goto cleanup; + } + ret = dxg_copy_to_user(output, &result.args, sizeof(result.args)); + +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_open_sync_object(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + d3dkmt_handle shared_handle, + d3dkmt_handle *host_handle) +{ + struct dxgkvmb_command_opensyncobject command = { }; + struct dxgkvmb_command_opensyncobject_return result = { }; + int ret = 0; + + command_vm_to_host_init2(&command.hdr, DXGK_VMBCOMMAND_OPENSYNCOBJECT, + process->host_handle); + command.global_sync_object = shared_handle; + + ret = dxgglobal_acquire_channel_lock(); + if (ret) + goto cleanup; + + ret = dxgvmb_send_sync_msg(channel, &command, sizeof(command), + &result, sizeof(result)); + + dxgglobal_release_channel_lock(); + + if (ret) + goto cleanup; + + if (!NT_SUCCESS(result.status)) { + ret = result.status; + goto cleanup; + } + + *host_handle = result.sync_object; + +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_open_sync_object_nt(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + struct d3dkmt_opensyncobjectfromnthandle2 + *args, struct dxgsyncobject *syncobj) +{ + struct dxgkvmb_command_opensyncobject command = { }; + struct dxgkvmb_command_opensyncobject_return result = { }; + int ret = 0; + + command_vm_to_host_init2(&command.hdr, DXGK_VMBCOMMAND_OPENSYNCOBJECT, + process->host_handle); + command.device = args->device; + command.global_sync_object = syncobj->shared_owner->host_shared_handle; + command.flags = args->flags; + if (syncobj->monitored_fence) + command.engine_affinity = args->monitored_fence.engine_affinity; + + ret = dxgglobal_acquire_channel_lock(); + if (ret) + goto cleanup; + + ret = dxgvmb_send_sync_msg(channel, &command, sizeof(command), + &result, sizeof(result)); + + dxgglobal_release_channel_lock(); + + if (ret) + goto cleanup; + + if (!NT_SUCCESS(result.status)) { + ret = result.status; + goto cleanup; + } + + args->sync_object = result.sync_object; + if (syncobj->monitored_fence) { + void *va = dxg_map_iospace(result.guest_cpu_physical_address, + PAGE_SIZE, PROT_READ | PROT_WRITE, + true); + if (va == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + args->monitored_fence.fence_value_cpu_va = va; + args->monitored_fence.fence_value_gpu_va = + result.gpu_virtual_address; + syncobj->mapped_address = va; + } + +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_create_nt_shared_object(struct dxgprocess *process, + d3dkmt_handle object, + d3dkmt_handle *shared_handle) +{ + struct dxgkvmb_command_createntsharedobject command = { }; + int ret; + + command_vm_to_host_init2(&command.hdr, + DXGK_VMBCOMMAND_CREATENTSHAREDOBJECT, + process->host_handle); + command.object = object; + + ret = dxgglobal_acquire_channel_lock(); + if (ret) + goto cleanup; + + ret = dxgvmb_send_sync_msg(dxgglobal_get_dxgvmbuschannel(), + &command, sizeof(command), shared_handle, + sizeof(*shared_handle)); + + dxgglobal_release_channel_lock(); + + if (ret) + goto cleanup; + if (*shared_handle == 0) { + pr_err("failed to create NT shared object"); + ret = STATUS_INTERNAL_ERROR; + } + +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_destroy_nt_shared_object(d3dkmt_handle shared_handle) +{ + struct dxgkvmb_command_destroyntsharedobject command = { }; + int ret; + + command_vm_to_host_init1(&command.hdr, + DXGK_VMBCOMMAND_DESTROYNTSHAREDOBJECT); + command.shared_handle = shared_handle; + + ret = dxgglobal_acquire_channel_lock(); + if (ret) + goto cleanup; + + ret = dxgvmb_send_sync_msg_ntstatus(dxgglobal_get_dxgvmbuschannel(), + &command, sizeof(command)); + + dxgglobal_release_channel_lock(); + +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +int dxgvmb_send_open_resource(struct dxgprocess *process, + struct dxgvmbuschannel *channel, + d3dkmt_handle device, + bool nt_security_sharing, + d3dkmt_handle global_share, + uint allocation_count, + uint total_priv_drv_data_size, + d3dkmt_handle *resource_handle, + d3dkmt_handle *alloc_handles) +{ + struct dxgkvmb_command_openresource command = { }; + struct dxgkvmb_command_openresource_return *result = NULL; + d3dkmt_handle *handles; + int ret = 0; + int i; + uint result_size = allocation_count * sizeof(d3dkmt_handle) + + sizeof(*result); + + result = dxgmem_alloc(process, DXGMEM_VMBUS, result_size); + if (result == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + command_vgpu_to_host_init2(&command.hdr, DXGK_VMBCOMMAND_OPENRESOURCE, + process->host_handle); + command.device = device; + command.nt_security_sharing = nt_security_sharing; + command.global_share = global_share; + command.allocation_count = allocation_count; + command.total_priv_drv_data_size = total_priv_drv_data_size; + + ret = dxgvmb_send_sync_msg(channel, &command, sizeof(command), + result, result_size); + if (ret) + goto cleanup; + if (!NT_SUCCESS(result->status)) { + ret = result->status; + goto cleanup; + } + + *resource_handle = result->resource; + handles = (d3dkmt_handle *) &result[1]; + for (i = 0; i < allocation_count; i++) + alloc_handles[i] = handles[i]; + +cleanup: + TRACE_FUNC_EXIT_ERR(__func__, ret); + if (result) + dxgmem_free(process, DXGMEM_VMBUS, result); + return ret; +} + +int dxgvmb_send_get_standard_alloc_priv_data(struct dxgdevice *device, + enum d3dkmdt_standardallocationtype + alloc_type, + struct d3dkmdt_gdisurfacedata + *alloc_data, + uint physical_adapter_index, + uint *alloc_priv_driver_size, + void *priv_alloc_data) +{ + struct dxgkvmb_command_getstandardallocprivdata command = { }; + struct dxgkvmb_command_getstandardallocprivdata_return *result = NULL; + uint result_size = sizeof(*result); + int ret = 0; + + TRACE_DEBUG(1, "%s", __func__); + + if (priv_alloc_data) { + result_size = *alloc_priv_driver_size; + } + result = dxgmem_alloc(device->process, DXGMEM_VMBUS, result_size); + if (result == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + command_vgpu_to_host_init2(&command.hdr, + DXGK_VMBCOMMAND_DDIGETSTANDARDALLOCATIONDRIVERDATA, + device->process->host_handle); + + command.alloc_type = alloc_type; + command.priv_driver_data_size = *alloc_priv_driver_size; + command.physical_adapter_index = physical_adapter_index; + switch (alloc_type) { + case D3DKMDT_STANDARDALLOCATION_GDISURFACE: + command.gdi_surface = *alloc_data; + break; + case D3DKMDT_STANDARDALLOCATION_SHAREDPRIMARYSURFACE: + case D3DKMDT_STANDARDALLOCATION_SHADOWSURFACE: + case D3DKMDT_STANDARDALLOCATION_STAGINGSURFACE: + default: + pr_err("Invalid standard alloc type"); + goto cleanup; + } + + ret = dxgvmb_send_sync_msg(&device->adapter->channel, + &command, sizeof(command), result, + result_size); + if (ret) + goto cleanup; + if (!NT_SUCCESS(result->status)) { + ret = result->status; + goto cleanup; + } + if (*alloc_priv_driver_size && + result->priv_driver_data_size != *alloc_priv_driver_size) { + pr_err("Priv data size mismatch"); + goto cleanup; + } + *alloc_priv_driver_size = result->priv_driver_data_size; + if (priv_alloc_data) { + memcpy(priv_alloc_data, &result[1], + result->priv_driver_data_size); + } + +cleanup: + + if (result) + dxgmem_free(device->process, DXGMEM_VMBUS, result); + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} diff --git a/drivers/gpu/dxgkrnl/dxgvmbus.h b/drivers/gpu/dxgkrnl/dxgvmbus.h new file mode 100644 index 000000000000..70f7b36d3245 --- /dev/null +++ b/drivers/gpu/dxgkrnl/dxgvmbus.h @@ -0,0 +1,859 @@ +// SPDX-License-Identifier: GPL-2.0 + +/* + * Copyright (c) 2019, Microsoft Corporation. + * + * Author: + * Iouri Tarassov iourit@microsoft.com + * + * Dxgkrnl Graphics Port Driver + * VM bus interface with the host definitions + * + */ + +#ifndef _DXGVMBUS_H +#define _DXGVMBUS_H + +#include "d3dkmthk.h" + +struct dxgprocess; +struct dxgadapter; + +#define DXG_MAX_VM_BUS_PACKET_SIZE (1024 * 128) + +#define DXGK_DECL_VMBUS_OUTPUTSIZE(Type)\ + ((sizeof(##Type) + 0x7) & ~(uint)0x7) +#define DXGK_DECL_VMBUS_ALIGN_FOR_OUTPUT(Size) (((Size) + 0x7) & ~(uint)0x7) +/* + * Defines a structure, which has the size, multiple of 8 bytes. + */ +#define DXGK_DECL_ALIGNED8_STRUCT(Type, Name, OutputSize) \ + const uint _Size = DXGK_DECL_VMBUS_OUTPUTSIZE(Type); \ + uint8_t _AlignedStruct[_Size]; \ + ##Type & Name = (##Type &)_AlignedStruct; \ + uint OutputSize = _Size + +#define DXGK_BUFFER_VMBUS_ALIGNED(Buffer) (((Buffer) & 7) == 0) + +enum dxgkvmb_commandchanneltype { + DXGKVMB_VGPU_TO_HOST, + DXGKVMB_VM_TO_HOST, + DXGKVMB_HOST_TO_VM +}; + +/* + * + * Commands, sent to the host via the guest global VM bus channel + * DXG_GUEST_GLOBAL_VMBUS + * + */ + +#define DXG_VM_PROCESS_NAME_LENGTH 260 + +enum dxgkvmb_commandtype_global { + DXGK_VMBCOMMAND_VM_TO_HOST_FIRST = 1000, + DXGK_VMBCOMMAND_CREATEPROCESS = DXGK_VMBCOMMAND_VM_TO_HOST_FIRST, + DXGK_VMBCOMMAND_DESTROYPROCESS = 1001, + DXGK_VMBCOMMAND_OPENSYNCOBJECT = 1002, + DXGK_VMBCOMMAND_DESTROYSYNCOBJECT = 1003, + DXGK_VMBCOMMAND_CREATENTSHAREDOBJECT = 1004, + DXGK_VMBCOMMAND_DESTROYNTSHAREDOBJECT = 1005, + DXGK_VMBCOMMAND_SIGNALFENCE = 1006, + DXGK_VMBCOMMAND_NOTIFYPROCESSFREEZE = 1007, + DXGK_VMBCOMMAND_NOTIFYPROCESSTHAW = 1008, + DXGK_VMBCOMMAND_QUERYETWSESSION = 1009, + DXGK_VMBCOMMAND_SETIOSPACEREGION = 1010, + DXGK_VMBCOMMAND_COMPLETETRANSACTION = 1011, + DXGK_VMBCOMMAND_INVALID_VM_TO_HOST +}; + +/* + * + * Commands, sent to the host via the per adapter VM bus channel + * DXG_GUEST_VGPU_VMBUS + * + */ + +enum dxgkvmb_commandtype { + DXGK_VMBCOMMAND_CREATEDEVICE = 0, + DXGK_VMBCOMMAND_DESTROYDEVICE = 1, + DXGK_VMBCOMMAND_QUERYADAPTERINFO = 2, + DXGK_VMBCOMMAND_DDIQUERYADAPTERINFO = 3, + DXGK_VMBCOMMAND_CREATEALLOCATION = 4, + DXGK_VMBCOMMAND_DESTROYALLOCATION = 5, + DXGK_VMBCOMMAND_CREATECONTEXTVIRTUAL = 6, + DXGK_VMBCOMMAND_DESTROYCONTEXT = 7, + DXGK_VMBCOMMAND_CREATESYNCOBJECT = 8, + DXGK_VMBCOMMAND_CREATEPAGINGQUEUE = 9, + DXGK_VMBCOMMAND_DESTROYPAGINGQUEUE = 10, + DXGK_VMBCOMMAND_MAKERESIDENT = 11, + DXGK_VMBCOMMAND_EVICT = 12, + DXGK_VMBCOMMAND_ESCAPE = 13, + DXGK_VMBCOMMAND_OPENADAPTER = 14, + DXGK_VMBCOMMAND_CLOSEADAPTER = 15, + DXGK_VMBCOMMAND_FREEGPUVIRTUALADDRESS = 16, + DXGK_VMBCOMMAND_MAPGPUVIRTUALADDRESS = 17, + DXGK_VMBCOMMAND_RESERVEGPUVIRTUALADDRESS = 18, + DXGK_VMBCOMMAND_UPDATEGPUVIRTUALADDRESS = 19, + DXGK_VMBCOMMAND_SUBMITCOMMAND = 20, + dxgk_vmbcommand_queryvideomemoryinfo = 21, + DXGK_VMBCOMMAND_WAITFORSYNCOBJECTFROMCPU = 22, + DXGK_VMBCOMMAND_LOCK2 = 23, + DXGK_VMBCOMMAND_UNLOCK2 = 24, + DXGK_VMBCOMMAND_WAITFORSYNCOBJECTFROMGPU = 25, + DXGK_VMBCOMMAND_SIGNALSYNCOBJECT = 26, + DXGK_VMBCOMMAND_SIGNALFENCENTSHAREDBYREF = 27, + dxgk_vmbcommand_getdevicestate = 28, + DXGK_VMBCOMMAND_MARKDEVICEASERROR = 29, + DXGK_VMBCOMMAND_ADAPTERSTOP = 30, + DXGK_VMBCOMMAND_SETQUEUEDLIMIT = 31, + DXGK_VMBCOMMAND_OPENRESOURCE = 32, + DXGK_VMBCOMMAND_SETCONTEXTSCHEDULINGPRIORITY = 33, + DXGK_VMBCOMMAND_PRESENTHISTORYTOKEN = 34, + DXGK_VMBCOMMAND_SETREDIRECTEDFLIPFENCEVALUE = 35, + DXGK_VMBCOMMAND_GETINTERNALADAPTERINFO = 36, + DXGK_VMBCOMMAND_FLUSHHEAPTRANSITIONS = 37, + DXGK_VMBCOMMAND_BLT = 38, + DXGK_VMBCOMMAND_DDIGETSTANDARDALLOCATIONDRIVERDATA = 39, + DXGK_VMBCOMMAND_CDDGDICOMMAND = 40, + DXGK_VMBCOMMAND_QUERYALLOCATIONRESIDENCY = 41, + DXGK_VMBCOMMAND_FLUSHDEVICE = 42, + DXGK_VMBCOMMAND_FLUSHADAPTER = 43, + DXGK_VMBCOMMAND_DDIGETNODEMETADATA = 44, + DXGK_VMBCOMMAND_SETEXISTINGSYSMEMSTORE = 45, + DXGK_VMBCOMMAND_ISSYNCOBJECTSIGNALED = 46, + DXGK_VMBCOMMAND_CDDSYNCGPUACCESS = 47, + DXGK_VMBCOMMAND_QUERYSTATISTICS = 48, + DXGK_VMBCOMMAND_CHANGEVIDEOMEMORYRESERVATION = 49, + DXGK_VMBCOMMAND_CREATEHWQUEUE = 50, + DXGK_VMBCOMMAND_DESTROYHWQUEUE = 51, + DXGK_VMBCOMMAND_SUBMITCOMMANDTOHWQUEUE = 52, + DXGK_VMBCOMMAND_GETDRIVERSTOREFILE = 53, + DXGK_VMBCOMMAND_READDRIVERSTOREFILE = 54, + DXGK_VMBCOMMAND_GETNEXTHARDLINK = 55, + DXGK_VMBCOMMAND_UPDATEALLOCATIONPROPERTY = 56, + DXGK_VMBCOMMAND_OFFERALLOCATIONS = 57, + DXGK_VMBCOMMAND_RECLAIMALLOCATIONS = 58, + DXGK_VMBCOMMAND_SETALLOCATIONPRIORITY = 59, + DXGK_VMBCOMMAND_GETALLOCATIONPRIORITY = 60, + DXGK_VMBCOMMAND_GETCONTEXTSCHEDULINGPRIORITY = 61, + DXGK_VMBCOMMAND_QUERYCLOCKCALIBRATION = 62, + DXGK_VMBCOMMAND_QUERYRESOURCEINFO = 64, + DXGK_VMBCOMMAND_INVALID +}; + +enum dxgkvmb_commandtype_host_to_vm { + DXGK_VMBCOMMAND_SIGNALGUESTEVENT, + DXGK_VMBCOMMAND_PROPAGATEPRESENTHISTORYTOKEN, + DXGK_VMBCOMMAND_SETGUESTDATA, + DXGK_VMBCOMMAND_SIGNALGUESTEVENTPASSIVE, + DXGK_VMBCOMMAND_SENDWNFNOTIFICATION, + DXGK_VMBCOMMAND_INVALID_HOST_TO_VM +}; + +struct dxgkvmb_command_vm_to_host { + uint64_t command_id; + d3dkmt_handle process; + enum dxgkvmb_commandchanneltype channel_type; + enum dxgkvmb_commandtype_global command_type; +}; + +struct dxgkvmb_command_vgpu_to_host { + uint64_t command_id; + d3dkmt_handle process; + u32 channel_type; + enum dxgkvmb_commandtype command_type; +}; + +struct dxgkvmb_command_host_to_vm { + uint64_t command_id; + d3dkmt_handle process; + enum dxgkvmb_commandchanneltype channel_type; + enum dxgkvmb_commandtype_host_to_vm command_type; +}; + +struct dxgkvmb_command_signalguestevent { + struct dxgkvmb_command_host_to_vm hdr; + uint64_t event; + winhandle process_id; + bool dereference_event; +}; + +struct dxgkvmb_command_opensyncobject { + struct dxgkvmb_command_vm_to_host hdr; + d3dkmt_handle device; + d3dkmt_handle global_sync_object; + uint engine_affinity; + struct d3dddi_synchronizationobject_flags flags; +}; + +struct dxgkvmb_command_opensyncobject_return { + d3dkmt_handle sync_object; + ntstatus status; + d3dgpu_virtual_address gpu_virtual_address; + uint64_t guest_cpu_physical_address; +}; + +/* + * The command returns d3dkmt_handle of a shared object for the + * given pre-process object + */ +struct dxgkvmb_command_createntsharedobject { + struct dxgkvmb_command_vm_to_host hdr; + d3dkmt_handle object; +}; + +/* The command returns ntstatus */ +struct dxgkvmb_command_destroyntsharedobject { + struct dxgkvmb_command_vm_to_host hdr; + d3dkmt_handle shared_handle; +}; + +/* Returns ntstatus */ +struct dxgkvmb_command_setiospaceregion { + struct dxgkvmb_command_vm_to_host hdr; + uint64_t start; + uint64_t length; + uint shared_page_gpadl; +}; + +/* Returns ntstatus */ +struct dxgkvmb_command_setexistingsysmemstore { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle device; + d3dkmt_handle allocation; + uint gpadl; +}; + +struct dxgkvmb_command_createprocess { + struct dxgkvmb_command_vm_to_host hdr; + void *process; + winhandle process_id; + winwchar process_name[DXG_VM_PROCESS_NAME_LENGTH + 1]; + bool csrss_process:1; + bool dwm_process:1; + bool wow64_process:1; + bool linux_process:1; +}; + +struct dxgkvmb_command_createprocess_return { + d3dkmt_handle hprocess; +}; + +// The command returns ntstatus +struct dxgkvmb_command_destroyprocess { + struct dxgkvmb_command_vm_to_host hdr; +}; + +struct dxgkvmb_command_openadapter { + struct dxgkvmb_command_vgpu_to_host hdr; + uint vmbus_interface_version; + uint vmbus_last_compatible_interface_version; + struct winluid guest_adapter_luid; +}; + +struct dxgkvmb_command_openadapter_return { + d3dkmt_handle host_adapter_handle; + ntstatus status; + uint vmbus_interface_version; + uint vmbus_last_compatible_interface_version; +}; + +struct dxgkvmb_command_closeadapter { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle host_handle; +}; + +struct dxgkvmb_command_getinternaladapterinfo { + struct dxgkvmb_command_vgpu_to_host hdr; +}; + +struct dxgkvmb_command_getinternaladapterinfo_return { + struct dxgk_device_types device_types; + uint driver_store_copy_mode; + uint driver_ddi_version; + uint secure_virtual_machine:1; + uint virtual_machine_reset:1; + uint is_vail_supported:1; + struct winluid host_adapter_luid; + winwchar device_description[80]; + winwchar device_instance_id[W_MAX_PATH]; +}; + +struct dxgkvmb_command_queryadapterinfo { + struct dxgkvmb_command_vgpu_to_host hdr; + enum kmtqueryadapterinfotype query_type; + uint private_data_size; + uint8_t private_data[1]; +}; + +/* Returns ntstatus */ +struct dxgkvmb_command_setallocationpriority { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle device; + d3dkmt_handle resource; + uint allocation_count; + /* d3dkmt_handle allocations[allocation_count or 0]; */ + /* uint priorities[allocation_count or 1]; */ +}; + +struct dxgkvmb_command_getallocationpriority { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle device; + d3dkmt_handle resource; + uint allocation_count; + /* d3dkmt_handle allocations[allocation_count or 0]; */ +}; + +struct dxgkvmb_command_getallocationpriority_return { + ntstatus status; + /* uint priorities[allocation_count or 1]; */ +}; + +/* Returns ntstatus */ +struct dxgkvmb_command_setcontextschedulingpriority { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle context; + int priority; +}; + +/* Returns ntstatus */ +struct dxgkvmb_command_setcontextschedulingpriority2 { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle context; + int priority; + bool in_process; +}; + +struct dxgkvmb_command_getcontextschedulingpriority { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle context; + bool in_process; +}; + +struct dxgkvmb_command_getcontextschedulingpriority_return { + ntstatus status; + int priority; +}; + +struct dxgkvmb_command_createdevice { + struct dxgkvmb_command_vgpu_to_host hdr; + struct d3dkmt_createdeviceflags flags; + bool cdd_device; + void *error_code; +}; + +struct dxgkvmb_command_createdevice_return { + d3dkmt_handle device; +}; + +struct dxgkvmb_command_destroydevice { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle device; +}; + +struct dxgkvmb_command_makeresident { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle device; + d3dkmt_handle paging_queue; + struct d3dddi_makeresident_flags flags; + uint alloc_count; + d3dkmt_handle allocations[1]; +}; + +struct dxgkvmb_command_makeresident_return { + uint64_t paging_fence_value; + uint64_t num_bytes_to_trim; + ntstatus status; +}; + +struct dxgkvmb_command_evict { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle device; + struct d3dddi_evict_flags flags; + uint alloc_count; + d3dkmt_handle allocations[1]; +}; + +struct dxgkvmb_command_evict_return { + uint64_t num_bytes_to_trim; +}; + +struct dxgkvmb_command_submitcommand { + struct dxgkvmb_command_vgpu_to_host hdr; + struct d3dkmt_submitcommand args; + /* HistoryBufferHandles */ + /* PrivateDriverData */ +}; + +struct dxgkvmb_command_submitcommandtohwqueue { + struct dxgkvmb_command_vgpu_to_host hdr; + struct d3dkmt_submitcommandtohwqueue args; + /* Written primaries */ + /* PrivateDriverData */ +}; + +/* Returns ntstatus */ +struct dxgkvmb_command_flushheaptransitions { + struct dxgkvmb_command_vgpu_to_host hdr; +}; + +struct dxgkvmb_command_freegpuvirtualaddress { + struct dxgkvmb_command_vgpu_to_host hdr; + struct d3dkmt_freegpuvirtualaddress args; +}; + +struct dxgkvmb_command_mapgpuvirtualaddress { + struct dxgkvmb_command_vgpu_to_host hdr; + struct d3dddi_mapgpuvirtualaddress args; + d3dkmt_handle device; +}; + +struct dxgkvmb_command_mapgpuvirtualaddress_return { + d3dgpu_virtual_address virtual_address; + uint64_t paging_fence_value; + ntstatus status; +}; + +struct dxgkvmb_command_reservegpuvirtualaddress { + struct dxgkvmb_command_vgpu_to_host hdr; + struct d3dddi_reservegpuvirtualaddress args; +}; + +struct dxgkvmb_command_reservegpuvirtualaddress_return { + d3dgpu_virtual_address virtual_address; + uint64_t paging_fence_value; +}; + +struct dxgkvmb_command_updategpuvirtualaddress { + struct dxgkvmb_command_vgpu_to_host hdr; + uint64_t fence_value; + d3dkmt_handle device; + d3dkmt_handle context; + d3dkmt_handle fence_object; + uint num_operations; + uint flags; + struct d3dddi_updategpuvirtualaddress_operation operations[1]; +}; + +struct dxgkvmb_command_queryclockcalibration { + struct dxgkvmb_command_vgpu_to_host hdr; + struct d3dkmt_queryclockcalibration args; +}; + +struct dxgkvmb_command_queryclockcalibration_return { + ntstatus status; + struct dxgk_gpuclockdata clock_data; +}; + +struct dxgkvmb_command_createallocation_allocinfo { + uint flags; + uint priv_drv_data_size; + uint vidpn_source_id; +}; + +struct dxgkvmb_command_createallocation { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle device; + d3dkmt_handle resource; + uint private_runtime_data_size; + uint priv_drv_data_size; + uint alloc_count; + struct d3dkmt_createallocationflags flags; + winhandle private_runtime_resource_handle; + bool make_resident; +/* dxgkvmb_command_createallocation_allocinfo alloc_info[alloc_count]; */ +/* uint8_t private_rutime_data[private_runtime_data_size] */ +/* uint8_t priv_drv_data[] for each alloc_info */ +}; + +struct dxgkvmb_command_openresource { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle device; + bool nt_security_sharing; + d3dkmt_handle global_share; + uint allocation_count; + uint total_priv_drv_data_size; +}; + +struct dxgkvmb_command_openresource_return { + d3dkmt_handle resource; + ntstatus status; +/* d3dkmt_handle allocation[allocation_count]; */ +}; + +struct dxgkvmb_command_getstandardallocprivdata { + struct dxgkvmb_command_vgpu_to_host hdr; + enum d3dkmdt_standardallocationtype alloc_type; + uint priv_driver_data_size; + uint priv_driver_resource_size; + uint physical_adapter_index; + union { + struct d3dkmdt_sharedprimarysurfacedata primary; + struct d3dkmdt_shadowsurfacedata shadow; + struct d3dkmtd_stagingsurfacedata staging; + struct d3dkmdt_gdisurfacedata gdi_surface; + }; +}; + +struct dxgkvmb_command_getstandardallocprivdata_return { + ntstatus status; + uint priv_driver_data_size; + uint priv_driver_resource_size; + union { + struct d3dkmdt_sharedprimarysurfacedata primary; + struct d3dkmdt_shadowsurfacedata shadow; + struct d3dkmtd_stagingsurfacedata staging; + struct d3dkmdt_gdisurfacedata gdi_surface; + }; +/* char alloc_priv_data[priv_driver_data_size]; */ +/* char resource_priv_data[priv_driver_resource_size]; */ +}; + +struct dxgkarg_describeallocation { + winhandle allocation; + uint width; + uint height; + uint format; + uint multisample_method; + struct d3dddi_rational refresh_rate; + uint private_driver_attribute; + uint flags; + uint rotation; +}; + +struct dxgkvmb_allocflags { + union { + uint flags; + struct { + uint primary:1; + uint cdd_primary:1; + uint dod_primary:1; + uint overlay:1; + uint reserved6:1; + uint capture:1; + uint reserved0:4; + uint reserved1:1; + uint existing_sysmem:1; + uint stereo:1; + uint direct_flip:1; + uint hardware_protected:1; + uint reserved2:1; + uint reserved3:1; + uint reserved4:1; + uint protected:1; + uint cached:1; + uint independent_primary:1; + uint reserved:11; + }; + }; +}; + +struct dxgkvmb_command_allocinfo_return { + d3dkmt_handle allocation; + uint priv_drv_data_size; + struct dxgkvmb_allocflags allocation_flags; + uint64_t allocation_size; + struct dxgkarg_describeallocation driver_info; +}; + +struct dxgkvmb_command_createallocation_return { + struct d3dkmt_createallocationflags flags; + d3dkmt_handle resource; + d3dkmt_handle global_share; + uint vgpu_flags; + struct dxgkvmb_command_allocinfo_return allocation_info[1]; + /* Private driver data for allocations */ +}; + +/* The command returns ntstatus */ +struct dxgkvmb_command_destroyallocation { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle device; + d3dkmt_handle resource; + uint alloc_count; + struct d3dddicb_destroyallocation2flags flags; + d3dkmt_handle allocations[1]; +}; + +struct dxgkvmb_command_createcontextvirtual { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle context; + d3dkmt_handle device; + uint node_ordinal; + uint engine_affinity; + struct d3dddi_createcontextflags flags; + enum d3dkmt_clienthint client_hint; + uint priv_drv_data_size; + uint8_t priv_drv_data[1]; +}; + +/* The command returns ntstatus */ +struct dxgkvmb_command_destroycontext { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle context; +}; + +struct dxgkvmb_command_createpagingqueue { + struct dxgkvmb_command_vgpu_to_host hdr; + struct d3dkmt_createpagingqueue args; +}; + +struct dxgkvmb_command_createpagingqueue_return { + d3dkmt_handle paging_queue; + d3dkmt_handle sync_object; + uint64_t fence_storage_physical_address; + uint64_t fence_storage_offset; +}; + +struct dxgkvmb_command_destroypagingqueue { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle paging_queue; +}; + +struct dxgkvmb_command_createsyncobject { + struct dxgkvmb_command_vgpu_to_host hdr; + struct d3dkmt_createsynchronizationobject2 args; + uint client_hint; +}; + +struct dxgkvmb_command_createsyncobject_return { + d3dkmt_handle sync_object; + d3dkmt_handle global_sync_object; + d3dgpu_virtual_address fence_gpu_va; + uint64_t fence_storage_address; + uint fence_storage_offset; +}; + +/* The command returns ntstatus */ +struct dxgkvmb_command_destroysyncobject { + struct dxgkvmb_command_vm_to_host hdr; + d3dkmt_handle sync_object; +}; + +/* The command returns ntstatus */ +struct dxgkvmb_command_signalsyncobject { + struct dxgkvmb_command_vgpu_to_host hdr; + uint object_count; + struct d3dddicb_signalflags flags; + uint context_count; + uint64_t fence_value; + union { + /* Pointer to the guest event object */ + winhandle cpu_event_handle; + /* Non zero when signal from CPU is done */ + d3dkmt_handle device; + }; + /* d3dkmt_handle ObjectHandleArray[object_count] */ + /* d3dkmt_handle ContextArray[context_count] */ + /* uint64_t MonitoredFenceValueArray[object_count] */ +}; + +/* The command returns ntstatus */ +struct dxgkvmb_command_waitforsyncobjectfromcpu { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle device; + uint object_count; + struct d3dddi_waitforsynchronizationobjectfromcpu_flags flags; + uint64_t guest_event_pointer; + bool dereference_event; + /* d3dkmt_handle ObjectHandleArray[object_count] */ + /* uint64_t FenceValueArray [object_count] */ +}; + +/* The command returns ntstatus */ +struct dxgkvmb_command_waitforsyncobjectfromgpu { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle context; + /* Must be 1 when bLegacyFenceObject is TRUE */ + uint object_count; + bool legacy_fence_object; + uint64_t fence_values[1]; + /* d3dkmt_handle ObjectHandles[object_count] */ +}; + +struct dxgkvmb_command_lock2 { + struct dxgkvmb_command_vgpu_to_host hdr; + struct d3dkmt_lock2 args; + bool use_legacy_lock; + uint flags; + uint priv_drv_data; +}; + +struct dxgkvmb_command_lock2_return { + ntstatus status; + void *cpu_visible_buffer_offset; +}; + +struct dxgkvmb_command_unlock2 { + struct dxgkvmb_command_vgpu_to_host hdr; + struct d3dkmt_unlock2 args; + bool use_legacy_unlock; +}; + +struct dxgkvmb_command_updateallocationproperty { + struct dxgkvmb_command_vgpu_to_host hdr; + struct d3dddi_updateallocproperty args; +}; + +struct dxgkvmb_command_updateallocationproperty_return { + uint64_t paging_fence_value; + ntstatus status; +}; + +struct dxgkvmb_command_markdeviceaserror { + struct dxgkvmb_command_vgpu_to_host hdr; + struct d3dkmt_markdeviceaserror args; +}; + +/* Returns ntstatus */ +struct dxgkvmb_command_offerallocations { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle device; + uint allocation_count; + enum d3dkmt_offer_priority priority; + struct d3dkmt_offer_flags flags; + bool resources; + d3dkmt_handle allocations[1]; +}; + +struct dxgkvmb_command_reclaimallocations { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle device; + d3dkmt_handle paging_queue; + uint allocation_count; + bool resources; + bool write_results; + d3dkmt_handle allocations[1]; +}; + +struct dxgkvmb_command_reclaimallocations_return { + uint64_t paging_fence_value; + ntstatus status; + enum d3dddi_reclaim_result discarded[1]; +}; + +/* Returns ntstatus */ +struct dxgkvmb_command_changevideomemoryreservation { + struct dxgkvmb_command_vgpu_to_host hdr; + struct d3dkmt_changevideomemoryreservation args; +}; + +/* Returns the same structure */ +struct dxgkvmb_command_createhwqueue { + struct dxgkvmb_command_vgpu_to_host hdr; + ntstatus status; + d3dkmt_handle hwqueue; + d3dkmt_handle hwqueue_progress_fence; + void *hwqueue_progress_fence_cpuva; + d3dgpu_virtual_address hwqueue_progress_fence_gpuva; + d3dkmt_handle context; + struct d3dddi_createhwqueueflags flags; + uint priv_drv_data_size; + uint8_t priv_drv_data[1]; +}; + +/* The command returns ntstatus */ +struct dxgkvmb_command_destroyhwqueue { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle hwqueue; +}; + +struct dxgkvmb_command_queryallocationresidency { + struct dxgkvmb_command_vgpu_to_host hdr; + struct d3dkmt_queryallocationresidency args; + /* d3dkmt_handle allocations[0 or number of allocations] */ +}; + +struct dxgkvmb_command_queryallocationresidency_return { + ntstatus status; + /* d3dkmt_allocationresidencystatus[NumAllocations] */ +}; + +/* Returns only private data */ +struct dxgkvmb_command_escape { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle adapter; + d3dkmt_handle device; + enum d3dkmt_escapetype type; + struct d3dddi_escapeflags flags; + uint priv_drv_data_size; + d3dkmt_handle context; + uint8_t priv_drv_data[1]; +}; + +struct dxgkvmb_command_queryvideomemoryinfo { + struct dxgkvmb_command_vgpu_to_host hdr; + d3dkmt_handle adapter; + enum d3dkmt_memory_segment_group memory_segment_group; + uint physical_adapter_index; +}; + +struct dxgkvmb_command_queryvideomemoryinfo_return { + uint64_t budget; + uint64_t current_usage; + uint64_t current_reservation; + uint64_t available_for_reservation; +}; + +struct dxgkvmb_command_getdevicestate { + struct dxgkvmb_command_vgpu_to_host hdr; + struct d3dkmt_getdevicestate args; +}; + +struct dxgkvmb_command_getdevicestate_return { + struct d3dkmt_getdevicestate args; + ntstatus status; +}; + +/* + * Helper functions + */ +static inline void command_vm_to_host_init2(struct dxgkvmb_command_vm_to_host + *command, + enum dxgkvmb_commandtype_global + type, d3dkmt_handle process) +{ + command->command_type = type; + command->process = process; + command->command_id = 0; + command->channel_type = DXGKVMB_VM_TO_HOST; +} + +static inline void command_vgpu_to_host_init0(struct dxgkvmb_command_vm_to_host + *command) +{ + command->command_type = DXGK_VMBCOMMAND_INVALID; + command->process = 0; + command->command_id = 0; + command->channel_type = DXGKVMB_VGPU_TO_HOST; +} + +static inline void command_vgpu_to_host_init1(struct + dxgkvmb_command_vgpu_to_host + *command, + enum dxgkvmb_commandtype type) +{ + command->command_type = type; + command->process = 0; + command->command_id = 0; + command->channel_type = DXGKVMB_VGPU_TO_HOST; +} + +static inline void command_vgpu_to_host_init2(struct + dxgkvmb_command_vgpu_to_host + *command, + enum dxgkvmb_commandtype type, + d3dkmt_handle process_handle) +{ + command->command_type = type; + command->process = process_handle; + command->command_id = 0; + command->channel_type = DXGKVMB_VGPU_TO_HOST; +} + +int dxgvmb_send_sync_msg(struct dxgvmbuschannel *channel, + void *command, + u32 command_size, void *result, u32 result_size); + +#endif /* _DXGVMBUS_H */ diff --git a/drivers/gpu/dxgkrnl/hmgr.c b/drivers/gpu/dxgkrnl/hmgr.c new file mode 100644 index 000000000000..450fb67e703f --- /dev/null +++ b/drivers/gpu/dxgkrnl/hmgr.c @@ -0,0 +1,593 @@ +// SPDX-License-Identifier: GPL-2.0 + +/* + * Copyright (c) 2019, Microsoft Corporation. + * + * Author: + * Iouri Tarassov iourit@microsoft.com + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License version 2 as published + * by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or + * NON INFRINGEMENT. See the GNU General Public License for more + * details. + */ + +/* + * Dxgkrnl Graphics Port Driver handle manager + * + */ + +#include <linux/kernel.h> +#include <linux/mutex.h> +#include <linux/rwsem.h> + +#include "misc.h" +#include "dxgkrnl.h" +#include "hmgr.h" + +/* + * Handle parameters + */ +#define HMGRHANDLE_INSTANCE_BITS 6 +#define HMGRHANDLE_INDEX_BITS 24 +#define HMGRHANDLE_UNIQUE_BITS 2 + +#define HMGRHANDLE_INSTANCE_SHIFT 0 +#define HMGRHANDLE_INDEX_SHIFT \ + (HMGRHANDLE_INSTANCE_BITS + HMGRHANDLE_INSTANCE_SHIFT) +#define HMGRHANDLE_UNIQUE_SHIFT \ + (HMGRHANDLE_INDEX_BITS + HMGRHANDLE_INDEX_SHIFT) + +#define HMGRHANDLE_INSTANCE_MASK \ + (((1 << HMGRHANDLE_INSTANCE_BITS) - 1) << HMGRHANDLE_INSTANCE_SHIFT) +#define HMGRHANDLE_INDEX_MASK \ + (((1 << HMGRHANDLE_INDEX_BITS) - 1) << HMGRHANDLE_INDEX_SHIFT) +#define HMGRHANDLE_UNIQUE_MASK \ + (((1 << HMGRHANDLE_UNIQUE_BITS) - 1) << HMGRHANDLE_UNIQUE_SHIFT) + +#define HMGRHANDLE_INSTANCE_MAX ((1 << HMGRHANDLE_INSTANCE_BITS) - 1) +#define HMGRHANDLE_INDEX_MAX ((1 << HMGRHANDLE_INDEX_BITS) - 1) +#define HMGRHANDLE_UNIQUE_MAX ((1 << HMGRHANDLE_UNIQUE_BITS) - 1) + +/* + * Handle entry + */ +struct hmgrentry { + union { + void *object; + struct { + uint prev_free_index; + uint next_free_index; + }; + }; + uint type:HMGRENTRY_TYPE_BITS + 1; + uint unique:HMGRHANDLE_UNIQUE_BITS; + uint instance:HMGRHANDLE_INSTANCE_BITS; + uint destroyed:1; +}; + +#define HMGRTABLE_SIZE_INCREMENT 1024 +#define HMGRTABLE_MIN_FREE_ENTRIES 128 +#define HMGRTABLE_INVALID_INDEX (~((1 << HMGRHANDLE_INDEX_BITS) - 1)) +#define HMGRTABLE_SIZE_MAX 0xFFFFFFF + +static uint table_size_increment = HMGRTABLE_SIZE_INCREMENT; + +static inline uint get_unique(d3dkmt_handle h) +{ + return (h & HMGRHANDLE_UNIQUE_MASK) >> HMGRHANDLE_UNIQUE_SHIFT; +} + +static uint get_index(d3dkmt_handle h) +{ + return (h & HMGRHANDLE_INDEX_MASK) >> HMGRHANDLE_INDEX_SHIFT; +} + +uint get_instance(d3dkmt_handle h) +{ + return (h & HMGRHANDLE_INSTANCE_MASK) >> HMGRHANDLE_INSTANCE_SHIFT; +} + +static bool is_handle_valid(struct hmgrtable *table, d3dkmt_handle h, + bool ignore_destroyed, enum hmgrentry_type t) +{ + uint index = get_index(h); + uint unique = get_unique(h); + struct hmgrentry *entry; + + if (index >= table->table_size) { + pr_err("%s Invalid index %x %d\n", __func__, h, index); + return false; + } + + entry = &table->entry_table[index]; + if (unique != entry->unique) { + pr_err("%s Invalid unique %x %d %d %d %p", + __func__, h, unique, entry->unique, + index, entry->object); + return false; + } + + if (entry->destroyed && !ignore_destroyed) { + pr_err("%s Invalid destroyed", __func__); + return false; + } + + if (entry->type == HMGRENTRY_TYPE_FREE) { + pr_err("%s Entry is freed %x %d", __func__, h, index); + return false; + } + + if (t != HMGRENTRY_TYPE_FREE && t != entry->type) { + pr_err("%s type mismatch %x %d %d", __func__, h, + t, entry->type); + return false; + } + + return true; +} + +static d3dkmt_handle build_handle(uint index, uint unique, uint instance) +{ + uint handle_bits; + + handle_bits = (index << HMGRHANDLE_INDEX_SHIFT) & HMGRHANDLE_INDEX_MASK; + handle_bits |= (unique << HMGRHANDLE_UNIQUE_SHIFT) & + HMGRHANDLE_UNIQUE_MASK; + handle_bits |= (instance << HMGRHANDLE_INSTANCE_SHIFT) & + HMGRHANDLE_INSTANCE_MASK; + + return (d3dkmt_handle) handle_bits; +} + +inline uint hmgrtable_get_used_entry_count(struct hmgrtable *table) +{ + DXGKRNL_ASSERT(table->table_size >= table->free_count); + return (table->table_size - table->free_count); +} + +bool hmgrtable_mark_destroyed(struct hmgrtable *table, d3dkmt_handle h) +{ + if (!is_handle_valid(table, h, false, HMGRENTRY_TYPE_FREE)) + return false; + + table->entry_table[get_index(h)].destroyed = true; + return true; +} + +bool hmgrtable_unmark_destroyed(struct hmgrtable *table, d3dkmt_handle h) +{ + if (!is_handle_valid(table, h, true, HMGRENTRY_TYPE_FREE)) + return true; + + DXGKRNL_ASSERT(table->entry_table[get_index(h)].destroyed); + table->entry_table[get_index(h)].destroyed = 0; + return true; +} + +static inline bool is_empty(struct hmgrtable *table) +{ + return (table->free_count == table->table_size); +} + +void print_status(struct hmgrtable *table) +{ + int i; + + TRACE_DEBUG(1, "hmgrtable head, tail %p %d %d\n", + table, table->free_handle_list_head, + table->free_handle_list_tail); + if (table->entry_table == NULL) + return; + for (i = 0; i < 3; i++) { + if (table->entry_table[i].type != HMGRENTRY_TYPE_FREE) + TRACE_DEBUG(1, "hmgrtable entry %p %d %p\n", + table, i, table->entry_table[i].object); + else + TRACE_DEBUG(1, "hmgrtable entry %p %d %d %d\n", + table, i, + table->entry_table[i].next_free_index, + table->entry_table[i].prev_free_index); + } +} + +static bool expand_table(struct hmgrtable *table, uint NumEntries) +{ + uint new_table_size; + struct hmgrentry *new_entry; + uint table_index; + uint new_free_count; + uint prev_free_index; + uint tail_index = table->free_handle_list_tail; + + TRACE_DEBUG(1, "%s\n", __func__); + + /* The tail should point to the last free element in the list */ + if (!(table->free_count == 0 || + table->entry_table[tail_index].next_free_index == + HMGRTABLE_INVALID_INDEX)) { + pr_err("%s:corruption\n", __func__); + return false; + } + + new_table_size = table->table_size + table_size_increment; + if (new_table_size < NumEntries) + new_table_size = NumEntries; + + if (new_table_size > HMGRHANDLE_INDEX_MAX) { + pr_err("%s:corruption\n", __func__); + return false; + } + + new_entry = (struct hmgrentry *) + dxgmem_alloc(table->process, DXGMEM_HANDLE_TABLE, + new_table_size * sizeof(struct hmgrentry)); + if (new_entry == NULL) { + pr_err("%s:allocation failed\n", __func__); + return false; + } + + if (table->entry_table) { + memcpy(new_entry, table->entry_table, + table->table_size * sizeof(struct hmgrentry)); + dxgmem_free(table->process, DXGMEM_HANDLE_TABLE, + table->entry_table); + } else { + table->free_handle_list_head = 0; + } + + table->entry_table = new_entry; + + /* Initialize new table entries and add to the free list */ + table_index = table->table_size; + new_free_count = table->free_count + table_size_increment; + + prev_free_index = table->free_handle_list_tail; + + while (table_index < new_table_size) { + struct hmgrentry *entry = &table->entry_table[table_index]; + + entry->prev_free_index = prev_free_index; + entry->next_free_index = table_index + 1; + entry->type = HMGRENTRY_TYPE_FREE; + entry->unique = 1; + entry->instance = 0; + prev_free_index = table_index; + + table_index++; + } + + table->entry_table[table_index - 1].next_free_index = + (uint) HMGRTABLE_INVALID_INDEX; + + if (table->free_count != 0) { + /* Link the current free list with the new entries */ + struct hmgrentry *entry; + + entry = &table->entry_table[table->free_handle_list_tail]; + entry->next_free_index = table->table_size; + } + table->free_handle_list_tail = new_table_size - 1; + if (table->free_handle_list_head == HMGRTABLE_INVALID_INDEX) + table->free_handle_list_head = table->table_size; + + table->table_size = new_table_size; + table->free_count = new_free_count; + + TRACE_DEBUG(1, "%s end\n", __func__); + return true; +} + +void hmgrtable_init(struct hmgrtable *table, struct dxgprocess *process) +{ + table->process = process; + table->entry_table = NULL; + table->table_size = 0; + table->free_handle_list_head = HMGRTABLE_INVALID_INDEX; + table->free_handle_list_tail = HMGRTABLE_INVALID_INDEX; + table->free_count = 0; + init_rwsem(&table->table_lock); +} + +void hmgrtable_destroy(struct hmgrtable *table) +{ + if (table->entry_table) { + dxgmem_free(table->process, DXGMEM_HANDLE_TABLE, + table->entry_table); + table->entry_table = NULL; + } +} + +void hmgrtable_lock(struct hmgrtable *table, enum dxglockstate state) +{ + dxglockorder_acquire(DXGLOCK_HANDLETABLE); + if (state == DXGLOCK_EXCL) + down_write(&table->table_lock); + else + down_read(&table->table_lock); +} + +void hmgrtable_unlock(struct hmgrtable *table, enum dxglockstate state) +{ + if (state == DXGLOCK_EXCL) + up_write(&table->table_lock); + else + up_read(&table->table_lock); + dxglockorder_release(DXGLOCK_HANDLETABLE); +} + +d3dkmt_handle hmgrtable_alloc_handle(struct hmgrtable *table, void *object, + enum hmgrentry_type type, bool make_valid) +{ + uint index; + struct hmgrentry *entry; + uint unique; + + DXGKRNL_ASSERT(type <= HMGRENTRY_TYPE_LIMIT); + DXGKRNL_ASSERT(type > HMGRENTRY_TYPE_FREE); + + if (table->free_count <= HMGRTABLE_MIN_FREE_ENTRIES) { + if (!expand_table(table, 0)) { + pr_err("hmgrtable expand_table failed\n"); + return 0; + } + } + + if (table->free_handle_list_head >= table->table_size) { + pr_err("hmgrtable corrupted handle table head\n"); + return 0; + } + + index = table->free_handle_list_head; + entry = &table->entry_table[index]; + + if (entry->type != HMGRENTRY_TYPE_FREE) { + pr_err("hmgrtable expected free handle\n"); + return 0; + } + + table->free_handle_list_head = entry->next_free_index; + + if (entry->next_free_index != table->free_handle_list_tail) { + if (entry->next_free_index >= table->table_size) { + pr_err("hmgrtable invalid next free index\n"); + return 0; + } + table->entry_table[entry->next_free_index].prev_free_index = + HMGRTABLE_INVALID_INDEX; + } + + unique = table->entry_table[index].unique; + + table->entry_table[index].object = object; + table->entry_table[index].type = type; + table->entry_table[index].instance = 0; + table->entry_table[index].destroyed = !make_valid; + table->free_count--; + + return build_handle(index, unique, table->entry_table[index].instance); +} + +int hmgrtable_assign_handle_safe(struct hmgrtable *table, void *object, + enum hmgrentry_type type, d3dkmt_handle h) +{ + int ret; + + hmgrtable_lock(table, DXGLOCK_EXCL); + ret = hmgrtable_assign_handle(table, object, type, h); + hmgrtable_unlock(table, DXGLOCK_EXCL); + return ret; +} + +int hmgrtable_assign_handle(struct hmgrtable *table, void *object, + enum hmgrentry_type type, d3dkmt_handle h) +{ + uint index = get_index(h); + uint unique = get_unique(h); + struct hmgrentry *entry = NULL; + + TRACE_DEBUG(1, "%s %x, %d %p, %p\n", + __func__, h, index, object, table); + + if (index >= HMGRHANDLE_INDEX_MAX) { + pr_err("handle index is too big: %x %d", h, index); + return STATUS_INVALID_PARAMETER; + } + + if (index >= table->table_size) { + uint new_size = index + HMGRTABLE_SIZE_INCREMENT; + + if (new_size > HMGRHANDLE_INDEX_MAX) + new_size = HMGRHANDLE_INDEX_MAX; + if (!expand_table(table, new_size)) { + pr_err("failed to expand table\n"); + return STATUS_NO_MEMORY; + } + } + + entry = &table->entry_table[index]; + + if (entry->type != HMGRENTRY_TYPE_FREE) { + pr_err("the entry is already busy: %d %x", + entry->type, + hmgrtable_build_entry_handle(table, index)); + return STATUS_INVALID_PARAMETER; + } + + if (index != table->free_handle_list_tail) { + if (entry->next_free_index >= table->table_size) { + pr_err("hmgr: invalid next free index %d\n", + entry->next_free_index); + return STATUS_INVALID_PARAMETER; + } + table->entry_table[entry->next_free_index].prev_free_index = + entry->prev_free_index; + } else { + table->free_handle_list_tail = entry->prev_free_index; + } + + if (index != table->free_handle_list_head) { + if (entry->prev_free_index >= table->table_size) { + pr_err("hmgr: invalid next prev index %d\n", + entry->prev_free_index); + return STATUS_INVALID_PARAMETER; + } + table->entry_table[entry->prev_free_index].next_free_index = + entry->next_free_index; + } else { + table->free_handle_list_head = entry->next_free_index; + } + + entry->prev_free_index = HMGRTABLE_INVALID_INDEX; + entry->next_free_index = HMGRTABLE_INVALID_INDEX; + entry->object = object; + entry->type = type; + entry->instance = 0; + entry->unique = unique; + entry->destroyed = false; + + table->free_count--; + return 0; +} + +d3dkmt_handle hmgrtable_alloc_handle_safe(struct hmgrtable *table, void *obj, + enum hmgrentry_type type, + bool make_valid) +{ + d3dkmt_handle h; + + hmgrtable_lock(table, DXGLOCK_EXCL); + h = hmgrtable_alloc_handle(table, obj, type, make_valid); + hmgrtable_unlock(table, DXGLOCK_EXCL); + return h; +} + +void hmgrtable_free_handle(struct hmgrtable *table, enum hmgrentry_type t, + d3dkmt_handle h) +{ + struct hmgrentry *entry; + uint i = get_index(h); + + DXGKRNL_ASSERT(table->free_count < table->table_size); + DXGKRNL_ASSERT(table->free_count >= HMGRTABLE_MIN_FREE_ENTRIES); + + TRACE_DEBUG(1, "%s: %p %x\n", __func__, table, h); + + /* Ignore the destroyed flag when checking the handle */ + if (is_handle_valid(table, h, true, t)) { + entry = &table->entry_table[i]; + entry->unique = 1; + entry->type = HMGRENTRY_TYPE_FREE; + entry->destroyed = 0; + if (entry->unique != HMGRHANDLE_UNIQUE_MAX) + entry->unique += 1; + else + entry->unique = 1; + + table->free_count++; + + /* + * Insert the index to the free list at the tail. + */ + entry->next_free_index = HMGRTABLE_INVALID_INDEX; + entry->prev_free_index = table->free_handle_list_tail; + entry = &table->entry_table[table->free_handle_list_tail]; + entry->next_free_index = i; + table->free_handle_list_tail = i; + } else { + pr_err("%s:error: %d %x\n", __func__, i, h); + } +} + +void hmgrtable_free_handle_safe(struct hmgrtable *table, enum hmgrentry_type t, + d3dkmt_handle h) +{ + hmgrtable_lock(table, DXGLOCK_EXCL); + hmgrtable_free_handle(table, t, h); + hmgrtable_unlock(table, DXGLOCK_EXCL); +} + +d3dkmt_handle hmgrtable_build_entry_handle(struct hmgrtable *table, uint index) +{ + DXGKRNL_ASSERT(index < table->table_size); + + return build_handle(index, table->entry_table[index].unique, + table->entry_table[index].instance); +} + +void *hmgrtable_get_object(struct hmgrtable *table, d3dkmt_handle h) +{ + if (!is_handle_valid(table, h, false, HMGRENTRY_TYPE_FREE)) + return NULL; + + return table->entry_table[get_index(h)].object; +} + +void *hmgrtable_get_object_by_type(struct hmgrtable *table, + enum hmgrentry_type type, d3dkmt_handle h) +{ + if (!is_handle_valid(table, h, false, type)) { + pr_err("%s invalid handle %x\n", __func__, h); + return NULL; + } + return table->entry_table[get_index(h)].object; +} + +void *hmgrtable_get_entry_object(struct hmgrtable *table, uint index) +{ + DXGKRNL_ASSERT(index < table->table_size); + DXGKRNL_ASSERT(table->entry_table[index].type != HMGRENTRY_TYPE_FREE); + + return table->entry_table[index].object; +} + +enum hmgrentry_type hmgrtable_get_entry_type(struct hmgrtable *table, + uint index) +{ + DXGKRNL_ASSERT(index < table->table_size); + return (enum hmgrentry_type)table->entry_table[index].type; +} + +enum hmgrentry_type hmgrtable_get_object_type(struct hmgrtable *table, + d3dkmt_handle h) +{ + if (!is_handle_valid(table, h, false, HMGRENTRY_TYPE_FREE)) + return HMGRENTRY_TYPE_FREE; + + return hmgrtable_get_entry_type(table, get_index(h)); +} + +void *hmgrtable_get_object_ignore_destroyed(struct hmgrtable *table, + d3dkmt_handle h, + enum hmgrentry_type type) +{ + if (!is_handle_valid(table, h, true, type)) + return NULL; + return table->entry_table[get_index(h)].object; +} + +bool hmgrtable_next_entry(struct hmgrtable *tbl, uint *index, + enum hmgrentry_type *type, d3dkmt_handle *handle, + void **object) +{ + uint i; + struct hmgrentry *entry; + + for (i = *index; i < tbl->table_size; i++) { + entry = &tbl->entry_table[i]; + if (entry->type != HMGRENTRY_TYPE_FREE) { + *index = i + 1; + *object = entry->object; + *handle = build_handle(i, entry->unique, + entry->instance); + *type = entry->type; + return true; + } + } + return false; +} diff --git a/drivers/gpu/dxgkrnl/hmgr.h b/drivers/gpu/dxgkrnl/hmgr.h new file mode 100644 index 000000000000..2b508c2f7fcd --- /dev/null +++ b/drivers/gpu/dxgkrnl/hmgr.h @@ -0,0 +1,107 @@ +// SPDX-License-Identifier: GPL-2.0 + +/* + * Copyright (c) 2019, Microsoft Corporation. + * + * Author: + * Iouri Tarassov iourit@microsoft.com + * + * Dxgkrnl Graphics Port Driver + * Handle manager definitions + * + */ + +#ifndef _HMGR_H_ +#define _HMGR_H_ + +#include "misc.h" + +struct hmgrentry; + +/* + * Handle manager table. + * + * Implementation notes: + * A list of free handles is built on top of the array of table entries. + * free_handle_list_head is the index of the first entry in the list. + * m_FreeHandleListTail is the index of an entry in the list, which is + * HMGRTABLE_MIN_FREE_ENTRIES from the head. It means that when a handle is + * freed, the next time the handle can be re-used is after allocating + * HMGRTABLE_MIN_FREE_ENTRIES number of handles. + * Handles are allocated from the start of the list and free handles are + * inserted after the tail of the list. + * + */ +struct hmgrtable { + struct dxgprocess *process; + struct hmgrentry *entry_table; + uint free_handle_list_head; + uint free_handle_list_tail; + uint table_size; + uint free_count; + struct rw_semaphore table_lock; +}; + +/* + * Handle entry data types. + */ +#define HMGRENTRY_TYPE_BITS 5 + +enum hmgrentry_type { + HMGRENTRY_TYPE_FREE = 0, + HMGRENTRY_TYPE_DXGADAPTER = 1, + HMGRENTRY_TYPE_DXGSHAREDRESOURCE = 2, + HMGRENTRY_TYPE_DXGDEVICE = 3, + HMGRENTRY_TYPE_DXGRESOURCE = 4, + HMGRENTRY_TYPE_DXGALLOCATION = 5, + HMGRENTRY_TYPE_DXGOVERLAY = 6, + HMGRENTRY_TYPE_DXGCONTEXT = 7, + HMGRENTRY_TYPE_DXGSYNCOBJECT = 8, + HMGRENTRY_TYPE_DXGKEYEDMUTEX = 9, + HMGRENTRY_TYPE_DXGPAGINGQUEUE = 10, + HMGRENTRY_TYPE_DXGDEVICESYNCOBJECT = 11, + HMGRENTRY_TYPE_DXGPROCESS = 12, + HMGRENTRY_TYPE_DXGSHAREDVMOBJECT = 13, + HMGRENTRY_TYPE_DXGPROTECTEDSESSION = 14, + HMGRENTRY_TYPE_DXGHWQUEUE = 15, + HMGRENTRY_TYPE_DXGREMOTEBUNDLEOBJECT = 16, + HMGRENTRY_TYPE_DXGCOMPOSITIONSURFACEOBJECT = 17, + HMGRENTRY_TYPE_DXGCOMPOSITIONSURFACEPROXY = 18, + HMGRENTRY_TYPE_DXGTRACKEDWORKLOAD = 19, + HMGRENTRY_TYPE_LIMIT = ((1 << HMGRENTRY_TYPE_BITS) - 1), + HMGRENTRY_TYPE_MONITOREDFENCE = HMGRENTRY_TYPE_LIMIT + 1, +}; + +void hmgrtable_init(struct hmgrtable *tbl, struct dxgprocess *process); +void hmgrtable_destroy(struct hmgrtable *tbl); +void hmgrtable_lock(struct hmgrtable *tbl, enum dxglockstate state); +void hmgrtable_unlock(struct hmgrtable *tbl, enum dxglockstate state); +d3dkmt_handle hmgrtable_alloc_handle(struct hmgrtable *tbl, void *object, + enum hmgrentry_type t, bool make_valid); +d3dkmt_handle hmgrtable_alloc_handle_safe(struct hmgrtable *tbl, void *obj, + enum hmgrentry_type t, bool reserve); +int hmgrtable_assign_handle(struct hmgrtable *tbl, void *obj, + enum hmgrentry_type, d3dkmt_handle h); +int hmgrtable_assign_handle_safe(struct hmgrtable *tbl, void *obj, + enum hmgrentry_type t, d3dkmt_handle h); +void hmgrtable_free_handle(struct hmgrtable *tbl, enum hmgrentry_type t, + d3dkmt_handle h); +void hmgrtable_free_handle_safe(struct hmgrtable *tbl, enum hmgrentry_type t, + d3dkmt_handle h); +d3dkmt_handle hmgrtable_build_entry_handle(struct hmgrtable *tbl, uint index); +enum hmgrentry_type hmgrtable_get_object_type(struct hmgrtable *tbl, + d3dkmt_handle h); +void *hmgrtable_get_object(struct hmgrtable *tbl, d3dkmt_handle h); +void *hmgrtable_get_object_by_type(struct hmgrtable *tbl, enum hmgrentry_type t, + d3dkmt_handle h); +void *hmgrtable_get_object_ignore_destroyed(struct hmgrtable *tbl, + d3dkmt_handle h, + enum hmgrentry_type t); +bool hmgrtable_mark_destroyed(struct hmgrtable *tbl, d3dkmt_handle h); +bool hmgrtable_unmark_destroyed(struct hmgrtable *tbl, d3dkmt_handle h); +void *hmgrtable_get_entry_object(struct hmgrtable *tbl, uint index); +bool hmgrtable_next_entry(struct hmgrtable *tbl, uint *start_index, + enum hmgrentry_type *type, d3dkmt_handle *handle, + void **object); + +#endif diff --git a/drivers/gpu/dxgkrnl/ioctl.c b/drivers/gpu/dxgkrnl/ioctl.c new file mode 100644 index 000000000000..7d585b2cece8 --- /dev/null +++ b/drivers/gpu/dxgkrnl/ioctl.c @@ -0,0 +1,5269 @@ +// SPDX-License-Identifier: GPL-2.0 + +/* + * Copyright (c) 2019, Microsoft Corporation. + * + * Author: + * Iouri Tarassov iourit@microsoft.com + * + * Dxgkrnl Graphics Driver + * Ioctl implementation + * + */ + +#include <linux/eventfd.h> +#include <linux/file.h> +#include <linux/fs.h> +#include <linux/anon_inodes.h> +#include <linux/mman.h> + +#include "dxgkrnl.h" +#include "dxgvmbus.h" + +struct ioctl_desc { + ntstatus(*ioctl_callback) (struct dxgprocess *p, void *__user arg); + u32 ioctl; + u32 arg_size; +}; +static struct ioctl_desc ioctls[LX_IO_MAX + 1]; + +static int dxgsyncobj_release(struct inode *inode, struct file *file) +{ + struct dxgsharedsyncobject *syncobj = file->private_data; + struct dxgthreadinfo *thread = dxglockorder_get_thread(); + + TRACE_DEBUG(1, "%s: %p", __func__, syncobj); + dxgmutex_lock(&syncobj->fd_mutex); + dxgsharedsyncobj_acquire_reference(syncobj); + syncobj->host_shared_handle_nt_reference--; + if (syncobj->host_shared_handle_nt_reference == 0) { + if (syncobj->host_shared_handle_nt) { + dxgvmb_send_destroy_nt_shared_object( + syncobj->host_shared_handle_nt); + TRACE_DEBUG(1, "Syncobj host_handle_nt destroyed: %x", + syncobj->host_shared_handle_nt); + syncobj->host_shared_handle_nt = 0; + } + dxgsharedsyncobj_release_reference(syncobj); + } + dxgmutex_unlock(&syncobj->fd_mutex); + dxgsharedsyncobj_release_reference(syncobj); + dxglockorder_put_thread(thread); + return 0; +} + +static const struct file_operations dxg_syncobj_fops = { + .release = dxgsyncobj_release, +}; + +static int dxgsharedresource_release(struct inode *inode, struct file *file) +{ + struct dxgsharedresource *resource = file->private_data; + struct dxgthreadinfo *thread = dxglockorder_get_thread(); + + TRACE_DEBUG(1, "%s: %p", __func__, resource); + dxgmutex_lock(&resource->fd_mutex); + dxgsharedresource_acquire_reference(resource); + resource->host_shared_handle_nt_reference--; + if (resource->host_shared_handle_nt_reference == 0) { + if (resource->host_shared_handle_nt) { + dxgvmb_send_destroy_nt_shared_object( + resource->host_shared_handle_nt); + TRACE_DEBUG(1, "Resource host_handle_nt destroyed: %x", + resource->host_shared_handle_nt); + resource->host_shared_handle_nt = 0; + } + dxgsharedresource_release_reference(resource); + } + dxgmutex_unlock(&resource->fd_mutex); + dxgsharedresource_release_reference(resource); + dxglockorder_put_thread(thread); + return 0; +} + +static const struct file_operations dxg_resource_fops = { + .release = dxgsharedresource_release, +}; + +static int dxgk_open_adapter_from_luid(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_openadapterfromluid args; + int ret = 0; + struct dxgadapter *entry; + struct dxgadapter *adapter = NULL; + struct d3dkmt_openadapterfromluid *__user result = inargs; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + dxgglobal_acquire_adapter_list_lock(DXGLOCK_SHARED); + dxgglobal_acquire_process_adapter_lock(); + + list_for_each_entry(entry, &dxgglobal->adapter_list_head, + adapter_list_entry) { + if (dxgadapter_acquire_lock_shared(entry) == 0) { + TRACE_DEBUG(1, "Compare luids: %d:%d %d:%d", + entry->luid.b, entry->luid.a, + args.adapter_luid.b, args.adapter_luid.a); + if (*(u64 *) &entry->luid == + *(u64 *) &args.adapter_luid) { + ret = + dxgprocess_open_adapter(process, entry, + &args.adapter_handle); + + if (NT_SUCCESS(ret)) { + ret = dxg_copy_to_user( + &result->adapter_handle, + &args.adapter_handle, + sizeof(d3dkmt_handle)); + } + adapter = entry; + } + dxgadapter_release_lock_shared(entry); + if (adapter) + break; + } + } + + dxgglobal_release_process_adapter_lock(); + dxgglobal_release_adapter_list_lock(DXGLOCK_SHARED); + + if (args.adapter_handle == 0) + ret = STATUS_INVALID_PARAMETER; + +cleanup: + + if (ret) + dxgprocess_close_adapter(process, args.adapter_handle); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgkp_enum_adapters(struct dxgprocess *process, + union d3dkmt_enumadapters_filter filter, + uint adapter_count_max, + struct d3dkmt_adapterinfo *__user info_out, + uint * __user adapter_count_out) +{ + int ret = 0; + struct dxgadapter *entry; + struct d3dkmt_adapterinfo *info = NULL; + struct dxgadapter **adapters = NULL; + int adapter_count = 0; + int i; + + TRACE_FUNC_ENTER(__func__); + if (info_out == NULL || adapter_count_max == 0) { + ret = 0; + TRACE_DEBUG(1, "buffer is NULL"); + ret = dxg_copy_to_user(adapter_count_out, + &dxgglobal->num_adapters, sizeof(uint)); + goto cleanup; + } + + if (adapter_count_max > 0xFFFF) { + pr_err("too many adapters"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + info = dxgmem_alloc(process, DXGMEM_TMP, + sizeof(struct d3dkmt_adapterinfo) * + adapter_count_max); + if (info == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + adapters = dxgmem_alloc(process, DXGMEM_TMP, + sizeof(struct dxgadapter *) * + adapter_count_max); + if (adapters == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + dxgglobal_acquire_adapter_list_lock(DXGLOCK_SHARED); + dxgglobal_acquire_process_adapter_lock(); + + list_for_each_entry(entry, &dxgglobal->adapter_list_head, + adapter_list_entry) { + if (dxgadapter_acquire_lock_shared(entry) == 0) { + struct d3dkmt_adapterinfo *inf = &info[adapter_count]; + + ret = dxgprocess_open_adapter(process, entry, + &inf->adapter_handle); + if (NT_SUCCESS(ret)) { + inf->adapter_luid = entry->luid; + adapters[adapter_count] = entry; + TRACE_DEBUG(1, "adapter: %x %x:%x", + inf->adapter_handle, + inf->adapter_luid.b, + inf->adapter_luid.a); + adapter_count++; + } + dxgadapter_release_lock_shared(entry); + } + if (ret) + break; + } + + dxgglobal_release_process_adapter_lock(); + dxgglobal_release_adapter_list_lock(DXGLOCK_SHARED); + + if (adapter_count > adapter_count_max) { + ret = STATUS_BUFFER_TOO_SMALL; + TRACE_DEBUG(1, "Too many adapters"); + ret = dxg_copy_to_user(adapter_count_out, + &dxgglobal->num_adapters, sizeof(uint)); + goto cleanup; + } + + ret = dxg_copy_to_user(adapter_count_out, &adapter_count, + sizeof(adapter_count)); + if (ret) + goto cleanup; + ret = dxg_copy_to_user(info_out, info, sizeof(info[0]) * adapter_count); + +cleanup: + + if (NT_SUCCESS(ret)) { + TRACE_DEBUG(1, "found %d adapters", adapter_count); + goto success; + } + if (info) { + for (i = 0; i < adapter_count; i++) + dxgprocess_close_adapter(process, + info[i].adapter_handle); + } +success: + if (info) + dxgmem_free(process, DXGMEM_TMP, info); + if (adapters) + dxgmem_free(process, DXGMEM_TMP, adapters); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgsharedresource_seal(struct dxgsharedresource *shared_resource) +{ + int ret = 0; + int i = 0; + uint8_t *private_data; + uint data_size; + struct dxgresource *resource; + struct dxgallocation *alloc; + + TRACE_DEBUG(1, "Sealing resource: %p", shared_resource); + + down_write(&shared_resource->adapter->shared_resource_list_lock); + if (shared_resource->sealed) { + TRACE_DEBUG(1, "Resource already sealed"); + goto cleanup; + } + shared_resource->sealed = 1; + if (!list_empty(&shared_resource->resource_list_head)) { + resource = + list_first_entry(&shared_resource->resource_list_head, + struct dxgresource, + shared_resource_list_entry); + TRACE_DEBUG(1, "First resource: %p", resource); + dxgmutex_lock(&resource->resource_mutex); + list_for_each_entry(alloc, &resource->alloc_list_head, + alloc_list_entry) { + TRACE_DEBUG(1, "Resource alloc: %p %d", alloc, + alloc->priv_drv_data->data_size); + shared_resource->allocation_count++; + shared_resource->alloc_private_data_size += + alloc->priv_drv_data->data_size; + if (shared_resource->alloc_private_data_size < + alloc->priv_drv_data->data_size) { + pr_err("alloc private data overflow"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup1; + } + } + if (shared_resource->alloc_private_data_size == 0) + goto cleanup1; + shared_resource->alloc_private_data = + dxgmem_alloc(NULL, DXGMEM_ALLOCPRIVATE, + shared_resource->alloc_private_data_size); + if (shared_resource->alloc_private_data == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup1; + } + shared_resource->alloc_private_data_sizes = + dxgmem_alloc(NULL, DXGMEM_ALLOCPRIVATE, + sizeof(uint)*shared_resource->allocation_count); + if (shared_resource->alloc_private_data_sizes == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup1; + } + private_data = shared_resource->alloc_private_data; + data_size = shared_resource->alloc_private_data_size; + i = 0; + list_for_each_entry(alloc, &resource->alloc_list_head, + alloc_list_entry) { + uint alloc_data_size = alloc->priv_drv_data->data_size; + + if (alloc_data_size) { + if (data_size < alloc_data_size) { + pr_err("Invalid private data size"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup1; + } + shared_resource->alloc_private_data_sizes[i] = + alloc_data_size; + memcpy(private_data, + alloc->priv_drv_data->data, + alloc_data_size); + dxgmem_free(alloc->process, DXGMEM_ALLOCPRIVATE, + alloc->priv_drv_data); + alloc->priv_drv_data = NULL; + private_data += alloc_data_size; + data_size -= alloc_data_size; + } + i++; + } + if (data_size != 0) { + pr_err("Data size mismatch"); + ret = STATUS_INVALID_PARAMETER; + } +cleanup1: + dxgmutex_unlock(&resource->resource_mutex); + } +cleanup: + up_write(&shared_resource->adapter->shared_resource_list_lock); + return ret; +} + +static int dxgk_enum_adapters(struct dxgprocess *process, void *__user inargs) +{ + struct d3dkmt_enumadapters2 args; + int ret = 0; + struct dxgadapter *entry; + struct d3dkmt_adapterinfo *info = NULL; + struct dxgadapter **adapters = NULL; + int adapter_count = 0; + int i; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.adapters == NULL) { + ret = 0; + TRACE_DEBUG(1, "buffer is NULL"); + args.num_adapters = dxgglobal->num_adapters; + ret = dxg_copy_to_user(inargs, &args, sizeof(args)); + goto cleanup; + } + if (args.num_adapters < dxgglobal->num_adapters) { + args.num_adapters = dxgglobal->num_adapters; + TRACE_DEBUG(1, "buffer is too small"); + ret = STATUS_BUFFER_TOO_SMALL; + goto cleanup; + } + + if (args.num_adapters > D3DKMT_ADAPTERS_MAX) { + TRACE_DEBUG(1, "too many adapters"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + info = dxgmem_alloc(process, DXGMEM_TMP, + sizeof(struct d3dkmt_adapterinfo) * + args.num_adapters); + if (info == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + adapters = dxgmem_alloc(process, DXGMEM_TMP, + sizeof(struct dxgadapter *) * + args.num_adapters); + if (adapters == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + dxgglobal_acquire_adapter_list_lock(DXGLOCK_SHARED); + dxgglobal_acquire_process_adapter_lock(); + + list_for_each_entry(entry, &dxgglobal->adapter_list_head, + adapter_list_entry) { + if (dxgadapter_acquire_lock_shared(entry) == 0) { + struct d3dkmt_adapterinfo *inf = &info[adapter_count]; + + ret = dxgprocess_open_adapter(process, entry, + &inf->adapter_handle); + if (NT_SUCCESS(ret)) { + inf->adapter_luid = entry->luid; + adapters[adapter_count] = entry; + TRACE_DEBUG(1, "adapter: %x %llx", + inf->adapter_handle, + *(u64 *) &inf->adapter_luid); + adapter_count++; + } + dxgadapter_release_lock_shared(entry); + } + if (ret) + break; + } + + dxgglobal_release_process_adapter_lock(); + dxgglobal_release_adapter_list_lock(DXGLOCK_SHARED); + + args.num_adapters = adapter_count; + + ret = dxg_copy_to_user(inargs, &args, sizeof(args)); + if (ret) + goto cleanup; + ret = dxg_copy_to_user(args.adapters, info, + sizeof(info[0]) * args.num_adapters); + if (ret) + goto cleanup; + +cleanup: + + if (ret) { + if (info) { + for (i = 0; i < args.num_adapters; i++) { + dxgprocess_close_adapter(process, + info[i]. + adapter_handle); + } + } + } else { + TRACE_DEBUG(1, "found %d adapters", args.num_adapters); + } + + if (info) + dxgmem_free(process, DXGMEM_TMP, info); + if (adapters) + dxgmem_free(process, DXGMEM_TMP, adapters); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_enum_adapters3(struct dxgprocess *process, void *__user inargs) +{ + struct d3dkmt_enumadapters3 args; + int ret = 0; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + ret = dxgkp_enum_adapters(process, args.filter, + args.adapter_count, + args.adapters, + &((struct d3dkmt_enumadapters3 *)inargs)-> + adapter_count); + +cleanup: + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_close_adapter(struct dxgprocess *process, void *__user inargs) +{ + d3dkmt_handle args; + int ret = 0; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + ret = dxgprocess_close_adapter(process, args); + if (ret) + pr_err("%s failed", __func__); + +cleanup: + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_query_adapter_info(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_queryadapterinfo args; + int ret = 0; + struct dxgadapter *adapter = NULL; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.private_data_size > DXG_MAX_VM_BUS_PACKET_SIZE || + args.private_data_size == 0) { + pr_err("invalid private data size"); + goto cleanup; + } + + TRACE_DEBUG(1, "Type: %d Size: %x", args.type, args.private_data_size); + + adapter = dxgprocess_adapter_by_handle(process, args.adapter); + if (adapter == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) + goto cleanup; + + ret = dxgvmb_send_query_adapter_info(process, &adapter->channel, &args); + + dxgadapter_release_lock_shared(adapter); + +cleanup: + + if (adapter) + dxgadapter_release_reference(adapter); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_create_device(struct dxgprocess *process, void *__user inargs) +{ + struct d3dkmt_createdevice args; + int ret = 0; + struct dxgadapter *adapter = NULL; + struct dxgdevice *device = NULL; + d3dkmt_handle host_device_handle = 0; + bool adapter_locked = false; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + /* The call acquires reference on the adapter */ + adapter = dxgprocess_adapter_by_handle(process, args.adapter); + if (adapter == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + device = dxgdevice_create(adapter, process); + if (device == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) + goto cleanup; + + adapter_locked = true; + + host_device_handle = dxgvmb_send_create_device(adapter, process, &args); + if (host_device_handle) { + ret = + dxg_copy_to_user(&((struct d3dkmt_createdevice *)inargs)-> + device, &host_device_handle, + sizeof(d3dkmt_handle)); + if (ret) + goto cleanup; + + hmgrtable_lock(&process->handle_table, DXGLOCK_EXCL); + ret = hmgrtable_assign_handle(&process->handle_table, device, + HMGRENTRY_TYPE_DXGDEVICE, + host_device_handle); + if (!ret) { + device->handle = host_device_handle; + device->handle_valid = 1; + device->object_state = DXGOBJECTSTATE_ACTIVE; + } + hmgrtable_unlock(&process->handle_table, DXGLOCK_EXCL); + } + +cleanup: + + if (ret) { + if (host_device_handle) + dxgvmb_send_destroy_device(adapter, process, + host_device_handle); + if (device) + dxgdevice_destroy(device); + } + + if (adapter_locked) + dxgadapter_release_lock_shared(adapter); + + if (adapter) + dxgadapter_release_reference(adapter); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_destroy_device(struct dxgprocess *process, void *__user inargs) +{ + struct d3dkmt_destroydevice args; + int ret = 0; + struct dxgadapter *adapter = NULL; + struct dxgdevice *device = NULL; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + hmgrtable_lock(&process->handle_table, DXGLOCK_EXCL); + device = hmgrtable_get_object_by_type(&process->handle_table, + HMGRENTRY_TYPE_DXGDEVICE, + args.device); + if (device) { + hmgrtable_free_handle(&process->handle_table, + HMGRENTRY_TYPE_DXGDEVICE, args.device); + device->handle_valid = 0; + } + hmgrtable_unlock(&process->handle_table, DXGLOCK_EXCL); + + if (device == NULL) { + pr_err("invalid device handle: %x", args.device); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + + dxgdevice_destroy(device); + + if (dxgadapter_acquire_lock_shared(adapter) == 0) { + dxgvmb_send_destroy_device(adapter, process, args.device); + dxgadapter_release_lock_shared(adapter); + } + +cleanup: + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_create_context_virtual(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_createcontextvirtual args; + int ret = 0; + struct dxgadapter *adapter = NULL; + struct dxgdevice *device = NULL; + struct dxgcontext *context = NULL; + d3dkmt_handle host_context_handle = 0; + bool device_lock_acquired = false; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + /* + * The call acquires reference on the device. It is safe to access the + * adapter, because the device holds reference on it. + */ + device = dxgprocess_device_by_handle(process, args.device); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + ret = dxgdevice_acquire_lock_shared(device); + if (ret) + goto cleanup; + + device_lock_acquired = true; + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + context = dxgcontext_create(device); + if (context == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + host_context_handle = dxgvmb_send_create_context(adapter, + process, &args); + if (host_context_handle) { + hmgrtable_lock(&process->handle_table, DXGLOCK_EXCL); + ret = hmgrtable_assign_handle(&process->handle_table, context, + HMGRENTRY_TYPE_DXGCONTEXT, + host_context_handle); + if (!ret) + context->handle = host_context_handle; + hmgrtable_unlock(&process->handle_table, DXGLOCK_EXCL); + if (ret) + goto cleanup; + ret = + dxg_copy_to_user(& + ((struct d3dkmt_createcontextvirtual *) + inargs)->context, &host_context_handle, + sizeof(d3dkmt_handle)); + } else { + pr_err("invalid host handle"); + ret = STATUS_INVALID_PARAMETER; + } + +cleanup: + + if (ret) { + if (host_context_handle) { + dxgvmb_send_destroy_context(adapter, process, + host_context_handle); + } + if (context) + dxgcontext_destroy_safe(process, context); + } + + if (adapter) + dxgadapter_release_lock_shared(adapter); + + if (device) { + if (device_lock_acquired) + dxgdevice_release_lock_shared(device); + dxgdevice_release_reference(device); + } + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_destroy_context(struct dxgprocess *process, void *__user inargs) +{ + struct d3dkmt_destroycontext args; + int ret = 0; + struct dxgadapter *adapter = NULL; + struct dxgcontext *context = NULL; + struct dxgdevice *device = NULL; + d3dkmt_handle device_handle = 0; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + hmgrtable_lock(&process->handle_table, DXGLOCK_EXCL); + context = hmgrtable_get_object_by_type(&process->handle_table, + HMGRENTRY_TYPE_DXGCONTEXT, + args.context); + if (context) { + hmgrtable_free_handle(&process->handle_table, + HMGRENTRY_TYPE_DXGCONTEXT, args.context); + context->handle = 0; + device_handle = context->device_handle; + context->object_state = DXGOBJECTSTATE_DESTROYED; + } + hmgrtable_unlock(&process->handle_table, DXGLOCK_EXCL); + + if (context == NULL) { + pr_err("invalid context handle: %x", args.context); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + /* + * The call acquires reference on the device. It is safe to access the + * adapter, because the device holds reference on it. + */ + device = dxgprocess_device_by_handle(process, device_handle); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_destroy_context(adapter, process, args.context); + + dxgcontext_destroy_safe(process, context); + +cleanup: + + if (adapter) + dxgadapter_release_lock_shared(adapter); + + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_create_hwcontext(struct dxgprocess *process, + void *__user inargs) +{ + /* This is obsolete entry point */ + return STATUS_NOT_SUPPORTED; +} + +static int dxgk_destroy_hwcontext(struct dxgprocess *process, + void *__user inargs) +{ + /* This is obsolete entry point */ + return STATUS_NOT_SUPPORTED; +} + +static int dxgk_create_hwqueue(struct dxgprocess *process, void *__user inargs) +{ + struct d3dkmt_createhwqueue args; + struct dxgdevice *device = NULL; + struct dxgcontext *context = NULL; + struct dxgadapter *adapter = NULL; + struct dxghwqueue *hwqueue = NULL; + int ret = 0; + bool device_lock_acquired = false; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + /* + * The call acquires reference on the device. It is safe to access the + * adapter, because the device holds reference on it. + */ + device = dxgprocess_device_by_object_handle(process, + HMGRENTRY_TYPE_DXGCONTEXT, + args.context); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + ret = dxgdevice_acquire_lock_shared(device); + if (ret) + goto cleanup; + + device_lock_acquired = true; + + hmgrtable_lock(&process->handle_table, DXGLOCK_SHARED); + context = hmgrtable_get_object_by_type(&process->handle_table, + HMGRENTRY_TYPE_DXGCONTEXT, + args.context); + hmgrtable_unlock(&process->handle_table, DXGLOCK_SHARED); + + if (context == NULL) { + pr_err("Invalid context handle %x", args.context); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + hwqueue = dxghwqueue_create(context); + if (hwqueue == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_create_hwqueue(process, &adapter->channel, &args, + inargs, hwqueue); + +cleanup: + + if (ret && hwqueue) + dxghwqueue_destroy(process, hwqueue); + + if (adapter) + dxgadapter_release_lock_shared(adapter); + + if (device_lock_acquired) + dxgdevice_release_lock_shared(device); + + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_destroy_hwqueue(struct dxgprocess *process, void *__user inargs) +{ + struct d3dkmt_destroyhwqueue args; + int ret = 0; + struct dxgadapter *adapter = NULL; + struct dxgdevice *device = NULL; + struct dxghwqueue *hwqueue = NULL; + d3dkmt_handle device_handle = 0; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + hmgrtable_lock(&process->handle_table, DXGLOCK_EXCL); + hwqueue = hmgrtable_get_object_by_type(&process->handle_table, + HMGRENTRY_TYPE_DXGHWQUEUE, + args.queue); + if (hwqueue) { + hmgrtable_free_handle(&process->handle_table, + HMGRENTRY_TYPE_DXGHWQUEUE, args.queue); + hwqueue->handle = 0; + device_handle = hwqueue->device_handle; + } + hmgrtable_unlock(&process->handle_table, DXGLOCK_EXCL); + + if (hwqueue == NULL) { + pr_err("invalid hwqueue handle: %x", args.queue); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + /* + * The call acquires reference on the device. It is safe to access the + * adapter, because the device holds reference on it. + */ + device = dxgprocess_device_by_handle(process, device_handle); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_destroy_hwqueue(process, &adapter->channel, + args.queue); + + dxghwqueue_destroy(process, hwqueue); + +cleanup: + + if (adapter) + dxgadapter_release_lock_shared(adapter); + + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_create_paging_queue(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_createpagingqueue args; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + struct dxgpagingqueue *pqueue = NULL; + int ret = 0; + d3dkmt_handle host_handle = 0; + bool device_lock_acquired = false; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + /* + * The call acquires reference on the device. It is safe to access the + * adapter, because the device holds reference on it. + */ + device = dxgprocess_device_by_handle(process, args.device); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + ret = dxgdevice_acquire_lock_shared(device); + if (ret) + goto cleanup; + + device_lock_acquired = true; + adapter = device->adapter; + + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + pqueue = dxgpagingqueue_create(device); + if (pqueue == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + ret = dxgvmb_send_create_paging_queue(process, &adapter->channel, + device, &args, pqueue); + if (NT_SUCCESS(ret)) { + host_handle = args.paging_queue; + + ret = dxg_copy_to_user(inargs, &args, sizeof(args)); + if (ret) + goto cleanup; + + hmgrtable_lock(&process->handle_table, DXGLOCK_EXCL); + ret = hmgrtable_assign_handle(&process->handle_table, pqueue, + HMGRENTRY_TYPE_DXGPAGINGQUEUE, + host_handle); + if (NT_SUCCESS(ret)) { + pqueue->handle = host_handle; + ret = hmgrtable_assign_handle(&process->handle_table, + NULL, + HMGRENTRY_TYPE_MONITOREDFENCE, + args.sync_object); + if (NT_SUCCESS(ret)) + pqueue->syncobj_handle = args.sync_object; + } + hmgrtable_unlock(&process->handle_table, DXGLOCK_EXCL); + /* should not fail after this */ + } + +cleanup: + + if (ret) { + if (pqueue) + dxgpagingqueue_destroy(pqueue); + if (host_handle) + dxgvmb_send_destroy_paging_queue(process, + &adapter->channel, + host_handle); + } + + if (adapter) + dxgadapter_release_lock_shared(adapter); + + if (device) { + if (device_lock_acquired) + dxgdevice_release_lock_shared(device); + dxgdevice_release_reference(device); + } + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_destroy_paging_queue(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dddi_destroypagingqueue args; + struct dxgpagingqueue *paging_queue = NULL; + int ret = 0; + d3dkmt_handle device_handle = 0; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + hmgrtable_lock(&process->handle_table, DXGLOCK_EXCL); + paging_queue = hmgrtable_get_object_by_type(&process->handle_table, + HMGRENTRY_TYPE_DXGPAGINGQUEUE, + args.paging_queue); + if (paging_queue) { + device_handle = paging_queue->device_handle; + hmgrtable_free_handle(&process->handle_table, + HMGRENTRY_TYPE_DXGPAGINGQUEUE, + args.paging_queue); + hmgrtable_free_handle(&process->handle_table, + HMGRENTRY_TYPE_MONITOREDFENCE, + paging_queue->syncobj_handle); + paging_queue->syncobj_handle = 0; + paging_queue->handle = 0; + } + hmgrtable_unlock(&process->handle_table, DXGLOCK_EXCL); + + /* + * The call acquires reference on the device. It is safe to access the + * adapter, because the device holds reference on it. + */ + if (device_handle) + device = dxgprocess_device_by_handle(process, device_handle); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + ret = dxgdevice_acquire_lock_shared(device); + if (ret) { + dxgdevice_release_reference(device); + device = NULL; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_destroy_paging_queue(process, &adapter->channel, + args.paging_queue); + + dxgpagingqueue_destroy(paging_queue); + +cleanup: + + if (adapter) + dxgadapter_release_lock_shared(adapter); + + if (device) { + dxgdevice_release_lock_shared(device); + dxgdevice_release_reference(device); + } + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int get_standard_alloc_priv_data(struct dxgdevice *device, + struct d3dkmt_createstandardallocation + *alloc_info, + uint *standard_alloc_priv_data_size, + void **standard_alloc_priv_data) +{ + int ret = 0; + struct d3dkmdt_gdisurfacedata gdi_data = { }; + uint priv_data_size = 0; + void *priv_data = NULL; + + TRACE_DEBUG(1, "%s", __func__); + + gdi_data.type = D3DKMDT_GDISURFACE_TEXTURE_CROSSADAPTER; + gdi_data.width = alloc_info->existing_heap_data.size; + gdi_data.height = 1; + gdi_data.format = D3DDDIFMT_UNKNOWN; + + *standard_alloc_priv_data_size = 0; + ret = dxgvmb_send_get_standard_alloc_priv_data(device, + D3DKMDT_STANDARDALLOCATION_GDISURFACE, + &gdi_data, 0, + &priv_data_size, NULL); + if (ret) + goto cleanup; + TRACE_DEBUG(1, "Priv data size: %d", priv_data_size); + if (priv_data_size == 0) + goto cleanup; + priv_data = dxgmem_alloc(device->process, DXGMEM_TMP, priv_data_size); + if (priv_data == NULL) { + ret = STATUS_NO_MEMORY; + pr_err("failed to allocate memory for priv data: %d", + priv_data_size); + goto cleanup; + } + ret = dxgvmb_send_get_standard_alloc_priv_data(device, + D3DKMDT_STANDARDALLOCATION_GDISURFACE, + &gdi_data, 0, + &priv_data_size, + priv_data); + if (ret) + goto cleanup; + *standard_alloc_priv_data_size = priv_data_size; + *standard_alloc_priv_data = priv_data; + priv_data = NULL; + +cleanup: + if (priv_data) + dxgmem_free(device->process, DXGMEM_TMP, priv_data); + TRACE_FUNC_EXIT_ERR(__func__, ret); + return ret; +} + +static int dxgk_create_allocation(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_createallocation args; + int ret = 0; + struct dxgadapter *adapter = NULL; + struct dxgdevice *device = NULL; + struct d3dddi_allocationinfo2 *alloc_info = NULL; + struct d3dkmt_createstandardallocation standard_alloc; + uint alloc_info_size = 0; + struct dxgresource *resource = NULL; + struct dxgallocation **dxgalloc = NULL; + struct dxgsharedresource *shared_resource = NULL; + bool resource_mutex_acquired = false; + uint standard_alloc_priv_data_size = 0; + void *standard_alloc_priv_data = NULL; + int i; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.alloc_count > D3DKMT_CREATEALLOCATION_MAX || + args.alloc_count == 0) { + pr_err("invalid number of allocations to create"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + alloc_info_size = sizeof(struct d3dddi_allocationinfo2) * + args.alloc_count; + alloc_info = dxgmem_alloc(process, DXGMEM_TMP, alloc_info_size); + if (alloc_info == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + ret = dxg_copy_from_user(alloc_info, args.allocation_info, + alloc_info_size); + if (ret) + goto cleanup; + + for (i = 0; i < args.alloc_count; i++) { + if (args.flags.standard_allocation) { + if (alloc_info[i].priv_drv_data_size != 0) { + pr_err("private data size is not zero"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + } + if (alloc_info[i].priv_drv_data_size >= + DXG_MAX_VM_BUS_PACKET_SIZE) { + pr_err("private data size is too big: %d %d %ld", + i, alloc_info[i].priv_drv_data_size, + sizeof(alloc_info[0])); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + } + + if (args.flags.existing_section || args.flags.create_protected) { + pr_err("invalid allocation flags"); + goto cleanup; + } + + if (args.flags.standard_allocation) { + if (args.standard_allocation == NULL) { + pr_err("invalid standard allocation"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + ret = dxg_copy_from_user(&standard_alloc, + args.standard_allocation, + sizeof(standard_alloc)); + if (ret) + goto cleanup; + if (alloc_info[0].sysmem == NULL || + args.priv_drv_data_size != 0 || + args.alloc_count != 1 || + standard_alloc.type != + D3DKMT_STANDARDALLOCATIONTYPE_EXISTINGHEAP || + standard_alloc.existing_heap_data.size == 0 || + standard_alloc.existing_heap_data.size & (PAGE_SIZE - 1) || + (unsigned long)alloc_info[0].sysmem & (PAGE_SIZE - 1)) { + pr_err("invalid standard allocation"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + args.priv_drv_data_size = + sizeof(struct d3dkmt_createstandardallocation); + } + + if (args.flags.create_shared && !args.flags.create_resource) { + pr_err("create_resource must be set for create_shared"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + /* + * The call acquires reference on the device. It is safe to access the + * adapter, because the device holds reference on it. + */ + device = dxgprocess_device_by_handle(process, args.device); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + ret = dxgdevice_acquire_lock_shared(device); + if (ret) { + dxgdevice_release_reference(device); + device = NULL; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + if (args.flags.standard_allocation) { + ret = get_standard_alloc_priv_data(device, + &standard_alloc, + &standard_alloc_priv_data_size, + &standard_alloc_priv_data); + if (ret) + goto cleanup; + TRACE_DEBUG(1, "Alloc private data: %d", + standard_alloc_priv_data_size); + } + + if (args.flags.create_resource) { + resource = dxgresource_create(device); + if (resource == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + resource->private_runtime_handle = + args.private_runtime_resource_handle; + if (args.flags.create_shared) { + shared_resource = dxgsharedresource_create(adapter); + if (shared_resource == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + shared_resource->runtime_private_data_size = + args.priv_drv_data_size; + shared_resource->resource_private_data_size = + args.priv_drv_data_size; + if (args.flags.nt_security_sharing) + shared_resource->nt_security = 1; + + shared_resource->runtime_private_data_size = + args.private_runtime_data_size; + shared_resource->resource_private_data_size = + args.priv_drv_data_size; + dxgsharedresource_add_resource(shared_resource, + resource); + if (args.private_runtime_data_size) { + shared_resource->runtime_private_data = + dxgmem_alloc(NULL, + DXGMEM_RUNTIMEPRIVATE, + args. + private_runtime_data_size); + if (shared_resource->runtime_private_data == + NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + ret = + dxg_copy_from_user(shared_resource-> + runtime_private_data, + args. + private_runtime_data, + args. + private_runtime_data_size); + if (ret) + goto cleanup; + } + if (args.priv_drv_data_size) { + shared_resource->resource_private_data = + dxgmem_alloc(NULL, + DXGMEM_RESOURCEPRIVATE, + args.priv_drv_data_size); + if (shared_resource->resource_private_data == + NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + ret = + dxg_copy_from_user(shared_resource-> + resource_private_data, + args.priv_drv_data, + args.priv_drv_data_size); + if (ret) + goto cleanup; + } + } + } else { + if (args.resource) { + /* Adding new allocations to the given resource */ + + dxgprocess_ht_lock_shared_down(process); + resource = + hmgrtable_get_object_by_type(&process->handle_table, + HMGRENTRY_TYPE_DXGRESOURCE, + args.resource); + dxgresource_acquire_reference(resource); + dxgprocess_ht_lock_shared_up(process); + + if (resource == NULL || resource->device != device) { + pr_err("invalid resource handle %x", + args.resource); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + if (resource->shared_owner && + resource->shared_owner->sealed) { + pr_err("Resource is sealed"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + /* Synchronize with resource destruction */ + dxgmutex_lock(&resource->resource_mutex); + if (!dxgresource_is_active(resource)) { + dxgmutex_unlock(&resource->resource_mutex); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + resource_mutex_acquired = true; + } + } + + dxgalloc = dxgmem_alloc(process, DXGMEM_TMP, + sizeof(struct dxgallocation *) * + args.alloc_count); + if (dxgalloc == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + for (i = 0; i < args.alloc_count; i++) { + struct dxgallocation *alloc; + uint priv_data_size = alloc_info[i].priv_drv_data_size; + + if (alloc_info[i].sysmem && !args.flags.standard_allocation) { + if ((unsigned long) + alloc_info[i].sysmem & (PAGE_SIZE - 1)) { + pr_err("invalid sysmem alloc %d, %p", + i, alloc_info[i].sysmem); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + } + if ((alloc_info[0].sysmem == NULL) != + (alloc_info[i].sysmem == NULL)) { + pr_err("All allocations must have sysmem pointer"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + dxgalloc[i] = dxgallocation_create(process); + if (dxgalloc[i] == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + alloc = dxgalloc[i]; + + if (resource) + dxgresource_add_alloc(resource, alloc); + else + dxgdevice_add_alloc(device, alloc); + if (args.flags.create_shared) { + /* Remember alloc private data to use it during open */ + alloc->priv_drv_data = dxgmem_alloc(process, + DXGMEM_ALLOCPRIVATE, + priv_data_size + + offsetof(struct + privdata, + data) - 1); + if (alloc->priv_drv_data == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + if (args.flags.standard_allocation) { + memcpy(alloc->priv_drv_data->data, + standard_alloc_priv_data, + standard_alloc_priv_data_size); + alloc->priv_drv_data->data_size = + standard_alloc_priv_data_size; + } else { + ret = + dxg_copy_from_user(alloc->priv_drv_data-> + data, + alloc_info[i]. + priv_drv_data, + priv_data_size); + if (ret) + goto cleanup; + alloc->priv_drv_data->data_size = + priv_data_size; + } + } + } + + ret = dxgvmb_send_create_allocation(process, device, &args, inargs, + resource, dxgalloc, alloc_info, + &standard_alloc); + if (ret) + goto cleanup; + +cleanup: + + if (resource_mutex_acquired) { + dxgmutex_unlock(&resource->resource_mutex); + dxgresource_release_reference(resource); + } + if (ret) { + if (dxgalloc) { + for (i = 0; i < args.alloc_count; i++) { + if (dxgalloc[i]) + dxgallocation_destroy(dxgalloc[i]); + } + } + if (resource && args.flags.create_resource) { + if (shared_resource) { + dxgsharedresource_remove_resource + (shared_resource, resource); + } + dxgresource_destroy(resource); + } + } + if (shared_resource) + dxgsharedresource_release_reference(shared_resource); + if (dxgalloc) + dxgmem_free(process, DXGMEM_TMP, dxgalloc); + if (standard_alloc_priv_data) + dxgmem_free(process, DXGMEM_TMP, standard_alloc_priv_data); + if (alloc_info) + dxgmem_free(process, DXGMEM_TMP, alloc_info); + + if (adapter) + dxgadapter_release_lock_shared(adapter); + + if (device) { + dxgdevice_release_lock_shared(device); + dxgdevice_release_reference(device); + } + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +int validate_alloc(struct dxgallocation *alloc0, + struct dxgallocation *alloc, + struct dxgdevice *device, d3dkmt_handle alloc_handle) +{ + uint fail_reason; + + if (alloc == NULL) { + fail_reason = 1; + goto cleanup; + } + if (alloc->resource_owner != alloc0->resource_owner) { + fail_reason = 2; + goto cleanup; + } + if (alloc->resource_owner) { + if (alloc->owner.resource != alloc0->owner.resource) { + fail_reason = 3; + goto cleanup; + } + if (alloc->owner.resource->device != device) { + fail_reason = 4; + goto cleanup; + } + if (alloc->owner.resource->shared_owner) { + fail_reason = 5; + goto cleanup; + } + } else { + if (alloc->owner.device != device) { + fail_reason = 6; + goto cleanup; + } + } + return 0; +cleanup: + pr_err("Alloc validation failed: reason: %d %x", + fail_reason, alloc_handle); + return STATUS_INVALID_PARAMETER; +} + +static int dxgk_destroy_allocation(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_destroyallocation2 args; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + int ret = 0; + d3dkmt_handle *alloc_handles = NULL; + struct dxgallocation **allocs = NULL; + struct dxgresource *resource = NULL; + int i; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.alloc_count > D3DKMT_CREATEALLOCATION_MAX || + ((args.alloc_count == 0) == (args.resource == 0))) { + pr_err("invalid number of allocations"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + if (args.alloc_count) { + uint handle_size = sizeof(d3dkmt_handle) * args.alloc_count; + + alloc_handles = dxgmem_alloc(process, DXGMEM_TMP, handle_size); + if (alloc_handles == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + allocs = dxgmem_alloc(process, DXGMEM_TMP, + sizeof(struct dxgallocation *) * + args.alloc_count); + if (allocs == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + ret = dxg_copy_from_user(alloc_handles, args.allocations, + handle_size); + if (ret) + goto cleanup; + } + + /* + * The call acquires reference on the device. It is safe to access the + * adapter, because the device holds reference on it. + */ + device = dxgprocess_device_by_handle(process, args.device); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + /* Acquire the device lock to synchronize with the device destriction */ + ret = dxgdevice_acquire_lock_shared(device); + if (ret) { + dxgdevice_release_reference(device); + device = NULL; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + /* + * Destroy the local allocation handles first. If the host handle + * is destroyed first, another object could be assigned to the process + * table at he same place as the allocation handle and it will fail. + */ + if (args.alloc_count) { + dxgprocess_ht_lock_exclusive_down(process); + for (i = 0; i < args.alloc_count; i++) { + allocs[i] = + hmgrtable_get_object_by_type(&process->handle_table, + HMGRENTRY_TYPE_DXGALLOCATION, + alloc_handles[i]); + ret = + validate_alloc(allocs[0], allocs[i], device, + alloc_handles[i]); + if (ret) { + dxgprocess_ht_lock_exclusive_up(process); + goto cleanup; + } + } + dxgprocess_ht_lock_exclusive_up(process); + for (i = 0; i < args.alloc_count; i++) + dxgallocation_free_handle(allocs[i]); + } else { + struct dxgallocation *alloc; + + dxgprocess_ht_lock_exclusive_down(process); + resource = hmgrtable_get_object_by_type(&process->handle_table, + HMGRENTRY_TYPE_DXGRESOURCE, + args.resource); + if (resource == NULL) { + pr_err("Invalid resource handle: %x", + args.resource); + ret = STATUS_INVALID_PARAMETER; + } else if (resource->device != device) { + pr_err("Resource belongs to wrong device: %x", + args.resource); + ret = STATUS_INVALID_PARAMETER; + } else { + hmgrtable_free_handle(&process->handle_table, + HMGRENTRY_TYPE_DXGRESOURCE, + args.resource); + resource->object_state = DXGOBJECTSTATE_DESTROYED; + resource->handle = 0; + resource->handle_valid = 0; + } + dxgprocess_ht_lock_exclusive_up(process); + + if (ret) + goto cleanup; + + dxgdevice_acquire_alloc_list_lock_shared(device); + list_for_each_entry(alloc, &resource->alloc_list_head, + alloc_list_entry) { + dxgallocation_free_handle(alloc); + } + dxgdevice_release_alloc_list_lock_shared(device); + } + + if (args.alloc_count && allocs[0]->resource_owner) + resource = allocs[0]->owner.resource; + + if (resource) { + dxgresource_acquire_reference(resource); + dxgmutex_lock(&resource->resource_mutex); + } + + ret = dxgvmb_send_destroy_allocation(process, device, &adapter->channel, + &args, alloc_handles); + + /* + * Destroy the allocations after the host destroyed it. + * The allocation gpadl teardown will wait until the host unmaps its + * gpadl. + */ + dxgdevice_acquire_alloc_list_lock(device); + if (args.alloc_count) { + for (i = 0; i < args.alloc_count; i++) { + if (allocs[i]) { + allocs[i]->alloc_handle = 0; + dxgallocation_destroy(allocs[i]); + } + } + } else { + dxgresource_destroy(resource); + } + dxgdevice_release_alloc_list_lock(device); + + if (resource) { + dxgmutex_unlock(&resource->resource_mutex); + dxgresource_release_reference(resource); + } + +cleanup: + + if (adapter) + dxgadapter_release_lock_shared(adapter); + + if (device) { + dxgdevice_release_lock_shared(device); + dxgdevice_release_reference(device); + } + + if (alloc_handles) + dxgmem_free(process, DXGMEM_TMP, alloc_handles); + + if (allocs) + dxgmem_free(process, DXGMEM_TMP, allocs); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_make_resident(struct dxgprocess *process, void *__user inargs) +{ + int ret = 0, ret2; + struct d3dddi_makeresident args; + struct d3dddi_makeresident *input = inargs; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.alloc_count > D3DKMT_CREATEALLOCATION_MAX || + args.alloc_count == 0) { + pr_err("invalid number of allocations"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + if (args.paging_queue == 0) { + pr_err("paging queue is missing"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + device = dxgprocess_device_by_object_handle(process, + HMGRENTRY_TYPE_DXGPAGINGQUEUE, + args.paging_queue); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_make_resident(process, NULL, &adapter->channel, + &args); + if (ret && ret != STATUS_PENDING) + goto cleanup; + + ret2 = dxg_copy_to_user(&input->paging_fence_value, + &args.paging_fence_value, sizeof(uint64_t)); + if (ret2) { + ret = ret2; + goto cleanup; + } + + ret2 = dxg_copy_to_user(&input->num_bytes_to_trim, + &args.num_bytes_to_trim, sizeof(uint64_t)); + if (ret2) { + ret = ret2; + goto cleanup; + } + +cleanup: + + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + + return ret; +} + +static int dxgk_evict(struct dxgprocess *process, void *__user inargs) +{ + int ret = 0; + struct d3dkmt_evict args; + struct d3dkmt_evict *input = inargs; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + + TRACE_FUNC_ENTER(__func__); + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.alloc_count > D3DKMT_CREATEALLOCATION_MAX || + args.alloc_count == 0) { + pr_err("invalid number of allocations"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + device = dxgprocess_device_by_handle(process, args.device); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_evict(process, &adapter->channel, &args); + if (ret) + goto cleanup; + + ret = dxg_copy_to_user(&input->num_bytes_to_trim, + &args.num_bytes_to_trim, sizeof(uint64_t)); + if (ret) + goto cleanup; + +cleanup: + + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_offer_allocations(struct dxgprocess *process, + void *__user inargs) +{ + int ret = 0; + struct d3dkmt_offerallocations args; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + + TRACE_FUNC_ENTER(__func__); + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.allocation_count > D3DKMT_CREATEALLOCATION_MAX || + args.allocation_count == 0) { + pr_err("invalid number of allocations"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + if ((args.resources == NULL) == (args.allocations == NULL)) { + pr_err("invalid pointer to resources/allocations"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + device = dxgprocess_device_by_handle(process, args.device); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_offer_allocations(process, &adapter->channel, &args); + +cleanup: + + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_reclaim_allocations(struct dxgprocess *process, + void *__user inargs) +{ + int ret = 0; + struct d3dkmt_reclaimallocations2 args; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + + TRACE_FUNC_ENTER(__func__); + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.allocation_count > D3DKMT_CREATEALLOCATION_MAX || + args.allocation_count == 0) { + pr_err("invalid number of allocations"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + if ((args.resources == NULL) == (args.allocations == NULL)) { + pr_err("invalid pointer to resources/allocations"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + device = dxgprocess_device_by_object_handle(process, + HMGRENTRY_TYPE_DXGPAGINGQUEUE, + args.paging_queue); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_reclaim_allocations(process, &adapter->channel, + device->handle, &args, + &((struct + d3dkmt_reclaimallocations2 *) + inargs)->paging_fence_value); + +cleanup: + + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_submit_command(struct dxgprocess *process, void *__user inargs) +{ + int ret = 0; + struct d3dkmt_submitcommand args; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + + TRACE_FUNC_ENTER(__func__); + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.broadcast_context_count > D3DDDI_MAX_BROADCAST_CONTEXT || + args.broadcast_context_count == 0) { + pr_err("invalid number of contexts"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + if (args.priv_drv_data_size > DXG_MAX_VM_BUS_PACKET_SIZE) { + pr_err("invalid private data size"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + if (args.num_history_buffers > 1024) { + pr_err("invalid number of history buffers"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + if (args.num_primaries > DXG_MAX_VM_BUS_PACKET_SIZE) { + pr_err("invalid number of primaries"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + device = dxgprocess_device_by_object_handle(process, + HMGRENTRY_TYPE_DXGCONTEXT, + args.broadcast_context[0]); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_submit_command(process, &adapter->channel, &args); + +cleanup: + + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_submit_command_to_hwqueue(struct dxgprocess *process, + void *__user inargs) +{ + int ret = 0; + struct d3dkmt_submitcommandtohwqueue args; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + + TRACE_FUNC_ENTER(__func__); + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.priv_drv_data_size > DXG_MAX_VM_BUS_PACKET_SIZE) { + pr_err("invalid private data size"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + if (args.num_primaries > DXG_MAX_VM_BUS_PACKET_SIZE) { + pr_err("invalid number of primaries"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + device = dxgprocess_device_by_object_handle(process, + HMGRENTRY_TYPE_DXGHWQUEUE, + args.hwqueue); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_submit_command_to_hwqueue(process, &adapter->channel, + &args); + +cleanup: + + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_submit_signal_to_hwqueue(struct dxgprocess *process, + void *__user inargs) +{ + int ret = 0; + struct d3dkmt_submitsignalsyncobjectstohwqueue args; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + d3dkmt_handle hwqueue; + + TRACE_FUNC_ENTER(__func__); + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.hwqueue_count > D3DDDI_MAX_BROADCAST_CONTEXT || + args.hwqueue_count == 0) { + pr_err("invalid hwqueue count"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + if (args.object_count > D3DDDI_MAX_OBJECT_SIGNALED || + args.object_count == 0) { + pr_err("invalid number of syn cobject"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + ret = dxg_copy_from_user(&hwqueue, args.hwqueues, + sizeof(d3dkmt_handle)); + if (ret) + goto cleanup; + + device = dxgprocess_device_by_object_handle(process, + HMGRENTRY_TYPE_DXGHWQUEUE, + hwqueue); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_signal_sync_object(process, &adapter->channel, + args.flags, 0, 0, + args.object_count, args.objects, + args.hwqueue_count, args.hwqueues, + args.object_count, + args.fence_values, NULL, 0); + +cleanup: + + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_submit_wait_to_hwqueue(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_submitwaitforsyncobjectstohwqueue args; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + int ret = 0; + d3dkmt_handle *objects = NULL; + uint object_size; + uint64_t *fences = NULL; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.object_count > D3DDDI_MAX_OBJECT_WAITED_ON || + args.object_count == 0) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + object_size = sizeof(d3dkmt_handle) * args.object_count; + objects = dxgmem_alloc(process, DXGMEM_TMP, object_size); + if (objects == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + ret = dxg_copy_from_user(objects, args.objects, object_size); + if (ret) + goto cleanup; + + object_size = sizeof(uint64_t) * args.object_count; + fences = dxgmem_alloc(process, DXGMEM_TMP, object_size); + if (fences == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + ret = dxg_copy_from_user(fences, args.fence_values, object_size); + if (ret) + goto cleanup; + + device = dxgprocess_device_by_object_handle(process, + HMGRENTRY_TYPE_DXGHWQUEUE, + args.hwqueue); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_wait_sync_object_gpu(process, &adapter->channel, + args.hwqueue, args.object_count, + objects, fences, false); + +cleanup: + + if (objects) + dxgmem_free(process, DXGMEM_TMP, objects); + if (fences) + dxgmem_free(process, DXGMEM_TMP, fences); + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_map_gpu_va(struct dxgprocess *process, void *__user inargs) +{ + int ret = 0, ret2; + struct d3dddi_mapgpuvirtualaddress args; + struct d3dddi_mapgpuvirtualaddress *input = inargs; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + + TRACE_FUNC_ENTER(__func__); + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + device = dxgprocess_device_by_object_handle(process, + HMGRENTRY_TYPE_DXGPAGINGQUEUE, + args.paging_queue); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_map_gpu_va(process, 0, &adapter->channel, &args); + if (ret && ret != (int)STATUS_PENDING) + goto cleanup; + + ret2 = dxg_copy_to_user(&input->paging_fence_value, + &args.paging_fence_value, sizeof(uint64_t)); + if (ret2) { + ret = ret2; + goto cleanup; + } + + ret2 = dxg_copy_to_user(&input->virtual_address, &args.virtual_address, + sizeof(args.virtual_address)); + if (ret2) { + ret = ret2; + goto cleanup; + } + +cleanup: + + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_reserve_gpu_va(struct dxgprocess *process, void *__user inargs) +{ + int ret = 0; + struct d3dddi_reservegpuvirtualaddress args; + struct d3dddi_reservegpuvirtualaddress *input = inargs; + struct dxgadapter *adapter = NULL; + struct dxgdevice *device = NULL; + + TRACE_FUNC_ENTER(__func__); + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + adapter = dxgprocess_adapter_by_handle(process, args.adapter); + if (adapter == NULL) { + device = dxgprocess_device_by_object_handle(process, + HMGRENTRY_TYPE_DXGPAGINGQUEUE, + args.adapter); + if (device == NULL) { + pr_err("invalid adapter or paging queue: 0x%x", + args.adapter); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + adapter = device->adapter; + dxgadapter_acquire_reference(adapter); + dxgdevice_release_reference(device); + } else { + args.adapter = adapter->host_handle; + } + + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + dxgadapter_release_reference(adapter); + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_reserve_gpu_va(process, &adapter->channel, &args); + if (ret) + goto cleanup; + + ret = dxg_copy_to_user(&input->virtual_address, &args.virtual_address, + sizeof(args.virtual_address)); + if (ret) + goto cleanup; + +cleanup: + + if (adapter) { + dxgadapter_release_lock_shared(adapter); + dxgadapter_release_reference(adapter); + } + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_free_gpu_va(struct dxgprocess *process, void *__user inargs) +{ + int ret = 0; + struct d3dkmt_freegpuvirtualaddress args; + struct dxgadapter *adapter = NULL; + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + adapter = dxgprocess_adapter_by_handle(process, args.adapter); + if (adapter == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + dxgadapter_release_reference(adapter); + adapter = NULL; + goto cleanup; + } + + args.adapter = adapter->host_handle; + ret = dxgvmb_send_free_gpu_va(process, &adapter->channel, &args); + +cleanup: + + if (adapter) { + dxgadapter_release_lock_shared(adapter); + dxgadapter_release_reference(adapter); + } + + return ret; +} + +static int dxgk_update_gpu_va(struct dxgprocess *process, void *__user inargs) +{ + int ret = 0; + struct d3dkmt_updategpuvirtualaddress args; + struct d3dkmt_updategpuvirtualaddress *input = inargs; + struct dxgadapter *adapter = NULL; + struct dxgdevice *device = NULL; + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + device = dxgprocess_device_by_handle(process, args.device); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_update_gpu_va(process, &adapter->channel, &args); + if (ret) + goto cleanup; + + ret = dxg_copy_to_user(&input->fence_value, &args.fence_value, + sizeof(args.fence_value)); + if (ret) + goto cleanup; + +cleanup: + + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + + return ret; +} + +static int dxgk_create_sync_object(struct dxgprocess *process, + void *__user inargs) +{ + int ret = 0; + struct d3dkmt_createsynchronizationobject2 args; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + struct eventfd_ctx *event = NULL; + struct dxgsyncobject *syncobj = NULL; + bool host_event_added = false; + bool device_lock_acquired = false; + struct dxgsharedsyncobject *syncobjgbl = NULL; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + device = dxgprocess_device_by_handle(process, args.device); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + ret = dxgdevice_acquire_lock_shared(device); + if (ret) + goto cleanup; + + device_lock_acquired = true; + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + syncobj = dxgsyncobject_create(process, device, adapter, args.info.type, + args.info.flags); + if (syncobj == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + if (args.info.flags.shared && syncobj->monitored_fence && + !args.info.flags.nt_security_sharing) { + pr_err("monitored fence requires nt_security_sharing"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + if (args.info.type == D3DDDI_CPU_NOTIFICATION) { + event = eventfd_ctx_fdget((int) + args.info.cpu_notification.event); + if (IS_ERR(event)) { + pr_err("failed to reference the event"); + event = NULL; + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + syncobj->host_event->event_id = dxgglobal_new_host_event_id(); + syncobj->host_event->cpu_event = event; + syncobj->host_event->remove_from_list = false; + syncobj->host_event->destroy_after_signal = false; + dxgglobal_add_host_event(syncobj->host_event); + host_event_added = true; + args.info.cpu_notification.event = + syncobj->host_event->event_id; + TRACE_DEBUG(1, "creating CPU notification event: %lld", + args.info.cpu_notification.event); + } + + ret = dxgvmb_send_create_sync_object(process, &adapter->channel, &args, + syncobj); + if (ret) + goto cleanup; + + if (args.info.flags.shared) { + if (args.info.shared_handle == 0) { + pr_err("shared handle should not be 0"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + syncobjgbl = dxgsharedsyncobj_create(device->adapter, syncobj); + if (syncobjgbl == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + dxgsharedsyncobj_add_syncobj(syncobjgbl, syncobj); + + syncobjgbl->host_shared_handle = args.info.shared_handle; + if (!args.info.flags.nt_security_sharing) { + hmgrtable_lock(&dxgglobal->handle_table, DXGLOCK_EXCL); + syncobjgbl->global_shared_handle = + hmgrtable_alloc_handle(&dxgglobal->handle_table, + syncobjgbl, + HMGRENTRY_TYPE_DXGSYNCOBJECT, + true); + if (syncobjgbl->global_shared_handle) { + args.info.shared_handle = + syncobjgbl->global_shared_handle; + } else { + ret = STATUS_NO_MEMORY; + } + hmgrtable_unlock(&dxgglobal->handle_table, + DXGLOCK_EXCL); + if (ret) + goto cleanup; + } + } + + ret = dxg_copy_to_user(inargs, &args, sizeof(args)); + if (ret) + goto cleanup; + + hmgrtable_lock(&process->handle_table, DXGLOCK_EXCL); + ret = hmgrtable_assign_handle(&process->handle_table, syncobj, + HMGRENTRY_TYPE_DXGSYNCOBJECT, + args.sync_object); + if (!ret) + syncobj->handle = args.sync_object; + hmgrtable_unlock(&process->handle_table, DXGLOCK_EXCL); + +cleanup: + + if (ret) { + if (syncobj) { + dxgsyncobject_destroy(process, syncobj); + if (args.sync_object) + dxgvmb_send_destroy_sync_object(process, + args. + sync_object); + event = NULL; + } + if (event) + eventfd_ctx_put(event); + } + if (syncobjgbl) + dxgsharedsyncobj_release_reference(syncobjgbl); + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device_lock_acquired) + dxgdevice_release_lock_shared(device); + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_destroy_sync_object(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_destroysynchronizationobject args; + struct dxgsyncobject *syncobj = NULL; + int ret = 0; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + TRACE_DEBUG(1, "handle 0x%x", args.sync_object); + hmgrtable_lock(&process->handle_table, DXGLOCK_EXCL); + syncobj = hmgrtable_get_object_by_type(&process->handle_table, + HMGRENTRY_TYPE_DXGSYNCOBJECT, + args.sync_object); + if (syncobj) { + TRACE_DEBUG(1, "syncobj 0x%p", syncobj); + syncobj->handle = 0; + hmgrtable_free_handle(&process->handle_table, + HMGRENTRY_TYPE_DXGSYNCOBJECT, + args.sync_object); + } + hmgrtable_unlock(&process->handle_table, DXGLOCK_EXCL); + + if (syncobj == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + dxgsyncobject_destroy(process, syncobj); + + ret = dxgvmb_send_destroy_sync_object(process, args.sync_object); + +cleanup: + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_open_sync_object_nt(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_opensyncobjectfromnthandle2 args; + struct dxgsyncobject *syncobj = NULL; + struct dxgsharedsyncobject *syncobj_fd = NULL; + struct file *file = NULL; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + struct d3dddi_synchronizationobject_flags flags = { }; + int ret; + bool device_lock_acquired = false; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + args.sync_object = 0; + + if (args.device) { + device = dxgprocess_device_by_handle(process, args.device); + if (device == NULL) + goto cleanup; + } else { + pr_err("device handle is missing"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + ret = dxgdevice_acquire_lock_shared(device); + if (ret) + goto cleanup; + + device_lock_acquired = true; + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + file = fget(args.nt_handle); + if (!file) { + pr_err("failed to get file from handle: %llx", + args.nt_handle); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + if (file->f_op != &dxg_syncobj_fops) { + pr_err("invalid fd: %llx", args.nt_handle); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + syncobj_fd = file->private_data; + if (syncobj_fd == NULL) { + pr_err("invalid private data: %llx", args.nt_handle); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + flags.shared = 1; + flags.nt_security_sharing = 1; + syncobj = dxgsyncobject_create(process, device, adapter, + syncobj_fd->type, flags); + if (syncobj == NULL) { + pr_err("failed to create sync object"); + goto cleanup; + } + + dxgsharedsyncobj_add_syncobj(syncobj_fd, syncobj); + + ret = dxgvmb_send_open_sync_object_nt(process, &dxgglobal->channel, + &args, syncobj); + if (ret) { + pr_err("failed to open sync object on host: %x", + syncobj_fd->host_shared_handle); + goto cleanup; + } + + hmgrtable_lock(&process->handle_table, DXGLOCK_EXCL); + ret = hmgrtable_assign_handle(&process->handle_table, syncobj, + HMGRENTRY_TYPE_DXGSYNCOBJECT, + args.sync_object); + if (!ret) { + syncobj->handle = args.sync_object; + dxgsyncobject_acquire_reference(syncobj); + } + hmgrtable_unlock(&process->handle_table, DXGLOCK_EXCL); + + if (ret) + goto cleanup; + + ret = dxg_copy_to_user(inargs, &args, sizeof(args)); + if (!ret) + goto success; + +cleanup: + + if (syncobj) { + dxgsyncobject_destroy(process, syncobj); + syncobj = NULL; + } + + if (args.sync_object) + dxgvmb_send_destroy_sync_object(process, args.sync_object); + +success: + + if (file) + fput(file); + if (syncobj) + dxgsyncobject_release_reference(syncobj); + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device_lock_acquired) + dxgdevice_release_lock_shared(device); + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_open_sync_object(struct dxgprocess *process, + void *__user inargs) +{ + d3dkmt_handle shared_handle = 0; + d3dkmt_handle new_handle = 0; + struct d3dkmt_opensynchronizationobject *__user inp = inargs; + struct dxgsyncobject *syncobj = NULL; + struct dxgsharedsyncobject *syncobjgbl = NULL; + struct d3dddi_synchronizationobject_flags flags = { }; + int ret; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&shared_handle, &inp->shared_handle, + sizeof(shared_handle)); + if (ret) + goto cleanup; + + hmgrtable_lock(&dxgglobal->handle_table, DXGLOCK_SHARED); + syncobjgbl = hmgrtable_get_object_by_type(&dxgglobal->handle_table, + HMGRENTRY_TYPE_DXGSYNCOBJECT, + shared_handle); + if (syncobjgbl) + dxgsharedsyncobj_acquire_reference(syncobjgbl); + hmgrtable_unlock(&dxgglobal->handle_table, DXGLOCK_SHARED); + + if (syncobjgbl == NULL) { + pr_err("invalid sync object shared handle: %x", + shared_handle); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + if (syncobjgbl->monitored_fence) { + pr_err("Open monitored fence using global handle"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + flags.shared = 1; + syncobj = dxgsyncobject_create(process, NULL, syncobjgbl->adapter, + syncobjgbl->type, flags); + if (syncobj == NULL) { + pr_err("failed to create sync object"); + goto cleanup; + } + + dxgsharedsyncobj_add_syncobj(syncobjgbl, syncobj); + + ret = dxgvmb_send_open_sync_object(process, &dxgglobal->channel, + syncobjgbl->host_shared_handle, + &new_handle); + if (ret) { + pr_err("failed to open sync object on host: %x", + syncobjgbl->host_shared_handle); + goto cleanup; + } + + hmgrtable_lock(&process->handle_table, DXGLOCK_EXCL); + ret = hmgrtable_assign_handle(&process->handle_table, syncobj, + HMGRENTRY_TYPE_DXGSYNCOBJECT, new_handle); + if (!ret) { + syncobj->handle = new_handle; + dxgsyncobject_acquire_reference(syncobj); + } + hmgrtable_unlock(&process->handle_table, DXGLOCK_EXCL); + + if (ret) + goto cleanup; + + ret = dxg_copy_to_user(&inp->sync_object, &new_handle, + sizeof(new_handle)); + if (!ret) + goto success; + +cleanup: + + if (syncobj) { + dxgsyncobject_destroy(process, syncobj); + syncobj = NULL; + } + + if (new_handle) + dxgvmb_send_destroy_sync_object(process, new_handle); + +success: + + if (syncobj) + dxgsyncobject_release_reference(syncobj); + if (syncobjgbl) + dxgsharedsyncobj_release_reference(syncobjgbl); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_signal_sync_object(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_signalsynchronizationobject2 args; + struct d3dkmt_signalsynchronizationobject2 *__user in_args = inargs; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + int ret = 0; + uint fence_count = 1; + struct eventfd_ctx *event = NULL; + struct dxghostevent *host_event = NULL; + bool host_event_added = false; + u64 host_event_id = 0; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.context_count >= D3DDDI_MAX_BROADCAST_CONTEXT || + args.object_count > D3DDDI_MAX_OBJECT_SIGNALED) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + if (args.flags.enqueue_cpu_event) { + host_event = dxgmem_alloc(process, DXGMEM_EVENT, + sizeof(*host_event)); + if (host_event == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + host_event->process = process; + event = eventfd_ctx_fdget((int)args.cpu_event_handle); + if (IS_ERR(event)) { + pr_err("failed to reference the event"); + event = NULL; + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + fence_count = 0; + host_event->cpu_event = event; + host_event_id = dxgglobal_new_host_event_id(); + host_event->event_id = host_event_id; + host_event->remove_from_list = true; + host_event->destroy_after_signal = true; + dxgglobal_add_host_event(host_event); + host_event_added = true; + } + + device = dxgprocess_device_by_object_handle(process, + HMGRENTRY_TYPE_DXGCONTEXT, + args.context); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_signal_sync_object(process, &adapter->channel, + args.flags, args.fence.fence_value, + args.context, args.object_count, + in_args->object_array, + args.context_count, + in_args->contexts, fence_count, + NULL, (void *)host_event_id, 0); + + /* + * When the send operation succeeds, the host event will be destroyed + * after signal from the host + */ + +cleanup: + + if (ret) { + if (host_event_added) { + /* The event might be signaled and destroyed by host */ + host_event = dxgglobal_get_host_event(host_event_id); + if (host_event) { + eventfd_ctx_put(event); + event = NULL; + dxgmem_free(process, DXGMEM_EVENT, host_event); + host_event = NULL; + } + } + if (event) + eventfd_ctx_put(event); + if (host_event) + dxgmem_free(process, DXGMEM_EVENT, host_event); + } + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_signal_sync_object_cpu(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_signalsynchronizationobjectfromcpu args; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + int ret = 0; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + device = dxgprocess_device_by_handle(process, args.device); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_signal_sync_object(process, &adapter->channel, + args.flags, 0, 0, + args.object_count, args.objects, 0, + NULL, args.object_count, + args.fence_values, NULL, + args.device); + +cleanup: + + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_signal_sync_object_gpu(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_signalsynchronizationobjectfromgpu args; + struct d3dkmt_signalsynchronizationobjectfromgpu *__user user_args = + inargs; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + struct d3dddicb_signalflags flags = { }; + int ret = 0; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.object_count == 0 || + args.object_count > DXG_MAX_VM_BUS_PACKET_SIZE) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + device = dxgprocess_device_by_object_handle(process, + HMGRENTRY_TYPE_DXGCONTEXT, + args.context); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_signal_sync_object(process, &adapter->channel, + flags, 0, 0, args.object_count, + args.objects, 1, + &user_args->context, + args.object_count, + args.monitored_fence_values, NULL, + 0); + +cleanup: + + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_signal_sync_object_gpu2(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_signalsynchronizationobjectfromgpu2 args; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + d3dkmt_handle context_handle; + struct eventfd_ctx *event = NULL; + uint64_t *fences = NULL; + uint fence_count = 0; + int ret = 0; + struct dxghostevent *host_event = NULL; + bool host_event_added = false; + u64 host_event_id = 0; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.flags.enqueue_cpu_event) { + if (args.object_count != 0 || args.cpu_event_handle == 0) { + pr_err("Bad input for EnqueueCpuEvent: %d %lld", + args.object_count, args.cpu_event_handle); + + } + } else if (args.object_count == 0 || + args.object_count > DXG_MAX_VM_BUS_PACKET_SIZE || + args.context_count == 0 || + args.context_count > DXG_MAX_VM_BUS_PACKET_SIZE) { + pr_err("Invalid input: %d %d", + args.object_count, args.context_count); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + ret = dxg_copy_from_user(&context_handle, args.contexts, + sizeof(d3dkmt_handle)); + if (ret) + goto cleanup; + + if (args.flags.enqueue_cpu_event) { + host_event = dxgmem_alloc(process, DXGMEM_EVENT, + sizeof(*host_event)); + if (host_event == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + host_event->process = process; + event = eventfd_ctx_fdget((int)args.cpu_event_handle); + if (IS_ERR(event)) { + pr_err("failed to reference the event"); + event = NULL; + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + fence_count = 0; + host_event->cpu_event = event; + host_event_id = dxgglobal_new_host_event_id(); + host_event->event_id = host_event_id; + host_event->remove_from_list = true; + host_event->destroy_after_signal = true; + dxgglobal_add_host_event(host_event); + host_event_added = true; + } else { + fences = args.monitored_fence_values; + fence_count = args.object_count; + } + + device = dxgprocess_device_by_object_handle(process, + HMGRENTRY_TYPE_DXGCONTEXT, + context_handle); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_signal_sync_object(process, &adapter->channel, + args.flags, 0, 0, + args.object_count, args.objects, + args.context_count, args.contexts, + fence_count, fences, + (void *)host_event_id, 0); + +cleanup: + + if (ret) { + if (host_event_added) { + /* The event might be signaled and destroyed by host */ + host_event = dxgglobal_get_host_event(host_event_id); + if (host_event) { + eventfd_ctx_put(event); + event = NULL; + dxgmem_free(process, DXGMEM_EVENT, host_event); + host_event = NULL; + } + } + if (event) + eventfd_ctx_put(event); + if (host_event) + dxgmem_free(process, DXGMEM_EVENT, host_event); + } + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_wait_sync_object(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_waitforsynchronizationobject2 args; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + int ret = 0; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.object_count > D3DDDI_MAX_OBJECT_WAITED_ON || + args.object_count == 0) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + device = dxgprocess_device_by_object_handle(process, + HMGRENTRY_TYPE_DXGCONTEXT, + args.context); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + TRACE_DEBUG(1, "Fence value: %lld", args.fence.fence_value); + ret = dxgvmb_send_wait_sync_object_gpu(process, &adapter->channel, + args.context, args.object_count, + args.object_array, + &args.fence.fence_value, true); + +cleanup: + + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_wait_sync_object_cpu(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_waitforsynchronizationobjectfromcpu args; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + struct eventfd_ctx *event = NULL; + struct dxghostevent host_event = { }; + struct dxghostevent *async_host_event = NULL; + struct completion local_event = { }; + u64 event_id = 0; + int ret = 0; + unsigned long t; + bool host_event_added = false; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.object_count > DXG_MAX_VM_BUS_PACKET_SIZE || + args.object_count == 0) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + if (args.async_event) { + async_host_event = dxgmem_alloc(process, DXGMEM_EVENT, + sizeof(*async_host_event)); + if (async_host_event == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + async_host_event->process = process; + event = eventfd_ctx_fdget((int)args.async_event); + if (IS_ERR(event)) { + pr_err("failed to reference the event"); + event = NULL; + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + async_host_event->cpu_event = event; + async_host_event->event_id = dxgglobal_new_host_event_id(); + async_host_event->destroy_after_signal = true; + dxgglobal_add_host_event(async_host_event); + event_id = async_host_event->event_id; + host_event_added = true; + } else { + init_completion(&local_event); + host_event.completion_event = &local_event; + host_event.event_id = dxgglobal_new_host_event_id(); + dxgglobal_add_host_event(&host_event); + event_id = host_event.event_id; + } + + device = dxgprocess_device_by_handle(process, args.device); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_wait_sync_object_cpu(process, &adapter->channel, + &args, event_id); + if (ret) + goto cleanup; + + if (args.async_event == 0) { + t = wait_for_completion_timeout(&local_event, (10 * HZ)); + if (!t) { + TRACE_DEBUG(1, "timeout waiting for completion"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + } + +cleanup: + + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + if (host_event.event_id) + dxgglobal_remove_host_event(&host_event); + if (ret) { + if (host_event_added) { + async_host_event = dxgglobal_get_host_event(event_id); + if (async_host_event) { + eventfd_ctx_put(event); + event = NULL; + dxgmem_free(process, DXGMEM_EVENT, + async_host_event); + async_host_event = NULL; + } + } + if (event) + eventfd_ctx_put(event); + if (async_host_event) + dxgmem_free(process, DXGMEM_EVENT, async_host_event); + } + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_wait_sync_object_gpu(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_waitforsynchronizationobjectfromgpu args; + struct dxgcontext *context = NULL; + d3dkmt_handle device_handle = 0; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + struct dxgsyncobject *syncobj = NULL; + d3dkmt_handle *objects = NULL; + uint object_size; + uint64_t *fences = NULL; + int ret = 0; + enum hmgrentry_type syncobj_type = HMGRENTRY_TYPE_FREE; + bool monitored_fence = false; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.object_count > DXG_MAX_VM_BUS_PACKET_SIZE || + args.object_count == 0) { + pr_err("Invalid object count: %d", args.object_count); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + object_size = sizeof(d3dkmt_handle) * args.object_count; + objects = dxgmem_alloc(process, DXGMEM_TMP, object_size); + if (objects == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + ret = dxg_copy_from_user(objects, args.objects, object_size); + if (ret) + goto cleanup; + + hmgrtable_lock(&process->handle_table, DXGLOCK_SHARED); + context = hmgrtable_get_object_by_type(&process->handle_table, + HMGRENTRY_TYPE_DXGCONTEXT, + args.context); + if (context) { + device_handle = context->device_handle; + syncobj_type = + hmgrtable_get_object_type(&process->handle_table, + objects[0]); + } + if (device_handle == 0) { + pr_err("Invalid context handle: %x", args.context); + ret = STATUS_INVALID_PARAMETER; + } else { + if (syncobj_type == HMGRENTRY_TYPE_MONITOREDFENCE) { + monitored_fence = true; + } else if (syncobj_type == HMGRENTRY_TYPE_DXGSYNCOBJECT) { + syncobj = + hmgrtable_get_object_by_type(&process->handle_table, + HMGRENTRY_TYPE_DXGSYNCOBJECT, + objects[0]); + if (syncobj == NULL) { + pr_err("Invalid syncobj: %x", objects[0]); + ret = STATUS_INVALID_PARAMETER; + } else { + monitored_fence = syncobj->monitored_fence; + } + } else { + pr_err("Invalid syncobj type: %x", objects[0]); + ret = STATUS_INVALID_PARAMETER; + } + } + hmgrtable_unlock(&process->handle_table, DXGLOCK_SHARED); + + if (ret) + goto cleanup; + + if (monitored_fence) { + object_size = sizeof(uint64_t) * args.object_count; + fences = dxgmem_alloc(process, DXGMEM_TMP, object_size); + if (fences == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + ret = dxg_copy_from_user(fences, args.monitored_fence_values, + object_size); + if (ret) + goto cleanup; + } else { + fences = &args.fence_value; + } + + device = dxgprocess_device_by_handle(process, device_handle); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_wait_sync_object_gpu(process, &adapter->channel, + args.context, args.object_count, + objects, fences, + !monitored_fence); + +cleanup: + + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + if (objects) + dxgmem_free(process, DXGMEM_TMP, objects); + if (fences && fences != &args.fence_value) + dxgmem_free(process, DXGMEM_TMP, fences); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_lock2(struct dxgprocess *process, void *__user inargs) +{ + struct d3dkmt_lock2 args; + struct d3dkmt_lock2 *__user result = inargs; + int ret = 0; + struct dxgadapter *adapter = NULL; + struct dxgdevice *device = NULL; + struct dxgallocation *alloc = NULL; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + args.data = NULL; + hmgrtable_lock(&process->handle_table, DXGLOCK_SHARED); + alloc = hmgrtable_get_object_by_type(&process->handle_table, + HMGRENTRY_TYPE_DXGALLOCATION, + args.allocation); + if (alloc == NULL) { + ret = STATUS_INVALID_PARAMETER; + } else { + if (alloc->cpu_address) { + ret = dxg_copy_to_user(&result->data, + &alloc->cpu_address, + sizeof(args.data)); + if (NT_SUCCESS(ret)) { + args.data = alloc->cpu_address; + if (alloc->cpu_address_mapped) + alloc->cpu_address_refcount++; + } + } + } + hmgrtable_unlock(&process->handle_table, DXGLOCK_SHARED); + if (ret) + goto cleanup; + if (args.data) + goto success; + + /* + * The call acquires reference on the device. It is safe to access the + * adapter, because the device holds reference on it. + */ + device = dxgprocess_device_by_handle(process, args.device); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_lock2(process, &adapter->channel, &args, result); + +cleanup: + + if (adapter) + dxgadapter_release_lock_shared(adapter); + + if (device) + dxgdevice_release_reference(device); + +success: + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_unlock2(struct dxgprocess *process, void *__user inargs) +{ + struct d3dkmt_unlock2 args; + int ret = 0; + struct dxgadapter *adapter = NULL; + struct dxgdevice *device = NULL; + struct dxgallocation *alloc = NULL; + bool done = false; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + hmgrtable_lock(&process->handle_table, DXGLOCK_SHARED); + alloc = hmgrtable_get_object_by_type(&process->handle_table, + HMGRENTRY_TYPE_DXGALLOCATION, + args.allocation); + if (alloc == NULL) { + ret = STATUS_INVALID_PARAMETER; + } else { + if (alloc->cpu_address == NULL) { + pr_err("Allocation is not locked: %p", alloc); + ret = STATUS_INVALID_PARAMETER; + } else if (alloc->cpu_address_mapped) { + if (alloc->cpu_address_refcount > 0) { + alloc->cpu_address_refcount--; + if (alloc->cpu_address_refcount != 0) { + done = true; + } else { + dxg_unmap_iospace(alloc->cpu_address, + alloc-> + num_pages << + PAGE_SHIFT); + alloc->cpu_address_mapped = false; + alloc->cpu_address = NULL; + } + } else { + pr_err("Bad cpu access refcount"); + done = true; + } + } + } + hmgrtable_unlock(&process->handle_table, DXGLOCK_SHARED); + if (done) + goto success; + if (ret) + goto cleanup; + + /* + * The call acquires reference on the device. It is safe to access the + * adapter, because the device holds reference on it. + */ + device = dxgprocess_device_by_handle(process, args.device); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_unlock2(process, &adapter->channel, &args); + if (ret) + goto cleanup; + +cleanup: + if (adapter) + dxgadapter_release_lock_shared(adapter); + + if (device) + dxgdevice_release_reference(device); + +success: + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_update_alloc_property(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dddi_updateallocproperty args; + int ret = 0; + struct dxgadapter *adapter = NULL; + struct dxgdevice *device = NULL; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + device = dxgprocess_device_by_object_handle(process, + HMGRENTRY_TYPE_DXGPAGINGQUEUE, + args.paging_queue); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_update_alloc_property(process, &adapter->channel, + &args, inargs); + +cleanup: + if (adapter) + dxgadapter_release_lock_shared(adapter); + + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_mark_device_as_error(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_markdeviceaserror args; + struct dxgadapter *adapter = NULL; + struct dxgdevice *device = NULL; + int ret; + + TRACE_FUNC_ENTER(__func__); + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + device = dxgprocess_device_by_handle(process, args.device); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + ret = dxgvmb_send_mark_device_as_error(process, &adapter->channel, + &args); +cleanup: + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_query_alloc_residency(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_queryallocationresidency args; + struct dxgadapter *adapter = NULL; + struct dxgdevice *device = NULL; + int ret; + + TRACE_FUNC_ENTER(__func__); + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if ((args.allocation_count == 0) == (args.resource == 0)) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + device = dxgprocess_device_by_handle(process, args.device); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + ret = dxgvmb_send_query_alloc_residency(process, &adapter->channel, + &args); +cleanup: + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_set_allocation_priority(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_setallocationpriority args; + struct dxgadapter *adapter = NULL; + struct dxgdevice *device = NULL; + int ret; + + TRACE_FUNC_ENTER(__func__); + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + device = dxgprocess_device_by_handle(process, args.device); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + ret = dxgvmb_send_set_allocation_priority(process, &adapter->channel, + &args); +cleanup: + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_get_allocation_priority(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_getallocationpriority args; + struct dxgadapter *adapter = NULL; + struct dxgdevice *device = NULL; + int ret; + + TRACE_FUNC_ENTER(__func__); + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + device = dxgprocess_device_by_handle(process, args.device); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + ret = dxgvmb_send_get_allocation_priority(process, &adapter->channel, + &args); +cleanup: + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static long set_context_scheduling_priority(struct dxgprocess *process, + d3dkmt_handle hcontext, + int priority, bool in_process) +{ + int ret = 0; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + + device = dxgprocess_device_by_object_handle(process, + HMGRENTRY_TYPE_DXGCONTEXT, + hcontext); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + ret = dxgvmb_send_set_context_scheduling_priority(process, + &adapter->channel, + hcontext, priority, + in_process); + if (ret) + pr_err("send_set_context_scheduling_priority failed"); +cleanup: + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + + return ret; +} + +static int dxgk_set_context_scheduling_priority(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_setcontextschedulingpriority args; + int ret; + + TRACE_FUNC_ENTER(__func__); + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + ret = set_context_scheduling_priority(process, args.context, + args.priority, false); +cleanup: + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static long get_context_scheduling_priority(struct dxgprocess *process, + d3dkmt_handle hcontext, + __user int *priority, + bool in_process) +{ + int ret = 0; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + int pri = 0; + + device = dxgprocess_device_by_object_handle(process, + HMGRENTRY_TYPE_DXGCONTEXT, + hcontext); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + ret = dxgvmb_send_get_context_scheduling_priority(process, + &adapter->channel, + hcontext, &pri, + in_process); + if (ret) + goto cleanup; + ret = dxg_copy_to_user(priority, &pri, sizeof(pri)); + +cleanup: + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + + return ret; +} + +static int dxgk_get_context_scheduling_priority(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_getcontextschedulingpriority args; + int ret; + + TRACE_FUNC_ENTER(__func__); + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + ret = get_context_scheduling_priority(process, args.context, + &((struct + d3dkmt_getcontextschedulingpriority + *) + inargs)->priority, false); +cleanup: + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_set_context_process_scheduling_priority(struct dxgprocess + *process, + void *__user inargs) +{ + struct d3dkmt_setcontextinprocessschedulingpriority args; + int ret; + + TRACE_FUNC_ENTER(__func__); + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + ret = set_context_scheduling_priority(process, args.context, + args.priority, true); +cleanup: + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_get_context_process_scheduling_priority(struct dxgprocess + *process, + void *__user inargs) +{ + struct d3dkmt_getcontextinprocessschedulingpriority args; + int ret; + + TRACE_FUNC_ENTER(__func__); + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + ret = get_context_scheduling_priority(process, args.context, + &((struct + d3dkmt_getcontextinprocessschedulingpriority + *) + inargs)->priority, true); +cleanup: + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_change_vidmem_reservation(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_changevideomemoryreservation args; + int ret; + struct dxgadapter *adapter = NULL; + bool adapter_locked = false; + + TRACE_FUNC_ENTER(__func__); + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.process != 0) { + pr_err("setting memory reservation for other process"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = dxgprocess_adapter_by_handle(process, args.adapter); + if (adapter == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + adapter_locked = true; + args.adapter = 0; + ret = dxgvmb_send_change_vidmem_reservation(process, &adapter->channel, + 0, &args); + +cleanup: + + if (adapter_locked) + dxgadapter_release_lock_shared(adapter); + if (adapter) + dxgadapter_release_reference(adapter); + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_query_clock_calibration(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_queryclockcalibration args; + int ret; + struct dxgadapter *adapter = NULL; + bool adapter_locked = false; + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + adapter = dxgprocess_adapter_by_handle(process, args.adapter); + if (adapter == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + adapter_locked = true; + + args.adapter = adapter->host_handle; + ret = dxgvmb_send_query_clock_calibration(process, &adapter->channel, + &args, inargs); + if (ret) + goto cleanup; + ret = dxg_copy_to_user(inargs, &args, sizeof(args)); + +cleanup: + + if (adapter_locked) + dxgadapter_release_lock_shared(adapter); + if (adapter) + dxgadapter_release_reference(adapter); + return ret; +} + +static int dxgk_flush_heap_transitions(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_flushheaptransitions args; + int ret; + struct dxgadapter *adapter = NULL; + bool adapter_locked = false; + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + adapter = dxgprocess_adapter_by_handle(process, args.adapter); + if (adapter == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + adapter_locked = true; + + args.adapter = adapter->host_handle; + ret = dxgvmb_send_flush_heap_transitions(process, &adapter->channel, + &args); + if (ret) + goto cleanup; + ret = dxg_copy_to_user(inargs, &args, sizeof(args)); + +cleanup: + + if (adapter_locked) + dxgadapter_release_lock_shared(adapter); + if (adapter) + dxgadapter_release_reference(adapter); + return ret; +} + +static int handle_table_escape(struct dxgprocess *process, + struct d3dkmt_escape *args, + struct d3dkmt_ht_desc *cmdin) +{ + int ret = 0; + struct d3dkmt_ht_desc cmd; + struct hmgrtable *table; + + dxgmutex_lock(&process->process_mutex); + cmd = *cmdin; + if (cmd.index >= 2) { + pr_err("invalid table index"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + table = process->test_handle_table[cmd.index]; + if (table == NULL) { + table = dxgmem_alloc(process, DXGMEM_HANDLE_TABLE, + sizeof(*table)); + if (table == NULL) { + pr_err("failed to allocate handle table"); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + hmgrtable_init(table, process); + process->test_handle_table[cmd.index] = table; + } + switch (cmd.command) { + case D3DKMT_HT_COMMAND_ALLOC: + cmd.handle = hmgrtable_alloc_handle_safe(table, cmd.object, + (enum hmgrentry_type) + cmd.object_type, true); + ret = dxg_copy_to_user(args->priv_drv_data, &cmd, sizeof(cmd)); + break; + case D3DKMT_HT_COMMAND_FREE: + hmgrtable_free_handle_safe(table, + (enum hmgrentry_type)cmd.object_type, + cmd.handle); + break; + break; + case D3DKMT_HT_COMMAND_ASSIGN: + ret = hmgrtable_assign_handle_safe(table, cmd.object, + (enum hmgrentry_type)cmd. + object_type, cmd.handle); + break; + case D3DKMT_HT_COMMAND_GET: + hmgrtable_lock(table, DXGLOCK_SHARED); + cmd.object = hmgrtable_get_object_by_type(table, + (enum hmgrentry_type) + cmd.object_type, + cmd.handle); + hmgrtable_unlock(table, DXGLOCK_SHARED); + ret = dxg_copy_to_user(args->priv_drv_data, &cmd, sizeof(cmd)); + break; + case D3DKMT_HT_COMMAND_DESTROY: + if (table) { + hmgrtable_destroy(table); + dxgmem_free(process, DXGMEM_HANDLE_TABLE, table); + } + process->test_handle_table[cmd.index] = NULL; + break; + default: + ret = STATUS_INVALID_PARAMETER; + pr_err("unknoen handle table command"); + break; + } + +cleanup: + dxgmutex_unlock(&process->process_mutex); + return ret; +} + +static int dxgk_escape(struct dxgprocess *process, void *__user inargs) +{ + struct d3dkmt_escape args; + int ret; + struct dxgadapter *adapter = NULL; + bool adapter_locked = false; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + adapter = dxgprocess_adapter_by_handle(process, args.adapter); + if (adapter == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + adapter_locked = true; + + if (args.type == D3DKMT_ESCAPE_DRT_TEST) { + struct d3dkmt_ht_desc drtcmd; + + if (args.priv_drv_data_size >= sizeof(drtcmd)) { + ret = dxg_copy_from_user(&drtcmd, + args.priv_drv_data, + sizeof(drtcmd)); + if (ret) + goto cleanup; + if (drtcmd.head.command == + D3DKMT_DRT_TEST_COMMAND_HANDLETABLE) { + dxgadapter_release_lock_shared(adapter); + adapter_locked = false; + ret = handle_table_escape(process, &args, + &drtcmd); + goto cleanup; + } + } + } + + args.adapter = adapter->host_handle; + ret = dxgvmb_send_escape(process, &adapter->channel, &args); + if (ret) + goto cleanup; + +cleanup: + + if (adapter_locked) + dxgadapter_release_lock_shared(adapter); + if (adapter) + dxgadapter_release_reference(adapter); + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_query_vidmem_info(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_queryvideomemoryinfo args; + int ret; + struct dxgadapter *adapter = NULL; + bool adapter_locked = false; + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.process != 0) { + pr_err("query vidmem info from another process "); + goto cleanup; + } + + adapter = dxgprocess_adapter_by_handle(process, args.adapter); + if (adapter == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + adapter_locked = true; + + args.adapter = adapter->host_handle; + ret = dxgvmb_send_query_vidmem_info(process, &adapter->channel, + &args, inargs); + +cleanup: + + if (adapter_locked) + dxgadapter_release_lock_shared(adapter); + if (adapter) + dxgadapter_release_reference(adapter); + if (ret) + pr_err("%s failed: %x", __func__, ret); + return ret; +} + +static int dxgk_get_device_state(struct dxgprocess *process, + void *__user inargs) +{ + int ret = 0; + struct d3dkmt_getdevicestate args; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + device = dxgprocess_device_by_handle(process, args.device); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgvmb_send_get_device_state(process, &adapter->channel, + &args, inargs); + +cleanup: + + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_reference(device); + if (ret) + pr_err("%s failed %x", __func__, ret); + + return ret; +} + +static int dxgsharedsyncobj_get_host_nt_handle(struct dxgsharedsyncobject + *syncobj, + struct dxgprocess *process, + d3dkmt_handle object_handle) +{ + int ret = 0; + + dxgmutex_lock(&syncobj->fd_mutex); + if (syncobj->host_shared_handle_nt_reference == 0) { + ret = dxgvmb_send_create_nt_shared_object(process, + object_handle, + &syncobj-> + host_shared_handle_nt); + if (ret) + goto cleanup; + TRACE_DEBUG(1, "Host_shared_handle_ht: %x", + syncobj->host_shared_handle_nt); + dxgsharedsyncobj_acquire_reference(syncobj); + } + syncobj->host_shared_handle_nt_reference++; +cleanup: + dxgmutex_unlock(&syncobj->fd_mutex); + return ret; +} + +static int dxgsharedresource_get_host_nt_handle(struct dxgsharedresource + *resource, + struct dxgprocess *process, + d3dkmt_handle object_handle) +{ + int ret = 0; + + dxgmutex_lock(&resource->fd_mutex); + if (resource->host_shared_handle_nt_reference == 0) { + ret = dxgvmb_send_create_nt_shared_object(process, + object_handle, + &resource-> + host_shared_handle_nt); + if (ret) + goto cleanup; + TRACE_DEBUG(1, "Resource host_shared_handle_ht: %x", + resource->host_shared_handle_nt); + dxgsharedresource_acquire_reference(resource); + } + resource->host_shared_handle_nt_reference++; +cleanup: + dxgmutex_unlock(&resource->fd_mutex); + return ret; +} + +enum dxg_sharedobject_type { + DXG_SHARED_SYNCOBJECT, + DXG_SHARED_RESOURCE +}; + +static int get_object_fd(enum dxg_sharedobject_type type, + void *object, int *fdout) +{ + struct file *file; + int fd; + + fd = get_unused_fd_flags(O_CLOEXEC); + if (fd < 0) { + pr_err("get_unused_fd_flags failed: %x", fd); + return STATUS_INTERNAL_ERROR; + } + + switch (type) { + case DXG_SHARED_SYNCOBJECT: + file = anon_inode_getfile("dxgsyncobj", + &dxg_syncobj_fops, object, 0); + break; + case DXG_SHARED_RESOURCE: + file = anon_inode_getfile("dxgresource", + &dxg_resource_fops, object, 0); + break; + default: + return STATUS_INVALID_PARAMETER; + }; + if (IS_ERR(file)) { + pr_err("anon_inode_getfile failed: %x", fd); + put_unused_fd(fd); + return STATUS_INTERNAL_ERROR; + } + + fd_install(fd, file); + *fdout = fd; + return 0; +} + +static int dxgk_share_objects(struct dxgprocess *process, void *__user inargs) +{ + struct d3dkmt_shareobjects args; + enum hmgrentry_type object_type; + struct dxgsyncobject *syncobj = NULL; + struct dxgresource *resource = NULL; + struct dxgsharedsyncobject *shared_syncobj = NULL; + struct dxgsharedresource *shared_resource = NULL; + d3dkmt_handle *handles = NULL; + int object_fd = 0; + void *obj = NULL; + uint handle_size; + int ret; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + if (args.object_count == 0 || args.object_count > 1) { + pr_err("invalid object count %d", args.object_count); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + handle_size = args.object_count * sizeof(d3dkmt_handle); + + handles = dxgmem_alloc(process, DXGMEM_TMP, handle_size); + if (handles == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + ret = dxg_copy_from_user(handles, args.objects, handle_size); + if (ret) + goto cleanup; + + TRACE_DEBUG(1, "Sharing handle: %x", handles[0]); + + hmgrtable_lock(&process->handle_table, DXGLOCK_SHARED); + object_type = hmgrtable_get_object_type(&process->handle_table, + handles[0]); + obj = hmgrtable_get_object(&process->handle_table, handles[0]); + if (obj == NULL) { + pr_err("invalid object handle %x", handles[0]); + ret = STATUS_INVALID_PARAMETER; + } else { + switch (object_type) { + case HMGRENTRY_TYPE_DXGSYNCOBJECT: + syncobj = obj; + if (syncobj->shared) { + dxgsyncobject_acquire_reference(syncobj); + shared_syncobj = syncobj->shared_owner; + } else { + pr_err("sync object is not shared"); + syncobj = NULL; + ret = STATUS_INVALID_PARAMETER; + } + break; + case HMGRENTRY_TYPE_DXGRESOURCE: + resource = obj; + if (resource->shared_owner) { + dxgresource_acquire_reference(resource); + shared_resource = resource->shared_owner; + } else { + resource = NULL; + pr_err("resource object is not shared"); + ret = STATUS_INVALID_PARAMETER; + } + break; + default: + pr_err("invalid object type %d", object_type); + ret = STATUS_INVALID_PARAMETER; + break; + } + } + hmgrtable_unlock(&process->handle_table, DXGLOCK_SHARED); + + if (ret) + goto cleanup; + + switch (object_type) { + case HMGRENTRY_TYPE_DXGSYNCOBJECT: + ret = get_object_fd(DXG_SHARED_SYNCOBJECT, shared_syncobj, + &object_fd); + if (!ret) + ret = + dxgsharedsyncobj_get_host_nt_handle(shared_syncobj, + process, + handles[0]); + break; + case HMGRENTRY_TYPE_DXGRESOURCE: + ret = get_object_fd(DXG_SHARED_RESOURCE, shared_resource, + &object_fd); + if (!ret) + ret = + dxgsharedresource_get_host_nt_handle + (shared_resource, process, handles[0]); + break; + default: + ret = STATUS_INVALID_PARAMETER; + break; + } + + if (ret) + goto cleanup; + + TRACE_DEBUG(1, "Object FD: %x", object_fd); + + { + winhandle tmp = (winhandle) object_fd; + + ret = dxg_copy_to_user(args.shared_handle, &tmp, + sizeof(winhandle)); + } + +cleanup: + if (ret) { + if (object_fd > 0) + put_unused_fd(object_fd); + } + + if (handles) + dxgmem_free(process, DXGMEM_TMP, handles); + + if (syncobj) + dxgsyncobject_release_reference(syncobj); + + if (resource) + dxgresource_release_reference(resource); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_invalidate_cache(struct dxgprocess *process, + void *__user inargs) +{ + pr_err("%s is not implemented", __func__); + return STATUS_NOT_IMPLEMENTED; +} + +static int dxgk_query_resource_info(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_queryresourceinfo args; + struct dxgdevice *device = NULL; + struct dxgsharedresource *shared_resource = NULL; + int ret; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + hmgrtable_lock(&dxgglobal->handle_table, DXGLOCK_SHARED); + shared_resource = hmgrtable_get_object_by_type(&dxgglobal->handle_table, + HMGRENTRY_TYPE_DXGSHAREDRESOURCE, + args.global_share); + if (shared_resource) { + if (!dxgsharedresource_acquire_reference(shared_resource)) + shared_resource = NULL; + } + hmgrtable_unlock(&dxgglobal->handle_table, DXGLOCK_SHARED); + + if (shared_resource == NULL) { + pr_err("Invalid shared resource handle: %x", + args.global_share); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + device = dxgprocess_device_by_handle(process, args.device); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + ret = dxgdevice_acquire_lock_shared(device); + if (ret) { + dxgdevice_release_reference(device); + device = NULL; + goto cleanup; + } + + ret = dxgsharedresource_seal(shared_resource); + if (ret) + goto cleanup; + + args.private_runtime_data_size = + shared_resource->runtime_private_data_size; + args.resource_priv_drv_data_size = + shared_resource->resource_private_data_size; + args.allocation_count = shared_resource->allocation_count; + args.total_priv_drv_data_size = + shared_resource->alloc_private_data_size; + + ret = dxg_copy_to_user(inargs, &args, sizeof(args)); + +cleanup: + + if (shared_resource) + dxgsharedresource_release_reference(shared_resource); + if (device) + dxgdevice_release_lock_shared(device); + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_query_resource_info_nt(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_queryresourceinfofromnthandle args; + int ret; + struct dxgdevice *device = NULL; + struct dxgsharedresource *shared_resource = NULL; + struct file *file = NULL; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + file = fget(args.nt_handle); + if (!file) { + pr_err("failed to get file from handle: %llx", + args.nt_handle); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + if (file->f_op != &dxg_resource_fops) { + pr_err("invalid fd: %llx", args.nt_handle); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + shared_resource = file->private_data; + if (shared_resource == NULL) { + pr_err("invalid private data: %llx", args.nt_handle); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + device = dxgprocess_device_by_handle(process, args.device); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + ret = dxgdevice_acquire_lock_shared(device); + if (ret) { + dxgdevice_release_reference(device); + device = NULL; + goto cleanup; + } + + ret = dxgsharedresource_seal(shared_resource); + if (ret) + goto cleanup; + + args.private_runtime_data_size = + shared_resource->runtime_private_data_size; + args.resource_priv_drv_data_size = + shared_resource->resource_private_data_size; + args.allocation_count = shared_resource->allocation_count; + args.total_priv_drv_data_size = + shared_resource->alloc_private_data_size; + + ret = dxg_copy_to_user(inargs, &args, sizeof(args)); + +cleanup: + + if (file) + fput(file); + if (device) + dxgdevice_release_lock_shared(device); + if (device) + dxgdevice_release_reference(device); + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +int assign_resource_handles(struct dxgprocess *process, + struct dxgsharedresource *shared_resource, + struct d3dkmt_openresourcefromnthandle *args, + d3dkmt_handle resource_handle, + struct dxgresource *resource, + struct dxgallocation **allocs, + d3dkmt_handle *handles) +{ + int ret = 0; + int i; + uint8_t *cur_priv_data; + struct d3dddi_openallocationinfo2 open_alloc_info = { }; + + TRACE_DEBUG(1, "%s", __func__); + + hmgrtable_lock(&process->handle_table, DXGLOCK_EXCL); + ret = hmgrtable_assign_handle(&process->handle_table, resource, + HMGRENTRY_TYPE_DXGRESOURCE, + resource_handle); + if (ret) + goto cleanup; + resource->handle = resource_handle; + resource->handle_valid = 1; + cur_priv_data = shared_resource->alloc_private_data; + for (i = 0; i < args->allocation_count; i++) { + ret = hmgrtable_assign_handle(&process->handle_table, allocs[i], + HMGRENTRY_TYPE_DXGALLOCATION, + handles[i]); + if (ret) + goto cleanup; + allocs[i]->alloc_handle = handles[i]; + allocs[i]->handle_valid = 1; + open_alloc_info.allocation = handles[i]; + if (shared_resource->alloc_private_data_sizes) + open_alloc_info.priv_drv_data_size = + shared_resource->alloc_private_data_sizes[i]; + else + open_alloc_info.priv_drv_data_size = 0; + + open_alloc_info.priv_drv_data = cur_priv_data; + cur_priv_data += open_alloc_info.priv_drv_data_size; + + ret = dxg_copy_to_user(&args->open_alloc_info[i], + &open_alloc_info, + sizeof(open_alloc_info)); + if (ret) + goto cleanup; + } +cleanup: + hmgrtable_unlock(&process->handle_table, DXGLOCK_EXCL); + if (ret) { + for (i = 0; i < args->allocation_count; i++) + dxgallocation_free_handle(allocs[i]); + dxgresource_free_handle(resource); + } + TRACE_DEBUG(1, "%s end %x", __func__, ret); + return ret; +} + +int open_resource(struct dxgprocess *process, + struct d3dkmt_openresourcefromnthandle *args, + bool nt_handle, __user d3dkmt_handle *res_out) +{ + int ret = 0; + int i; + d3dkmt_handle *alloc_handles = NULL; + int alloc_handles_size = sizeof(d3dkmt_handle) * args->allocation_count; + struct dxgsharedresource *shared_resource = NULL; + struct dxgresource *resource = NULL; + struct dxgallocation **allocs = NULL; + d3dkmt_handle global_share; + struct dxgdevice *device = NULL; + struct dxgadapter *adapter = NULL; + d3dkmt_handle resource_handle = 0; + struct file *file = NULL; + + TRACE_DEBUG(1, "Opening resource handle: %llx", args->nt_handle); + + if (nt_handle) { + file = fget(args->nt_handle); + if (!file) { + pr_err("failed to get file from handle: %llx", + args->nt_handle); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + if (file->f_op != &dxg_resource_fops) { + pr_err("invalid fd type: %llx", args->nt_handle); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + shared_resource = file->private_data; + if (shared_resource == NULL) { + pr_err("invalid private data: %llx", + args->nt_handle); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + if (!dxgsharedresource_acquire_reference(shared_resource)) + shared_resource = NULL; + else + global_share = shared_resource->host_shared_handle_nt; + } else { + hmgrtable_lock(&dxgglobal->handle_table, DXGLOCK_SHARED); + shared_resource = + hmgrtable_get_object_by_type(&dxgglobal->handle_table, + HMGRENTRY_TYPE_DXGSHAREDRESOURCE, + (d3dkmt_handle) args-> + nt_handle); + if (shared_resource) { + if (!dxgsharedresource_acquire_reference + (shared_resource)) + shared_resource = NULL; + else + global_share = + shared_resource->host_shared_handle; + } + hmgrtable_unlock(&dxgglobal->handle_table, DXGLOCK_SHARED); + } + + if (shared_resource == NULL) { + pr_err("Invalid shared resource handle: %x", + (d3dkmt_handle) args->nt_handle); + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + TRACE_DEBUG(1, "Shared resource: %p %x", shared_resource, + global_share); + + device = dxgprocess_device_by_handle(process, args->device); + if (device == NULL) { + ret = STATUS_INVALID_PARAMETER; + goto cleanup; + } + + ret = dxgdevice_acquire_lock_shared(device); + if (ret) { + dxgdevice_release_reference(device); + device = NULL; + goto cleanup; + } + + adapter = device->adapter; + ret = dxgadapter_acquire_lock_shared(adapter); + if (ret) { + adapter = NULL; + goto cleanup; + } + + ret = dxgsharedresource_seal(shared_resource); + if (ret) + goto cleanup; + + if (args->allocation_count != shared_resource->allocation_count || + args->private_runtime_data_size < + shared_resource->runtime_private_data_size || + args->resource_priv_drv_data_size < + shared_resource->resource_private_data_size || + args->total_priv_drv_data_size < + shared_resource->alloc_private_data_size) { + ret = STATUS_INVALID_PARAMETER; + pr_err("Invalid data sizes"); + goto cleanup; + } + + alloc_handles = dxgmem_alloc(process, DXGMEM_TMP, alloc_handles_size); + if (alloc_handles == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + allocs = dxgmem_alloc(process, DXGMEM_TMP, + sizeof(void *) * args->allocation_count); + if (allocs == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + + resource = dxgresource_create(device); + if (resource == NULL) { + ret = STATUS_NO_MEMORY; + goto cleanup; + } + dxgsharedresource_add_resource(shared_resource, resource); + + for (i = 0; i < args->allocation_count; i++) { + allocs[i] = dxgallocation_create(process); + if (allocs[i] == NULL) + goto cleanup; + dxgresource_add_alloc(resource, allocs[i]); + } + + ret = dxgvmb_send_open_resource(process, &adapter->channel, + device->handle, nt_handle, global_share, + args->allocation_count, + args->total_priv_drv_data_size, + &resource_handle, alloc_handles); + if (ret) { + pr_err("dxgvmb_send_open_resource failed"); + goto cleanup; + } + + if (shared_resource->runtime_private_data_size) { + ret = dxg_copy_to_user(args->private_runtime_data, + shared_resource->runtime_private_data, + shared_resource-> + runtime_private_data_size); + if (ret) + goto cleanup; + } + + if (shared_resource->resource_private_data_size) { + ret = dxg_copy_to_user(args->resource_priv_drv_data, + shared_resource->resource_private_data, + shared_resource-> + resource_private_data_size); + if (ret) + goto cleanup; + } + + if (shared_resource->alloc_private_data_size) { + ret = dxg_copy_to_user(args->total_priv_drv_data, + shared_resource->alloc_private_data, + shared_resource-> + alloc_private_data_size); + if (ret) + goto cleanup; + } + + ret = assign_resource_handles(process, shared_resource, args, + resource_handle, resource, allocs, + alloc_handles); + if (ret) + goto cleanup; + + ret = dxg_copy_to_user(res_out, &resource_handle, + sizeof(d3dkmt_handle)); + +cleanup: + + if (ret) { + if (resource_handle) { + struct d3dkmt_destroyallocation2 tmp = { }; + + tmp.flags.assume_not_in_use = 1; + tmp.device = args->device; + tmp.resource = resource_handle; + ret = dxgvmb_send_destroy_allocation(process, device, + &adapter->channel, + &tmp, NULL); + } + if (resource) + dxgresource_destroy(resource); + } + + if (file) + fput(file); + if (allocs) + dxgmem_free(process, DXGMEM_TMP, allocs); + if (shared_resource) + dxgsharedresource_release_reference(shared_resource); + if (alloc_handles) + dxgmem_free(process, DXGMEM_TMP, alloc_handles); + if (adapter) + dxgadapter_release_lock_shared(adapter); + if (device) + dxgdevice_release_lock_shared(device); + if (device) + dxgdevice_release_reference(device); + + return ret; +} + +static int dxgk_open_resource(struct dxgprocess *process, void *__user inargs) +{ + struct d3dkmt_openresource args; + struct d3dkmt_openresourcefromnthandle args_nt = { }; + int ret; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + args_nt.device = args.device; + args_nt.nt_handle = (winhandle) args.global_share; + args_nt.allocation_count = args.allocation_count; + args_nt.open_alloc_info = args.open_alloc_info; + args_nt.private_runtime_data_size = args.private_runtime_data_size; + args_nt.private_runtime_data = args.private_runtime_data; + args_nt.resource_priv_drv_data_size = args.resource_priv_drv_data_size; + args_nt.resource_priv_drv_data = args.resource_priv_drv_data; + args_nt.total_priv_drv_data_size = args.total_priv_drv_data_size; + args_nt.total_priv_drv_data = args.total_priv_drv_data; + + ret = open_resource(process, &args_nt, false, + &((struct d3dkmt_openresource *)inargs)->resource); + +cleanup: + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static ntstatus dxgk_open_resource_nt(struct dxgprocess *process, + void *__user inargs) +{ + struct d3dkmt_openresourcefromnthandle args; + int ret; + + TRACE_FUNC_ENTER(__func__); + + ret = dxg_copy_from_user(&args, inargs, sizeof(args)); + if (ret) + goto cleanup; + + ret = open_resource(process, &args, true, + &((struct d3dkmt_openresourcefromnthandle *) + inargs)->resource); + +cleanup: + + TRACE_FUNC_EXIT(__func__, ret); + return ret; +} + +static int dxgk_render(struct dxgprocess *process, void *__user inargs) +{ + pr_err("%s is not implemented", __func__); + return STATUS_NOT_IMPLEMENTED; +} + +static int dxgk_create_context(struct dxgprocess *process, void *__user inargs) +{ + pr_err("%s is not implemented", __func__); + return STATUS_NOT_IMPLEMENTED; +} + +static int dxgk_get_shared_resource_adapter_luid(struct dxgprocess *process, + void *__user inargs) +{ + pr_err("shared_resource_adapter_luid is not implemented"); + return STATUS_NOT_IMPLEMENTED; +} + +/* + * IOCTL processing + */ +static int dxgk_ioctl(struct file *f, unsigned int p1, unsigned long p2) +{ + struct dxgthreadinfo *thread; + int code = _IOC_NR(p1); + ntstatus status; + struct dxgprocess *process; + + if (code < 1 || code > LX_IO_MAX) { + pr_err("bad ioctl %x %x %x %x", + code, _IOC_TYPE(p1), _IOC_SIZE(p1), _IOC_DIR(p1)); + return STATUS_INVALID_PARAMETER; + } + if (ioctls[code].ioctl_callback == NULL) { + pr_err("ioctl callback is NULL %x", code); + return STATUS_INTERNAL_ERROR; + } + if (ioctls[code].ioctl != p1) { + pr_err("ioctl mismatch. Code: %x User: %x Kernel: %x", + code, p1, ioctls[code].ioctl); + return STATUS_INTERNAL_ERROR; + } + process = (struct dxgprocess *)f->private_data; + if (process->tgid != current->tgid) { + pr_err("Call from a wrong process: %d %d", + process->tgid, current->tgid); + return STATUS_INVALID_PARAMETER; + } + thread = dxglockorder_get_thread(); + status = ioctls[code].ioctl_callback(process, (void *__user)p2); + dxglockorder_check_empty(thread); + dxglockorder_put_thread(thread); + return status; +} + +long dxgk_compat_ioctl(struct file *f, unsigned int p1, unsigned long p2) +{ + TRACE_DEBUG(2, "compat ioctl %x", p1); + return dxgk_ioctl(f, p1, p2); +} + +long dxgk_unlocked_ioctl(struct file *f, unsigned int p1, unsigned long p2) +{ + TRACE_DEBUG(2, "unlocked ioctl %x Code:%d", p1, _IOC_NR(p1)); + return dxgk_ioctl(f, p1, p2); +} + +#define SET_IOCTL(callback, v) \ + ioctls[_IOC_NR(v)].ioctl_callback = callback; \ + ioctls[_IOC_NR(v)].ioctl = v + +void ioctl_desc_init(void) +{ + memset(ioctls, 0, sizeof(ioctls)); + SET_IOCTL(/*0x1 */ dxgk_open_adapter_from_luid, + LX_DXOPENADAPTERFROMLUID); + SET_IOCTL(/*0x2 */ dxgk_create_device, + LX_DXCREATEDEVICE); + SET_IOCTL(/*0x3 */ dxgk_create_context, + LX_DXCREATECONTEXT); + SET_IOCTL(/*0x4 */ dxgk_create_context_virtual, + LX_DXCREATECONTEXTVIRTUAL); + SET_IOCTL(/*0x5 */ dxgk_destroy_context, + LX_DXDESTROYCONTEXT); + SET_IOCTL(/*0x6 */ dxgk_create_allocation, + LX_DXCREATEALLOCATION); + SET_IOCTL(/*0x7 */ dxgk_create_paging_queue, + LX_DXCREATEPAGINGQUEUE); + SET_IOCTL(/*0x8 */ dxgk_reserve_gpu_va, + LX_DXRESERVEGPUVIRTUALADDRESS); + SET_IOCTL(/*0x9 */ dxgk_query_adapter_info, + LX_DXQUERYADAPTERINFO); + SET_IOCTL(/*0xa */ dxgk_query_vidmem_info, + LX_DXQUERYVIDEOMEMORYINFO); + SET_IOCTL(/*0xb */ dxgk_make_resident, + LX_DXMAKERESIDENT); + SET_IOCTL(/*0xc */ dxgk_map_gpu_va, + LX_DXMAPGPUVIRTUALADDRESS); + SET_IOCTL(/*0xd */ dxgk_escape, + LX_DXESCAPE); + SET_IOCTL(/*0xe */ dxgk_get_device_state, + LX_DXGETDEVICESTATE); + SET_IOCTL(/*0xf */ dxgk_submit_command, + LX_DXSUBMITCOMMAND); + SET_IOCTL(/*0x10 */ dxgk_create_sync_object, + LX_DXCREATESYNCHRONIZATIONOBJECT); + SET_IOCTL(/*0x11 */ dxgk_signal_sync_object, + LX_DXSIGNALSYNCHRONIZATIONOBJECT); + SET_IOCTL(/*0x12 */ dxgk_wait_sync_object, + LX_DXWAITFORSYNCHRONIZATIONOBJECT); + SET_IOCTL(/*0x13 */ dxgk_destroy_allocation, + LX_DXDESTROYALLOCATION2); + SET_IOCTL(/*0x14 */ dxgk_enum_adapters, + LX_DXENUMADAPTERS2); + SET_IOCTL(/*0x15 */ dxgk_close_adapter, + LX_DXCLOSEADAPTER); + SET_IOCTL(/*0x16 */ dxgk_change_vidmem_reservation, + LX_DXCHANGEVIDEOMEMORYRESERVATION); + SET_IOCTL(/*0x17 */ dxgk_create_hwcontext, + LX_DXCREATEHWCONTEXT); + SET_IOCTL(/*0x18 */ dxgk_create_hwqueue, + LX_DXCREATEHWQUEUE); + SET_IOCTL(/*0x19 */ dxgk_destroy_device, + LX_DXDESTROYDEVICE); + SET_IOCTL(/*0x1a */ dxgk_destroy_hwcontext, + LX_DXDESTROYHWCONTEXT); + SET_IOCTL(/*0x1b */ dxgk_destroy_hwqueue, + LX_DXDESTROYHWQUEUE); + SET_IOCTL(/*0x1c */ dxgk_destroy_paging_queue, + LX_DXDESTROYPAGINGQUEUE); + SET_IOCTL(/*0x1d */ dxgk_destroy_sync_object, + LX_DXDESTROYSYNCHRONIZATIONOBJECT); + SET_IOCTL(/*0x1e */ dxgk_evict, + LX_DXEVICT); + SET_IOCTL(/*0x1f */ dxgk_flush_heap_transitions, + LX_DXFLUSHHEAPTRANSITIONS); + SET_IOCTL(/*0x20 */ dxgk_free_gpu_va, + LX_DXFREEGPUVIRTUALADDRESS); + SET_IOCTL(/*0x21 */ dxgk_get_context_process_scheduling_priority, + LX_DXGETCONTEXTINPROCESSSCHEDULINGPRIORITY); + SET_IOCTL(/*0x22 */ dxgk_get_context_scheduling_priority, + LX_DXGETCONTEXTSCHEDULINGPRIORITY); + SET_IOCTL(/*0x23 */ dxgk_get_shared_resource_adapter_luid, + LX_DXGETSHAREDRESOURCEADAPTERLUID); + SET_IOCTL(/*0x24 */ dxgk_invalidate_cache, + LX_DXINVALIDATECACHE); + SET_IOCTL(/*0x25 */ dxgk_lock2, + LX_DXLOCK2); + SET_IOCTL(/*0x26 */ dxgk_mark_device_as_error, + LX_DXMARKDEVICEASERROR); + SET_IOCTL(/*0x27 */ dxgk_offer_allocations, + LX_DXOFFERALLOCATIONS); + SET_IOCTL(/*0x28 */ dxgk_open_resource, + LX_DXOPENRESOURCE); + SET_IOCTL(/*0x29 */ dxgk_open_sync_object, + LX_DXOPENSYNCHRONIZATIONOBJECT); + SET_IOCTL(/*0x2a */ dxgk_query_alloc_residency, + LX_DXQUERYALLOCATIONRESIDENCY); + SET_IOCTL(/*0x2b */ dxgk_query_resource_info, + LX_DXQUERYRESOURCEINFO); + SET_IOCTL(/*0x2c */ dxgk_reclaim_allocations, + LX_DXRECLAIMALLOCATIONS2); + SET_IOCTL(/*0x2d */ dxgk_render, + LX_DXRENDER); + SET_IOCTL(/*0x2e */ dxgk_set_allocation_priority, + LX_DXSETALLOCATIONPRIORITY); + SET_IOCTL(/*0x2f */ dxgk_set_context_process_scheduling_priority, + LX_DXSETCONTEXTINPROCESSSCHEDULINGPRIORITY); + SET_IOCTL(/*0x30 */ dxgk_set_context_scheduling_priority, + LX_DXSETCONTEXTSCHEDULINGPRIORITY); + SET_IOCTL(/*0x31 */ dxgk_signal_sync_object_cpu, + LX_DXSIGNALSYNCHRONIZATIONOBJECTFROMCPU); + SET_IOCTL(/*0x32 */ dxgk_signal_sync_object_gpu, + LX_DXSIGNALSYNCHRONIZATIONOBJECTFROMGPU); + SET_IOCTL(/*0x33 */ dxgk_signal_sync_object_gpu2, + LX_DXSIGNALSYNCHRONIZATIONOBJECTFROMGPU2); + SET_IOCTL(/*0x34 */ dxgk_submit_command_to_hwqueue, + LX_DXSUBMITCOMMANDTOHWQUEUE); + SET_IOCTL(/*0x35 */ dxgk_submit_wait_to_hwqueue, + LX_DXSUBMITWAITFORSYNCOBJECTSTOHWQUEUE); + SET_IOCTL(/*0x36 */ dxgk_submit_signal_to_hwqueue, + LX_DXSUBMITSIGNALSYNCOBJECTSTOHWQUEUE); + SET_IOCTL(/*0x37 */ dxgk_unlock2, + LX_DXUNLOCK2); + SET_IOCTL(/*0x38 */ dxgk_update_alloc_property, + LX_DXUPDATEALLOCPROPERTY); + SET_IOCTL(/*0x39 */ dxgk_update_gpu_va, + LX_DXUPDATEGPUVIRTUALADDRESS); + SET_IOCTL(/*0x3a */ dxgk_wait_sync_object_cpu, + LX_DXWAITFORSYNCHRONIZATIONOBJECTFROMCPU); + SET_IOCTL(/*0x3b */ dxgk_wait_sync_object_gpu, + LX_DXWAITFORSYNCHRONIZATIONOBJECTFROMGPU); + SET_IOCTL(/*0x3c */ dxgk_get_allocation_priority, + LX_DXGETALLOCATIONPRIORITY); + SET_IOCTL(/*0x3d */ dxgk_query_clock_calibration, + LX_DXQUERYCLOCKCALIBRATION); + SET_IOCTL(/*0x3e */ dxgk_enum_adapters3, + LX_DXENUMADAPTERS3); + SET_IOCTL(/*0x3f */ dxgk_share_objects, + LX_DXSHAREOBJECTS); + SET_IOCTL(/*0x40 */ dxgk_open_sync_object_nt, + LX_DXOPENSYNCOBJECTFROMNTHANDLE2); + SET_IOCTL(/*0x41 */ dxgk_query_resource_info_nt, + LX_DXQUERYRESOURCEINFOFROMNTHANDLE); + SET_IOCTL(/*0x42 */ dxgk_open_resource_nt, + LX_DXOPENRESOURCEFROMNTHANDLE); +} diff --git a/drivers/gpu/dxgkrnl/misc.c b/drivers/gpu/dxgkrnl/misc.c new file mode 100644 index 000000000000..6f88e0108612 --- /dev/null +++ b/drivers/gpu/dxgkrnl/misc.c @@ -0,0 +1,280 @@ +// SPDX-License-Identifier: GPL-2.0 + +/* + * Copyright (c) 2019, Microsoft Corporation. + * + * Author: + * Iouri Tarassov iourit@microsoft.com + * + * Dxgkrnl Graphics Driver + * Helper functions + * + */ + +#include <linux/slab.h> +#include <linux/vmalloc.h> +#include <linux/uaccess.h> + +#include "dxgkrnl.h" +#include "misc.h" + +static atomic_t dxg_memory[DXGMEM_LAST]; + +winwchar *wcsncpy(winwchar *dest, const winwchar *src, size_t n) +{ + int i; + + for (i = 0; i < n; i++) { + dest[i] = src[i]; + if (src[i] == 0) { + i++; + break; + } + } + dest[i - 1] = 0; + return dest; +} + +int dxg_copy_from_user(void *to, const void __user *from, unsigned long len) +{ + int ret = copy_from_user(to, from, len); + + if (ret) { + pr_err("copy_from_user failed: %p %p %lx\n", + to, from, len); + ret = STATUS_INVALID_PARAMETER; + } + return ret; +} + +int dxg_copy_to_user(void *to, const void __user *from, unsigned long len) +{ + int ret = copy_to_user(to, from, len); + + if (ret) { + pr_err("copy_to_user failed: %p %p %lx\n", to, from, len); + ret = STATUS_INVALID_PARAMETER; + } + return ret; +} + +void *dxgmem_alloc(struct dxgprocess *process, enum dxgk_memory_tag tag, + size_t size) +{ + void *address = vzalloc(size); + + if (address) { + if (process) + atomic_inc(&process->dxg_memory[tag]); + else + atomic_inc(&dxg_memory[tag]); + } + return address; +} + +void dxgmem_free(struct dxgprocess *process, enum dxgk_memory_tag tag, + void *address) +{ + vfree(address); + if (process) { + if (atomic_read(&process->dxg_memory[tag]) == 0) { + pr_err("%s process error: %d\n", + __func__, tag); + dump_stack(); + } else { + atomic_dec(&process->dxg_memory[tag]); + } + } else { + if (atomic_read(&dxg_memory[tag]) == 0) { + pr_err("%s error: %d\n", __func__, tag); + dump_stack(); + } else { + atomic_dec(&dxg_memory[tag]); + } + } +} + +void *dxgmem_kalloc(enum dxgk_memory_tag tag, size_t size, gfp_t flags) +{ + void *address = kzalloc(size, flags); + + if (address) + atomic_inc(&dxg_memory[tag]); + return address; +} + +void dxgmem_kfree(enum dxgk_memory_tag tag, void *address) +{ + kfree(address); + if (atomic_read(&dxg_memory[tag]) == 0) + pr_err("dxgmem_free error: %d\n", tag); + atomic_dec(&dxg_memory[tag]); +} + +void dxgmem_check(struct dxgprocess *process, enum dxgk_memory_tag ignore_tag) +{ + int i; + int v; + + for (i = 0; i < DXGMEM_LAST; i++) { + if (process) + v = atomic_read(&process->dxg_memory[i]); + else + v = atomic_read(&dxg_memory[i]); + if (i != ignore_tag && v != 0) + pr_err("memory leak: %2d %3d\n", i, v); + } +} + +void dxgmutex_init(struct dxgmutex *m, enum dxgk_lockorder order) +{ + mutex_init(&m->mutex); + m->lock_order = order; +} + +void dxgmutex_lock(struct dxgmutex *m) +{ + dxglockorder_acquire(m->lock_order); + mutex_lock(&m->mutex); +} + +void dxgmutex_unlock(struct dxgmutex *m) +{ + mutex_unlock(&m->mutex); + dxglockorder_release(m->lock_order); +} + +void dxglockorder_acquire(enum dxgk_lockorder order) +{ + struct dxgthreadinfo *info = dxglockorder_get_thread(); + int index = info->current_lock_index; + + TRACE_LOCK_ORDER(1, "%s %p %d %d", + __func__, info->thread, index, order); + if (index == DXGK_MAX_LOCK_DEPTH) { + pr_err("Lock depth exceeded"); + dxg_panic(); + } + if (index != 0) { + struct dxgk_lockinfo *lock_info = &info->lock_info[index - 1]; + + if (lock_info->lock_order <= order) { + pr_err("Invalid lock order: %d %d %d", index, + lock_info->lock_order, order); + dxg_panic(); + } + } + info->lock_info[index].lock_order = order; + info->current_lock_index++; + dxglockorder_put_thread(info); +} + +void dxglockorder_release(enum dxgk_lockorder order) +{ + struct dxgthreadinfo *info = dxglockorder_get_thread(); + int index; + int i; + + info->current_lock_index--; + index = info->current_lock_index; + TRACE_LOCK_ORDER(1, "dxglockorder release %p %d %d", + info->thread, index, order); + if (index < 0) { + pr_err("Lock depth underflow"); + dxg_panic(); + } + for (i = index; i >= 0; i--) { + if (info->lock_info[i].lock_order == order) { + memmove(&info->lock_info[i], &info->lock_info[index], + (index - i) * sizeof(info->lock_info[0])); + break; + } + } + if (i < 0) { + pr_err("Failed to find lock order to release"); + dxg_panic(); + } + memset(&info->lock_info[index], 0, sizeof(info->lock_info[0])); + dxglockorder_put_thread(info); +} + +struct dxgthreadinfo *dxglockorder_get_thread(void) +{ + struct dxgthreadinfo *info = NULL; + struct dxgthreadinfo *entry = NULL; + struct task_struct *thread = current; + + mutex_lock(&dxgglobal->thread_info_mutex); + list_for_each_entry(entry, &dxgglobal->thread_info_list_head, + thread_info_list_entry) { + if (entry->thread == thread) { + info = entry; + TRACE_LOCK_ORDER(1, "dxglockorder found thread %p %d", + thread, info->refcount + 1); + break; + } + } + if (info == NULL) { + info = dxgmem_kalloc(DXGMEM_THREADINFO, + sizeof(struct dxgthreadinfo), GFP_ATOMIC); + if (info) { + TRACE_LOCK_ORDER(1, "dxglockorder new thread %p", + thread); + info->thread = thread; + list_add(&info->thread_info_list_entry, + &dxgglobal->thread_info_list_head); + } + } + if (info) + info->refcount++; + mutex_unlock(&dxgglobal->thread_info_mutex); + return info; +} + +void dxglockorder_put_thread(struct dxgthreadinfo *info) +{ + if (info) { + TRACE_LOCK_ORDER(1, "dxglockorder remove thread %p %d", + info->thread, info->refcount); + if (info->refcount <= 0) { + pr_err("Invalid refcount for thread info: %p %d", + info, info->refcount); + dxg_panic(); + return; + } + info->refcount--; + if (info->refcount == 0) { + TRACE_LOCK_ORDER(1, "dxglockorder remove thread %p", + info->thread); + if (info->current_lock_index != 0) { + pr_err("A lock is not released: %d %d", + info->current_lock_index, + info->lock_info[ + info->current_lock_index].lock_order); + dxg_panic(); + } + + if (!info->lock_held) { + mutex_lock(&dxgglobal->thread_info_mutex); + list_del(&info->thread_info_list_entry); + mutex_unlock(&dxgglobal->thread_info_mutex); + } + + dxgmem_kfree(DXGMEM_THREADINFO, info); + } + } +} + +void dxglockorder_check_empty(struct dxgthreadinfo *info) +{ + if (info->refcount != 1) { + pr_err("A lock is not released"); + dxg_panic(); + } +} + +void dxg_panic(void) +{ + dump_stack(); + BUG_ON(true); +} diff --git a/drivers/gpu/dxgkrnl/misc.h b/drivers/gpu/dxgkrnl/misc.h new file mode 100644 index 000000000000..b5130cce767f --- /dev/null +++ b/drivers/gpu/dxgkrnl/misc.h @@ -0,0 +1,288 @@ +// SPDX-License-Identifier: GPL-2.0 + +/* + * Copyright (c) 2019, Microsoft Corporation. + * + * Author: + * Iouri Tarassov iourit@microsoft.com + * + * Dxgkrnl Graphics Driver + * Misc definitions + * + */ + +#ifndef _MISC_H_ +#define _MISC_H_ + +struct dxgprocess; + +enum dxgk_memory_tag { + DXGMEM_GLOBAL = 0, + DXGMEM_ADAPTER = 1, + DXGMEM_THREADINFO = 2, + DXGMEM_PROCESS = 3, + DXGMEM_VMBUS = 4, + DXGMEM_DEVICE = 5, + DXGMEM_RESOURCE = 6, + DXGMEM_CONTEXT = 7, + DXGMEM_PQUEUE = 8, + DXGMEM_SYNCOBJ = 9, + DXGMEM_PROCESS_ADAPTER = 10, + DXGMEM_HWQUEUE = 11, + DXGMEM_HANDLE_TABLE = 12, + DXGMEM_TMP = 13, + DXGMEM_ALLOCATION = 14, + DXGMEM_EVENT = 15, + DXGMEM_HOSTEVENT = 16, + DXGMEM_SHAREDSYNCOBJ = 17, + DXGMEM_SHAREDRESOURCE = 18, + DXGMEM_ALLOCPRIVATE = 19, + DXGMEM_RUNTIMEPRIVATE = 20, + DXGMEM_RESOURCEPRIVATE = 21, + DXGMEM_LAST +}; + +void dxgmem_check(struct dxgprocess *process, enum dxgk_memory_tag ignore_tag); +void *dxgmem_alloc(struct dxgprocess *process, enum dxgk_memory_tag tag, + size_t size); +void dxgmem_free(struct dxgprocess *process, enum dxgk_memory_tag tag, + void *address); +void *dxgmem_kalloc(enum dxgk_memory_tag tag, size_t size, gfp_t flags); +void dxgmem_kfree(enum dxgk_memory_tag tag, void *address); + +#define DXGK_MAX_LOCK_DEPTH 64 +#define W_MAX_PATH 260 + +#define d3dkmt_handle u32 +#define d3dgpu_virtual_address u64 +#define winwchar u16 +#define winhandle u64 +#define ntstatus int +#define winbool u32 +#define d3dgpu_size_t u64 + +struct winluid { + uint a; + uint b; +}; + +/* + * Synchronization lock hierarchy. + * dxgadapter->adapter_process_list_lock + * + * The higher enum value, the higher is the lock order. + * When a lower lock ois held, the higher lock should not be acquired. + */ +enum dxgk_lockorder { + DXGLOCK_INVALID = 0, + DXGLOCK_PROCESSADAPTERDEVICELIST, /* device_list_mutex */ + DXGLOCK_GLOBAL_HOSTEVENTLIST, /* host_event_list_mutex */ + DXGLOCK_GLOBAL_CHANNEL, /* channel_lock */ + DXGLOCK_FDMUTEX, /* fd_mutex */ + DXGLOCK_PROCESSLIST, /* plistmutex */ + DXGLOCK_HANDLETABLE, /* table_lock */ + DXGLOCK_DEVICE_CONTEXTLIST, /* context_list_lock */ + DXGLOCK_DEVICE_ALLOCLIST, /* alloc_list_lock */ + DXGLOCK_RESOURCE, /* resource_mutex */ + DXGLOCK_SHAREDRESOURCELIST, /* shared_resource_list_lock */ + DXGLOCK_ADAPTER, /* core_lock */ + DXGLOCK_DEVICE, /* device_lock */ + DXGLOCK_PROCESSMUTEX, /* process->process_mutex */ + DXGLOCK_PROCESSADAPTER, /* process_adapter_mutex */ + DXGLOCK_GLOBAL_ADAPTERLIST, /* adapter_list_lock */ + DXGLOCK_GLOBAL_DEVICE, /* device_mutex */ +}; + +struct dxgk_lockinfo { + enum dxgk_lockorder lock_order; +}; + +struct dxgthreadinfo { + struct list_head thread_info_list_entry; + struct task_struct *thread; + int refcount; + int current_lock_index; + struct dxgk_lockinfo lock_info[DXGK_MAX_LOCK_DEPTH]; + bool lock_held; +}; + +void dxglockorder_acquire(enum dxgk_lockorder order); +void dxglockorder_release(enum dxgk_lockorder order); +void dxglockorder_check_empty(struct dxgthreadinfo *info); +struct dxgthreadinfo *dxglockorder_get_thread(void); +void dxglockorder_put_thread(struct dxgthreadinfo *info); +void dxg_panic(void); + +struct dxgmutex { + struct mutex mutex; + enum dxgk_lockorder lock_order; +}; +void dxgmutex_init(struct dxgmutex *m, enum dxgk_lockorder order); +void dxgmutex_lock(struct dxgmutex *m); +void dxgmutex_unlock(struct dxgmutex *m); + +winwchar *wcsncpy(winwchar *dest, const winwchar *src, size_t n); + +enum dxglockstate { + DXGLOCK_SHARED, + DXGLOCK_EXCL +}; + +#define STATUS_SUCCESS ((ntstatus)0) +#define STATUS_GRAPHICS_DRIVER_MISMATCH ((ntstatus)0xC01E0009L) +#define STATUS_OBJECT_NAME_INVALID ((ntstatus)0xC0000033L) +#define STATUS_OBJECT_PATH_INVALID ((ntstatus)0xC0000039L) +#define STATUS_DEVICE_REMOVED ((ntstatus)0xC00002B6L) +#define STATUS_DISK_FULL ((ntstatus)0xC000007FL) +#define STATUS_GRAPHICS_GPU_EXCEPTION_ON_DEVICE ((ntstatus)0xC01E0200L) +#define STATUS_GRAPHICS_ALLOCATION_CONTENT_LOST ((ntstatus)0xC01E0116L) +#define STATUS_GRAPHICS_ALLOCATION_CLOSED ((ntstatus)0xC01E0112L) +#define STATUS_GRAPHICS_INVALID_ALLOCATION_INSTANCE ((ntstatus)0xC01E0113L) +#define STATUS_GRAPHICS_INVALID_ALLOCATION_HANDLE ((ntstatus)0xC01E0114L) +#define STATUS_ILLEGAL_CHARACTER ((ntstatus)0xC0000161L) +#define STATUS_INVALID_HANDLE ((ntstatus)0xC0000008L) +#define STATUS_ILLEGAL_INSTRUCTION ((ntstatus)0xC000001DL) +#define STATUS_INVALID_PARAMETER_1 ((ntstatus)0xC00000EFL) +#define STATUS_INVALID_PARAMETER_2 ((ntstatus)0xC00000F0L) +#define STATUS_INVALID_PARAMETER_3 ((ntstatus)0xC00000F1L) +#define STATUS_INVALID_PARAMETER_4 ((ntstatus)0xC00000F2L) +#define STATUS_INVALID_PARAMETER_5 ((ntstatus)0xC00000F3L) +#define STATUS_INVALID_PARAMETER_6 ((ntstatus)0xC00000F4L) +#define STATUS_INVALID_PARAMETER_7 ((ntstatus)0xC00000F5L) +#define STATUS_INVALID_PARAMETER_8 ((ntstatus)0xC00000F6L) +#define STATUS_INVALID_PARAMETER_9 ((ntstatus)0xC00000F7L) +#define STATUS_INVALID_PARAMETER_10 ((ntstatus)0xC00000F8L) +#define STATUS_INVALID_PARAMETER_11 ((ntstatus)0xC00000F9L) +#define STATUS_INVALID_PARAMETER_12 ((ntstatus)0xC00000FAL) +#define STATUS_IN_PAGE_ERROR ((ntstatus)0xC0000006L) +#define STATUS_NOT_IMPLEMENTED ((ntstatus)0xC0000002L) +#define STATUS_PENDING ((ntstatus)0x00000103L) +#define STATUS_ACCESS_DENIED ((ntstatus)0xC0000022L) +#define STATUS_BUFFER_TOO_SMALL ((ntstatus)0xC0000023L) +#define STATUS_OBJECT_PATH_SYNTAX_BAD ((ntstatus)0xC000003BL) +#define STATUS_OBJECT_TYPE_MISMATCH ((ntstatus)0xC0000024L) +#define STATUS_GRAPHICS_ALLOCATION_BUSY ((ntstatus)0xC01E0102L) +#define STATUS_GRAPHICS_WRONG_ALLOCATION_DEVICE ((ntstatus)0xC01E0115L) +#define STATUS_PRIVILEGED_INSTRUCTION ((ntstatus)0xC0000096L) +#define STATUS_SHARING_VIOLATION ((ntstatus)0xC0000043L) +#define STATUS_BUFFER_OVERFLOW ((ntstatus)0x80000005L) +#define STATUS_MEDIA_WRITE_PROTECTED ((ntstatus)0xC00000A2L) +#define STATUS_INTEGER_OVERFLOW ((ntstatus)0xC0000095L) +#define STATUS_PRIVILEGE_NOT_HELD ((ntstatus)0xC0000061L) +#define STATUS_NOT_SUPPORTED ((ntstatus)0xC00000BBL) +#define STATUS_HOST_UNREACHABLE ((ntstatus)0xC000023DL) +#define STATUS_NETWORK_UNREACHABLE ((ntstatus)0xC000023CL) +#define STATUS_CONNECTION_REFUSED ((ntstatus)0xC0000236L) +#define STATUS_CONNECTION_REFUSED ((ntstatus)0xC0000236L) +#define STATUS_TIMEOUT ((ntstatus)0x00000102L) +#define STATUS_WRONG_VOLUME ((ntstatus)0xC0000012L) +#define STATUS_IO_TIMEOUT ((ntstatus)0xC00000B5L) +#define STATUS_RETRY ((ntstatus)0xC000022DL) +#define STATUS_CANCELLED ((ntstatus)0xC0000120L) +#define STATUS_CONNECTION_DISCONNECTED ((ntstatus)0xC000020CL) +#define STATUS_CONNECTION_RESET ((ntstatus)0xC000020DL) +#define STATUS_CONNECTION_ABORTED ((ntstatus)0xC0000241L) +#define STATUS_INVALID_PARAMETER ((ntstatus)0xC000000DL) +#define STATUS_INVALID_DEVICE_REQUEST ((ntstatus)0xC0000010L) +#define STATUS_OBJECT_NAME_NOT_FOUND ((ntstatus)0xC0000034L) +#define STATUS_OBJECT_PATH_NOT_FOUND ((ntstatus)0xC000003AL) +#define STATUS_NOT_FOUND ((ntstatus)0xC0000225L) +#define STATUS_DELETE_PENDING ((ntstatus)0xC0000056L) +#define STATUS_BAD_NETWORK_NAME ((ntstatus)0xC00000CCL) +#define STATUS_CANNOT_DELETE ((ntstatus)0xC0000121L) +#define STATUS_INTERNAL_ERROR ((ntstatus)0xC00000E5L) +#define STATUS_OBJECTID_EXISTS ((ntstatus)0xC000022BL) +#define STATUS_DUPLICATE_OBJECTID ((ntstatus)0xC000022AL) +#define STATUS_ADDRESS_ALREADY_EXISTS ((ntstatus)0xC000020AL) +#define STATUS_ACCESS_VIOLATION ((ntstatus)0xC0000005L) +#define STATUS_INSUFFICIENT_RESOURCES ((ntstatus)0xC000009AL) +#define STATUS_NO_MEMORY ((ntstatus)0xC0000017L) +#define STATUS_COMMITMENT_LIMIT ((ntstatus)0xC000012DL) +#define STATUS_GRAPHICS_NO_VIDEO_MEMORY ((ntstatus)0xC01E0100L) +#define STATUS_OBJECTID_NOT_FOUND ((ntstatus)0xC00002F0L) +#define STATUS_DIRECTORY_NOT_EMPTY ((ntstatus)0xC0000101L) +#define STATUS_OBJECT_NAME_EXISTS ((ntstatus)0x40000000L) +#define STATUS_OBJECT_NAME_COLLISION ((ntstatus)0xC0000035L) +#define STATUS_UNSUCCESSFUL ((ntstatus)0xC0000001L) +#define STATUS_NOT_IMPLEMENTED ((ntstatus)0xC0000002L) +#define NT_SUCCESS(status) ((int)status >= 0) + +#define DXGKRNL_ASSERT(exp) +#define UNUSED(x) (void)(x) + +#undef pr_fmt +#define pr_fmt(fmt) "dxgk:err: " fmt +#define pr_fmt2(fmt) "dxgk:err: " fmt + +#define DXGKDEBUG 1 +/* #define USEPRINTK 1 */ + +#ifndef DXGKDEBUG +#define TRACE_DEBUG(...) +#define TRACE_DEFINE(...) +#define TRACE_FUNC_ENTER(...) +#define TRACE_FUNC_EXIT(...) +#else +#ifdef USEPRINTK +#define TRACE_DEBUG(level, fmt, ...)\ + printk(KERN_DEBUG pr_fmt2(fmt), ##__VA_ARGS__); + +#define TRACE_DEBUG2(level, offset, fmt, ...) \ +do { \ + if (offset == 0) \ + printk(KERN_DEBUG pr_fmt(fmt), ##__VA_ARGS__); \ + else \ + printk(KERN_DEBUG pr_fmt2(fmt), ##__VA_ARGS__); \ +} while (false) + +#define TRACE_FUNC_ENTER(msg) \ + printk(KERN_DEBUG "dxgk: %s", msg) +#define TRACE_FUNC_EXIT(msg, ret) \ +do { \ + if (!NT_SUCCESS(ret)) \ + dxg_pr_err("%s %x %d", msg, ret, ret); \ + else \ + printk(KERN_DEBUG "dxgk: %s end", msg); \ +} while (false) +#define TRACE_FUNC_EXIT_ERR(msg, ret) \ +do { \ + if (!NT_SUCCESS(ret)) \ + dxg_pr_err("%s %x", msg, ret); \ +} while (false) +#else +#define TRACE_DEBUG(level, fmt, ...)\ + dev_dbg(dxgglobaldev, pr_fmt2(fmt), ##__VA_ARGS__) + +#define TRACE_DEBUG2(level, offset, fmt, ...) \ +do { \ + if (offset == 0) \ + dev_dbg(dxgglobaldev, pr_fmt(fmt), ##__VA_ARGS__); \ + else \ + dev_dbg(dxgglobaldev, pr_fmt2(fmt), ##__VA_ARGS__);\ +} while (false) + +#define TRACE_FUNC_ENTER(msg) \ + dev_dbg(dxgglobaldev, "dxgk: %s", msg) +#define TRACE_FUNC_EXIT(msg, ret) \ +do { \ + if (!NT_SUCCESS(ret)) \ + dev_dbg(dxgglobaldev, "dxgk:err: %s %x", msg, ret); \ + else \ + dev_dbg(dxgglobaldev, "dxgk: %s end", msg); \ +} while (false) +#define TRACE_FUNC_EXIT_ERR(msg, ret) \ +do { \ + if (!NT_SUCCESS(ret)) \ + dev_dbg(dxgglobaldev, "dxgk:err: %s %x", msg, ret); \ +} while (false) +#endif /* USEPRINTK */ +#define TRACE_DEFINE(arg) arg +#endif + +#ifdef DXGKDEBUGLOCKORDER +#define TRACE_LOCK_ORDER(...) TRACE_DEBUG(...) +#else +#define TRACE_LOCK_ORDER(...) +#endif + +#endif
On Tue, May 19, 2020 at 12:32:31PM -0400, Sasha Levin wrote:
+/*
- Dxgkrnl Graphics Port Driver ioctl definitions
- */
+#define LX_IOCTL_DIR_WRITE 0x1 +#define LX_IOCTL_DIR_READ 0x2
+#define LX_IOCTL_DIR(_ioctl) (((_ioctl) >> 30) & 0x3) +#define LX_IOCTL_SIZE(_ioctl) (((_ioctl) >> 16) & 0x3FFF) +#define LX_IOCTL_TYPE(_ioctl) (((_ioctl) >> 8) & 0xFF) +#define LX_IOCTL_CODE(_ioctl) (((_ioctl) >> 0) & 0xFF)
Why create new ioctl macros, can't the "normal" kernel macros work properly?
+#define LX_IOCTL(_dir, _size, _type, _code) ( \
- (((uint)(_dir) & 0x3) << 30) | \
- (((uint)(_size) & 0x3FFF) << 16) | \
- (((uint)(_type) & 0xFF) << 8) | \
- (((uint)(_code) & 0xFF) << 0))
+#define LX_IO(_type, _code) LX_IOCTL(0, 0, (_type), (_code)) +#define LX_IOR(_type, _code, _size) \
- LX_IOCTL(LX_IOCTL_DIR_READ, (_size), (_type), (_code))
+#define LX_IOW(_type, _code, _size) \
- LX_IOCTL(LX_IOCTL_DIR_WRITE, (_size), (_type), (_code))
+#define LX_IOWR(_type, _code, _size) \
- LX_IOCTL(LX_IOCTL_DIR_WRITE | \
- LX_IOCTL_DIR_READ, (_size), (_type), (_code))
+#define LX_DXOPENADAPTERFROMLUID \
- LX_IOWR(0x47, 0x01, sizeof(struct d3dkmt_openadapterfromluid))
<snip>
These structures do not seem to be all using the correct types for a "real" ioctl in the kernel, so you will have to fix them all up before this will work properly.
+void ioctl_desc_init(void);
Very odd global name you are using here :)
Anyway, neat stuff, glad to see it posted, great work!
greg k-h
On Tue, May 19, 2020 at 12:32:31PM -0400, Sasha Levin wrote:
+#define DXGK_MAX_LOCK_DEPTH 64 +#define W_MAX_PATH 260
We already have a max path number, why use a different one?
+#define d3dkmt_handle u32 +#define d3dgpu_virtual_address u64 +#define winwchar u16 +#define winhandle u64 +#define ntstatus int +#define winbool u32 +#define d3dgpu_size_t u64
These are all ripe for a simple search/replace in your editor before you do your next version :)
thanks,
greg k-h
On Tue, May 19, 2020 at 07:21:05PM +0200, Greg KH wrote:
On Tue, May 19, 2020 at 12:32:31PM -0400, Sasha Levin wrote:
+#define DXGK_MAX_LOCK_DEPTH 64 +#define W_MAX_PATH 260
We already have a max path number, why use a different one?
It's max path for Windows, not Linux (thus the "W_" prefix) :)
Maybe changing it to WIN_MAX_PATH or such will make it better?
+#define d3dkmt_handle u32 +#define d3dgpu_virtual_address u64 +#define winwchar u16 +#define winhandle u64 +#define ntstatus int +#define winbool u32 +#define d3dgpu_size_t u64
These are all ripe for a simple search/replace in your editor before you do your next version :)
I've actually attempted that, and reverted that change, mostly because the whole 'handle' thing became very confusing.
Note that we have a few 'handles', each with a different size, and thus calling get_something_something_handle() type of functions becase very confusing since it's not clear what handle we're working with in that case.
With regards to the rest, I wanted to leave stuff like 'winbool' to document the expected ABI between the Windows and Linux side of things. Ideally it would be 'bool' or 'u8', but as you see we had to use 'u32' here which I feel lessens our ability to have the code document itself.
I don't feel too strongly against doing the conversion, and I won't object to doing it if you do, but just be aware that I've tried it and preferred to go back (even though our coding style doesn't like this) :)
On Tue, May 19, 2020 at 01:45:53PM -0400, Sasha Levin wrote:
On Tue, May 19, 2020 at 07:21:05PM +0200, Greg KH wrote:
On Tue, May 19, 2020 at 12:32:31PM -0400, Sasha Levin wrote:
+#define DXGK_MAX_LOCK_DEPTH 64 +#define W_MAX_PATH 260
We already have a max path number, why use a different one?
It's max path for Windows, not Linux (thus the "W_" prefix) :)
Ah, not obvious :)
Maybe changing it to WIN_MAX_PATH or such will make it better?
Probably.
+#define d3dkmt_handle u32 +#define d3dgpu_virtual_address u64 +#define winwchar u16 +#define winhandle u64 +#define ntstatus int +#define winbool u32 +#define d3dgpu_size_t u64
These are all ripe for a simple search/replace in your editor before you do your next version :)
I've actually attempted that, and reverted that change, mostly because the whole 'handle' thing became very confusing.
Yeah, "handles" in windows can be a mess, with some being pointers and others just integers. Trying to make a specific typedef for it is usually the better way overall, that way you can get the compiler to check for mistakes. These #defines will not really help with that.
But, 'ntstatus' should be ok to just make "int" everywhere, right?
Note that we have a few 'handles', each with a different size, and thus calling get_something_something_handle() type of functions becase very confusing since it's not clear what handle we're working with in that case.
Yeah, typedefs can help there.
With regards to the rest, I wanted to leave stuff like 'winbool' to document the expected ABI between the Windows and Linux side of things. Ideally it would be 'bool' or 'u8', but as you see we had to use 'u32' here which I feel lessens our ability to have the code document itself.
'bool' probably will not work as I think it's compiler dependent, __u8 is probably best.
thanks,
greg k-h
On Tue, May 19, 2020 at 12:32:31PM -0400, Sasha Levin wrote:
+static int dxgglobal_init_global_channel(struct hv_device *hdev) +{
- int ret = 0;
- TRACE_DEBUG(1, "%s %x %x", __func__, hdev->vendor_id, hdev->device_id);
- {
TRACE_DEBUG(1, "device type : %pUb\n", &hdev->dev_type);TRACE_DEBUG(1, "device channel: %pUb %p primary: %p\n",&hdev->channel->offermsg.offer.if_type,hdev->channel, hdev->channel->primary_channel);- }
- if (dxgglobal->hdev) {
/* This device should appear only once */pr_err("dxgglobal already initialized\n");ret = -EBADE;goto error;- }
- dxgglobal->hdev = hdev;
- ret = dxgvmbuschannel_init(&dxgglobal->channel, hdev);
- if (ret) {
pr_err("dxgvmbuschannel_init failed: %d\n", ret);goto error;- }
- ret = dxgglobal_getiospace(dxgglobal);
- if (ret) {
pr_err("getiospace failed: %d\n", ret);goto error;- }
- ret = dxgvmb_send_set_iospace_region(dxgglobal->mmiospace_base,
dxgglobal->mmiospace_size, 0);- if (ret) {
pr_err("send_set_iospace_region failed\n");goto error;- }
- hv_set_drvdata(hdev, dxgglobal);
- if (alloc_chrdev_region(&dxgglobal->device_devt, 0, 1, "dxgkrnl") < 0) {
pr_err("alloc_chrdev_region failed\n");ret = -ENODEV;goto error;- }
- dxgglobal->devt_initialized = true;
- dxgglobal->device_class = class_create(THIS_MODULE, "dxgkdrv");
- if (dxgglobal->device_class == NULL) {
pr_err("class_create failed\n");ret = -ENODEV;goto error;- }
- dxgglobal->device_class->devnode = dxg_devnode;
- dxgglobal->device = device_create(dxgglobal->device_class, NULL,
dxgglobal->device_devt, NULL, "dxg");- if (dxgglobal->device == NULL) {
pr_err("device_create failed\n");ret = -ENODEV;goto error;- }
- dxgglobaldev = dxgglobal->device;
- cdev_init(&dxgglobal->device_cdev, &dxgk_fops);
- ret = cdev_add(&dxgglobal->device_cdev, dxgglobal->device_devt, 1);
- if (ret < 0) {
pr_err("cdev_add failed: %d\n", ret);goto error;- }
- dxgglobal->cdev_initialized = true;
+error:
- return ret;
+}
As you only are asking for a single char dev node, please just use the misc device api instead of creating your own class and major number on the fly. It's much simpler and easier overall to make sure you got all of the above logic correct.
thanks,
greg k-h
Connect the dxgkrnl module to the drivers/gpu/ makefile and Kconfig.
Signed-off-by: Sasha Levin sashal@kernel.org --- drivers/gpu/Makefile | 2 +- drivers/video/Kconfig | 2 ++ 2 files changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/Makefile b/drivers/gpu/Makefile index 835c88318cec..28c22c814494 100644 --- a/drivers/gpu/Makefile +++ b/drivers/gpu/Makefile @@ -3,6 +3,6 @@ # taken to initialize them in the correct order. Link order is the only way # to ensure this currently. obj-$(CONFIG_TEGRA_HOST1X) += host1x/ -obj-y += drm/ vga/ +obj-y += drm/ vga/ dxgkrnl/ obj-$(CONFIG_IMX_IPUV3_CORE) += ipu-v3/ obj-$(CONFIG_TRACE_GPU_MEM) += trace/ diff --git a/drivers/video/Kconfig b/drivers/video/Kconfig index 427a993c7f57..362c08778a54 100644 --- a/drivers/video/Kconfig +++ b/drivers/video/Kconfig @@ -19,6 +19,8 @@ source "drivers/gpu/ipu-v3/Kconfig"
source "drivers/gpu/drm/Kconfig"
+source "drivers/gpu/dxgkrnl/Kconfig" + menu "Frame buffer Devices" source "drivers/video/fbdev/Kconfig" endmenu
Register a new device type with vmbus.
Signed-off-by: Sasha Levin sashal@kernel.org --- include/linux/hyperv.h | 16 ++++++++++++++++ 1 file changed, 16 insertions(+)
diff --git a/include/linux/hyperv.h b/include/linux/hyperv.h index 692c89ccf5df..ad16e9bc676a 100644 --- a/include/linux/hyperv.h +++ b/include/linux/hyperv.h @@ -1352,6 +1352,22 @@ void vmbus_free_mmio(resource_size_t start, resource_size_t size); .guid = GUID_INIT(0xda0a7802, 0xe377, 0x4aac, 0x8e, 0x77, \ 0x05, 0x58, 0xeb, 0x10, 0x73, 0xf8)
+/* + * GPU paravirtualization global DXGK channel + * {DDE9CBC0-5060-4436-9448-EA1254A5D177} + */ +#define HV_GPUP_DXGK_GLOBAL_GUID \ + .guid = GUID_INIT(0xdde9cbc0, 0x5060, 0x4436, 0x94, 0x48, \ + 0xea, 0x12, 0x54, 0xa5, 0xd1, 0x77) + +/* + * GPU paravirtualization per virtual GPU DXGK channel + * {6E382D18-3336-4F4B-ACC4-2B7703D4DF4A} + */ +#define HV_GPUP_DXGK_VGPU_GUID \ + .guid = GUID_INIT(0x6e382d18, 0x3336, 0x4f4b, 0xac, 0xc4, \ + 0x2b, 0x77, 0x3, 0xd4, 0xdf, 0x4a) + /* * Synthetic FC GUID * {2f9bcc4a-0069-4af3-b76b-6fd0be528cda}
Signed-off-by: Sasha Levin sashal@kernel.org --- MAINTAINERS | 7 +++++++ 1 file changed, 7 insertions(+)
diff --git a/MAINTAINERS b/MAINTAINERS index e64e5db31497..dccdfadda5df 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -4997,6 +4997,13 @@ F: Documentation/filesystems/dnotify.txt F: fs/notify/dnotify/ F: include/linux/dnotify.h
+DirectX GPU DRIVER +M: Sasha Levin sashal@kernel.org +M: Iouri Tarassov iourit@microsoft.com +L: linux-hyperv@vger.kernel.org +S: Supported +F: drivers/gpu/dxgcore/ + DISK GEOMETRY AND PARTITION HANDLING M: Andries Brouwer aeb@cwi.nl S: Maintained
Hi Sasha
So obviously great that Microsoft is trying to upstream all this, and very much welcome and all that.
But I guess there's a bunch of rather fundamental issues before we look into any kind of code details. And that might make this quite a hard sell for upstream to drivers/gpu subsystem:
- From the blog it sounds like the userspace is all closed. That includes the hw specific part and compiler chunks, all stuff we've generally expected to be able to look in the past for any kind of other driver. It's event documented here:
https://dri.freedesktop.org/docs/drm/gpu/drm-uapi.html#open-source-userspace...
What's your plan here?
btw since the main goal here (at least at first) seems to be get compute and ML going the official work-around here is to relabel your driver as an accelerator driver (just sed -e s/vGPU/vaccel/ over the entire thing or so) and then Olof and Greg will take it into drivers/accel ...
- Next up (but that's not really a surprise for a fresh vendor driver) at a more technical level, this seems to reinvent the world, from device enumeration (why is this not exposed as /dev/dri/card0 so it better integrates with existing linux desktop stuff, in case that becomes a goal ever) down to reinvented kref_put_mutex (and please look at drm_device->struct_mutex for an example of how bad of a nightmare that locking pattern is and how many years it took us to untangle that one.
- Why DX12 on linux? Looking at this feels like classic divide and conquer (or well triple E from the 90s), we have vk, we have drm_syncobj, we have an entire ecosystem of winsys layers that work across vendors. Is the plan here that we get a dx12 driver for other hw mesa drivers from you guys, so this is all consistent and we have a nice linux platform? How does this integrate everywhere else with linux winsys standards, like dma-buf for passing stuff around, dma-fence/sync_file/drm_syncobj for syncing, drm_fourcc/modifiers for some idea how it all meshes together?
- There's been a pile of hallway track/private discussions about moving on from the buffer-based memory managed model to something more modern. That relates to your DXLOCK2 question, but there's a lot more to userspace managed gpu memory residency than just that. monitored fences are another part. Also, to avoid a platform split we need to figure out how to tie this back into the dma-buf and dma-fence (including various uapi flavours) or it'll be made of fail. dx12 has all that in some form, except 0 integration with the linux stuff we have (no surprise, since linux isn't windows). Finally if we go to the trouble of a completely revamped I think ioctls aren't a great idea, something like iouring (the gossip name is drm_uring) would be a lot better. Also for easier paravirt we'd need 0 cpu pointers in any such new interface. Adding a few people who've been involved in these discussions thus far, mostly under a drm/hmm.ko heading iirc.
I think the above are the really big ticket items around what's the plan here and are we solving even the right problem.
Cheers, Daniel
On Tue, May 19, 2020 at 6:33 PM Sasha Levin sashal@kernel.org wrote:
There is a blog post that goes into more detail about the bigger picture, and walks through all the required pieces to make this work. It is available here: https://devblogs.microsoft.com/directx/directx-heart-linux . The rest of this cover letter will focus on the Linux Kernel bits.
Overview
This is the first draft of the Microsoft Virtual GPU (vGPU) driver. The driver exposes a paravirtualized GPU to user mode applications running in a virtual machine on a Windows host. This enables hardware acceleration in environment such as WSL (Windows Subsystem for Linux) where the Linux virtual machine is able to share the GPU with the Windows host.
The projection is accomplished by exposing the WDDM (Windows Display Driver Model) interface as a set of IOCTL. This allows APIs and user mode driver written against the WDDM GPU abstraction on Windows to be ported to run within a Linux environment. This enables the port of the D3D12 and DirectML APIs as well as their associated user mode driver to Linux. This also enables third party APIs, such as the popular NVIDIA Cuda compute API, to be hardware accelerated within a WSL environment.
Only the rendering/compute aspect of the GPU are projected to the virtual machine, no display functionality is exposed. Further, at this time there are no presentation integration. So although the D3D12 API can be use to render graphics offscreen, there is no path (yet) for pixel to flow from the Linux environment back onto the Windows host desktop. This GPU stack is effectively side-by-side with the native Linux graphics stack.
The driver creates the /dev/dxg device, which can be opened by user mode application and handles their ioctls. The IOCTL interface to the driver is defined in dxgkmthk.h (Dxgkrnl Graphics Port Driver ioctl definitions). The interface matches the D3DKMT interface on Windows. Ioctls are implemented in ioctl.c.
When a VM starts, hyper-v on the host adds virtual GPU devices to the VM via the hyper-v driver. The host offers several VM bus channels to the VM: the global channel and one channel per virtual GPU, assigned to the VM.
The driver registers with the hyper-v driver (hv_driver) for the arrival of VM bus channels. dxg_probe_device recognizes the vGPU channels and creates the corresponding objects (dxgadapter for vGPUs and dxgglobal for the global channel).
The driver uses the hyper-V VM bus interface to communicate with the host. dxgvmbus.c implements the communication interface.
The global channel has 8GB of IO space assigned by the host. This space is managed by the host and used to give the guest direct CPU access to some allocations. Video memory is allocated on the host except in the case of existing_sysmem allocations. The Windows host allocates memory for the GPU on behalf of the guest. The Linux guest can access that memory by mapping GPU virtual address to allocations and then referencing those GPU virtual address from within GPU command buffers submitted to the GPU. For allocations which require CPU access, the allocation is mapped by the host into a location in the 8GB of IO space reserved in the guest for that purpose. The Windows host uses the nested CPU page table to ensure that this guest IO space always map to the correct location for the allocation as it may migrate between dedicated GPU memory (e.g. VRAM, firmware reserved DDR) and shared system memory (regular DDR) over its lifetime. The Linux guest maps a user mode CPU virtual address to an allocation IO space range for direct access by user mode APIs and drivers.
Implementation of LX_DXLOCK2 ioctl
We would appreciate your feedback on the implementation of the LX_DXLOCK2 ioctl.
This ioctl is used to get a CPU address to an allocation, which is resident in video/system memory on the host. The way it works:
The driver sends the Lock message to the host
The host allocates space in the VM IO space and maps it to the
allocation memory
The host returns the address in IO space for the mapped allocation
The driver (in dxg_map_iospace) allocates a user mode virtual address
range using vm_mmap and maps it to the IO space using io_remap_ofn_range)
- The VA is returned to the application
Internal objects
The following objects are created by the driver (defined in dxgkrnl.h):
dxgadapter - represents a virtual GPU
dxgprocess - tracks per process state (handle table of created objects, list of objects, etc.)
dxgdevice - a container for other objects (contexts, paging queues, allocations, GPU synchronization objects)
dxgcontext - represents thread of GPU execution for packet scheduling.
dxghwqueue - represents thread of GPU execution of hardware scheduling
dxgallocation - represents a GPU accessible allocation
dxgsyncobject - represents a GPU synchronization object
dxgresource - collection of dxgalloction objects
dxgsharedresource, dxgsharedsyncobj - helper objects to share objects between different dxgdevice objects, which can belong to different
processes
Object handles
All GPU objects, created by the driver, are accessible by a handle (d3dkmt_handle). Each process has its own handle table, which is implemented in hmgr.c. For each API visible object, created by the driver, there is an object, created on the host. For example, the is a dxgprocess object on the host for each dxgprocess object in the VM, etc. The object handles have the same value in the host and the VM, which is done to avoid translation from the guest handles to the host handles.
Signaling CPU events by the host
The WDDM interface provides a way to signal CPU event objects when execution of a context reached certain point. The way it is implemented:
application sends an event_fd via ioctl to the driver
eventfd_ctx_get is used to get a pointer to the file object (eventfd_ctx)
the pointer to sent the host via a VM bus message
when GPU execution reaches a certain point, the host sends a message to the VM with the event pointer
signal_guest_event() handles the messages and eventually eventfd_signal() is called.
Sasha Levin (4): gpu: dxgkrnl: core code gpu: dxgkrnl: hook up dxgkrnl Drivers: hv: vmbus: hook up dxgkrnl gpu: dxgkrnl: create a MAINTAINERS entry
MAINTAINERS | 7 + drivers/gpu/Makefile | 2 +- drivers/gpu/dxgkrnl/Kconfig | 10 + drivers/gpu/dxgkrnl/Makefile | 12 + drivers/gpu/dxgkrnl/d3dkmthk.h | 1635 +++++++++ drivers/gpu/dxgkrnl/dxgadapter.c | 1399 ++++++++ drivers/gpu/dxgkrnl/dxgkrnl.h | 913 ++++++ drivers/gpu/dxgkrnl/dxgmodule.c | 692 ++++ drivers/gpu/dxgkrnl/dxgprocess.c | 355 ++ drivers/gpu/dxgkrnl/dxgvmbus.c | 2955 +++++++++++++++++ drivers/gpu/dxgkrnl/dxgvmbus.h | 859 +++++ drivers/gpu/dxgkrnl/hmgr.c | 593 ++++ drivers/gpu/dxgkrnl/hmgr.h | 107 + drivers/gpu/dxgkrnl/ioctl.c | 5269 ++++++++++++++++++++++++++++++ drivers/gpu/dxgkrnl/misc.c | 280 ++ drivers/gpu/dxgkrnl/misc.h | 288 ++ drivers/video/Kconfig | 2 + include/linux/hyperv.h | 16 + 18 files changed, 15393 insertions(+), 1 deletion(-) create mode 100644 drivers/gpu/dxgkrnl/Kconfig create mode 100644 drivers/gpu/dxgkrnl/Makefile create mode 100644 drivers/gpu/dxgkrnl/d3dkmthk.h create mode 100644 drivers/gpu/dxgkrnl/dxgadapter.c create mode 100644 drivers/gpu/dxgkrnl/dxgkrnl.h create mode 100644 drivers/gpu/dxgkrnl/dxgmodule.c create mode 100644 drivers/gpu/dxgkrnl/dxgprocess.c create mode 100644 drivers/gpu/dxgkrnl/dxgvmbus.c create mode 100644 drivers/gpu/dxgkrnl/dxgvmbus.h create mode 100644 drivers/gpu/dxgkrnl/hmgr.c create mode 100644 drivers/gpu/dxgkrnl/hmgr.h create mode 100644 drivers/gpu/dxgkrnl/ioctl.c create mode 100644 drivers/gpu/dxgkrnl/misc.c create mode 100644 drivers/gpu/dxgkrnl/misc.h
-- 2.25.1
dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel
-- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch
Hi Daniel,
On Tue, May 19, 2020 at 09:21:15PM +0200, Daniel Vetter wrote:
Hi Sasha
So obviously great that Microsoft is trying to upstream all this, and very much welcome and all that.
But I guess there's a bunch of rather fundamental issues before we look into any kind of code details. And that might make this quite a hard sell for upstream to drivers/gpu subsystem:
Let me preface my answers by saying that speaking personally I very much dislike that the userspace is closed and wish I could do something about it.
- From the blog it sounds like the userspace is all closed. That
includes the hw specific part and compiler chunks, all stuff we've generally expected to be able to look in the past for any kind of other driver. It's event documented here:
https://dri.freedesktop.org/docs/drm/gpu/drm-uapi.html#open-source-userspace...
What's your plan here?
Let me answer with a (genuine) question: does this driver have anything to do with DRM even after we enable graphics on it? I'm still trying to figure it out.
There is an open source DX12 Galluim driver (that lives here: https://gitlab.freedesktop.org/kusma/mesa/-/tree/msclc-d3d12) with open source compiler and so on.
The plan is for Microsoft to provide shims to allow the existing Linux userspace interact with DX12; I'll explain below why we had to pipe DX12 all the way into the Linux guest, but this is *not* to introduce DX12 into the Linux world as competition. There is no intent for anyone in the Linux world to start coding for the DX12 API.
This is why I'm not sure whether this touches DRM on the Linux side of things. Nothing is actually rendered on Linux but rather piped to Windows to be done there.
btw since the main goal here (at least at first) seems to be get compute and ML going the official work-around here is to relabel your driver as an accelerator driver (just sed -e s/vGPU/vaccel/ over the entire thing or so) and then Olof and Greg will take it into drivers/accel ...
This submission is not a case of "we want it upstream NOW" but rather "let's work together to figure out how to do it right" :)
I thought about placing this driver in drivers/hyper-v/ given that it's basically just a pipe between the host and the guest. There is no fancy logic in this drivers. Maybe the right place is indeed drivers/accel or drivers/hyper-v but I'd love if we agree on that rather than doing that as a workaround and 6 months down the road enabling graphics.
- Next up (but that's not really a surprise for a fresh vendor driver)
at a more technical level, this seems to reinvent the world, from device enumeration (why is this not exposed as /dev/dri/card0 so it better integrates with existing linux desktop stuff, in case that becomes a goal ever) down to reinvented kref_put_mutex (and please look at drm_device->struct_mutex for an example of how bad of a nightmare that locking pattern is and how many years it took us to untangle that one.
I'd maybe note that neither of us here at Microsoft is an expert in the Linux DRM world. Stuff might have been done in a certain way because we didn't know better.
- Why DX12 on linux? Looking at this feels like classic divide and
There is a single usecase for this: WSL2 developer who wants to run machine learning on his GPU. The developer is working on his laptop, which is running Windows and that laptop has a single GPU that Windows is using.
Since the GPU is being used by Windows, we can't assign it directly to the Linux guest, but instead we can use GPU Partitioning to give the guest access to the GPU. This means that the guest needs to be able to "speak" DX12, which is why we pulled DX12 into Linux.
conquer (or well triple E from the 90s), we have vk, we have drm_syncobj, we have an entire ecosystem of winsys layers that work across vendors. Is the plan here that we get a dx12 driver for other hw mesa drivers from you guys, so this is all consistent and we have a nice linux platform? How does this integrate everywhere else with linux winsys standards, like dma-buf for passing stuff around, dma-fence/sync_file/drm_syncobj for syncing, drm_fourcc/modifiers for some idea how it all meshes together?
Let me point you to this blog post that has more information about the graphics side of things: https://www.collabora.com/news-and-blog/news-and-events/introducing-opencl-a... .
The intent is to wrap DX12 with shims to work with the existing ecosystem; DX12 isn't a new player on it's own and thus isn't trying to divide/conquer anything.
- There's been a pile of hallway track/private discussions about
moving on from the buffer-based memory managed model to something more modern. That relates to your DXLOCK2 question, but there's a lot more to userspace managed gpu memory residency than just that. monitored fences are another part. Also, to avoid a platform split we need to figure out how to tie this back into the dma-buf and dma-fence (including various uapi flavours) or it'll be made of fail. dx12 has all that in some form, except 0 integration with the linux stuff we have (no surprise, since linux isn't windows). Finally if we go to the trouble of a completely revamped I think ioctls aren't a great idea, something like iouring (the gossip name is drm_uring) would be a lot better. Also for easier paravirt we'd need 0 cpu pointers in any such new interface. Adding a few people who've been involved in these discussions thus far, mostly under a drm/hmm.ko heading iirc.
I think the above are the really big ticket items around what's the plan here and are we solving even the right problem.
Part of the reason behind this implementation is simplicity. Again, no objections around moving to uring and doing other improvements.
On Tuesday 2020-05-19 22:36, Sasha Levin wrote:
- Why DX12 on linux? Looking at this feels like classic divide and
There is a single usecase for this: WSL2 developer who wants to run machine learning on his GPU. The developer is working on his laptop, which is running Windows and that laptop has a single GPU that Windows is using.
It does not feel right conceptually. If the target is a Windows API (DX12/ML), why bother with Linux environments? Make it a Windows executable, thereby skipping the WSL translation layer and passthrough.
On Tue, May 19, 2020 at 2:36 PM Sasha Levin sashal@kernel.org wrote:
Hi Daniel,
On Tue, May 19, 2020 at 09:21:15PM +0200, Daniel Vetter wrote:
Hi Sasha
So obviously great that Microsoft is trying to upstream all this, and very much welcome and all that.
But I guess there's a bunch of rather fundamental issues before we look into any kind of code details. And that might make this quite a hard sell for upstream to drivers/gpu subsystem:
Let me preface my answers by saying that speaking personally I very much dislike that the userspace is closed and wish I could do something about it.
- From the blog it sounds like the userspace is all closed. That
includes the hw specific part and compiler chunks, all stuff we've generally expected to be able to look in the past for any kind of other driver. It's event documented here:
https://dri.freedesktop.org/docs/drm/gpu/drm-uapi.html#open-source-userspace...
What's your plan here?
Let me answer with a (genuine) question: does this driver have anything to do with DRM even after we enable graphics on it? I'm still trying to figure it out.
There is an open source DX12 Galluim driver (that lives here: https://gitlab.freedesktop.org/kusma/mesa/-/tree/msclc-d3d12) with open source compiler and so on.
The plan is for Microsoft to provide shims to allow the existing Linux userspace interact with DX12; I'll explain below why we had to pipe DX12 all the way into the Linux guest, but this is *not* to introduce DX12 into the Linux world as competition. There is no intent for anyone in the Linux world to start coding for the DX12 API.
If that really is the case why is microsoft recommending developers to break compatibility with native Linux and use the DX12 API's here: https://devblogs.microsoft.com/directx/in-the-works-opencl-and-opengl-mappin...
Quote: "Make it easier for developers to port their apps to D3D12. For developers looking to move from older OpenCL and OpenGL API versions to D3D12, the open source mapping layers will provide helpful example code on how to use the D3D12 Translation Layer library."
If developers of applications that use OpenCL and OpenGL API's were to follow this advice and transition to D3D12 their applications would no longer work on Linux systems unless using WSL2. Is Microsoft planning on creating a D3D12/DirectML frontend that doesn't depend on WSL2?
This is why I'm not sure whether this touches DRM on the Linux side of things. Nothing is actually rendered on Linux but rather piped to Windows to be done there.
btw since the main goal here (at least at first) seems to be get compute and ML going the official work-around here is to relabel your driver as an accelerator driver (just sed -e s/vGPU/vaccel/ over the entire thing or so) and then Olof and Greg will take it into drivers/accel ...
This submission is not a case of "we want it upstream NOW" but rather "let's work together to figure out how to do it right" :)
I thought about placing this driver in drivers/hyper-v/ given that it's basically just a pipe between the host and the guest. There is no fancy logic in this drivers. Maybe the right place is indeed drivers/accel or drivers/hyper-v but I'd love if we agree on that rather than doing that as a workaround and 6 months down the road enabling graphics.
- Next up (but that's not really a surprise for a fresh vendor driver)
at a more technical level, this seems to reinvent the world, from device enumeration (why is this not exposed as /dev/dri/card0 so it better integrates with existing linux desktop stuff, in case that becomes a goal ever) down to reinvented kref_put_mutex (and please look at drm_device->struct_mutex for an example of how bad of a nightmare that locking pattern is and how many years it took us to untangle that one.
I'd maybe note that neither of us here at Microsoft is an expert in the Linux DRM world. Stuff might have been done in a certain way because we didn't know better.
- Why DX12 on linux? Looking at this feels like classic divide and
There is a single usecase for this: WSL2 developer who wants to run machine learning on his GPU. The developer is working on his laptop, which is running Windows and that laptop has a single GPU that Windows is using.
Since the GPU is being used by Windows, we can't assign it directly to the Linux guest, but instead we can use GPU Partitioning to give the guest access to the GPU. This means that the guest needs to be able to "speak" DX12, which is why we pulled DX12 into Linux.
conquer (or well triple E from the 90s), we have vk, we have drm_syncobj, we have an entire ecosystem of winsys layers that work across vendors. Is the plan here that we get a dx12 driver for other hw mesa drivers from you guys, so this is all consistent and we have a nice linux platform? How does this integrate everywhere else with linux winsys standards, like dma-buf for passing stuff around, dma-fence/sync_file/drm_syncobj for syncing, drm_fourcc/modifiers for some idea how it all meshes together?
Let me point you to this blog post that has more information about the graphics side of things: https://www.collabora.com/news-and-blog/news-and-events/introducing-opencl-a... .
The intent is to wrap DX12 with shims to work with the existing ecosystem; DX12 isn't a new player on it's own and thus isn't trying to divide/conquer anything.
Shouldn't tensorflow/machine learning be going through the opencl compatibility layer/shims instead of talking directly to DX12/DirectML?
If tensorflow or any other machine learning software uses DX12 API's directly then they won't be compatible with Linux unless running on top of WSL2.
- There's been a pile of hallway track/private discussions about
moving on from the buffer-based memory managed model to something more modern. That relates to your DXLOCK2 question, but there's a lot more to userspace managed gpu memory residency than just that. monitored fences are another part. Also, to avoid a platform split we need to figure out how to tie this back into the dma-buf and dma-fence (including various uapi flavours) or it'll be made of fail. dx12 has all that in some form, except 0 integration with the linux stuff we have (no surprise, since linux isn't windows). Finally if we go to the trouble of a completely revamped I think ioctls aren't a great idea, something like iouring (the gossip name is drm_uring) would be a lot better. Also for easier paravirt we'd need 0 cpu pointers in any such new interface. Adding a few people who've been involved in these discussions thus far, mostly under a drm/hmm.ko heading iirc.
I think the above are the really big ticket items around what's the plan here and are we solving even the right problem.
Part of the reason behind this implementation is simplicity. Again, no objections around moving to uring and doing other improvements.
-- Thanks, Sasha
On Wed, 20 May 2020 at 02:33, Sasha Levin sashal@kernel.org wrote:
There is a blog post that goes into more detail about the bigger picture, and walks through all the required pieces to make this work. It is available here: https://devblogs.microsoft.com/directx/directx-heart-linux . The rest of this cover letter will focus on the Linux Kernel bits.
Overview
This is the first draft of the Microsoft Virtual GPU (vGPU) driver. The driver exposes a paravirtualized GPU to user mode applications running in a virtual machine on a Windows host. This enables hardware acceleration in environment such as WSL (Windows Subsystem for Linux) where the Linux virtual machine is able to share the GPU with the Windows host.
The projection is accomplished by exposing the WDDM (Windows Display Driver Model) interface as a set of IOCTL. This allows APIs and user mode driver written against the WDDM GPU abstraction on Windows to be ported to run within a Linux environment. This enables the port of the D3D12 and DirectML APIs as well as their associated user mode driver to Linux. This also enables third party APIs, such as the popular NVIDIA Cuda compute API, to be hardware accelerated within a WSL environment.
Only the rendering/compute aspect of the GPU are projected to the virtual machine, no display functionality is exposed. Further, at this time there are no presentation integration. So although the D3D12 API can be use to render graphics offscreen, there is no path (yet) for pixel to flow from the Linux environment back onto the Windows host desktop. This GPU stack is effectively side-by-side with the native Linux graphics stack.
Okay I've had some caffiene and absorbed some more of this.
This is a driver that connects a binary blob interface in the Windows kernel drivers to a binary blob that you run inside a Linux guest. It's a binary transport between two binary pieces. Personally this holds little of interest to me, I can see why it might be nice to have this upstream, but I don't forsee any other Linux distributor ever enabling it or having to ship it, it's purely a WSL2 pipe. I'm not saying I'd be happy to see this in the tree, since I don't see the value of maintaining it upstream, but it probably should just exists in a drivers/hyperv type area.
Having said that, I hit one stumbling block: "Further, at this time there are no presentation integration. "
If we upstream this driver as-is into some hyperv specific place, and you decide to add presentation integration this is more than likely going to mean you will want to interact with dma-bufs and dma-fences. If the driver is hidden away in a hyperv place it's likely we won't even notice that feature landing until it's too late.
I would like to see a coherent plan for presentation support (not code, just an architectural diagram), because I think when you contemplate how that works it will change the picture of how this driver looks and intergrates into the rest of the Linux graphics ecosystem.
As-is I'd rather this didn't land under my purview, since I don't see the value this adds to the Linux ecosystem at all, and I think it's important when putting a burden on upstream that you provide some value.
Dave.
On Wed, May 20, 2020 at 12:42 AM Dave Airlie airlied@gmail.com wrote:
On Wed, 20 May 2020 at 02:33, Sasha Levin sashal@kernel.org wrote:
There is a blog post that goes into more detail about the bigger picture, and walks through all the required pieces to make this work. It is available here: https://devblogs.microsoft.com/directx/directx-heart-linux . The rest of this cover letter will focus on the Linux Kernel bits.
Overview
This is the first draft of the Microsoft Virtual GPU (vGPU) driver. The driver exposes a paravirtualized GPU to user mode applications running in a virtual machine on a Windows host. This enables hardware acceleration in environment such as WSL (Windows Subsystem for Linux) where the Linux virtual machine is able to share the GPU with the Windows host.
The projection is accomplished by exposing the WDDM (Windows Display Driver Model) interface as a set of IOCTL. This allows APIs and user mode driver written against the WDDM GPU abstraction on Windows to be ported to run within a Linux environment. This enables the port of the D3D12 and DirectML APIs as well as their associated user mode driver to Linux. This also enables third party APIs, such as the popular NVIDIA Cuda compute API, to be hardware accelerated within a WSL environment.
Only the rendering/compute aspect of the GPU are projected to the virtual machine, no display functionality is exposed. Further, at this time there are no presentation integration. So although the D3D12 API can be use to render graphics offscreen, there is no path (yet) for pixel to flow from the Linux environment back onto the Windows host desktop. This GPU stack is effectively side-by-side with the native Linux graphics stack.
Okay I've had some caffiene and absorbed some more of this.
This is a driver that connects a binary blob interface in the Windows kernel drivers to a binary blob that you run inside a Linux guest. It's a binary transport between two binary pieces. Personally this holds little of interest to me, I can see why it might be nice to have this upstream, but I don't forsee any other Linux distributor ever enabling it or having to ship it, it's purely a WSL2 pipe. I'm not saying I'd be happy to see this in the tree, since I don't see the value of maintaining it upstream, but it probably should just exists in a drivers/hyperv type area.
Yup as-is (especially with the goal of this being aimed at ml/compute only) drivers/hyperv sounds a bunch more reasonable than drivers/gpu.
Having said that, I hit one stumbling block: "Further, at this time there are no presentation integration. "
If we upstream this driver as-is into some hyperv specific place, and you decide to add presentation integration this is more than likely going to mean you will want to interact with dma-bufs and dma-fences. If the driver is hidden away in a hyperv place it's likely we won't even notice that feature landing until it's too late.
I've recently added regex matches to MAINTAINERS so we'll see dma_buf/fence/anything show up on dri-devel. So that part is solved hopefully.
I would like to see a coherent plan for presentation support (not code, just an architectural diagram), because I think when you contemplate how that works it will change the picture of how this driver looks and intergrates into the rest of the Linux graphics ecosystem.
Yeah once we have the feature-creep to presentation support all the integration fun starts, with all the questions about "why does this not look like any other linux gpu driver". We have that already with nvidia insisting they just can't implement any of the upstream gpu uapi we have, but at least they're not in-tree, so not our problem from an upstream maintainership pov.
But once this dx12 pipe is landed and then we want to extend it it's still going to have all the "we can't ever release the sources to any of the parts we usually expect to be open for gpu drivers in upstream" problems. Then we're stuck at a rather awkward point of why one vendor gets an exception and all the others dont.
As-is I'd rather this didn't land under my purview, since I don't see the value this adds to the Linux ecosystem at all, and I think it's important when putting a burden on upstream that you provide some value.
Well there is some in the form of "more hw/platform support". But given that gpus evolved rather fast, including the entire integration ecosystem (it's by far not just the hw drivers that move quickly). So that value deprecates a lot faster than for other kernel subsystems. And all that's left is the pain of not breaking anything without actually being able to evolve the overall stack in any meaningful way. -Daniel
Hey guys,
Thanks for the discussion. I may not be able to immediately answer all of your questions, but I'll do my best 😊.
drivers/hyperv sounds like it could be a better location. We weren't too sure where to put this, we though /drivers/gpu would be appropriate given this deal with GPUs, but I get your point... this is a vGPU driver that really only works when being run under Hyper-V, so drivers/hyperv is likely more appropriate.
In term of presentation, I need to clarify a few things. We announced today that we're also adding support for Linux GUI applications. The way this will work is roughly as follow. We're writing a Wayland compositor that will essentially bridge over RDP-RAIL (RAIL=Remote Application Integrated Locally). We're starting from a Weston base. Weston already has an RDP Backend, but that's for a full desktop remoting scheme. Weston draws a desktop and remote it over RDP... and then you can peek at that desktop using an rdp client on the Windows side. RAIL works differently. In that case our wayland compositor no longer paint a desktop... instead it simply forward individual visual / wl_surface over the RDP RAIL channel such that these visual can be displayed on the Windows desktop. The RDP client create proxy window for each of these top level visual and their content is filled with the data coming over the RDP channel. All pixels are owned by the RDP server/WSL... so these windows looks different than native window are they are painted and themed by WSL. The proxy window on the host gather input and inject back over RDP... This is essentially how application remoting works on windows and this is all publicly documented as part of the various RDP protocol specification. As a matter of fact, for the RDP server on the Weston side we are looking at continue to leverage FreeRDP (and provide fixes/enhancement as needed to the public project). Further, we're looking at further improvement down this path to avoid having to copy the content over the RAIL channel and instead just share/swap buffer between the guest and the host. We have extension to the RDP protocol, called VAIL (Virtualized Application Integrated Locally) which does that today. Today this is only use in Windows on Windows for very specific scenario. We're looking at extending the public RDP protocol with these VAIL extension to make this an official Microsoft supported protocol which would allow us to target this in WSL. We have finished designing this part in details. Our goal would be to leverage something along the line of wl_drm, dma-buf, dma-fence, etc... This compositor and all our contribution to FreeRDP will be fully open source, including our design doc. We're not quite sure yet whether this will be offered as a separate project entirely distinct from it's Weston root... or if we'll propose an extension to Weston to operate in this mode. We would like to build it such that in theory any Wayland compositor could add support for this mode of operation if they want to remote application to a Windows host (over the network, or on the same box).
We see /dev/dxg really as a projection of the GPU when running in WSL such that the GPU can be shared between WSL and the host... not something that would coexist "at the same time" with a real DRM GPU.
We have consider the possibility of bringing DX to Linux with no Windows cord attached. I'm not ready to discuss this at this time 😊... but in the hypothetical that we were do this, DX would be running on top of DRI/DRM on native Linux. We likely would be contributing some changes to DRM to address area of divergence and get better mapping for our user mode driver, but we wouldn't try to shoehorn /dev/dxg into the picture. In that hypothetical world, we would essentially have DX target DRM on native Linux and DX continue to target DXG in WSL to share the GPU with the host. I think this further reinforce the point you guys were making that the right place for our current dxgkrnl driver to live in would be /drivers/hyperv/dxgkrnl. In insight, I totally agree 😊.
I think this cover all questions, let me know if I missed anything.
Thanks, Steve
-----Original Message----- From: Daniel Vetter daniel@ffwll.ch Sent: Tuesday, May 19, 2020 4:01 PM To: Dave Airlie airlied@gmail.com Cc: Sasha Levin sashal@kernel.org; linux-hyperv@vger.kernel.org; Stephen Hemminger sthemmin@microsoft.com; Ursulin, Tvrtko tvrtko.ursulin@intel.com; Greg Kroah-Hartman gregkh@linuxfoundation.org; Haiyang Zhang haiyangz@microsoft.com; LKML linux-kernel@vger.kernel.org; dri-devel dri-devel@lists.freedesktop.org; Chris Wilson chris@chris-wilson.co.uk; Steve Pronovost spronovo@microsoft.com; Linux Fbdev development list linux-fbdev@vger.kernel.org; Iouri Tarassov iourit@microsoft.com; Deucher, Alexander alexander.deucher@amd.com; KY Srinivasan kys@microsoft.com; Wei Liu wei.liu@kernel.org; Hawking Zhang Hawking.Zhang@amd.com Subject: [EXTERNAL] Re: [RFC PATCH 0/4] DirectX on Linux
On Wed, May 20, 2020 at 12:42 AM Dave Airlie airlied@gmail.com wrote:
On Wed, 20 May 2020 at 02:33, Sasha Levin sashal@kernel.org wrote:
There is a blog post that goes into more detail about the bigger picture, and walks through all the required pieces to make this work. It is available here: https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Fde vblogs.microsoft.com%2Fdirectx%2Fdirectx-heart-linux&data=02%7C01%7Cspronovo%40microsoft.com%7C3f18e46192b24cccf6a008d7fc489063%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C637255260910730243&sdata=IRRknzg%2F6Myzj3JXESN7GgmN6AcUV3DxhL95P%2ButtCw%3D&reserved=0 . The rest of this cover letter will focus on the Linux Kernel bits.
Overview
This is the first draft of the Microsoft Virtual GPU (vGPU) driver. The driver exposes a paravirtualized GPU to user mode applications running in a virtual machine on a Windows host. This enables hardware acceleration in environment such as WSL (Windows Subsystem for Linux) where the Linux virtual machine is able to share the GPU with the Windows host.
The projection is accomplished by exposing the WDDM (Windows Display Driver Model) interface as a set of IOCTL. This allows APIs and user mode driver written against the WDDM GPU abstraction on Windows to be ported to run within a Linux environment. This enables the port of the D3D12 and DirectML APIs as well as their associated user mode driver to Linux. This also enables third party APIs, such as the popular NVIDIA Cuda compute API, to be hardware accelerated within a WSL environment.
Only the rendering/compute aspect of the GPU are projected to the virtual machine, no display functionality is exposed. Further, at this time there are no presentation integration. So although the D3D12 API can be use to render graphics offscreen, there is no path (yet) for pixel to flow from the Linux environment back onto the Windows host desktop. This GPU stack is effectively side-by-side with the native Linux graphics stack.
Okay I've had some caffiene and absorbed some more of this.
This is a driver that connects a binary blob interface in the Windows kernel drivers to a binary blob that you run inside a Linux guest. It's a binary transport between two binary pieces. Personally this holds little of interest to me, I can see why it might be nice to have this upstream, but I don't forsee any other Linux distributor ever enabling it or having to ship it, it's purely a WSL2 pipe. I'm not saying I'd be happy to see this in the tree, since I don't see the value of maintaining it upstream, but it probably should just exists in a drivers/hyperv type area.
Yup as-is (especially with the goal of this being aimed at ml/compute only) drivers/hyperv sounds a bunch more reasonable than drivers/gpu.
Having said that, I hit one stumbling block: "Further, at this time there are no presentation integration. "
If we upstream this driver as-is into some hyperv specific place, and you decide to add presentation integration this is more than likely going to mean you will want to interact with dma-bufs and dma-fences. If the driver is hidden away in a hyperv place it's likely we won't even notice that feature landing until it's too late.
I've recently added regex matches to MAINTAINERS so we'll see dma_buf/fence/anything show up on dri-devel. So that part is solved hopefully.
I would like to see a coherent plan for presentation support (not code, just an architectural diagram), because I think when you contemplate how that works it will change the picture of how this driver looks and intergrates into the rest of the Linux graphics ecosystem.
Yeah once we have the feature-creep to presentation support all the integration fun starts, with all the questions about "why does this not look like any other linux gpu driver". We have that already with nvidia insisting they just can't implement any of the upstream gpu uapi we have, but at least they're not in-tree, so not our problem from an upstream maintainership pov.
But once this dx12 pipe is landed and then we want to extend it it's still going to have all the "we can't ever release the sources to any of the parts we usually expect to be open for gpu drivers in upstream" problems. Then we're stuck at a rather awkward point of why one vendor gets an exception and all the others dont.
As-is I'd rather this didn't land under my purview, since I don't see the value this adds to the Linux ecosystem at all, and I think it's important when putting a burden on upstream that you provide some value.
Well there is some in the form of "more hw/platform support". But given that gpus evolved rather fast, including the entire integration ecosystem (it's by far not just the hw drivers that move quickly). So that value deprecates a lot faster than for other kernel subsystems. And all that's left is the pain of not breaking anything without actually being able to evolve the overall stack in any meaningful way. -Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - +https://nam06.safelinks.protection.outlook.com/?url=http%3A%2F%2Fblog.f +fwll.ch%2F&data=02%7C01%7Cspronovo%40microsoft.com%7C3f18e46192b24c +ccf6a008d7fc489063%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C6372552 +60910735230&sdata=hAIV1wJ29WF9IXTvJm3dr4StCwPzF0GdO2iWPyfnElg%3D&am +p;reserved=0
Hi Steve,
Sounds all good, some more comments and details below.
On Wed, May 20, 2020 at 5:47 AM Steve Pronovost spronovo@microsoft.com wrote:
Hey guys,
Thanks for the discussion. I may not be able to immediately answer all of your questions, but I'll do my best 😊.
drivers/hyperv sounds like it could be a better location. We weren't too sure where to put this, we though /drivers/gpu would be appropriate given this deal with GPUs, but I get your point... this is a vGPU driver that really only works when being run under Hyper-V, so drivers/hyperv is likely more appropriate.
I think "it's a virtual gpu" is the wrong sales pitch, as is "only runs on $platform". We have lots of drm drivers in drivers/gpu that fit that bill. The better pitch I think is "it's a not a gpu, it's a dx12 protocol pipe" and "we actually do not want to integrate with the linux gpu ecosystem and primitives, we want to integrate with dx12 ecosystem and primitives to make the seamless rdp/rail/vail stuff work nicely". Below some more thoughts on the technical said.
In term of presentation, I need to clarify a few things. We announced today that we're also adding support for Linux GUI applications. The way this will work is roughly as follow. We're writing a Wayland compositor that will essentially bridge over RDP-RAIL (RAIL=Remote Application Integrated Locally). We're starting from a Weston base. Weston already has an RDP Backend, but that's for a full desktop remoting scheme. Weston draws a desktop and remote it over RDP... and then you can peek at that desktop using an rdp client on the Windows side. RAIL works differently. In that case our wayland compositor no longer paint a desktop... instead it simply forward individual visual / wl_surface over the RDP RAIL channel such that these visual can be displayed on the Windows desktop. The RDP client create proxy window for each of these top level visual and their content is filled with the data coming over the RDP channel. All pixels are owned by the RDP server/WSL... so these windows looks different than native window are they are painted and themed by WSL. The proxy window on the host gather input and inject back over RDP... This is essentially how application remoting works on windows and this is all publicly documented as part of the various RDP protocol specification. As a matter of fact, for the RDP server on the Weston side we are looking at continue to leverage FreeRDP (and provide fixes/enhancement as needed to the public project). Further, we're looking at further improvement down this path to avoid having to copy the content over the RAIL channel and instead just share/swap buffer between the guest and the host. We have extension to the RDP protocol, called VAIL (Virtualized Application Integrated Locally) which does that today. Today this is only use in Windows on Windows for very specific scenario. We're looking at extending the public RDP protocol with these VAIL extension to make this an official Microsoft supported protocol which would allow us to target this in WSL. We have finished designing this part in details. Our goal would be to leverage something along the line of wl_drm, dma-buf, dma-fence, etc... This compositor and all our contribution to FreeRDP will be fully open source, including our design doc. We're not quite sure yet whether this will be offered as a separate project entirely distinct from it's Weston root... or if we'll propose an extension to Weston to operate in this mode. We would like to build it such that in theory any Wayland compositor could add support for this mode of operation if they want to remote application to a Windows host (over the network, or on the same box).
Sounds like a solid plan for presentation. I think this is all up to wayland/weston folks to figure out with you, from the kernel side I have only one concern (and I discussed that with a few folks already on irc, I think they're at least on this thread involved within microsoft too in some form): If we do integrate with linux concepts like wl_drm/dma-buf/fence and so on then we end up with a normal gpu driver, with with lots of blobby components all around that can't be opened (since large chunks written by hw vendors, so nothing microsoft can do about them). That's the awkward exception (why microsoft but not other gpu hw vendors/plaforms/whatever?) that we need to avoid.
But wayland doesn't really need dma-buf and the wl_drm protocols afaiui, as long as the egl extensions work you can have whatever private wayland protocol in your winsys code you want to shovel the buffers and syncobj from client to the wayland-rdp-rail compositor. If that uses dx12 native handles for these things we side-step the awkward exception question for linux gpu stack since it all stays 100% contained in drivers/hv. Just try to avoid the nvidia fail of insisting that you need your own set of egl extensions (egl_streams and egl_drm_kms_reinvented_but_badly and a few others iirc) for everything, that makes it very painful for all the compositor projects since they need duplicated code for no good reason :-)
Also looking at the fun the virtio folks have right now trying to get virgl dma-buf handles in the guest shared with other virtio devices or virgl instance in some other guest and getting it working ... If you avoid the dma-buf integration that might also sidestep a lot of technical headaches and be the simpler solution. At least as long as we're talking about wsl2 only.
We see /dev/dxg really as a projection of the GPU when running in WSL such that the GPU can be shared between WSL and the host... not something that would coexist "at the same time" with a real DRM GPU.
We have consider the possibility of bringing DX to Linux with no Windows cord attached. I'm not ready to discuss this at this time 😊... but in the hypothetical that we were do this, DX would be running on top of DRI/DRM on native Linux. We likely would be contributing some changes to DRM to address area of divergence and get better mapping for our user mode driver, but we wouldn't try to shoehorn /dev/dxg into the picture. In that hypothetical world, we would essentially have DX target DRM on native Linux and DX continue to target DXG in WSL to share the GPU with the host. I think this further reinforce the point you guys were making that the right place for our current dxgkrnl driver to live in would be /drivers/hyperv/dxgkrnl. In insight, I totally agree 😊.
We had a pile of discussions on irc about some of the ideas floating around for extending drm with some of the newer memory/sync concepts. So very much interested, but that's indeed a different fish and most likely will look a lot different to the dxgkrnl interface, while ofc still aiming to give the same power to apps. At least that's the goal, but aside from some prototype in some place this really doesn't exist yet anywhere.
I think this cover all questions, let me know if I missed anything.
I think it's all good.
Cheers, Daniel
Thanks, Steve
-----Original Message----- From: Daniel Vetter daniel@ffwll.ch Sent: Tuesday, May 19, 2020 4:01 PM To: Dave Airlie airlied@gmail.com Cc: Sasha Levin sashal@kernel.org; linux-hyperv@vger.kernel.org; Stephen Hemminger sthemmin@microsoft.com; Ursulin, Tvrtko < tvrtko.ursulin@intel.com>; Greg Kroah-Hartman gregkh@linuxfoundation.org; Haiyang Zhang haiyangz@microsoft.com; LKML linux-kernel@vger.kernel.org; dri-devel dri-devel@lists.freedesktop.org; Chris Wilson < chris@chris-wilson.co.uk>; Steve Pronovost spronovo@microsoft.com; Linux Fbdev development list linux-fbdev@vger.kernel.org; Iouri Tarassov iourit@microsoft.com; Deucher, Alexander < alexander.deucher@amd.com>; KY Srinivasan kys@microsoft.com; Wei Liu < wei.liu@kernel.org>; Hawking Zhang Hawking.Zhang@amd.com Subject: [EXTERNAL] Re: [RFC PATCH 0/4] DirectX on Linux
On Wed, May 20, 2020 at 12:42 AM Dave Airlie airlied@gmail.com wrote:
On Wed, 20 May 2020 at 02:33, Sasha Levin sashal@kernel.org wrote:
There is a blog post that goes into more detail about the bigger picture, and walks through all the required pieces to make this work. It is available here: https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Fde vblogs.microsoft.com
%2Fdirectx%2Fdirectx-heart-linux&data=02%7C01%7Cspronovo% 40microsoft.com%7C3f18e46192b24cccf6a008d7fc489063%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C637255260910730243&sdata=IRRknzg%2F6Myzj3JXESN7GgmN6AcUV3DxhL95P%2ButtCw%3D&reserved=0 . The rest of this cover letter will focus on the Linux Kernel bits.
Overview
This is the first draft of the Microsoft Virtual GPU (vGPU) driver. The driver exposes a paravirtualized GPU to user mode applications running in a virtual machine on a Windows host. This enables hardware acceleration in environment such as WSL (Windows Subsystem for Linux) where the Linux virtual machine is able to share the GPU with the Windows host.
The projection is accomplished by exposing the WDDM (Windows Display Driver Model) interface as a set of IOCTL. This allows APIs and user mode driver written against the WDDM GPU abstraction on Windows to be ported to run within a Linux environment. This enables the port of the D3D12 and DirectML APIs as well as their associated user mode driver to Linux. This also enables third party APIs, such as the popular NVIDIA Cuda compute API, to be hardware accelerated within a WSL
environment.
Only the rendering/compute aspect of the GPU are projected to the virtual machine, no display functionality is exposed. Further, at this time there are no presentation integration. So although the D3D12 API can be use to render graphics offscreen, there is no path (yet) for pixel to flow from the Linux environment back onto the Windows host desktop. This GPU stack is effectively side-by-side with the native Linux graphics stack.
Okay I've had some caffiene and absorbed some more of this.
This is a driver that connects a binary blob interface in the Windows kernel drivers to a binary blob that you run inside a Linux guest. It's a binary transport between two binary pieces. Personally this holds little of interest to me, I can see why it might be nice to have this upstream, but I don't forsee any other Linux distributor ever enabling it or having to ship it, it's purely a WSL2 pipe. I'm not saying I'd be happy to see this in the tree, since I don't see the value of maintaining it upstream, but it probably should just exists in a drivers/hyperv type area.
Yup as-is (especially with the goal of this being aimed at ml/compute only) drivers/hyperv sounds a bunch more reasonable than drivers/gpu.
Having said that, I hit one stumbling block: "Further, at this time there are no presentation integration. "
If we upstream this driver as-is into some hyperv specific place, and you decide to add presentation integration this is more than likely going to mean you will want to interact with dma-bufs and dma-fences. If the driver is hidden away in a hyperv place it's likely we won't even notice that feature landing until it's too late.
I've recently added regex matches to MAINTAINERS so we'll see dma_buf/fence/anything show up on dri-devel. So that part is solved hopefully.
I would like to see a coherent plan for presentation support (not code, just an architectural diagram), because I think when you contemplate how that works it will change the picture of how this driver looks and intergrates into the rest of the Linux graphics ecosystem.
Yeah once we have the feature-creep to presentation support all the integration fun starts, with all the questions about "why does this not look like any other linux gpu driver". We have that already with nvidia insisting they just can't implement any of the upstream gpu uapi we have, but at least they're not in-tree, so not our problem from an upstream maintainership pov.
But once this dx12 pipe is landed and then we want to extend it it's still going to have all the "we can't ever release the sources to any of the parts we usually expect to be open for gpu drivers in upstream" problems. Then we're stuck at a rather awkward point of why one vendor gets an exception and all the others dont.
As-is I'd rather this didn't land under my purview, since I don't see the value this adds to the Linux ecosystem at all, and I think it's important when putting a burden on upstream that you provide some value.
Well there is some in the form of "more hw/platform support". But given that gpus evolved rather fast, including the entire integration ecosystem (it's by far not just the hw drivers that move quickly). So that value deprecates a lot faster than for other kernel subsystems. And all that's left is the pain of not breaking anything without actually being able to evolve the overall stack in any meaningful way.
-Daniel
Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - +https://nam06.safelinks.protection.outlook.com/?url=http%3A%2F%2Fblog.f +fwll.ch%2F&data=02%7C01%7Cspronovo%40microsoft.com%7C3f18e46192b24c +ccf6a008d7fc489063%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C6372552 https://nam06.safelinks.protection.outlook.com/?url=http%3A%2F%2Fblog.f+fwll.ch%2F&data=02%7C01%7Cspronovo%40microsoft.com%7C3f18e46192b24c+ccf6a008d7fc489063%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C6372552 +60910735230&sdata=hAIV1wJ29WF9IXTvJm3dr4StCwPzF0GdO2iWPyfnElg%3D&am +p;reserved=0
Thanks Daniel, more below.
From: Daniel Vetter daniel@ffwll.ch Sent: Wednesday, May 20, 2020 12:41 AM To: Steve Pronovost spronovo@microsoft.com Cc: Dave Airlie airlied@gmail.com; Sasha Levin sashal@kernel.org; linux-hyperv@vger.kernel.org; Stephen Hemminger sthemmin@microsoft.com; Ursulin, Tvrtko tvrtko.ursulin@intel.com; Greg Kroah-Hartman gregkh@linuxfoundation.org; Haiyang Zhang haiyangz@microsoft.com; LKML linux-kernel@vger.kernel.org; dri-devel dri-devel@lists.freedesktop.org; Chris Wilson chris@chris-wilson.co.uk; Linux Fbdev development list linux-fbdev@vger.kernel.org; Iouri Tarassov iourit@microsoft.com; Deucher, Alexander alexander.deucher@amd.com; KY Srinivasan kys@microsoft.com; Wei Liu wei.liu@kernel.org; Hawking Zhang Hawking.Zhang@amd.com Subject: Re: [EXTERNAL] Re: [RFC PATCH 0/4] DirectX on Linux
Hi Steve,
Sounds all good, some more comments and details below.
On Wed, May 20, 2020 at 5:47 AM Steve Pronovost <spronovo@microsoft.commailto:spronovo@microsoft.com> wrote: Hey guys,
Thanks for the discussion. I may not be able to immediately answer all of your questions, but I'll do my best 😊.
drivers/hyperv sounds like it could be a better location. We weren't too sure where to put this, we though /drivers/gpu would be appropriate given this deal with GPUs, but I get your point... this is a vGPU driver that really only works when being run under Hyper-V, so drivers/hyperv is likely more appropriate.
I think "it's a virtual gpu" is the wrong sales pitch, as is "only runs on $platform". We have lots of drm drivers in drivers/gpu that fit that bill. The better pitch I think is "it's a not a gpu, it's a dx12 protocol pipe" and "we actually do not want to integrate with the linux gpu ecosystem and primitives, we want to integrate with dx12 ecosystem and primitives to make the seamless rdp/rail/vail stuff work nicely". Below some more thoughts on the technical said.
[spronovo] Agreed. As I mentioned in another reply, that protocol isn’t tied to DX… but the point you are making is still valid. This is really a projection of the Windows native abstraction of a GPU that windows user mode driver (dx, gl, cl, vulkan, cuda, etc…) are familiar with and use to communicate with the GPU… This effectively enable porting of these user mode driver to Linux inside of WSL and allow them to share the GPU with the host. Our goal is to offer CL/EGL/GLX/CUDA/… API support for applications running inside of WSL and integrate their output on the Windows desktop through the Wayland compositor we are building. The fact that we are using layer to implement some of these APIs (to reduce our partners work among other thing) is just an implementation details that most application shouldn’t have to worry about… “it just works” 😊. From that perspective we’re fine moving the driver under a different node than /driver/gpu 😊.
In term of presentation, I need to clarify a few things. We announced today that we're also adding support for Linux GUI applications. The way this will work is roughly as follow. We're writing a Wayland compositor that will essentially bridge over RDP-RAIL (RAIL=Remote Application Integrated Locally). We're starting from a Weston base. Weston already has an RDP Backend, but that's for a full desktop remoting scheme. Weston draws a desktop and remote it over RDP... and then you can peek at that desktop using an rdp client on the Windows side. RAIL works differently. In that case our wayland compositor no longer paint a desktop... instead it simply forward individual visual / wl_surface over the RDP RAIL channel such that these visual can be displayed on the Windows desktop. The RDP client create proxy window for each of these top level visual and their content is filled with the data coming over the RDP channel. All pixels are owned by the RDP server/WSL... so these windows looks different than native window are they are painted and themed by WSL. The proxy window on the host gather input and inject back over RDP... This is essentially how application remoting works on windows and this is all publicly documented as part of the various RDP protocol specification. As a matter of fact, for the RDP server on the Weston side we are looking at continue to leverage FreeRDP (and provide fixes/enhancement as needed to the public project). Further, we're looking at further improvement down this path to avoid having to copy the content over the RAIL channel and instead just share/swap buffer between the guest and the host. We have extension to the RDP protocol, called VAIL (Virtualized Application Integrated Locally) which does that today. Today this is only use in Windows on Windows for very specific scenario. We're looking at extending the public RDP protocol with these VAIL extension to make this an official Microsoft supported protocol which would allow us to target this in WSL. We have finished designing this part in details. Our goal would be to leverage something along the line of wl_drm, dma-buf, dma-fence, etc... This compositor and all our contribution to FreeRDP will be fully open source, including our design doc. We're not quite sure yet whether this will be offered as a separate project entirely distinct from it's Weston root... or if we'll propose an extension to Weston to operate in this mode. We would like to build it such that in theory any Wayland compositor could add support for this mode of operation if they want to remote application to a Windows host (over the network, or on the same box).
Sounds like a solid plan for presentation. I think this is all up to wayland/weston folks to figure out with you, from the kernel side I have only one concern (and I discussed that with a few folks already on irc, I think they're at least on this thread involved within microsoft too in some form): If we do integrate with linux concepts like wl_drm/dma-buf/fence and so on then we end up with a normal gpu driver, with with lots of blobby components all around that can't be opened (since large chunks written by hw vendors, so nothing microsoft can do about them). That's the awkward exception (why microsoft but not other gpu hw vendors/plaforms/whatever?) that we need to avoid.
But wayland doesn't really need dma-buf and the wl_drm protocols afaiui, as long as the egl extensions work you can have whatever private wayland protocol in your winsys code you want to shovel the buffers and syncobj from client to the wayland-rdp-rail compositor. If that uses dx12 native handles for these things we side-step the awkward exception question for linux gpu stack since it all stays 100% contained in drivers/hv. Just try to avoid the nvidia fail of insisting that you need your own set of egl extensions (egl_streams and egl_drm_kms_reinvented_but_badly and a few others iirc) for everything, that makes it very painful for all the compositor projects since they need duplicated code for no good reason :-)
[spronovo] Agreed. I haven’t fully dig into the details for the VAIL enhancement to our Wayland compositor that enables buffer sharing/swapping. This is something we do on window and are quite familiar with the mechanic/concept around this… but new to doing this under Wayland. I’m sharing early thinking so apology if I get some stuff wrong 😊… I fundamentally agree with the spirit of what you are saying. Fundamentally our goal here is that we want to avoid having Wayland client do anything non-standard to interop with our projected graphics stack. We want Wayland client to work unmodified. It’s not entirely clear to me at the moment whether this buffer handshake is entirely hidden inside of Mesa (where we could do whatever we want as you say, and just reuse what we already have) or visible to some Wayland client, in which case we would want to ensure we present a public interface that is compatible. I gather from your suggestion that it is the former. In a nutshell we want to be as standard as possible so things just work and avoid having to chase a never ending set of app compat problem 😊.
Also looking at the fun the virtio folks have right now trying to get virgl dma-buf handles in the guest shared with other virtio devices or virgl instance in some other guest and getting it working ... If you avoid the dma-buf integration that might also sidestep a lot of technical headaches and be the simpler solution. At least as long as we're talking about wsl2 only.
[spronovo] Yeah, will definitely look into this.
We see /dev/dxg really as a projection of the GPU when running in WSL such that the GPU can be shared between WSL and the host... not something that would coexist "at the same time" with a real DRM GPU.
We have consider the possibility of bringing DX to Linux with no Windows cord attached. I'm not ready to discuss this at this time 😊... but in the hypothetical that we were do this, DX would be running on top of DRI/DRM on native Linux. We likely would be contributing some changes to DRM to address area of divergence and get better mapping for our user mode driver, but we wouldn't try to shoehorn /dev/dxg into the picture. In that hypothetical world, we would essentially have DX target DRM on native Linux and DX continue to target DXG in WSL to share the GPU with the host. I think this further reinforce the point you guys were making that the right place for our current dxgkrnl driver to live in would be /drivers/hyperv/dxgkrnl. In insight, I totally agree 😊.
We had a pile of discussions on irc about some of the ideas floating around for extending drm with some of the newer memory/sync concepts. So very much interested, but that's indeed a different fish and most likely will look a lot different to the dxgkrnl interface, while ofc still aiming to give the same power to apps. At least that's the goal, but aside from some prototype in some place this really doesn't exist yet anywhere.
[spronovo] There are indeed some very interesting and hard challenges if we wanted to do this 😊. But yeah, this would be totally separate from the dxgkrnl interface. Dxgkrnl is really meant for WSL (or Linux in a Windows hosted VM)... which we should have made clearer and avoid some of the confusion.
I think this cover all questions, let me know if I missed anything.
I think it's all good.
Cheers, Daniel
Thanks, Steve
-----Original Message----- From: Daniel Vetter <daniel@ffwll.chmailto:daniel@ffwll.ch> Sent: Tuesday, May 19, 2020 4:01 PM To: Dave Airlie <airlied@gmail.commailto:airlied@gmail.com> Cc: Sasha Levin <sashal@kernel.orgmailto:sashal@kernel.org>; linux-hyperv@vger.kernel.orgmailto:linux-hyperv@vger.kernel.org; Stephen Hemminger <sthemmin@microsoft.commailto:sthemmin@microsoft.com>; Ursulin, Tvrtko <tvrtko.ursulin@intel.commailto:tvrtko.ursulin@intel.com>; Greg Kroah-Hartman <gregkh@linuxfoundation.orgmailto:gregkh@linuxfoundation.org>; Haiyang Zhang <haiyangz@microsoft.commailto:haiyangz@microsoft.com>; LKML <linux-kernel@vger.kernel.orgmailto:linux-kernel@vger.kernel.org>; dri-devel <dri-devel@lists.freedesktop.orgmailto:dri-devel@lists.freedesktop.org>; Chris Wilson <chris@chris-wilson.co.ukmailto:chris@chris-wilson.co.uk>; Steve Pronovost <spronovo@microsoft.commailto:spronovo@microsoft.com>; Linux Fbdev development list <linux-fbdev@vger.kernel.orgmailto:linux-fbdev@vger.kernel.org>; Iouri Tarassov <iourit@microsoft.commailto:iourit@microsoft.com>; Deucher, Alexander <alexander.deucher@amd.commailto:alexander.deucher@amd.com>; KY Srinivasan <kys@microsoft.commailto:kys@microsoft.com>; Wei Liu <wei.liu@kernel.orgmailto:wei.liu@kernel.org>; Hawking Zhang <Hawking.Zhang@amd.commailto:Hawking.Zhang@amd.com> Subject: [EXTERNAL] Re: [RFC PATCH 0/4] DirectX on Linux
On Wed, May 20, 2020 at 12:42 AM Dave Airlie <airlied@gmail.commailto:airlied@gmail.com> wrote:
On Wed, 20 May 2020 at 02:33, Sasha Levin <sashal@kernel.orgmailto:sashal@kernel.org> wrote:
There is a blog post that goes into more detail about the bigger picture, and walks through all the required pieces to make this work. It is available here: https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Fde vblogs.microsoft.comhttps://nam06.safelinks.protection.outlook.com/?url=http%3A%2F%2Fvblogs.microsoft.com%2F&data=02%7C01%7Cspronovo%40microsoft.com%7C5e9c6fb0890844aef87c08d7fc91265e%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C637255572659005449&sdata=gVJvYw%2BRkrdugZw5LzWbYXzoAoT2KhVjRZLs%2Ft%2BG3aY%3D&reserved=0%2Fdirectx%2Fdirectx-heart-linux&data=02%7C01%7Cspronovo%40microsoft.comhttps://nam06.safelinks.protection.outlook.com/?url=http%3A%2F%2F40microsoft.com%2F&data=02%7C01%7Cspronovo%40microsoft.com%7C5e9c6fb0890844aef87c08d7fc91265e%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C637255572659015442&sdata=0TfOBiUpDLctMq7EZmK7N3xdPFeqauzsXJPODRCErXU%3D&reserved=0%7C3f18e46192b24cccf6a008d7fc489063%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C637255260910730243&sdata=IRRknzg%2F6Myzj3JXESN7GgmN6AcUV3DxhL95P%2ButtCw%3D&reserved=0 . The rest of this cover letter will focus on the Linux Kernel bits.
Overview
This is the first draft of the Microsoft Virtual GPU (vGPU) driver. The driver exposes a paravirtualized GPU to user mode applications running in a virtual machine on a Windows host. This enables hardware acceleration in environment such as WSL (Windows Subsystem for Linux) where the Linux virtual machine is able to share the GPU with the Windows host.
The projection is accomplished by exposing the WDDM (Windows Display Driver Model) interface as a set of IOCTL. This allows APIs and user mode driver written against the WDDM GPU abstraction on Windows to be ported to run within a Linux environment. This enables the port of the D3D12 and DirectML APIs as well as their associated user mode driver to Linux. This also enables third party APIs, such as the popular NVIDIA Cuda compute API, to be hardware accelerated within a WSL environment.
Only the rendering/compute aspect of the GPU are projected to the virtual machine, no display functionality is exposed. Further, at this time there are no presentation integration. So although the D3D12 API can be use to render graphics offscreen, there is no path (yet) for pixel to flow from the Linux environment back onto the Windows host desktop. This GPU stack is effectively side-by-side with the native Linux graphics stack.
Okay I've had some caffiene and absorbed some more of this.
This is a driver that connects a binary blob interface in the Windows kernel drivers to a binary blob that you run inside a Linux guest. It's a binary transport between two binary pieces. Personally this holds little of interest to me, I can see why it might be nice to have this upstream, but I don't forsee any other Linux distributor ever enabling it or having to ship it, it's purely a WSL2 pipe. I'm not saying I'd be happy to see this in the tree, since I don't see the value of maintaining it upstream, but it probably should just exists in a drivers/hyperv type area.
Yup as-is (especially with the goal of this being aimed at ml/compute only) drivers/hyperv sounds a bunch more reasonable than drivers/gpu.
Having said that, I hit one stumbling block: "Further, at this time there are no presentation integration. "
If we upstream this driver as-is into some hyperv specific place, and you decide to add presentation integration this is more than likely going to mean you will want to interact with dma-bufs and dma-fences. If the driver is hidden away in a hyperv place it's likely we won't even notice that feature landing until it's too late.
I've recently added regex matches to MAINTAINERS so we'll see dma_buf/fence/anything show up on dri-devel. So that part is solved hopefully.
I would like to see a coherent plan for presentation support (not code, just an architectural diagram), because I think when you contemplate how that works it will change the picture of how this driver looks and intergrates into the rest of the Linux graphics ecosystem.
Yeah once we have the feature-creep to presentation support all the integration fun starts, with all the questions about "why does this not look like any other linux gpu driver". We have that already with nvidia insisting they just can't implement any of the upstream gpu uapi we have, but at least they're not in-tree, so not our problem from an upstream maintainership pov.
But once this dx12 pipe is landed and then we want to extend it it's still going to have all the "we can't ever release the sources to any of the parts we usually expect to be open for gpu drivers in upstream" problems. Then we're stuck at a rather awkward point of why one vendor gets an exception and all the others dont.
As-is I'd rather this didn't land under my purview, since I don't see the value this adds to the Linux ecosystem at all, and I think it's important when putting a burden on upstream that you provide some value.
Well there is some in the form of "more hw/platform support". But given that gpus evolved rather fast, including the entire integration ecosystem (it's by far not just the hw drivers that move quickly). So that value deprecates a lot faster than for other kernel subsystems. And all that's left is the pain of not breaking anything without actually being able to evolve the overall stack in any meaningful way. -Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - +https://nam06.safelinks.protection.outlook.com/?url=http%3A%2F%2Fblog.f +fwll.ch%2F&data=02%7C01%7Cspronovo%40microsoft.com%7C3f18e46192b24c +ccf6a008d7fc489063%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C6372552https://nam06.safelinks.protection.outlook.com/?url=http%3A%2F%2Fblog.f+fwll.ch%2F&data=02%7C01%7Cspronovo%40microsoft.com%7C3f18e46192b24c+ccf6a008d7fc489063%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C6372552 +60910735230&sdata=hAIV1wJ29WF9IXTvJm3dr4StCwPzF0GdO2iWPyfnElg%3D&am +p;reserved=0
-- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.chhttps://nam06.safelinks.protection.outlook.com/?url=http%3A%2F%2Fblog.ffwll.ch%2F&data=02%7C01%7Cspronovo%40microsoft.com%7C5e9c6fb0890844aef87c08d7fc91265e%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C637255572659025436&sdata=me6tZuh9c4YKPTcOdY5%2FlgE%2B52usSI%2F5QRUCsCy%2FTfQ%3D&reserved=0
[resending as plain text, sorry about that]
Thanks Daniel, more below.
From: Daniel Vetter mailto:daniel@ffwll.ch Sent: Wednesday, May 20, 2020 12:41 AM To: Steve Pronovost mailto:spronovo@microsoft.com Cc: Dave Airlie mailto:airlied@gmail.com; Sasha Levin mailto:sashal@kernel.org; mailto:linux-hyperv@vger.kernel.org; Stephen Hemminger mailto:sthemmin@microsoft.com; Ursulin, Tvrtko mailto:tvrtko.ursulin@intel.com; Greg Kroah-Hartman mailto:gregkh@linuxfoundation.org; Haiyang Zhang mailto:haiyangz@microsoft.com; LKML mailto:linux-kernel@vger.kernel.org; dri-devel mailto:dri-devel@lists.freedesktop.org; Chris Wilson mailto:chris@chris-wilson.co.uk; Linux Fbdev development list mailto:linux-fbdev@vger.kernel.org; Iouri Tarassov mailto:iourit@microsoft.com; Deucher, Alexander mailto:alexander.deucher@amd.com; KY Srinivasan mailto:kys@microsoft.com; Wei Liu mailto:wei.liu@kernel.org; Hawking Zhang mailto:Hawking.Zhang@amd.com Subject: Re: [EXTERNAL] Re: [RFC PATCH 0/4] DirectX on Linux
Hi Steve,
Sounds all good, some more comments and details below.
On Wed, May 20, 2020 at 5:47 AM Steve Pronovost mailto:spronovo@microsoft.com wrote: Hey guys,
Thanks for the discussion. I may not be able to immediately answer all of your questions, but I'll do my best 😊.
drivers/hyperv sounds like it could be a better location. We weren't too sure where to put this, we though /drivers/gpu would be appropriate given this deal with GPUs, but I get your point... this is a vGPU driver that really only works when being run under Hyper-V, so drivers/hyperv is likely more appropriate.
I think "it's a virtual gpu" is the wrong sales pitch, as is "only runs on $platform". We have lots of drm drivers in drivers/gpu that fit that bill. The better pitch I think is "it's a not a gpu, it's a dx12 protocol pipe" and "we actually do not want to integrate with the linux gpu ecosystem and primitives, we want to integrate with dx12 ecosystem and primitives to make the seamless rdp/rail/vail stuff work nicely". Below some more thoughts on the technical said.
[spronovo] Agreed. As I mentioned in another reply, that protocol isn’t tied to DX… but the point you are making is still valid. This is really a projection of the Windows native abstraction of a GPU that windows user mode driver (dx, gl, cl, vulkan, cuda, etc…) are familiar with and use to communicate with the GPU… This effectively enable porting of these user mode driver to Linux inside of WSL and allow them to share the GPU with the host. Our goal is to offer CL/EGL/GLX/CUDA/… API support for applications running inside of WSL and integrate their output on the Windows desktop through the Wayland compositor we are building. The fact that we are using layer to implement some of these APIs (to reduce our partners work among other thing) is just an implementation details that most application shouldn’t have to worry about… “it just works” 😊. From that perspective we’re fine moving the driver under a different node than /driver/gpu 😊.
In term of presentation, I need to clarify a few things. We announced today that we're also adding support for Linux GUI applications. The way this will work is roughly as follow. We're writing a Wayland compositor that will essentially bridge over RDP-RAIL (RAIL=Remote Application Integrated Locally). We're starting from a Weston base. Weston already has an RDP Backend, but that's for a full desktop remoting scheme. Weston draws a desktop and remote it over RDP... and then you can peek at that desktop using an rdp client on the Windows side. RAIL works differently. In that case our wayland compositor no longer paint a desktop... instead it simply forward individual visual / wl_surface over the RDP RAIL channel such that these visual can be displayed on the Windows desktop. The RDP client create proxy window for each of these top level visual and their content is filled with the data coming over the RDP channel. All pixels are owned by the RDP server/WSL... so these windows looks different than native window are they are painted and themed by WSL. The proxy window on the host gather input and inject back over RDP... This is essentially how application remoting works on windows and this is all publicly documented as part of the various RDP protocol specification. As a matter of fact, for the RDP server on the Weston side we are looking at continue to leverage FreeRDP (and provide fixes/enhancement as needed to the public project). Further, we're looking at further improvement down this path to avoid having to copy the content over the RAIL channel and instead just share/swap buffer between the guest and the host. We have extension to the RDP protocol, called VAIL (Virtualized Application Integrated Locally) which does that today. Today this is only use in Windows on Windows for very specific scenario. We're looking at extending the public RDP protocol with these VAIL extension to make this an official Microsoft supported protocol which would allow us to target this in WSL. We have finished designing this part in details. Our goal would be to leverage something along the line of wl_drm, dma-buf, dma-fence, etc... This compositor and all our contribution to FreeRDP will be fully open source, including our design doc. We're not quite sure yet whether this will be offered as a separate project entirely distinct from it's Weston root... or if we'll propose an extension to Weston to operate in this mode. We would like to build it such that in theory any Wayland compositor could add support for this mode of operation if they want to remote application to a Windows host (over the network, or on the same box).
Sounds like a solid plan for presentation. I think this is all up to wayland/weston folks to figure out with you, from the kernel side I have only one concern (and I discussed that with a few folks already on irc, I think they're at least on this thread involved within microsoft too in some form): If we do integrate with linux concepts like wl_drm/dma-buf/fence and so on then we end up with a normal gpu driver, with with lots of blobby components all around that can't be opened (since large chunks written by hw vendors, so nothing microsoft can do about them). That's the awkward exception (why microsoft but not other gpu hw vendors/plaforms/whatever?) that we need to avoid.
But wayland doesn't really need dma-buf and the wl_drm protocols afaiui, as long as the egl extensions work you can have whatever private wayland protocol in your winsys code you want to shovel the buffers and syncobj from client to the wayland-rdp-rail compositor. If that uses dx12 native handles for these things we side-step the awkward exception question for linux gpu stack since it all stays 100% contained in drivers/hv. Just try to avoid the nvidia fail of insisting that you need your own set of egl extensions (egl_streams and egl_drm_kms_reinvented_but_badly and a few others iirc) for everything, that makes it very painful for all the compositor projects since they need duplicated code for no good reason :-)
[spronovo] Agreed. I haven’t fully dig into the details for the VAIL enhancement to our Wayland compositor that enables buffer sharing/swapping. This is something we do on window and are quite familiar with the mechanic/concept around this… but new to doing this under Wayland. I’m sharing early thinking so apology if I get some stuff wrong 😊… I fundamentally agree with the spirit of what you are saying. Fundamentally our goal here is that we want to avoid having Wayland client do anything non-standard to interop with our projected graphics stack. We want Wayland client to work unmodified. It’s not entirely clear to me at the moment whether this buffer handshake is entirely hidden inside of Mesa (where we could do whatever we want as you say, and just reuse what we already have) or visible to some Wayland client, in which case we would want to ensure we present a public interface that is compatible. I gather from your suggestion that it is the former. In a nutshell we want to be as standard as possible so things just work and avoid having to chase a never ending set of app compat problem 😊.
Also looking at the fun the virtio folks have right now trying to get virgl dma-buf handles in the guest shared with other virtio devices or virgl instance in some other guest and getting it working ... If you avoid the dma-buf integration that might also sidestep a lot of technical headaches and be the simpler solution. At least as long as we're talking about wsl2 only.
[spronovo] Yeah, will definitely look into this.
We see /dev/dxg really as a projection of the GPU when running in WSL such that the GPU can be shared between WSL and the host... not something that would coexist "at the same time" with a real DRM GPU.
We have consider the possibility of bringing DX to Linux with no Windows cord attached. I'm not ready to discuss this at this time 😊... but in the hypothetical that we were do this, DX would be running on top of DRI/DRM on native Linux. We likely would be contributing some changes to DRM to address area of divergence and get better mapping for our user mode driver, but we wouldn't try to shoehorn /dev/dxg into the picture. In that hypothetical world, we would essentially have DX target DRM on native Linux and DX continue to target DXG in WSL to share the GPU with the host. I think this further reinforce the point you guys were making that the right place for our current dxgkrnl driver to live in would be /drivers/hyperv/dxgkrnl. In insight, I totally agree 😊. We had a pile of discussions on irc about some of the ideas floating around for extending drm with some of the newer memory/sync concepts. So very much interested, but that's indeed a different fish and most likely will look a lot different to the dxgkrnl interface, while ofc still aiming to give the same power to apps. At least that's the goal, but aside from some prototype in some place this really doesn't exist yet anywhere.
[spronovo] There are indeed some very interesting and hard challenges if we wanted to do this 😊. But yeah, this would be totally separate from the dxgkrnl interface. Dxgkrnl is really meant for WSL (or Linux in a Windows hosted VM)... which we should have made clearer and avoid some of the confusion.
I think this cover all questions, let me know if I missed anything.
I think it's all good.
Cheers, Daniel
Thanks, Steve
-----Original Message----- From: Daniel Vetter mailto:daniel@ffwll.ch Sent: Tuesday, May 19, 2020 4:01 PM To: Dave Airlie mailto:airlied@gmail.com Cc: Sasha Levin mailto:sashal@kernel.org; mailto:linux-hyperv@vger.kernel.org; Stephen Hemminger mailto:sthemmin@microsoft.com; Ursulin, Tvrtko mailto:tvrtko.ursulin@intel.com; Greg Kroah-Hartman mailto:gregkh@linuxfoundation.org; Haiyang Zhang mailto:haiyangz@microsoft.com; LKML mailto:linux-kernel@vger.kernel.org; dri-devel mailto:dri-devel@lists.freedesktop.org; Chris Wilson mailto:chris@chris-wilson.co.uk; Steve Pronovost mailto:spronovo@microsoft.com; Linux Fbdev development list mailto:linux-fbdev@vger.kernel.org; Iouri Tarassov mailto:iourit@microsoft.com; Deucher, Alexander mailto:alexander.deucher@amd.com; KY Srinivasan mailto:kys@microsoft.com; Wei Liu mailto:wei.liu@kernel.org; Hawking Zhang mailto:Hawking.Zhang@amd.com Subject: [EXTERNAL] Re: [RFC PATCH 0/4] DirectX on Linux
On Wed, May 20, 2020 at 12:42 AM Dave Airlie mailto:airlied@gmail.com wrote:
On Wed, 20 May 2020 at 02:33, Sasha Levin mailto:sashal@kernel.org wrote:
There is a blog post that goes into more detail about the bigger picture, and walks through all the required pieces to make this work. It is available here: https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Fde https://nam06.safelinks.protection.outlook.com/?url=http%3A%2F%2Fvblogs.micr... . The rest of this cover letter will focus on the Linux Kernel bits.
Overview
This is the first draft of the Microsoft Virtual GPU (vGPU) driver. The driver exposes a paravirtualized GPU to user mode applications running in a virtual machine on a Windows host. This enables hardware acceleration in environment such as WSL (Windows Subsystem for Linux) where the Linux virtual machine is able to share the GPU with the Windows host.
The projection is accomplished by exposing the WDDM (Windows Display Driver Model) interface as a set of IOCTL. This allows APIs and user mode driver written against the WDDM GPU abstraction on Windows to be ported to run within a Linux environment. This enables the port of the D3D12 and DirectML APIs as well as their associated user mode driver to Linux. This also enables third party APIs, such as the popular NVIDIA Cuda compute API, to be hardware accelerated within a WSL environment.
Only the rendering/compute aspect of the GPU are projected to the virtual machine, no display functionality is exposed. Further, at this time there are no presentation integration. So although the D3D12 API can be use to render graphics offscreen, there is no path (yet) for pixel to flow from the Linux environment back onto the Windows host desktop. This GPU stack is effectively side-by-side with the native Linux graphics stack.
Okay I've had some caffiene and absorbed some more of this.
This is a driver that connects a binary blob interface in the Windows kernel drivers to a binary blob that you run inside a Linux guest. It's a binary transport between two binary pieces. Personally this holds little of interest to me, I can see why it might be nice to have this upstream, but I don't forsee any other Linux distributor ever enabling it or having to ship it, it's purely a WSL2 pipe. I'm not saying I'd be happy to see this in the tree, since I don't see the value of maintaining it upstream, but it probably should just exists in a drivers/hyperv type area.
Yup as-is (especially with the goal of this being aimed at ml/compute only) drivers/hyperv sounds a bunch more reasonable than drivers/gpu.
Having said that, I hit one stumbling block: "Further, at this time there are no presentation integration. "
If we upstream this driver as-is into some hyperv specific place, and you decide to add presentation integration this is more than likely going to mean you will want to interact with dma-bufs and dma-fences. If the driver is hidden away in a hyperv place it's likely we won't even notice that feature landing until it's too late.
I've recently added regex matches to MAINTAINERS so we'll see dma_buf/fence/anything show up on dri-devel. So that part is solved hopefully.
I would like to see a coherent plan for presentation support (not code, just an architectural diagram), because I think when you contemplate how that works it will change the picture of how this driver looks and intergrates into the rest of the Linux graphics ecosystem.
Yeah once we have the feature-creep to presentation support all the integration fun starts, with all the questions about "why does this not look like any other linux gpu driver". We have that already with nvidia insisting they just can't implement any of the upstream gpu uapi we have, but at least they're not in-tree, so not our problem from an upstream maintainership pov.
But once this dx12 pipe is landed and then we want to extend it it's still going to have all the "we can't ever release the sources to any of the parts we usually expect to be open for gpu drivers in upstream" problems. Then we're stuck at a rather awkward point of why one vendor gets an exception and all the others dont.
As-is I'd rather this didn't land under my purview, since I don't see the value this adds to the Linux ecosystem at all, and I think it's important when putting a burden on upstream that you provide some value.
Well there is some in the form of "more hw/platform support". But given that gpus evolved rather fast, including the entire integration ecosystem (it's by far not just the hw drivers that move quickly). So that value deprecates a lot faster than for other kernel subsystems. And all that's left is the pain of not breaking anything without actually being able to evolve the overall stack in any meaningful way. -Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - +https://nam06.safelinks.protection.outlook.com/?url=http%3A%2F%2Fblog.f+fwll... +60910735230&sdata=hAIV1wJ29WF9IXTvJm3dr4StCwPzF0GdO2iWPyfnElg%3D&am +p;reserved=0
Hi!
Thanks for the discussion. I may not be able to immediately answer all of your questions, but I'll do my best ????.
Could you do something with your email settings? Because this is not how you should use email on lkml. "[EXTERNAL]" in the subject, top-posting, unwrapped lines...
Thank you, Pavel
On Wed, 20 May 2020 at 08:42, Dave Airlie airlied@gmail.com wrote:
On Wed, 20 May 2020 at 02:33, Sasha Levin sashal@kernel.org wrote:
There is a blog post that goes into more detail about the bigger picture, and walks through all the required pieces to make this work. It is available here: https://devblogs.microsoft.com/directx/directx-heart-linux . The rest of this cover letter will focus on the Linux Kernel bits.
Overview
This is the first draft of the Microsoft Virtual GPU (vGPU) driver. The driver exposes a paravirtualized GPU to user mode applications running in a virtual machine on a Windows host. This enables hardware acceleration in environment such as WSL (Windows Subsystem for Linux) where the Linux virtual machine is able to share the GPU with the Windows host.
The projection is accomplished by exposing the WDDM (Windows Display Driver Model) interface as a set of IOCTL. This allows APIs and user mode driver written against the WDDM GPU abstraction on Windows to be ported to run within a Linux environment. This enables the port of the D3D12 and DirectML APIs as well as their associated user mode driver to Linux. This also enables third party APIs, such as the popular NVIDIA Cuda compute API, to be hardware accelerated within a WSL environment.
Only the rendering/compute aspect of the GPU are projected to the virtual machine, no display functionality is exposed. Further, at this time there are no presentation integration. So although the D3D12 API can be use to render graphics offscreen, there is no path (yet) for pixel to flow from the Linux environment back onto the Windows host desktop. This GPU stack is effectively side-by-side with the native Linux graphics stack.
Okay I've had some caffiene and absorbed some more of this.
This is a driver that connects a binary blob interface in the Windows kernel drivers to a binary blob that you run inside a Linux guest. It's a binary transport between two binary pieces. Personally this holds little of interest to me, I can see why it might be nice to have this upstream, but I don't forsee any other Linux distributor ever enabling it or having to ship it, it's purely a WSL2 pipe. I'm not saying I'd be happy to see this in the tree, since I don't see the value of maintaining it upstream, but it probably should just exists in a drivers/hyperv type area.
Having said that, I hit one stumbling block: "Further, at this time there are no presentation integration. "
If we upstream this driver as-is into some hyperv specific place, and you decide to add presentation integration this is more than likely going to mean you will want to interact with dma-bufs and dma-fences. If the driver is hidden away in a hyperv place it's likely we won't even notice that feature landing until it's too late.
I would like to see a coherent plan for presentation support (not code, just an architectural diagram), because I think when you contemplate how that works it will change the picture of how this driver looks and intergrates into the rest of the Linux graphics ecosystem.
As-is I'd rather this didn't land under my purview, since I don't see the value this adds to the Linux ecosystem at all, and I think it's important when putting a burden on upstream that you provide some value.
I also have another concern from a legal standpoint I'd rather not review the ioctl part of this. I'd probably request under DRI developers abstain as well.
This is a Windows kernel API being smashed into a Linux driver. I don't want to be tainted by knowledge of an API that I've no idea of the legal status of derived works. (it this all covered patent wise under OIN?)
I don't want to ever be accused of designing a Linux kernel API with illgotten D3DKMT knowledge, I feel tainting myself with knowledge of a properietary API might cause derived work issues.
Dave.
Having said that, I hit one stumbling block: "Further, at this time there are no presentation integration. "
If we upstream this driver as-is into some hyperv specific place, and you decide to add presentation integration this is more than likely going to mean you will want to interact with dma-bufs and dma-fences. If the driver is hidden away in a hyperv place it's likely we won't even notice that feature landing until it's too late.
I would like to see a coherent plan for presentation support (not code, just an architectural diagram), because I think when you contemplate how that works it will change the picture of how this driver looks and intergrates into the rest of the Linux graphics ecosystem.
As-is I'd rather this didn't land under my purview, since I don't see the value this adds to the Linux ecosystem at all, and I think it's important when putting a burden on upstream that you provide some value.
I also have another concern from a legal standpoint I'd rather not review the ioctl part of this. I'd probably request under DRI developers abstain as well.
This is a Windows kernel API being smashed into a Linux driver. I don't want to be tainted by knowledge of an API that I've no idea of the legal status of derived works. (it this all covered patent wise under OIN?)
If you can't look onto it, perhaps it is not suitable to merge into kernel...?
What would be legal requirements so this is "safe to look at"? We should really require submitter to meet them...
Pavel
On Tue, Jun 16, 2020 at 12:51:56PM +0200, Pavel Machek wrote:
Having said that, I hit one stumbling block: "Further, at this time there are no presentation integration. "
If we upstream this driver as-is into some hyperv specific place, and you decide to add presentation integration this is more than likely going to mean you will want to interact with dma-bufs and dma-fences. If the driver is hidden away in a hyperv place it's likely we won't even notice that feature landing until it's too late.
I would like to see a coherent plan for presentation support (not code, just an architectural diagram), because I think when you contemplate how that works it will change the picture of how this driver looks and intergrates into the rest of the Linux graphics ecosystem.
As-is I'd rather this didn't land under my purview, since I don't see the value this adds to the Linux ecosystem at all, and I think it's important when putting a burden on upstream that you provide some value.
I also have another concern from a legal standpoint I'd rather not review the ioctl part of this. I'd probably request under DRI developers abstain as well.
This is a Windows kernel API being smashed into a Linux driver. I don't want to be tainted by knowledge of an API that I've no idea of the legal status of derived works. (it this all covered patent wise under OIN?)
If you can't look onto it, perhaps it is not suitable to merge into kernel...?
What would be legal requirements so this is "safe to look at"? We should really require submitter to meet them...
Could you walk me through your view on what the function of the "Signed-off-by" tag is?
Hi
Am 19.05.20 um 18:32 schrieb Sasha Levin:
There is a blog post that goes into more detail about the bigger picture, and walks through all the required pieces to make this work. It is available here: https://devblogs.microsoft.com/directx/directx-heart-linux . The rest of this cover letter will focus on the Linux Kernel bits.
That's quite a surprise. Thanks for your efforts to contribute.
Overview
This is the first draft of the Microsoft Virtual GPU (vGPU) driver. The driver exposes a paravirtualized GPU to user mode applications running in a virtual machine on a Windows host. This enables hardware acceleration in environment such as WSL (Windows Subsystem for Linux) where the Linux virtual machine is able to share the GPU with the Windows host.
The projection is accomplished by exposing the WDDM (Windows Display Driver Model) interface as a set of IOCTL. This allows APIs and user mode driver written against the WDDM GPU abstraction on Windows to be ported to run within a Linux environment. This enables the port of the D3D12 and DirectML APIs as well as their associated user mode driver to Linux. This also enables third party APIs, such as the popular NVIDIA Cuda compute API, to be hardware accelerated within a WSL environment.
Only the rendering/compute aspect of the GPU are projected to the virtual machine, no display functionality is exposed. Further, at this time there are no presentation integration. So although the D3D12 API can be use to render graphics offscreen, there is no path (yet) for pixel to flow from the Linux environment back onto the Windows host desktop. This GPU stack is effectively side-by-side with the native Linux graphics stack.
The driver creates the /dev/dxg device, which can be opened by user mode application and handles their ioctls. The IOCTL interface to the driver is defined in dxgkmthk.h (Dxgkrnl Graphics Port Driver ioctl definitions). The interface matches the D3DKMT interface on Windows. Ioctls are implemented in ioctl.c.
Echoing what others said, you're not making a DRM driver. The driver should live outside of the DRM code.
I have one question about the driver API: on Windows, DirectX versions are loosly tied to Windows releases. So I guess you can change the kernel interface among DirectX versions?
If so, how would this work on Linux in the long term? If there ever is a DirectX 13 or 14 with incompatible kernel interfaces, how would you plan to update the Linux driver?
Best regards Thomas
When a VM starts, hyper-v on the host adds virtual GPU devices to the VM via the hyper-v driver. The host offers several VM bus channels to the VM: the global channel and one channel per virtual GPU, assigned to the VM.
The driver registers with the hyper-v driver (hv_driver) for the arrival of VM bus channels. dxg_probe_device recognizes the vGPU channels and creates the corresponding objects (dxgadapter for vGPUs and dxgglobal for the global channel).
The driver uses the hyper-V VM bus interface to communicate with the host. dxgvmbus.c implements the communication interface.
The global channel has 8GB of IO space assigned by the host. This space is managed by the host and used to give the guest direct CPU access to some allocations. Video memory is allocated on the host except in the case of existing_sysmem allocations. The Windows host allocates memory for the GPU on behalf of the guest. The Linux guest can access that memory by mapping GPU virtual address to allocations and then referencing those GPU virtual address from within GPU command buffers submitted to the GPU. For allocations which require CPU access, the allocation is mapped by the host into a location in the 8GB of IO space reserved in the guest for that purpose. The Windows host uses the nested CPU page table to ensure that this guest IO space always map to the correct location for the allocation as it may migrate between dedicated GPU memory (e.g. VRAM, firmware reserved DDR) and shared system memory (regular DDR) over its lifetime. The Linux guest maps a user mode CPU virtual address to an allocation IO space range for direct access by user mode APIs and drivers.
Implementation of LX_DXLOCK2 ioctl
We would appreciate your feedback on the implementation of the LX_DXLOCK2 ioctl.
This ioctl is used to get a CPU address to an allocation, which is resident in video/system memory on the host. The way it works:
The driver sends the Lock message to the host
The host allocates space in the VM IO space and maps it to the
allocation memory
The host returns the address in IO space for the mapped allocation
The driver (in dxg_map_iospace) allocates a user mode virtual address
range using vm_mmap and maps it to the IO space using io_remap_ofn_range)
- The VA is returned to the application
Internal objects
The following objects are created by the driver (defined in dxgkrnl.h):
dxgadapter - represents a virtual GPU
dxgprocess - tracks per process state (handle table of created objects, list of objects, etc.)
dxgdevice - a container for other objects (contexts, paging queues, allocations, GPU synchronization objects)
dxgcontext - represents thread of GPU execution for packet scheduling.
dxghwqueue - represents thread of GPU execution of hardware scheduling
dxgallocation - represents a GPU accessible allocation
dxgsyncobject - represents a GPU synchronization object
dxgresource - collection of dxgalloction objects
dxgsharedresource, dxgsharedsyncobj - helper objects to share objects between different dxgdevice objects, which can belong to different
processes
Object handles
All GPU objects, created by the driver, are accessible by a handle (d3dkmt_handle). Each process has its own handle table, which is implemented in hmgr.c. For each API visible object, created by the driver, there is an object, created on the host. For example, the is a dxgprocess object on the host for each dxgprocess object in the VM, etc. The object handles have the same value in the host and the VM, which is done to avoid translation from the guest handles to the host handles.
Signaling CPU events by the host
The WDDM interface provides a way to signal CPU event objects when execution of a context reached certain point. The way it is implemented:
application sends an event_fd via ioctl to the driver
eventfd_ctx_get is used to get a pointer to the file object (eventfd_ctx)
the pointer to sent the host via a VM bus message
when GPU execution reaches a certain point, the host sends a message to the VM with the event pointer
signal_guest_event() handles the messages and eventually eventfd_signal() is called.
Sasha Levin (4): gpu: dxgkrnl: core code gpu: dxgkrnl: hook up dxgkrnl Drivers: hv: vmbus: hook up dxgkrnl gpu: dxgkrnl: create a MAINTAINERS entry
MAINTAINERS | 7 + drivers/gpu/Makefile | 2 +- drivers/gpu/dxgkrnl/Kconfig | 10 + drivers/gpu/dxgkrnl/Makefile | 12 + drivers/gpu/dxgkrnl/d3dkmthk.h | 1635 +++++++++ drivers/gpu/dxgkrnl/dxgadapter.c | 1399 ++++++++ drivers/gpu/dxgkrnl/dxgkrnl.h | 913 ++++++ drivers/gpu/dxgkrnl/dxgmodule.c | 692 ++++ drivers/gpu/dxgkrnl/dxgprocess.c | 355 ++ drivers/gpu/dxgkrnl/dxgvmbus.c | 2955 +++++++++++++++++ drivers/gpu/dxgkrnl/dxgvmbus.h | 859 +++++ drivers/gpu/dxgkrnl/hmgr.c | 593 ++++ drivers/gpu/dxgkrnl/hmgr.h | 107 + drivers/gpu/dxgkrnl/ioctl.c | 5269 ++++++++++++++++++++++++++++++ drivers/gpu/dxgkrnl/misc.c | 280 ++ drivers/gpu/dxgkrnl/misc.h | 288 ++ drivers/video/Kconfig | 2 + include/linux/hyperv.h | 16 + 18 files changed, 15393 insertions(+), 1 deletion(-) create mode 100644 drivers/gpu/dxgkrnl/Kconfig create mode 100644 drivers/gpu/dxgkrnl/Makefile create mode 100644 drivers/gpu/dxgkrnl/d3dkmthk.h create mode 100644 drivers/gpu/dxgkrnl/dxgadapter.c create mode 100644 drivers/gpu/dxgkrnl/dxgkrnl.h create mode 100644 drivers/gpu/dxgkrnl/dxgmodule.c create mode 100644 drivers/gpu/dxgkrnl/dxgprocess.c create mode 100644 drivers/gpu/dxgkrnl/dxgvmbus.c create mode 100644 drivers/gpu/dxgkrnl/dxgvmbus.h create mode 100644 drivers/gpu/dxgkrnl/hmgr.c create mode 100644 drivers/gpu/dxgkrnl/hmgr.h create mode 100644 drivers/gpu/dxgkrnl/ioctl.c create mode 100644 drivers/gpu/dxgkrnl/misc.c create mode 100644 drivers/gpu/dxgkrnl/misc.h
Echoing what others said, you're not making a DRM driver. The driver should live outside of the DRM code.
Agreed, please see my earlier reply. We'll be moving the driver to drivers/hyperv node or something similar. Apology for the confusion here.
I have one question about the driver API: on Windows, DirectX versions are loosly tied to Windows releases. So I guess you can change the kernel interface among DirectX versions? If so, how would this work on Linux in the long term? If there ever is a DirectX 13 or 14 with incompatible kernel interfaces, how would you plan to update the Linux driver?
You should think of the communication over the VM Bus for the vGPU projection as a strongly versioned interface. We will be keeping compatibility with older version of that interface as it evolves over time so we can continue to run older guest (we already do). This protocol isn't actually tied to the DX API. It is a generic abstraction for the GPU that can be used for any APIs (for example the NVIDIA CUDA driver that we announced is going over the same protocol to access the GPU).
New version of user mode DX can either take advantage or sometime require new services from this kernel abstraction. This mean that pulling a new version of user mode DX can mean having to also pull a new version of this vGPU kernel driver. For WSL, these essentially ships together. The kernel driver ships as part of our WSL2 Linux Kernel integration. User mode DX bits ships with Windows.
-----Original Message----- From: Thomas Zimmermann tzimmermann@suse.de Sent: Wednesday, May 20, 2020 12:11 AM To: Sasha Levin sashal@kernel.org; alexander.deucher@amd.com; chris@chris-wilson.co.uk; ville.syrjala@linux.intel.com; Hawking.Zhang@amd.com; tvrtko.ursulin@intel.com Cc: linux-kernel@vger.kernel.org; linux-hyperv@vger.kernel.org; KY Srinivasan kys@microsoft.com; Haiyang Zhang haiyangz@microsoft.com; Stephen Hemminger sthemmin@microsoft.com; wei.liu@kernel.org; Steve Pronovost spronovo@microsoft.com; Iouri Tarassov iourit@microsoft.com; dri-devel@lists.freedesktop.org; linux-fbdev@vger.kernel.org; gregkh@linuxfoundation.org Subject: [EXTERNAL] Re: [RFC PATCH 0/4] DirectX on Linux
Hi
Am 19.05.20 um 18:32 schrieb Sasha Levin:
There is a blog post that goes into more detail about the bigger picture, and walks through all the required pieces to make this work. It is available here: https://devblogs.microsoft.com/directx/directx-heart-linux . The rest of this cover letter will focus on the Linux Kernel bits.
That's quite a surprise. Thanks for your efforts to contribute.
Overview
This is the first draft of the Microsoft Virtual GPU (vGPU) driver. The driver exposes a paravirtualized GPU to user mode applications running in a virtual machine on a Windows host. This enables hardware acceleration in environment such as WSL (Windows Subsystem for Linux) where the Linux virtual machine is able to share the GPU with the Windows host.
The projection is accomplished by exposing the WDDM (Windows Display Driver Model) interface as a set of IOCTL. This allows APIs and user mode driver written against the WDDM GPU abstraction on Windows to be ported to run within a Linux environment. This enables the port of the D3D12 and DirectML APIs as well as their associated user mode driver to Linux. This also enables third party APIs, such as the popular NVIDIA Cuda compute API, to be hardware accelerated within a WSL environment.
Only the rendering/compute aspect of the GPU are projected to the virtual machine, no display functionality is exposed. Further, at this time there are no presentation integration. So although the D3D12 API can be use to render graphics offscreen, there is no path (yet) for pixel to flow from the Linux environment back onto the Windows host desktop. This GPU stack is effectively side-by-side with the native Linux graphics stack.
The driver creates the /dev/dxg device, which can be opened by user mode application and handles their ioctls. The IOCTL interface to the driver is defined in dxgkmthk.h (Dxgkrnl Graphics Port Driver ioctl definitions). The interface matches the D3DKMT interface on Windows. Ioctls are implemented in ioctl.c.
Echoing what others said, you're not making a DRM driver. The driver should live outside of the DRM code.
I have one question about the driver API: on Windows, DirectX versions are loosly tied to Windows releases. So I guess you can change the kernel interface among DirectX versions?
If so, how would this work on Linux in the long term? If there ever is a DirectX 13 or 14 with incompatible kernel interfaces, how would you plan to update the Linux driver?
Best regards Thomas
When a VM starts, hyper-v on the host adds virtual GPU devices to the VM via the hyper-v driver. The host offers several VM bus channels to the VM: the global channel and one channel per virtual GPU, assigned to the VM.
The driver registers with the hyper-v driver (hv_driver) for the arrival of VM bus channels. dxg_probe_device recognizes the vGPU channels and creates the corresponding objects (dxgadapter for vGPUs and dxgglobal for the global channel).
The driver uses the hyper-V VM bus interface to communicate with the host. dxgvmbus.c implements the communication interface.
The global channel has 8GB of IO space assigned by the host. This space is managed by the host and used to give the guest direct CPU access to some allocations. Video memory is allocated on the host except in the case of existing_sysmem allocations. The Windows host allocates memory for the GPU on behalf of the guest. The Linux guest can access that memory by mapping GPU virtual address to allocations and then referencing those GPU virtual address from within GPU command buffers submitted to the GPU. For allocations which require CPU access, the allocation is mapped by the host into a location in the 8GB of IO space reserved in the guest for that purpose. The Windows host uses the nested CPU page table to ensure that this guest IO space always map to the correct location for the allocation as it may migrate between dedicated GPU memory (e.g. VRAM, firmware reserved DDR) and shared system memory (regular DDR) over its lifetime. The Linux guest maps a user mode CPU virtual address to an allocation IO space range for direct access by user mode APIs and drivers.
Implementation of LX_DXLOCK2 ioctl
We would appreciate your feedback on the implementation of the LX_DXLOCK2 ioctl.
This ioctl is used to get a CPU address to an allocation, which is resident in video/system memory on the host. The way it works:
The driver sends the Lock message to the host
The host allocates space in the VM IO space and maps it to the
allocation memory
The host returns the address in IO space for the mapped allocation
The driver (in dxg_map_iospace) allocates a user mode virtual
address range using vm_mmap and maps it to the IO space using io_remap_ofn_range)
- The VA is returned to the application
Internal objects
The following objects are created by the driver (defined in dxgkrnl.h):
dxgadapter - represents a virtual GPU
dxgprocess - tracks per process state (handle table of created objects, list of objects, etc.)
dxgdevice - a container for other objects (contexts, paging queues, allocations, GPU synchronization objects)
dxgcontext - represents thread of GPU execution for packet scheduling.
dxghwqueue - represents thread of GPU execution of hardware
scheduling
dxgallocation - represents a GPU accessible allocation
dxgsyncobject - represents a GPU synchronization object
dxgresource - collection of dxgalloction objects
dxgsharedresource, dxgsharedsyncobj - helper objects to share objects between different dxgdevice objects, which can belong to different
processes
Object handles
All GPU objects, created by the driver, are accessible by a handle (d3dkmt_handle). Each process has its own handle table, which is implemented in hmgr.c. For each API visible object, created by the driver, there is an object, created on the host. For example, the is a dxgprocess object on the host for each dxgprocess object in the VM, etc. The object handles have the same value in the host and the VM, which is done to avoid translation from the guest handles to the host handles.
Signaling CPU events by the host
The WDDM interface provides a way to signal CPU event objects when execution of a context reached certain point. The way it is implemented:
application sends an event_fd via ioctl to the driver
eventfd_ctx_get is used to get a pointer to the file object (eventfd_ctx)
the pointer to sent the host via a VM bus message
when GPU execution reaches a certain point, the host sends a message to the VM with the event pointer
signal_guest_event() handles the messages and eventually eventfd_signal() is called.
Sasha Levin (4): gpu: dxgkrnl: core code gpu: dxgkrnl: hook up dxgkrnl Drivers: hv: vmbus: hook up dxgkrnl gpu: dxgkrnl: create a MAINTAINERS entry
MAINTAINERS | 7 + drivers/gpu/Makefile | 2 +- drivers/gpu/dxgkrnl/Kconfig | 10 + drivers/gpu/dxgkrnl/Makefile | 12 + drivers/gpu/dxgkrnl/d3dkmthk.h | 1635 +++++++++ drivers/gpu/dxgkrnl/dxgadapter.c | 1399 ++++++++ drivers/gpu/dxgkrnl/dxgkrnl.h | 913 ++++++ drivers/gpu/dxgkrnl/dxgmodule.c | 692 ++++ drivers/gpu/dxgkrnl/dxgprocess.c | 355 ++ drivers/gpu/dxgkrnl/dxgvmbus.c | 2955 +++++++++++++++++ drivers/gpu/dxgkrnl/dxgvmbus.h | 859 +++++ drivers/gpu/dxgkrnl/hmgr.c | 593 ++++ drivers/gpu/dxgkrnl/hmgr.h | 107 + drivers/gpu/dxgkrnl/ioctl.c | 5269 ++++++++++++++++++++++++++++++ drivers/gpu/dxgkrnl/misc.c | 280 ++ drivers/gpu/dxgkrnl/misc.h | 288 ++ drivers/video/Kconfig | 2 + include/linux/hyperv.h | 16 + 18 files changed, 15393 insertions(+), 1 deletion(-) create mode 100644 drivers/gpu/dxgkrnl/Kconfig create mode 100644 drivers/gpu/dxgkrnl/Makefile create mode 100644 drivers/gpu/dxgkrnl/d3dkmthk.h create mode 100644 drivers/gpu/dxgkrnl/dxgadapter.c create mode 100644 drivers/gpu/dxgkrnl/dxgkrnl.h create mode 100644 drivers/gpu/dxgkrnl/dxgmodule.c create mode 100644 drivers/gpu/dxgkrnl/dxgprocess.c create mode 100644 drivers/gpu/dxgkrnl/dxgvmbus.c create mode 100644 drivers/gpu/dxgkrnl/dxgvmbus.h create mode 100644 drivers/gpu/dxgkrnl/hmgr.c create mode 100644 drivers/gpu/dxgkrnl/hmgr.h create mode 100644 drivers/gpu/dxgkrnl/ioctl.c create mode 100644 drivers/gpu/dxgkrnl/misc.c create mode 100644 drivers/gpu/dxgkrnl/misc.h
-- Thomas Zimmermann Graphics Driver Developer SUSE Software Solutions Germany GmbH Maxfeldstr. 5, 90409 Nürnberg, Germany (HRB 36809, AG Nürnberg) Geschäftsführer: Felix Imendörffer
Hi Steve,
thank you for the fast reply.
Am 20.05.20 um 09:42 schrieb Steve Pronovost:
Echoing what others said, you're not making a DRM driver. The driver should live outside of the DRM code.
Agreed, please see my earlier reply. We'll be moving the driver to drivers/hyperv node or something similar. Apology for the confusion here.
I have one question about the driver API: on Windows, DirectX versions are loosly tied to Windows releases. So I guess you can change the kernel interface among DirectX versions? If so, how would this work on Linux in the long term? If there ever is a DirectX 13 or 14 with incompatible kernel interfaces, how would you plan to update the Linux driver?
You should think of the communication over the VM Bus for the vGPU projection as a strongly versioned interface. We will be keeping compatibility with older version of that interface as it evolves over time so we can continue to run older guest (we already do). This protocol isn't actually tied to the DX API. It is a generic abstraction for the GPU that can be used for any APIs (for example the NVIDIA CUDA driver that we announced is going over the same protocol to access the GPU).
New version of user mode DX can either take advantage or sometime require new services from this kernel abstraction. This mean that pulling a new version of user mode DX can mean having to also pull a new version of this vGPU kernel driver. For WSL, these essentially ships together. The kernel driver ships as part of our WSL2 Linux Kernel integration. User mode DX bits ships with Windows.
Just a friendly advise: maintaining a proprietary component within a Linux environment is tough. You will need a good plan for long-term interface stability and compatibility with the other components.
Best regards Thomas
-----Original Message----- From: Thomas Zimmermann tzimmermann@suse.de Sent: Wednesday, May 20, 2020 12:11 AM To: Sasha Levin sashal@kernel.org; alexander.deucher@amd.com; chris@chris-wilson.co.uk; ville.syrjala@linux.intel.com; Hawking.Zhang@amd.com; tvrtko.ursulin@intel.com Cc: linux-kernel@vger.kernel.org; linux-hyperv@vger.kernel.org; KY Srinivasan kys@microsoft.com; Haiyang Zhang haiyangz@microsoft.com; Stephen Hemminger sthemmin@microsoft.com; wei.liu@kernel.org; Steve Pronovost spronovo@microsoft.com; Iouri Tarassov iourit@microsoft.com; dri-devel@lists.freedesktop.org; linux-fbdev@vger.kernel.org; gregkh@linuxfoundation.org Subject: [EXTERNAL] Re: [RFC PATCH 0/4] DirectX on Linux
Hi
Am 19.05.20 um 18:32 schrieb Sasha Levin:
There is a blog post that goes into more detail about the bigger picture, and walks through all the required pieces to make this work. It is available here: https://devblogs.microsoft.com/directx/directx-heart-linux . The rest of this cover letter will focus on the Linux Kernel bits.
That's quite a surprise. Thanks for your efforts to contribute.
Overview
This is the first draft of the Microsoft Virtual GPU (vGPU) driver. The driver exposes a paravirtualized GPU to user mode applications running in a virtual machine on a Windows host. This enables hardware acceleration in environment such as WSL (Windows Subsystem for Linux) where the Linux virtual machine is able to share the GPU with the Windows host.
The projection is accomplished by exposing the WDDM (Windows Display Driver Model) interface as a set of IOCTL. This allows APIs and user mode driver written against the WDDM GPU abstraction on Windows to be ported to run within a Linux environment. This enables the port of the D3D12 and DirectML APIs as well as their associated user mode driver to Linux. This also enables third party APIs, such as the popular NVIDIA Cuda compute API, to be hardware accelerated within a WSL environment.
Only the rendering/compute aspect of the GPU are projected to the virtual machine, no display functionality is exposed. Further, at this time there are no presentation integration. So although the D3D12 API can be use to render graphics offscreen, there is no path (yet) for pixel to flow from the Linux environment back onto the Windows host desktop. This GPU stack is effectively side-by-side with the native Linux graphics stack.
The driver creates the /dev/dxg device, which can be opened by user mode application and handles their ioctls. The IOCTL interface to the driver is defined in dxgkmthk.h (Dxgkrnl Graphics Port Driver ioctl definitions). The interface matches the D3DKMT interface on Windows. Ioctls are implemented in ioctl.c.
Echoing what others said, you're not making a DRM driver. The driver should live outside of the DRM code.
I have one question about the driver API: on Windows, DirectX versions are loosly tied to Windows releases. So I guess you can change the kernel interface among DirectX versions?
If so, how would this work on Linux in the long term? If there ever is a DirectX 13 or 14 with incompatible kernel interfaces, how would you plan to update the Linux driver?
Best regards Thomas
When a VM starts, hyper-v on the host adds virtual GPU devices to the VM via the hyper-v driver. The host offers several VM bus channels to the VM: the global channel and one channel per virtual GPU, assigned to the VM.
The driver registers with the hyper-v driver (hv_driver) for the arrival of VM bus channels. dxg_probe_device recognizes the vGPU channels and creates the corresponding objects (dxgadapter for vGPUs and dxgglobal for the global channel).
The driver uses the hyper-V VM bus interface to communicate with the host. dxgvmbus.c implements the communication interface.
The global channel has 8GB of IO space assigned by the host. This space is managed by the host and used to give the guest direct CPU access to some allocations. Video memory is allocated on the host except in the case of existing_sysmem allocations. The Windows host allocates memory for the GPU on behalf of the guest. The Linux guest can access that memory by mapping GPU virtual address to allocations and then referencing those GPU virtual address from within GPU command buffers submitted to the GPU. For allocations which require CPU access, the allocation is mapped by the host into a location in the 8GB of IO space reserved in the guest for that purpose. The Windows host uses the nested CPU page table to ensure that this guest IO space always map to the correct location for the allocation as it may migrate between dedicated GPU memory (e.g. VRAM, firmware reserved DDR) and shared system memory (regular DDR) over its lifetime. The Linux guest maps a user mode CPU virtual address to an allocation IO space range for direct access by user mode APIs and drivers.
Implementation of LX_DXLOCK2 ioctl
We would appreciate your feedback on the implementation of the LX_DXLOCK2 ioctl.
This ioctl is used to get a CPU address to an allocation, which is resident in video/system memory on the host. The way it works:
The driver sends the Lock message to the host
The host allocates space in the VM IO space and maps it to the
allocation memory
The host returns the address in IO space for the mapped allocation
The driver (in dxg_map_iospace) allocates a user mode virtual
address range using vm_mmap and maps it to the IO space using io_remap_ofn_range)
- The VA is returned to the application
Internal objects
The following objects are created by the driver (defined in dxgkrnl.h):
dxgadapter - represents a virtual GPU
dxgprocess - tracks per process state (handle table of created objects, list of objects, etc.)
dxgdevice - a container for other objects (contexts, paging queues, allocations, GPU synchronization objects)
dxgcontext - represents thread of GPU execution for packet scheduling.
dxghwqueue - represents thread of GPU execution of hardware
scheduling
dxgallocation - represents a GPU accessible allocation
dxgsyncobject - represents a GPU synchronization object
dxgresource - collection of dxgalloction objects
dxgsharedresource, dxgsharedsyncobj - helper objects to share objects between different dxgdevice objects, which can belong to different
processes
Object handles
All GPU objects, created by the driver, are accessible by a handle (d3dkmt_handle). Each process has its own handle table, which is implemented in hmgr.c. For each API visible object, created by the driver, there is an object, created on the host. For example, the is a dxgprocess object on the host for each dxgprocess object in the VM, etc. The object handles have the same value in the host and the VM, which is done to avoid translation from the guest handles to the host handles.
Signaling CPU events by the host
The WDDM interface provides a way to signal CPU event objects when execution of a context reached certain point. The way it is implemented:
application sends an event_fd via ioctl to the driver
eventfd_ctx_get is used to get a pointer to the file object (eventfd_ctx)
the pointer to sent the host via a VM bus message
when GPU execution reaches a certain point, the host sends a message to the VM with the event pointer
signal_guest_event() handles the messages and eventually eventfd_signal() is called.
Sasha Levin (4): gpu: dxgkrnl: core code gpu: dxgkrnl: hook up dxgkrnl Drivers: hv: vmbus: hook up dxgkrnl gpu: dxgkrnl: create a MAINTAINERS entry
MAINTAINERS | 7 + drivers/gpu/Makefile | 2 +- drivers/gpu/dxgkrnl/Kconfig | 10 + drivers/gpu/dxgkrnl/Makefile | 12 + drivers/gpu/dxgkrnl/d3dkmthk.h | 1635 +++++++++ drivers/gpu/dxgkrnl/dxgadapter.c | 1399 ++++++++ drivers/gpu/dxgkrnl/dxgkrnl.h | 913 ++++++ drivers/gpu/dxgkrnl/dxgmodule.c | 692 ++++ drivers/gpu/dxgkrnl/dxgprocess.c | 355 ++ drivers/gpu/dxgkrnl/dxgvmbus.c | 2955 +++++++++++++++++ drivers/gpu/dxgkrnl/dxgvmbus.h | 859 +++++ drivers/gpu/dxgkrnl/hmgr.c | 593 ++++ drivers/gpu/dxgkrnl/hmgr.h | 107 + drivers/gpu/dxgkrnl/ioctl.c | 5269 ++++++++++++++++++++++++++++++ drivers/gpu/dxgkrnl/misc.c | 280 ++ drivers/gpu/dxgkrnl/misc.h | 288 ++ drivers/video/Kconfig | 2 + include/linux/hyperv.h | 16 + 18 files changed, 15393 insertions(+), 1 deletion(-) create mode 100644 drivers/gpu/dxgkrnl/Kconfig create mode 100644 drivers/gpu/dxgkrnl/Makefile create mode 100644 drivers/gpu/dxgkrnl/d3dkmthk.h create mode 100644 drivers/gpu/dxgkrnl/dxgadapter.c create mode 100644 drivers/gpu/dxgkrnl/dxgkrnl.h create mode 100644 drivers/gpu/dxgkrnl/dxgmodule.c create mode 100644 drivers/gpu/dxgkrnl/dxgprocess.c create mode 100644 drivers/gpu/dxgkrnl/dxgvmbus.c create mode 100644 drivers/gpu/dxgkrnl/dxgvmbus.h create mode 100644 drivers/gpu/dxgkrnl/hmgr.c create mode 100644 drivers/gpu/dxgkrnl/hmgr.h create mode 100644 drivers/gpu/dxgkrnl/ioctl.c create mode 100644 drivers/gpu/dxgkrnl/misc.c create mode 100644 drivers/gpu/dxgkrnl/misc.h
-- Thomas Zimmermann Graphics Driver Developer SUSE Software Solutions Germany GmbH Maxfeldstr. 5, 90409 Nürnberg, Germany (HRB 36809, AG Nürnberg) Geschäftsführer: Felix Imendörffer
Hi!
The driver creates the /dev/dxg device, which can be opened by user mode application and handles their ioctls. The IOCTL interface to the driver is defined in dxgkmthk.h (Dxgkrnl Graphics Port Driver ioctl definitions). The interface matches the D3DKMT interface on Windows. Ioctls are implemented in ioctl.c.
Echoing what others said, you're not making a DRM driver. The driver should live outside of the DRM code.
Actually, this sounds to me like "this should not be merged into linux kernel". I mean, we already have DRM API on Linux. We don't want another one, do we?
And at the very least... this misses API docs for /dev/dxg. Code can't really be reviewed without that.
Best regards, Pavel
On Tue, Jun 16, 2020 at 12:51:13PM +0200, Pavel Machek wrote:
Hi!
The driver creates the /dev/dxg device, which can be opened by user mode application and handles their ioctls. The IOCTL interface to the driver is defined in dxgkmthk.h (Dxgkrnl Graphics Port Driver ioctl definitions). The interface matches the D3DKMT interface on Windows. Ioctls are implemented in ioctl.c.
Echoing what others said, you're not making a DRM driver. The driver should live outside of the DRM code.
Actually, this sounds to me like "this should not be merged into linux kernel". I mean, we already have DRM API on Linux. We don't want another one, do we?
This driver doesn't have any display functionality.
And at the very least... this misses API docs for /dev/dxg. Code can't really be reviewed without that.
The docs live here: https://docs.microsoft.com/en-us/windows-hardware/drivers/ddi/d3dkmthk/
On Tue 2020-06-16 09:28:19, Sasha Levin wrote:
On Tue, Jun 16, 2020 at 12:51:13PM +0200, Pavel Machek wrote:
Hi!
The driver creates the /dev/dxg device, which can be opened by user mode application and handles their ioctls. The IOCTL interface to the driver is defined in dxgkmthk.h (Dxgkrnl Graphics Port Driver ioctl definitions). The interface matches the D3DKMT interface on Windows. Ioctls are implemented in ioctl.c.
Echoing what others said, you're not making a DRM driver. The driver should live outside of the DRM code.
Actually, this sounds to me like "this should not be merged into linux kernel". I mean, we already have DRM API on Linux. We don't want another one, do we?
This driver doesn't have any display functionality.
Graphics cards without displays connected are quite common. I may be wrong, but I believe we normally handle them using DRM...
And at the very least... this misses API docs for /dev/dxg. Code can't really be reviewed without that.
The docs live here: https://docs.microsoft.com/en-us/windows-hardware/drivers/ddi/d3dkmthk/
I don't see "/dev/dxg" being metioned there. Plus, kernel API documentation should really go to Documentation, and be suitably licensed. Pavel
On Tue, Jun 16, 2020 at 04:41:22PM +0200, Pavel Machek wrote:
On Tue 2020-06-16 09:28:19, Sasha Levin wrote:
On Tue, Jun 16, 2020 at 12:51:13PM +0200, Pavel Machek wrote:
Hi!
The driver creates the /dev/dxg device, which can be opened by user mode application and handles their ioctls. The IOCTL interface to the driver is defined in dxgkmthk.h (Dxgkrnl Graphics Port Driver ioctl definitions). The interface matches the D3DKMT interface on Windows. Ioctls are implemented in ioctl.c.
Echoing what others said, you're not making a DRM driver. The driver should live outside of the DRM code.
Actually, this sounds to me like "this should not be merged into linux kernel". I mean, we already have DRM API on Linux. We don't want another one, do we?
This driver doesn't have any display functionality.
Graphics cards without displays connected are quite common. I may be wrong, but I believe we normally handle them using DRM...
This is more similar to the accelerators that live in drivers/misc/ right now.
And at the very least... this misses API docs for /dev/dxg. Code can't really be reviewed without that.
The docs live here: https://docs.microsoft.com/en-us/windows-hardware/drivers/ddi/d3dkmthk/
I don't see "/dev/dxg" being metioned there. Plus, kernel API
Right, this is because this entire codebase is just a pipe to the API I've linked, it doesn't implement anything new on it's own.
documentation should really go to Documentation, and be suitably licensed.
While I don't mind copying the docs into Documentation, I'm concerned that over time they will diverge from the docs on the website. This is similar to how other documentation (such as the virtio spec) live out of tree to avoid these issues.
w.r.t the licensing, again: this was sent under GPL2 (note the SPDX tags in each file), and the patches carry a S-O-B by someone who was a Microsoft employee at the time the patches were sent.
dri-devel@lists.freedesktop.org
-
Daniel Vetter -
Dave Airlie -
Greg KH -
James Hilliard -
Jan Engelhardt -
Pavel Machek -
Pavel Machek -
Sasha Levin -
Steve Pronovost -
Thomas Zimmermann