dri-devel May 2015

dri-devel@lists.freedesktop.org

191 participants
1132 discussions

[PATCH] drm/amdkfd: Initialize sdma vm when creating sdma queue
by Oded Gabbay 05 May '15

05 May '15

From: Xihan Zhang <xihan.zhang(a)amd.com> This patch fixes a bug where sdma vm wasn't initialized when an sdma queue was created in HWS mode. This caused GPUVM faults to appear on dmesg and it is one of the causes that SDMA queues are not working. Signed-off-by: Xihan Zhang <xihan.zhang(a)amd.com> Reviewed-by: Ben Goz <ben.goz(a)amd.comt> Signed-off-by: Oded Gabbay <oded.gabbay(a)gmail.com> --- drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c index 7b1d510..596ee5c 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c @@ -883,6 +883,8 @@ static int create_queue_cpsch(struct device_queue_manager *dqm, struct queue *q, return -ENOMEM; } + init_sdma_vm(dqm, q, qpd); + retval = mqd->init_mqd(mqd, &q->mqd, &q->mqd_mem_obj, &q->gart_mqd_addr, &q->properties); if (retval != 0) -- 1.9.1

2 1

[PATCH] drm/amdkfd: Don't report local memory size
by Oded Gabbay 05 May '15

05 May '15

This patch sets the local memory size that is reported to userspace to 0. This is done to make sure that userspace won't try to allocate local memory for HSA. As long as amdkfd doesn't support allocating local memory for HSA, we need this patch. Signed-off-by: Oded Gabbay <oded.gabbay(a)gmail.com> --- drivers/gpu/drm/amd/amdkfd/kfd_topology.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c index 661c660..e469c4b 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c @@ -728,9 +728,9 @@ static ssize_t node_show(struct kobject *kobj, struct attribute *attr, sysfs_show_32bit_prop(buffer, "max_engine_clk_fcompute", dev->gpu->kfd2kgd->get_max_engine_clock_in_mhz( dev->gpu->kgd)); + sysfs_show_64bit_prop(buffer, "local_mem_size", - dev->gpu->kfd2kgd->get_vmem_size( - dev->gpu->kgd)); + (unsigned long long int) 0); sysfs_show_32bit_prop(buffer, "fw_version", dev->gpu->kfd2kgd->get_fw_version( -- 1.9.1

2 1

[PATCH] drmPrime*: initialize output args to 0
by Guillaume Desmottes 05 May '15

05 May '15

Fix Valgrind errors because those memory was uninitialized. https://bugs.freedesktop.org/show_bug.cgi?id=90194 Signed-off-by: Guillaume Desmottes <guillaume.desmottes(a)collabora.co.uk> --- xf86drm.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/xf86drm.c b/xf86drm.c index f7c45f8..01c398e 100644 --- a/xf86drm.c +++ b/xf86drm.c @@ -2724,6 +2724,7 @@ int drmPrimeHandleToFD(int fd, uint32_t handle, uint32_t flags, int *prime_fd) struct drm_prime_handle args; int ret; + args.fd = -1; args.handle = handle; args.flags = flags; ret = drmIoctl(fd, DRM_IOCTL_PRIME_HANDLE_TO_FD, &args); @@ -2739,6 +2740,7 @@ int drmPrimeFDToHandle(int fd, int prime_fd, uint32_t *handle) struct drm_prime_handle args; int ret; + args.handle = 0; args.fd = prime_fd; args.flags = 0; ret = drmIoctl(fd, DRM_IOCTL_PRIME_FD_TO_HANDLE, &args); -- 2.1.0

2 2

[PATCH 0/9 v3] Helper to abstract vma handling in media layer
by Jan Kara 05 May '15

05 May '15

Hello, I'm sending the third version of my patch series to abstract vma handling from the various media drivers. After this patch set drivers have to know much less details about vmas, their types, and locking. Also quite some code is removed from them. As a bonus drivers get automatically VM_FAULT_RETRY handling. The primary motivation for this series is to remove knowledge about mmap_sem locking from as many places a possible so that we can change it with reasonable effort. The core of the series is the new helper get_vaddr_frames() which is given a virtual address and it fills in PFNs / struct page pointers (depending on VMA type) into the provided array. If PFNs correspond to normal pages it also grabs references to these pages. The difference from get_user_pages() is that this function can also deal with pfnmap, and io mappings which is what the media drivers need. I have tested the patches with vivid driver so at least vb2 code got some exposure. Conversion of other drivers was just compile-tested so I'd like to ask respective maintainers if they could have a look. Also I'd like to ask mm folks to check patch 2/9 implementing the helper. Thanks! Honza Changes since v2: * Renamed functions and structures as Mel suggested * Other minor changes suggested by Mel * Rebased on top of 4.1-rc2 * Changed functions to get pointer to array of pages / pfns to perform conversion if necessary. This fixes possible issue in the omap I may have introduced in v2 and generally makes the API less errorprone.

1 9

[PATCH] drm/radeon: fix userptr lockup
by Christian König 05 May '15

05 May '15

From: Christian König <christian.koenig(a)amd.com> We shouldn't try to reserve and wait for a BO that isn't bound. Otherwise we can run into a deadlock if we have a fault during binding the BO. Signed-off-by: Christian König <christian.koenig(a)amd.com> --- drivers/gpu/drm/radeon/radeon_mn.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/gpu/drm/radeon/radeon_mn.c b/drivers/gpu/drm/radeon/radeon_mn.c index 535bf40..eef006c 100644 --- a/drivers/gpu/drm/radeon/radeon_mn.c +++ b/drivers/gpu/drm/radeon/radeon_mn.c @@ -142,6 +142,9 @@ static void radeon_mn_invalidate_range_start(struct mmu_notifier *mn, list_for_each_entry(bo, &node->bos, mn_list) { + if (!bo->tbo.ttm || bo->tbo.ttm->state != tt_bound) + continue; + r = radeon_bo_reserve(bo, true); if (r) { DRM_ERROR("(%ld) failed to reserve user bo\n", r); -- 1.9.1

2 1

[PATCH] drm/radeon: fix userptr BO unpin bug v2
by Christian König 05 May '15

05 May '15

From: "monk.liu" <monk.liu(a)amd.com> Fixing a memory leak with userptrs. v2: clean up the loop, use an iterator instead Signed-off-by: monk.liu <monk.liu(a)amd.com> Signed-off-by: Christian König <christian.koenig(a)amd.com> CC: stable(a)vger.kernel.org --- drivers/gpu/drm/radeon/radeon_ttm.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c index b292aca..921dd0f 100644 --- a/drivers/gpu/drm/radeon/radeon_ttm.c +++ b/drivers/gpu/drm/radeon/radeon_ttm.c @@ -591,7 +591,7 @@ static void radeon_ttm_tt_unpin_userptr(struct ttm_tt *ttm) { struct radeon_device *rdev = radeon_get_rdev(ttm->bdev); struct radeon_ttm_tt *gtt = (void *)ttm; - struct scatterlist *sg; + struct sg_page_iter sg_iter; int i; int write = !(gtt->userflags & RADEON_GEM_USERPTR_READONLY); @@ -605,9 +605,8 @@ static void radeon_ttm_tt_unpin_userptr(struct ttm_tt *ttm) /* free the sg table and pages again */ dma_unmap_sg(rdev->dev, ttm->sg->sgl, ttm->sg->nents, direction); - for_each_sg(ttm->sg->sgl, sg, ttm->sg->nents, i) { - struct page *page = sg_page(sg); - + for_each_sg_page(ttm->sg->sgl, &sg_iter, ttm->sg->nents, 0) { + struct page *page = sg_page_iter_page(&sg_iter); if (!(gtt->userflags & RADEON_GEM_USERPTR_READONLY)) set_page_dirty(page); -- 1.9.1

2 1

[PATCH] drm/msm: fix unbalanced DRM framebuffer init/destroy
by Stephane Viau 05 May '15

05 May '15

When msm_framebuffer_init() fails before calling drm_framebuffer_init(), drm_framebuffer_cleanup() [called in msm_framebuffer_destroy()] is still being called even though drm_framebuffer_init() was not called for that buffer. Thus a NULL pointer derefencing: [ 247.529691] Unable to handle kernel NULL pointer dereference at virtual address 0000027c ... [ 247.563996] PC is at __mutex_lock_slowpath+0x94/0x3a8 ... [ 247.823025] [<c07c3c78>] (__mutex_lock_slowpath) from [<c07c3fac>] (mutex_lock+0x20/0x3c) [ 247.831186] [<c07c3fac>] (mutex_lock) from [<c0347cf0>] (drm_framebuffer_cleanup+0x18/0x38) [ 247.839520] [<c0347cf0>] (drm_framebuffer_cleanup) from [<c036d138>] (msm_framebuffer_destroy+0x48/0x100) [ 247.849066] [<c036d138>] (msm_framebuffer_destroy) from [<c036d580>] (msm_framebuffer_init+0x1e8/0x228) [ 247.858439] [<c036d580>] (msm_framebuffer_init) from [<c036d630>] (msm_framebuffer_create+0x70/0x134) [ 247.867642] [<c036d630>] (msm_framebuffer_create) from [<c03493ec>] (internal_framebuffer_create+0x67c/0x7b4) [ 247.877537] [<c03493ec>] (internal_framebuffer_create) from [<c034ce34>] (drm_mode_addfb2+0x20/0x98) [ 247.886650] [<c034ce34>] (drm_mode_addfb2) from [<c034071c>] (drm_ioctl+0x240/0x420) [ 247.894378] [<c034071c>] (drm_ioctl) from [<c011df7c>] (do_vfs_ioctl+0x4e4/0x5a4) ... Signed-off-by: Stephane Viau <sviau(a)codeaurora.org> --- drivers/gpu/drm/msm/msm_fb.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/gpu/drm/msm/msm_fb.c b/drivers/gpu/drm/msm/msm_fb.c index 6b573e6..c0f09d3 100644 --- a/drivers/gpu/drm/msm/msm_fb.c +++ b/drivers/gpu/drm/msm/msm_fb.c @@ -239,8 +239,7 @@ struct drm_framebuffer *msm_framebuffer_init(struct drm_device *dev, return fb; fail: - if (fb) - msm_framebuffer_destroy(fb); + kfree(msm_fb); return ERR_PTR(ret); } -- Qualcomm Innovation Center, Inc. The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project

1 0

Re: [PATCH] gma500:Remove functions that are now deprecated and move to the newer functions in drm_dp_helper.c
by Patrik Jakobsson 05 May '15

05 May '15

On Tue, May 5, 2015 at 12:29 AM, Nicholas Krause <xerofoify(a)gmail.com> wrote: > This removes the deprecated functions,i2c_dp_aux_add_bus and > i2c_dp_aux_prepare_bus and the only call in the function, > cdv_intel_dp_i2c_init to i2c_dp_aux_add_bus respectfully. > The call and use of these functions is now replaced alongside > the removal of setting other values in the function,cdv_intel_dp_i2c_init > to use the helper function, drm_dp_aux_register that can handle all this > work internally. You need to fill in the drm_dp_aux structure and implement a proper transfer function. This patch would only break things. But the cdv dp output is already broken on my system so it needs fixing first. -Patrik > Signed-off-by: Nicholas Krause <xerofoify(a)gmail.com> > --- > drivers/gpu/drm/gma500/cdv_intel_dp.c | 42 ++--------------------------------- > 1 file changed, 2 insertions(+), 40 deletions(-) > > diff --git a/drivers/gpu/drm/gma500/cdv_intel_dp.c b/drivers/gpu/drm/gma500/cdv_intel_dp.c > index 0fafb8e..c8c20fc 100644 > --- a/drivers/gpu/drm/gma500/cdv_intel_dp.c > +++ b/drivers/gpu/drm/gma500/cdv_intel_dp.c > @@ -200,38 +200,6 @@ static const struct i2c_algorithm i2c_dp_aux_algo = { > .functionality = i2c_algo_dp_aux_functionality, > }; > > -static void > -i2c_dp_aux_reset_bus(struct i2c_adapter *adapter) > -{ > - (void) i2c_algo_dp_aux_address(adapter, 0, false); > - (void) i2c_algo_dp_aux_stop(adapter, false); > -} > - > -static int > -i2c_dp_aux_prepare_bus(struct i2c_adapter *adapter) > -{ > - adapter->algo = &i2c_dp_aux_algo; > - adapter->retries = 3; > - i2c_dp_aux_reset_bus(adapter); > - return 0; > -} > - > -/* > - * FIXME: This is the old dp aux helper, gma500 is the last driver that needs to > - * be ported over to the new helper code in drm_dp_helper.c like i915 or radeon. > - */ > -static int __deprecated > -i2c_dp_aux_add_bus(struct i2c_adapter *adapter) > -{ > - int error; > - > - error = i2c_dp_aux_prepare_bus(adapter); > - if (error) > - return error; > - error = i2c_add_adapter(adapter); > - return error; > -} > - > #define _wait_for(COND, MS, W) ({ \ > unsigned long timeout__ = jiffies + msecs_to_jiffies(MS); \ > int ret__ = 0; \ > @@ -275,6 +243,7 @@ struct cdv_intel_dp { > int backlight_on_delay; > int backlight_off_delay; > struct drm_display_mode *panel_fixed_mode; /* for eDP */ > + struct drm_dp_aux aux; > bool panel_on; > }; > > @@ -855,18 +824,11 @@ cdv_intel_dp_i2c_init(struct gma_connector *connector, > intel_dp->algo.running = false; > intel_dp->algo.address = 0; > intel_dp->algo.aux_ch = cdv_intel_dp_i2c_aux_ch; > - > memset(&intel_dp->adapter, '\0', sizeof (intel_dp->adapter)); > - intel_dp->adapter.owner = THIS_MODULE; > - intel_dp->adapter.class = I2C_CLASS_DDC; > - strncpy (intel_dp->adapter.name, name, sizeof(intel_dp->adapter.name) - 1); > - intel_dp->adapter.name[sizeof(intel_dp->adapter.name) - 1] = '\0'; > - intel_dp->adapter.algo_data = &intel_dp->algo; > - intel_dp->adapter.dev.parent = connector->base.kdev; > > if (is_edp(encoder)) > cdv_intel_edp_panel_vdd_on(encoder); > - ret = i2c_dp_aux_add_bus(&intel_dp->adapter); > + ret = drm_dp_aux_register(&intel_dp->aux); > if (is_edp(encoder)) > cdv_intel_edp_panel_vdd_off(encoder); > > -- > 2.1.4 >

1 0

Re: [PATCH] gma500:Remove functions that are now deprecated and move to the newer functions in drm_dp_helper.c
by Alan Cox 05 May '15

05 May '15

On Mon, 2015-05-04 at 18:29 -0400, Nicholas Krause wrote: > This removes the deprecated functions,i2c_dp_aux_add_bus and > i2c_dp_aux_prepare_bus and the only call in the function, > cdv_intel_dp_i2c_init to i2c_dp_aux_add_bus respectfully. > The call and use of these functions is now replaced alongside > the removal of setting other values in the function,cdv_intel_dp_i2c_init > to use the helper function, drm_dp_aux_register that can handle all this > work internally. Which devices have you tested this on ? Alan

1 0

[PATCH 1/3] drm: drop unused 'magicfree' list
by David Herrmann 05 May '15

05 May '15

This list is write-only. It's never used for read-access, so no reason to keep it around. Drop it! Signed-off-by: David Herrmann <dh.herrmann(a)gmail.com> --- drivers/gpu/drm/drm_auth.c | 3 --- drivers/gpu/drm/drm_drv.c | 1 - include/drm/drmP.h | 2 -- 3 files changed, 6 deletions(-) diff --git a/drivers/gpu/drm/drm_auth.c b/drivers/gpu/drm/drm_auth.c index fc8e8aa..8a37524 100644 --- a/drivers/gpu/drm/drm_auth.c +++ b/drivers/gpu/drm/drm_auth.c @@ -37,7 +37,6 @@ #include "drm_internal.h" struct drm_magic_entry { - struct list_head head; struct drm_hash_item hash_item; struct drm_file *priv; }; @@ -93,7 +92,6 @@ static int drm_add_magic(struct drm_master *master, struct drm_file *priv, entry->hash_item.key = (unsigned long)magic; mutex_lock(&dev->struct_mutex); drm_ht_insert_item(&master->magiclist, &entry->hash_item); - list_add_tail(&entry->head, &master->magicfree); mutex_unlock(&dev->struct_mutex); return 0; @@ -123,7 +121,6 @@ int drm_remove_magic(struct drm_master *master, drm_magic_t magic) } pt = drm_hash_entry(hash, struct drm_magic_entry, hash_item); drm_ht_remove_item(&master->magiclist, hash); - list_del(&pt->head); mutex_unlock(&dev->struct_mutex); kfree(pt); diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c index 48f7359..26ed9fe 100644 --- a/drivers/gpu/drm/drm_drv.c +++ b/drivers/gpu/drm/drm_drv.c @@ -109,7 +109,6 @@ struct drm_master *drm_master_create(struct drm_minor *minor) kfree(master); return NULL; } - INIT_LIST_HEAD(&master->magicfree); master->minor = minor; return master; diff --git a/include/drm/drmP.h b/include/drm/drmP.h index 62c40777..80be07a 100644 --- a/include/drm/drmP.h +++ b/include/drm/drmP.h @@ -356,7 +356,6 @@ struct drm_lock_data { * @unique: Unique identifier: e.g. busid. Protected by drm_global_mutex. * @unique_len: Length of unique field. Protected by drm_global_mutex. * @magiclist: Hash of used authentication tokens. Protected by struct_mutex. - * @magicfree: List of used authentication tokens. Protected by struct_mutex. * @lock: DRI lock information. * @driver_priv: Pointer to driver-private information. */ @@ -366,7 +365,6 @@ struct drm_master { char *unique; int unique_len; struct drm_open_hash magiclist; - struct list_head magicfree; struct drm_lock_data lock; void *driver_priv; }; -- 2.3.7

3 17

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

dri-devel May 2015