dri-devel February 2021

dri-devel@lists.freedesktop.org

314 participants
684 discussions

Re: [PATCH v2 1/4] hmm: Device exclusive memory access
by Jason Gunthorpe 19 Feb '21

19 Feb '21

On Fri, Feb 19, 2021 at 09:47:41AM +0000, Christoph Hellwig wrote: > > diff --git a/include/linux/hmm.h b/include/linux/hmm.h > > index 866a0fa104c4..5d28ff6d4d80 100644 > > +++ b/include/linux/hmm.h > > @@ -109,6 +109,10 @@ struct hmm_range { > > */ > > int hmm_range_fault(struct hmm_range *range); > > > > +int hmm_exclusive_range(struct mm_struct *mm, unsigned long start, > > + unsigned long end, struct page **pages); > > +vm_fault_t hmm_remove_exclusive_entry(struct vm_fault *vmf); > > Can we avoid the hmm naming for new code (we should probably also kill > it off for the existing code)? Yes please, I'd prefer it if hmm.c was just the special page walker and not a grab bag of unrelated things Is there is a more natural place to put this in the mm for this idea? Jason

1 0

[PATCH v4 0/7] drm/sun4i: Add support for dual-link LVDS on the A20
by Maxime Ripard 19 Feb '21

19 Feb '21

Hi, This is a fourth attempt at supporting the LVDS dual-link output on the Allwinner A20. Let me know what you think, Maxime Changes from v3: - Reintroduce the endpoint sanity check (and make of_graph_get_port_by_id const in the process) - Make a few doc fixes - Fix device_node leakage - Changed allwinner,lvds-companion to more generic link-companion - Rebased on 5.11 Changes from v2: - Added the DT binding description - Split the patch to enable the A20 - Reworked a bit the error messages Changes from v1: - Reworked the DT bindings - Refactored a bit the panel registration in the tcon code. Maxime Ripard (7): of: Make of_graph_get_port_by_id take a const device_node drm/of: Change the prototype of drm_of_lvds_get_dual_link_pixel_order dt-bindings: display: sun4i: Add LVDS Dual-Link property drm/sun4i: tcon: Refactor the LVDS and panel probing drm/sun4i: tcon: Support the LVDS Dual-Link drm/sun4i: tcon: Enable the A20 dual-link output [DO NOT MERGE] ARM: dts: sun7i: Enable LVDS Dual-Link on the Cubieboard .../display/allwinner,sun4i-a10-tcon.yaml | 6 + arch/arm/boot/dts/sun7i-a20-cubieboard2.dts | 69 ++++++++ drivers/gpu/drm/drm_of.c | 138 +++++++++++---- drivers/gpu/drm/rcar-du/rcar_lvds.c | 8 +- drivers/gpu/drm/sun4i/sun4i_tcon.c | 164 ++++++++++-------- drivers/gpu/drm/sun4i/sun4i_tcon.h | 4 + drivers/of/property.c | 2 +- include/drm/drm_of.h | 16 +- include/linux/of_graph.h | 4 +- 9 files changed, 299 insertions(+), 112 deletions(-) -- 2.29.2

1 7

[PATCH] drm/amdgpu/swsmu/navi1x: Remove unnecessary conversion to bool
by Jiapeng Chong 19 Feb '21

19 Feb '21

Fix the following coccicheck warnings: ./drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c:900:47-52: WARNING: conversion to bool not needed here. Reported-by: Abaci Robot <abaci(a)linux.alibaba.com> Signed-off-by: Jiapeng Chong <jiapeng.chong(a)linux.alibaba.com> --- drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c index cd7efa9..dc9ce86 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c @@ -897,7 +897,7 @@ static bool navi10_is_support_fine_grained_dpm(struct smu_context *smu, enum smu dpm_desc = &pptable->DpmDescriptor[clk_index]; /* 0 - Fine grained DPM, 1 - Discrete DPM */ - return dpm_desc->SnapToDiscrete == 0 ? true : false; + return !(dpm_desc->SnapToDiscrete == 0); } static inline bool navi10_od_feature_is_supported(struct smu_11_0_overdrive_table *od_table, enum SMU_11_0_ODFEATURE_CAP cap) -- 1.8.3.1

2 1

[RFC v4 0/3] Introduce Virtio based Dmabuf driver(s)
by Vivek Kasireddy 19 Feb '21

19 Feb '21

The Virtual Dmabuf or Virtio based Dmabuf (Vdmabuf) driver can be used to "transfer" a page-backed dmabuf created in the Guest to the Host without making any copies. This is mostly accomplished by recreating the dmabuf on the Host using the PFNs and other meta-data shared by the guest. A use-case where this driver would be a good fit is a multi-GPU system (perhaps one discrete and one integrated) where one of the GPUs does not have access to the display/connectors/outputs. This could be an embedded system design decision or a restriction made at the firmware/BIOS level or perhaps the device is setup in UPT (Universal Passthrough) mode. When such a GPU is passthrough'd to a Guest OS, this driver can help in transferring the scanout buffer(s) (rendered using the native rendering stack) to the Host for the purpose of displaying them. Or, quite simply, this driver can be used to transfer a dmabuf created by an application running on the Guest to another application running on the Host. The userspace component running in the Guest that transfers the dmabuf is referred to as the producer or exporter and its counterpart running in the Host is referred to as importer or consumer. For instance, a Wayland compositor would potentially be a producer and Qemu UI would be a consumer. It is the producer's responsibility to not reuse or destroy the shared buffer while it is still being used by the consumer. The consumer would send a release cmd indicating that it is done after which the shared buffer can be safely used again by the producer. One way the producer can prevent accidental re-use of the shared buffer is to lock the buffer when it exports it and unlock it after it gets a release cmd. As an example, the GBM API provides a simple way to lock and unlock a surface's buffers. For each dmabuf that is to be shared with the Host, a 128-bit unique ID is generated that identifies this buffer across the whole system. This ID is a combination of the Qemu process ID, a counter and a randomizer. We could potentially use UUID API but we currently use the above mentioned combination to identify the source of the buffer at any given time for potential bookkeeping. A typical cycle starts with the producer or exporter calling the alloc_fd IOCTL to get a new fd/dmabuf from Vdmabuf. It would then import and render to it and finally exports it by calling the export IOCTL. A new unique ID is generated for this buffer and it gets registered with the Host. The Host then alerts the consumer or importer by raising an event and shares the ID. In response, the consumer calls the import IOCTL using the ID and gets a newly created dmabuf fd in return. After it is done using the dmabuf, the consumer finally calls the release IOCTL and the Guest is notified which in turn notifies the producer letting it know that the buffer is now safe to reuse. v2: - Added a notifier mechanism for getting the kvm pointer. - Added start and stop routines in the Vhost backend. - Augmented the cover letter and made some minor improvements. v3: - Refactored the code to make it similar to vsock. - Used two virtqueues instead of one for efficient two-way communication. v4: - Made Vdmabuf guest driver allocate the dma-buf and backing storage and exported it to Weston to be used as a render target. (Gerd) The Vdmabuf driver was tested using Weston (headless) and Qemu from here: https://gitlab.freedesktop.org/Vivek/weston/-/blob/vdmabuf/libweston/backen… https://lists.nongnu.org/archive/html/qemu-devel/2021-02/msg02976.html TODO: - Use dma_fences to improve synchronization for multiple importers. - Ensure that a process other than Qemu can also be the importer on Host. Other Considerations: - Should virtio-gpu be augmented to provide the same functionality as vdmabuf? - How can virtio-gpu/vdmabuf work with a Windows Guest? - Should there be a Vhost backend for virtio-gpu to reduce overhead? - Should a transfer of a dma-buf from Guest to Host be dependent on a DRM driver (virtio-gpu)? Vivek Kasireddy (3): kvm: Add a notifier for create and destroy VM events virtio: Introduce Vdmabuf driver vhost: Add Vdmabuf backend drivers/vhost/Kconfig | 9 + drivers/vhost/Makefile | 3 + drivers/vhost/vdmabuf.c | 1372 +++++++++++++++++++++++++++ drivers/virtio/Kconfig | 8 + drivers/virtio/Makefile | 1 + drivers/virtio/virtio_vdmabuf.c | 1105 +++++++++++++++++++++ include/linux/kvm_host.h | 5 + include/linux/virtio_vdmabuf.h | 287 ++++++ include/uapi/linux/vhost.h | 3 + include/uapi/linux/virtio_ids.h | 1 + include/uapi/linux/virtio_vdmabuf.h | 87 ++ virt/kvm/kvm_main.c | 20 +- 12 files changed, 2899 insertions(+), 2 deletions(-) create mode 100644 drivers/vhost/vdmabuf.c create mode 100644 drivers/virtio/virtio_vdmabuf.c create mode 100644 include/linux/virtio_vdmabuf.h create mode 100644 include/uapi/linux/virtio_vdmabuf.h -- 2.26.2

1 3

[PATCH] drm/drv: Remove initialization of static variables
by Tian Tao 19 Feb '21

19 Feb '21

Address the following checkpatch errors: ERROR: do not initialise statics to false Signed-off-by: Tian Tao <tiantao6(a)hisilicon.com> --- drivers/gpu/drm/drm_drv.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c index 20d22e4..c2f78de 100644 --- a/drivers/gpu/drm/drm_drv.c +++ b/drivers/gpu/drm/drm_drv.c @@ -61,7 +61,7 @@ static struct idr drm_minors_idr; * prefer to embed struct drm_device into their own device * structure and call drm_dev_init() themselves. */ -static bool drm_core_init_complete = false; +static bool drm_core_init_complete; static struct dentry *drm_debugfs_root; -- 2.7.4

2 1

Re: [PATCH] drm/ttm: Fix a memory leak
by Christian König 19 Feb '21

19 Feb '21

Hi Xinhui, you also need to CC dri-devel for TTM patches, but apart from that it is a really good catch. Am 19.02.21 um 05:25 schrieb xinhui pan: > Free the memory on failure. > Also no need to re-alloc memory on retry. > > Signed-off-by: xinhui pan <xinhui.pan(a)amd.com> Reviewed-by: Christian König <christian.koenig(a)amd.com> CC: stable(a)vger.kernel.org # 5.11 Going to push that to drm-misc-fixes in a Minute since that is a really important one. Thanks, Christian. > --- > drivers/gpu/drm/ttm/ttm_bo.c | 9 ++++++--- > 1 file changed, 6 insertions(+), 3 deletions(-) > > diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c > index e38102282fd5..a97d41f4ce3c 100644 > --- a/drivers/gpu/drm/ttm/ttm_bo.c > +++ b/drivers/gpu/drm/ttm/ttm_bo.c > @@ -906,8 +906,10 @@ static int ttm_bo_bounce_temp_buffer(struct ttm_buffer_object *bo, > return ret; > /* move to the bounce domain */ > ret = ttm_bo_handle_move_mem(bo, &hop_mem, false, ctx, NULL); > - if (ret) > + if (ret) { > + ttm_resource_free(bo, &hop_mem); > return ret; > + } > return 0; > } > > @@ -938,18 +940,19 @@ static int ttm_bo_move_buffer(struct ttm_buffer_object *bo, > * stop and the driver will be called to make > * the second hop. > */ > -bounce: > ret = ttm_bo_mem_space(bo, placement, &mem, ctx); > if (ret) > return ret; > +bounce: > ret = ttm_bo_handle_move_mem(bo, &mem, false, ctx, &hop); > if (ret == -EMULTIHOP) { > ret = ttm_bo_bounce_temp_buffer(bo, &mem, ctx, &hop); > if (ret) > - return ret; > + goto out; > /* try and move to final place now. */ > goto bounce; > } > +out: > if (ret) > ttm_resource_free(bo, &mem); > return ret;

1 0

[PATCH] drm/radeon/dpm: fix non-restricted types with le16_to_cpu()
by Yang Li 19 Feb '21

19 Feb '21

Fix the following sparse warnings: drivers/gpu/drm/radeon/rv6xx_dpm.c:1798:21: warning: cast to restricted __le32 drivers/gpu/drm/radeon/rv6xx_dpm.c:1799:22: warning: cast to restricted __le16 drivers/gpu/drm/radeon/rv6xx_dpm.c:1800:23: warning: cast to restricted __le16 Reported-by: Abaci Robot <abaci(a)linux.alibaba.com> Signed-off-by: Yang Li <yang.lee(a)linux.alibaba.com> --- drivers/gpu/drm/radeon/rv6xx_dpm.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/radeon/rv6xx_dpm.c b/drivers/gpu/drm/radeon/rv6xx_dpm.c index 69d380f..e6ab99e 100644 --- a/drivers/gpu/drm/radeon/rv6xx_dpm.c +++ b/drivers/gpu/drm/radeon/rv6xx_dpm.c @@ -1795,9 +1795,9 @@ static void rv6xx_parse_pplib_non_clock_info(struct radeon_device *rdev, struct radeon_ps *rps, struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info) { - rps->caps = le32_to_cpu(non_clock_info->ulCapsAndSettings); - rps->class = le16_to_cpu(non_clock_info->usClassification); - rps->class2 = le16_to_cpu(non_clock_info->usClassification2); + rps->caps = le32_to_cpu((__le32 __force)non_clock_info->ulCapsAndSettings); + rps->class = le16_to_cpu((__le16 __force)non_clock_info->usClassification); + rps->class2 = le16_to_cpu((__le16 __force)non_clock_info->usClassification2); if (r600_is_uvd_state(rps->class, rps->class2)) { rps->vclk = RV6XX_DEFAULT_VCLK_FREQ; -- 1.8.3.1

1 0

[PATCH v2 0/4] Add support for SVM atomics in Nouveau
by Alistair Popple 19 Feb '21

19 Feb '21

This is the second version of a series to add support to Nouveau for atomic memory operations on OpenCL shared virtual memory (SVM) regions. This is achieved using the atomic PTE bits on the GPU to only permit atomic operations to system memory when a page is not mapped in userspace on the CPU. The previous version of this series used an unmap and pin page migration, however this resulted in problems with ZONE_MOVABLE and other issues so this series uses a different approach. Instead exclusive device access is implemented by adding a new swap entry type (SWAP_DEVICE_EXCLUSIVE) which is similar to a migration entry. The main difference is that on fault the original entry is immediately restored by the fault handler instead of waiting. Restoring the entry triggers calls to MMU notifers which allows a device driver to revoke the atomic access permission from the GPU prior to the CPU finalising the entry. Patch 1 contains the bulk of the memory management changes required to implement the new entry type. Patch 2 contains some additions to the HMM selftests to ensure everything works as expected. Patch 3 was posted previously and has not changed. Patch 4 is similar to what was posted previously but updated to use the new swap entries instread of migration. This has been tested using the latest upstream Mesa userspace with a simple OpenCL test program which checks the results of atomic GPU operations on a SVM buffer whilst also writing to the same buffer from the CPU. Alistair Popple (4): hmm: Device exclusive memory access hmm: Selftests for exclusive device memory nouveau/svm: Refactor nouveau_range_fault nouveau/svm: Implement atomic SVM access Documentation/vm/hmm.rst | 15 ++ drivers/gpu/drm/nouveau/include/nvif/if000c.h | 1 + drivers/gpu/drm/nouveau/nouveau_svm.c | 118 +++++++--- drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmm.h | 1 + .../drm/nouveau/nvkm/subdev/mmu/vmmgp100.c | 6 + fs/proc/task_mmu.c | 7 + include/linux/hmm.h | 4 + include/linux/rmap.h | 1 + include/linux/swap.h | 10 +- include/linux/swapops.h | 32 +++ lib/test_hmm.c | 124 ++++++++++ lib/test_hmm_uapi.h | 2 + mm/hmm.c | 206 ++++++++++++++++ mm/memory.c | 34 ++- mm/mprotect.c | 7 + mm/page_vma_mapped.c | 14 +- mm/rmap.c | 29 ++- tools/testing/selftests/vm/hmm-tests.c | 219 ++++++++++++++++++ 18 files changed, 792 insertions(+), 38 deletions(-) -- 2.20.1

2 7

[PATCH] drm/ioctl: fix space prohibited after that open parenthesis
by Tian Tao 19 Feb '21

19 Feb '21

This patch resolves below errors reported by checkpath ERROR: space prohibited after that open parenthesis '(' Signed-off-by: Tian Tao <tiantao6(a)hisilicon.com> --- drivers/gpu/drm/drm_ioc32.c | 28 ++++++++++++++-------------- 1 file changed, 14 insertions(+), 14 deletions(-) diff --git a/drivers/gpu/drm/drm_ioc32.c b/drivers/gpu/drm/drm_ioc32.c index f86448a..5b77db4 100644 --- a/drivers/gpu/drm/drm_ioc32.c +++ b/drivers/gpu/drm/drm_ioc32.c @@ -43,35 +43,35 @@ #define DRM_IOCTL_GET_UNIQUE32 DRM_IOWR(0x01, drm_unique32_t) #define DRM_IOCTL_GET_MAP32 DRM_IOWR(0x04, drm_map32_t) #define DRM_IOCTL_GET_CLIENT32 DRM_IOWR(0x05, drm_client32_t) -#define DRM_IOCTL_GET_STATS32 DRM_IOR( 0x06, drm_stats32_t) +#define DRM_IOCTL_GET_STATS32 DRM_IOR(0x06, drm_stats32_t) -#define DRM_IOCTL_SET_UNIQUE32 DRM_IOW( 0x10, drm_unique32_t) +#define DRM_IOCTL_SET_UNIQUE32 DRM_IOW(0x10, drm_unique32_t) #define DRM_IOCTL_ADD_MAP32 DRM_IOWR(0x15, drm_map32_t) #define DRM_IOCTL_ADD_BUFS32 DRM_IOWR(0x16, drm_buf_desc32_t) -#define DRM_IOCTL_MARK_BUFS32 DRM_IOW( 0x17, drm_buf_desc32_t) +#define DRM_IOCTL_MARK_BUFS32 DRM_IOW(0x17, drm_buf_desc32_t) #define DRM_IOCTL_INFO_BUFS32 DRM_IOWR(0x18, drm_buf_info32_t) #define DRM_IOCTL_MAP_BUFS32 DRM_IOWR(0x19, drm_buf_map32_t) -#define DRM_IOCTL_FREE_BUFS32 DRM_IOW( 0x1a, drm_buf_free32_t) +#define DRM_IOCTL_FREE_BUFS32 DRM_IOW(0x1a, drm_buf_free32_t) -#define DRM_IOCTL_RM_MAP32 DRM_IOW( 0x1b, drm_map32_t) +#define DRM_IOCTL_RM_MAP32 DRM_IOW(0x1b, drm_map32_t) -#define DRM_IOCTL_SET_SAREA_CTX32 DRM_IOW( 0x1c, drm_ctx_priv_map32_t) +#define DRM_IOCTL_SET_SAREA_CTX32 DRM_IOW(0x1c, drm_ctx_priv_map32_t) #define DRM_IOCTL_GET_SAREA_CTX32 DRM_IOWR(0x1d, drm_ctx_priv_map32_t) #define DRM_IOCTL_RES_CTX32 DRM_IOWR(0x26, drm_ctx_res32_t) #define DRM_IOCTL_DMA32 DRM_IOWR(0x29, drm_dma32_t) -#define DRM_IOCTL_AGP_ENABLE32 DRM_IOW( 0x32, drm_agp_mode32_t) -#define DRM_IOCTL_AGP_INFO32 DRM_IOR( 0x33, drm_agp_info32_t) +#define DRM_IOCTL_AGP_ENABLE32 DRM_IOW(0x32, drm_agp_mode32_t) +#define DRM_IOCTL_AGP_INFO32 DRM_IOR(0x33, drm_agp_info32_t) #define DRM_IOCTL_AGP_ALLOC32 DRM_IOWR(0x34, drm_agp_buffer32_t) -#define DRM_IOCTL_AGP_FREE32 DRM_IOW( 0x35, drm_agp_buffer32_t) -#define DRM_IOCTL_AGP_BIND32 DRM_IOW( 0x36, drm_agp_binding32_t) -#define DRM_IOCTL_AGP_UNBIND32 DRM_IOW( 0x37, drm_agp_binding32_t) +#define DRM_IOCTL_AGP_FREE32 DRM_IOW(0x35, drm_agp_buffer32_t) +#define DRM_IOCTL_AGP_BIND32 DRM_IOW(0x36, drm_agp_binding32_t) +#define DRM_IOCTL_AGP_UNBIND32 DRM_IOW(0x37, drm_agp_binding32_t) -#define DRM_IOCTL_SG_ALLOC32 DRM_IOW( 0x38, drm_scatter_gather32_t) -#define DRM_IOCTL_SG_FREE32 DRM_IOW( 0x39, drm_scatter_gather32_t) +#define DRM_IOCTL_SG_ALLOC32 DRM_IOW(0x38, drm_scatter_gather32_t) +#define DRM_IOCTL_SG_FREE32 DRM_IOW(0x39, drm_scatter_gather32_t) -#define DRM_IOCTL_UPDATE_DRAW32 DRM_IOW( 0x3f, drm_update_draw32_t) +#define DRM_IOCTL_UPDATE_DRAW32 DRM_IOW(0x3f, drm_update_draw32_t) #define DRM_IOCTL_WAIT_VBLANK32 DRM_IOWR(0x3a, drm_wait_vblank32_t) -- 2.7.4

1 0

[PATCH 8/9] drm/panel: simple: add Jenson JT60249-01
by Giulio Benetti 18 Feb '21

18 Feb '21

From: Giulio Benetti <giulio.benetti(a)micronovasrl.com> This patch adds support for Jenson JT60249-01 800x480 5" panel to DRM simple panel driver. Signed-off-by: Giulio Benetti <giulio.benetti(a)micronovasrl.com> Signed-off-by: Giulio Benetti <giulio.benetti(a)benettiengineering.com> --- drivers/gpu/drm/panel/panel-simple.c | 27 +++++++++++++++++++++++++++ 1 file changed, 27 insertions(+) diff --git a/drivers/gpu/drm/panel/panel-simple.c b/drivers/gpu/drm/panel/panel-simple.c index f243087ce89b..995a581f8b2b 100644 --- a/drivers/gpu/drm/panel/panel-simple.c +++ b/drivers/gpu/drm/panel/panel-simple.c @@ -2444,6 +2444,30 @@ static const struct panel_desc jenson_jt60248_01 = { .bus_flags = DRM_BUS_FLAG_PIXDATA_SAMPLE_NEGEDGE, }; +static const struct drm_display_mode jenson_jt60249_01_mode = { + .clock = 25000, + .hdisplay = 800, + .hsync_start = 800 + 8, + .hsync_end = 800 + 8 + 4, + .htotal = 800 + 8 + 4 + 8, + .vdisplay = 480, + .vsync_start = 480 + 8, + .vsync_end = 480 + 8 + 4, + .vtotal = 480 + 8 + 4 + 8, + .flags = DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC, +}; + +static const struct panel_desc jenson_jt60249_01 = { + .modes = &jenson_jt60249_01_mode, + .num_modes = 1, + .bpc = 8, + .size = { + .width = 108, + .height = 65, + }, + .bus_flags = DRM_BUS_FLAG_PIXDATA_SAMPLE_NEGEDGE, +}; + static const struct drm_display_mode kingdisplay_kd116n21_30nv_a010_mode = { .clock = 81000, .hdisplay = 1366, @@ -4216,6 +4240,9 @@ static const struct of_device_id platform_of_match[] = { }, { .compatible = "jenson,jt60248-01", .data = &jenson_jt60248_01, + }, { + .compatible = "jenson,jt60249-01", + .data = &jenson_jt60249_01, }, { .compatible = "kingdisplay,kd116n21-30nv-a010", .data = &kingdisplay_kd116n21_30nv_a010, -- 2.25.1

1 1

← Newer
1
...
30
31
32
33
34
35
36
...
69
Older →

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

dri-devel February 2021