From 2943ef32c97666a6556c98496722c4b72420ad37 Mon Sep 17 00:00:00 2001 From: Andrzej Hajda Date: Tue, 15 Mar 2016 13:46:00 +0100 Subject: drm/atomic: use helper to get crtc state DRM core provide helper to access crtc state. Signed-off-by: Andrzej Hajda Signed-off-by: Daniel Vetter Link: http://patchwork.freedesktop.org/patch/msgid/1458045960-25193-1-git-send-email-a.hajda@samsung.com --- drivers/gpu/drm/drm_atomic_helper.c | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c index e93bb0b..9135111 100644 --- a/drivers/gpu/drm/drm_atomic_helper.c +++ b/drivers/gpu/drm/drm_atomic_helper.c @@ -67,7 +67,8 @@ drm_atomic_helper_plane_changed(struct drm_atomic_state *state, struct drm_crtc_state *crtc_state; if (plane->state->crtc) { - crtc_state = state->crtc_states[drm_crtc_index(plane->state->crtc)]; + crtc_state = drm_atomic_get_existing_crtc_state(state, + plane->state->crtc); if (WARN_ON(!crtc_state)) return; @@ -76,8 +77,8 @@ drm_atomic_helper_plane_changed(struct drm_atomic_state *state, } if (plane_state->crtc) { - crtc_state = - state->crtc_states[drm_crtc_index(plane_state->crtc)]; + crtc_state = drm_atomic_get_existing_crtc_state(state, + plane_state->crtc); if (WARN_ON(!crtc_state)) return; @@ -374,8 +375,8 @@ mode_fixup(struct drm_atomic_state *state) if (!conn_state->crtc || !conn_state->best_encoder) continue; - crtc_state = - state->crtc_states[drm_crtc_index(conn_state->crtc)]; + crtc_state = drm_atomic_get_existing_crtc_state(state, + conn_state->crtc); /* * Each encoder has at most one connector (since we always steal @@ -679,7 +680,8 @@ disable_outputs(struct drm_device *dev, struct drm_atomic_state *old_state) if (!old_conn_state->crtc) continue; - old_crtc_state = old_state->crtc_states[drm_crtc_index(old_conn_state->crtc)]; + old_crtc_state = drm_atomic_get_existing_crtc_state(old_state, + old_conn_state->crtc); if (!old_crtc_state->active || !drm_atomic_crtc_needs_modeset(old_conn_state->crtc->state)) -- cgit v1.1 From b47ff7e6a534bc285eb98f90eadd2a56e2e54056 Mon Sep 17 00:00:00 2001 From: Andrzej Hajda Date: Tue, 15 Mar 2016 13:46:28 +0100 Subject: drm/atmel-hlcdc: use helper to get crtc state DRM core provide helper to access crtc state. Signed-off-by: Andrzej Hajda Acked-by: Boris Brezillon Signed-off-by: Daniel Vetter Link: http://patchwork.freedesktop.org/patch/msgid/1458045988-25246-1-git-send-email-a.hajda@samsung.com --- drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_plane.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_plane.c b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_plane.c index 1ffe9c3..d65dcae 100644 --- a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_plane.c +++ b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_plane.c @@ -558,7 +558,7 @@ static int atmel_hlcdc_plane_atomic_check(struct drm_plane *p, if (!state->base.crtc || !fb) return 0; - crtc_state = s->state->crtc_states[drm_crtc_index(s->crtc)]; + crtc_state = drm_atomic_get_existing_crtc_state(s->state, s->crtc); mode = &crtc_state->adjusted_mode; state->src_x = s->src_x; -- cgit v1.1 From 18b862dcd57a3e23e34c8cd1e939f68548c1209a Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Fri, 18 Mar 2016 20:02:39 +0000 Subject: dma-buf, drm, ion: Propagate error code from dma_buf_start_cpu_access() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Drivers, especially i915.ko, can fail during the initial migration of a dma-buf for CPU access. However, the error code from the driver was not being propagated back to ioctl and so userspace was blissfully ignorant of the failure. Rendering corruption ensues. Whilst fixing the ioctl to return the error code from dma_buf_start_cpu_access(), also do the same for dma_buf_end_cpu_access(). For most drivers, dma_buf_end_cpu_access() cannot fail. i915.ko however, as most drivers would, wants to avoid being uninterruptible (as would be required to guarrantee no failure when flushing the buffer to the device). As userspace already has to handle errors from the SYNC_IOCTL, take advantage of this to be able to restart the syscall across signals. This fixes a coherency issue for i915.ko as well as reducing the uninterruptible hold upon its BKL, the struct_mutex. Fixes commit c11e391da2a8fe973c3c2398452000bed505851e Author: Daniel Vetter Date: Thu Feb 11 20:04:51 2016 -0200 dma-buf: Add ioctls to allow userspace to flush Testcase: igt/gem_concurrent_blit/*dmabuf*interruptible Testcase: igt/prime_mmap_coherency/ioctl-errors Signed-off-by: Chris Wilson Cc: Tiago Vignatti Cc: Stéphane Marchesin Cc: David Herrmann Cc: Sumit Semwal Cc: Daniel Vetter CC: linux-media@vger.kernel.org Cc: dri-devel@lists.freedesktop.org Cc: linaro-mm-sig@lists.linaro.org Cc: intel-gfx@lists.freedesktop.org Cc: devel@driverdev.osuosl.org Signed-off-by: Daniel Vetter Link: http://patchwork.freedesktop.org/patch/msgid/1458331359-2634-1-git-send-email-chris@chris-wilson.co.uk --- drivers/dma-buf/dma-buf.c | 17 +++++++++++------ drivers/gpu/drm/i915/i915_gem_dmabuf.c | 15 +++++---------- drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c | 5 +++-- drivers/gpu/drm/udl/udl_fb.c | 4 ++-- drivers/staging/android/ion/ion.c | 6 ++++-- include/linux/dma-buf.h | 6 +++--- 6 files changed, 28 insertions(+), 25 deletions(-) diff --git a/drivers/dma-buf/dma-buf.c b/drivers/dma-buf/dma-buf.c index 9810d1d..774a60f 100644 --- a/drivers/dma-buf/dma-buf.c +++ b/drivers/dma-buf/dma-buf.c @@ -259,6 +259,7 @@ static long dma_buf_ioctl(struct file *file, struct dma_buf *dmabuf; struct dma_buf_sync sync; enum dma_data_direction direction; + int ret; dmabuf = file->private_data; @@ -285,11 +286,11 @@ static long dma_buf_ioctl(struct file *file, } if (sync.flags & DMA_BUF_SYNC_END) - dma_buf_end_cpu_access(dmabuf, direction); + ret = dma_buf_end_cpu_access(dmabuf, direction); else - dma_buf_begin_cpu_access(dmabuf, direction); + ret = dma_buf_begin_cpu_access(dmabuf, direction); - return 0; + return ret; default: return -ENOTTY; } @@ -613,13 +614,17 @@ EXPORT_SYMBOL_GPL(dma_buf_begin_cpu_access); * * This call must always succeed. */ -void dma_buf_end_cpu_access(struct dma_buf *dmabuf, - enum dma_data_direction direction) +int dma_buf_end_cpu_access(struct dma_buf *dmabuf, + enum dma_data_direction direction) { + int ret = 0; + WARN_ON(!dmabuf); if (dmabuf->ops->end_cpu_access) - dmabuf->ops->end_cpu_access(dmabuf, direction); + ret = dmabuf->ops->end_cpu_access(dmabuf, direction); + + return ret; } EXPORT_SYMBOL_GPL(dma_buf_end_cpu_access); diff --git a/drivers/gpu/drm/i915/i915_gem_dmabuf.c b/drivers/gpu/drm/i915/i915_gem_dmabuf.c index 1f3eef6..0506016 100644 --- a/drivers/gpu/drm/i915/i915_gem_dmabuf.c +++ b/drivers/gpu/drm/i915/i915_gem_dmabuf.c @@ -228,25 +228,20 @@ static int i915_gem_begin_cpu_access(struct dma_buf *dma_buf, enum dma_data_dire return ret; } -static void i915_gem_end_cpu_access(struct dma_buf *dma_buf, enum dma_data_direction direction) +static int i915_gem_end_cpu_access(struct dma_buf *dma_buf, enum dma_data_direction direction) { struct drm_i915_gem_object *obj = dma_buf_to_obj(dma_buf); struct drm_device *dev = obj->base.dev; - struct drm_i915_private *dev_priv = to_i915(dev); - bool was_interruptible; int ret; - mutex_lock(&dev->struct_mutex); - was_interruptible = dev_priv->mm.interruptible; - dev_priv->mm.interruptible = false; + ret = i915_mutex_lock_interruptible(dev); + if (ret) + return ret; ret = i915_gem_object_set_to_gtt_domain(obj, false); - - dev_priv->mm.interruptible = was_interruptible; mutex_unlock(&dev->struct_mutex); - if (unlikely(ret)) - DRM_ERROR("unable to flush buffer following CPU access; rendering may be corrupt\n"); + return ret; } static const struct dma_buf_ops i915_dmabuf_ops = { diff --git a/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c b/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c index aebae1c..c75249d 100644 --- a/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c +++ b/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c @@ -93,11 +93,12 @@ static int omap_gem_dmabuf_begin_cpu_access(struct dma_buf *buffer, return omap_gem_get_pages(obj, &pages, true); } -static void omap_gem_dmabuf_end_cpu_access(struct dma_buf *buffer, - enum dma_data_direction dir) +static int omap_gem_dmabuf_end_cpu_access(struct dma_buf *buffer, + enum dma_data_direction dir) { struct drm_gem_object *obj = buffer->priv; omap_gem_put_pages(obj); + return 0; } diff --git a/drivers/gpu/drm/udl/udl_fb.c b/drivers/gpu/drm/udl/udl_fb.c index c427499..33239a2 100644 --- a/drivers/gpu/drm/udl/udl_fb.c +++ b/drivers/gpu/drm/udl/udl_fb.c @@ -423,8 +423,8 @@ static int udl_user_framebuffer_dirty(struct drm_framebuffer *fb, } if (ufb->obj->base.import_attach) { - dma_buf_end_cpu_access(ufb->obj->base.import_attach->dmabuf, - DMA_FROM_DEVICE); + ret = dma_buf_end_cpu_access(ufb->obj->base.import_attach->dmabuf, + DMA_FROM_DEVICE); } unlock: diff --git a/drivers/staging/android/ion/ion.c b/drivers/staging/android/ion/ion.c index 0754a37..49436b4 100644 --- a/drivers/staging/android/ion/ion.c +++ b/drivers/staging/android/ion/ion.c @@ -1075,14 +1075,16 @@ static int ion_dma_buf_begin_cpu_access(struct dma_buf *dmabuf, return PTR_ERR_OR_ZERO(vaddr); } -static void ion_dma_buf_end_cpu_access(struct dma_buf *dmabuf, - enum dma_data_direction direction) +static int ion_dma_buf_end_cpu_access(struct dma_buf *dmabuf, + enum dma_data_direction direction) { struct ion_buffer *buffer = dmabuf->priv; mutex_lock(&buffer->lock); ion_buffer_kmap_put(buffer); mutex_unlock(&buffer->lock); + + return 0; } static struct dma_buf_ops dma_buf_ops = { diff --git a/include/linux/dma-buf.h b/include/linux/dma-buf.h index 532108e..3fe90d4 100644 --- a/include/linux/dma-buf.h +++ b/include/linux/dma-buf.h @@ -94,7 +94,7 @@ struct dma_buf_ops { void (*release)(struct dma_buf *); int (*begin_cpu_access)(struct dma_buf *, enum dma_data_direction); - void (*end_cpu_access)(struct dma_buf *, enum dma_data_direction); + int (*end_cpu_access)(struct dma_buf *, enum dma_data_direction); void *(*kmap_atomic)(struct dma_buf *, unsigned long); void (*kunmap_atomic)(struct dma_buf *, unsigned long, void *); void *(*kmap)(struct dma_buf *, unsigned long); @@ -224,8 +224,8 @@ void dma_buf_unmap_attachment(struct dma_buf_attachment *, struct sg_table *, enum dma_data_direction); int dma_buf_begin_cpu_access(struct dma_buf *dma_buf, enum dma_data_direction dir); -void dma_buf_end_cpu_access(struct dma_buf *dma_buf, - enum dma_data_direction dir); +int dma_buf_end_cpu_access(struct dma_buf *dma_buf, + enum dma_data_direction dir); void *dma_buf_kmap_atomic(struct dma_buf *, unsigned long); void dma_buf_kunmap_atomic(struct dma_buf *, unsigned long, void *); void *dma_buf_kmap(struct dma_buf *, unsigned long); -- cgit v1.1 From 3c66f4aa7f3cb6579dfe977c05fcb5db25b41cd7 Mon Sep 17 00:00:00 2001 From: Luis de Bethencourt Date: Sat, 19 Mar 2016 17:22:36 +0000 Subject: drm: remove excess description Description of expected_size doesn't match any parameter of the function drm_atomic_replace_property_blob. Removing it. Signed-off-by: Luis de Bethencourt Signed-off-by: Daniel Vetter Link: http://patchwork.freedesktop.org/patch/msgid/1458408156-15990-1-git-send-email-luisbg@osg.samsung.com --- drivers/gpu/drm/drm_atomic.c | 1 - 1 file changed, 1 deletion(-) diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c index c7e7a95..d78f8ea 100644 --- a/drivers/gpu/drm/drm_atomic.c +++ b/drivers/gpu/drm/drm_atomic.c @@ -392,7 +392,6 @@ EXPORT_SYMBOL(drm_atomic_set_mode_prop_for_crtc); * drm_atomic_replace_property_blob - replace a blob property * @blob: a pointer to the member blob to be replaced * @new_blob: the new blob to replace with - * @expected_size: the expected size of the new blob * @replaced: whether the blob has been replaced * * RETURNS: -- cgit v1.1 From 87e332d56b2c7dc6733f690c9069f4fe75f21c9d Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Mon, 21 Mar 2016 08:24:22 +0100 Subject: dma-buf: Update docs for SYNC ioctl MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Just a bit of wording polish plus mentioning that it can fail and must be restarted. Requested by Sumit. v2: Fix them typos (Hans). Cc: Chris Wilson Cc: Tiago Vignatti Cc: Stéphane Marchesin Cc: David Herrmann Cc: Sumit Semwal Cc: Daniel Vetter CC: linux-media@vger.kernel.org Cc: dri-devel@lists.freedesktop.org Cc: linaro-mm-sig@lists.linaro.org Cc: intel-gfx@lists.freedesktop.org Cc: devel@driverdev.osuosl.org Cc: Hans Verkuil Acked-by: Sumit Semwal Acked-by: Hans Verkuil Signed-off-by: Daniel Vetter --- Documentation/dma-buf-sharing.txt | 11 ++++++----- drivers/dma-buf/dma-buf.c | 2 +- 2 files changed, 7 insertions(+), 6 deletions(-) diff --git a/Documentation/dma-buf-sharing.txt b/Documentation/dma-buf-sharing.txt index 32ac32e..ca44c58 100644 --- a/Documentation/dma-buf-sharing.txt +++ b/Documentation/dma-buf-sharing.txt @@ -352,7 +352,8 @@ Being able to mmap an export dma-buf buffer object has 2 main use-cases: No special interfaces, userspace simply calls mmap on the dma-buf fd, making sure that the cache synchronization ioctl (DMA_BUF_IOCTL_SYNC) is *always* - used when the access happens. This is discussed next paragraphs. + used when the access happens. Note that DMA_BUF_IOCTL_SYNC can fail with + -EAGAIN or -EINTR, in which case it must be restarted. Some systems might need some sort of cache coherency management e.g. when CPU and GPU domains are being accessed through dma-buf at the same time. To @@ -366,10 +367,10 @@ Being able to mmap an export dma-buf buffer object has 2 main use-cases: want (with the new data being consumed by the GPU or say scanout device) - munmap once you don't need the buffer any more - Therefore, for correctness and optimal performance, systems with the memory - cache shared by the GPU and CPU i.e. the "coherent" and also the - "incoherent" are always required to use SYNC_START and SYNC_END before and - after, respectively, when accessing the mapped address. + For correctness and optimal performance, it is always required to use + SYNC_START and SYNC_END before and after, respectively, when accessing the + mapped address. Userspace cannot rely on coherent access, even when there + are systems where it just works without calling these ioctls. 2. Supporting existing mmap interfaces in importers diff --git a/drivers/dma-buf/dma-buf.c b/drivers/dma-buf/dma-buf.c index 774a60f..4a2c07e 100644 --- a/drivers/dma-buf/dma-buf.c +++ b/drivers/dma-buf/dma-buf.c @@ -612,7 +612,7 @@ EXPORT_SYMBOL_GPL(dma_buf_begin_cpu_access); * @dmabuf: [in] buffer to complete cpu access for. * @direction: [in] length of range for cpu access. * - * This call must always succeed. + * Can return negative error values, returns 0 on success. */ int dma_buf_end_cpu_access(struct dma_buf *dmabuf, enum dma_data_direction direction) -- cgit v1.1 From b47bcb93bbf201e9c5af698945755efeb60c0bc8 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Mon, 29 Feb 2016 10:53:11 -0500 Subject: dma-buf/fence: fix fence_is_later v2 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit A fence is never later than itself. This caused a bunch of overhead for AMDGPU. v2: simplify check as suggested by Michel. Signed-off-by: Christian König Reviewed-by: Michel Dänzer Reviewed-by: Alex Deucher Reviewed-by: Gustavo Padovan Reviewed-by: Maarten Lankhorst Signed-off-by: Alex Deucher Signed-off-by: Sumit Semwal Signed-off-by: Daniel Vetter --- include/linux/fence.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/include/linux/fence.h b/include/linux/fence.h index bb52201..5aa95eb 100644 --- a/include/linux/fence.h +++ b/include/linux/fence.h @@ -292,7 +292,7 @@ static inline bool fence_is_later(struct fence *f1, struct fence *f2) if (WARN_ON(f1->context != f2->context)) return false; - return f1->seqno - f2->seqno < INT_MAX; + return (int)(f1->seqno - f2->seqno) > 0; } /** -- cgit v1.1 From a82d397bffc2ff3a72e9b8ca3850771ff5cf451f Mon Sep 17 00:00:00 2001 From: Colin Ian King Date: Fri, 18 Mar 2016 16:47:29 +0000 Subject: drm/amd/powerplay: fix memory leak of tdp_table tdp_table is being leaked on failed allocations of hwmgr->dyn_state.cac_dtp_table. kfree tdp_table on the error return path to fix the leak. Signed-off-by: Colin Ian King Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/powerplay/hwmgr/tonga_processpptables.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_processpptables.c b/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_processpptables.c index 34f4bef..b156481 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_processpptables.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_processpptables.c @@ -512,8 +512,10 @@ static int get_cac_tdp_table( hwmgr->dyn_state.cac_dtp_table = kzalloc(table_size, GFP_KERNEL); - if (NULL == hwmgr->dyn_state.cac_dtp_table) + if (NULL == hwmgr->dyn_state.cac_dtp_table) { + kfree(tdp_table); return -ENOMEM; + } memset(hwmgr->dyn_state.cac_dtp_table, 0x00, table_size); -- cgit v1.1 From 0ccbf11988d30c826810884fb9d2743e3923a464 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Tue, 15 Mar 2016 11:22:27 -0400 Subject: drm/amdgpu: clean up path handling for powerplay MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Use $(FULL_AMD_PATH) like everything else. Reviewed-by: Christian König Reviewed-by: Michel Dänzer Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/powerplay/Makefile | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/drivers/gpu/drm/amd/powerplay/Makefile b/drivers/gpu/drm/amd/powerplay/Makefile index e195bf5..043e6eb 100644 --- a/drivers/gpu/drm/amd/powerplay/Makefile +++ b/drivers/gpu/drm/amd/powerplay/Makefile @@ -1,17 +1,17 @@ subdir-ccflags-y += -Iinclude/drm \ - -Idrivers/gpu/drm/amd/powerplay/inc/ \ - -Idrivers/gpu/drm/amd/include/asic_reg \ - -Idrivers/gpu/drm/amd/include \ - -Idrivers/gpu/drm/amd/powerplay/smumgr\ - -Idrivers/gpu/drm/amd/powerplay/hwmgr \ - -Idrivers/gpu/drm/amd/powerplay/eventmgr + -I$(FULL_AMD_PATH)/powerplay/inc/ \ + -I$(FULL_AMD_PATH)/include/asic_reg \ + -I$(FULL_AMD_PATH)/include \ + -I$(FULL_AMD_PATH)/powerplay/smumgr\ + -I$(FULL_AMD_PATH)/powerplay/hwmgr \ + -I$(FULL_AMD_PATH)/powerplay/eventmgr AMD_PP_PATH = ../powerplay PP_LIBS = smumgr hwmgr eventmgr -AMD_POWERPLAY = $(addsuffix /Makefile,$(addprefix drivers/gpu/drm/amd/powerplay/,$(PP_LIBS))) +AMD_POWERPLAY = $(addsuffix /Makefile,$(addprefix $(FULL_AMD_PATH)/powerplay/,$(PP_LIBS))) include $(AMD_POWERPLAY) -- cgit v1.1 From 0d2b42b0bdba45c82d29d794ea30a4c90e3f4098 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Fri, 18 Mar 2016 19:29:51 +0100 Subject: drm/amdgpu: Revert "remove the userptr rmn->lock" MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This reverts commit c02196834456f2d5fad334088b70e98ce4967c34. In the meantime we moved get_user_pages() outside of the reservation lock, so that shouldn't be an issue any more Signed-off-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c | 22 ++++++++++++++-------- 1 file changed, 14 insertions(+), 8 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c index d7ec9bd..c47f2222 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c @@ -48,7 +48,8 @@ struct amdgpu_mn { /* protected by adev->mn_lock */ struct hlist_node node; - /* objects protected by mm->mmap_sem */ + /* objects protected by lock */ + struct mutex lock; struct rb_root objects; }; @@ -72,7 +73,7 @@ static void amdgpu_mn_destroy(struct work_struct *work) struct amdgpu_bo *bo, *next_bo; mutex_lock(&adev->mn_lock); - down_write(&rmn->mm->mmap_sem); + mutex_lock(&rmn->lock); hash_del(&rmn->node); rbtree_postorder_for_each_entry_safe(node, next_node, &rmn->objects, it.rb) { @@ -82,7 +83,7 @@ static void amdgpu_mn_destroy(struct work_struct *work) } kfree(node); } - up_write(&rmn->mm->mmap_sem); + mutex_unlock(&rmn->lock); mutex_unlock(&adev->mn_lock); mmu_notifier_unregister_no_release(&rmn->mn, rmn->mm); kfree(rmn); @@ -126,6 +127,8 @@ static void amdgpu_mn_invalidate_range_start(struct mmu_notifier *mn, /* notification is exclusive, but interval is inclusive */ end -= 1; + mutex_lock(&rmn->lock); + it = interval_tree_iter_first(&rmn->objects, start, end); while (it) { struct amdgpu_mn_node *node; @@ -160,6 +163,8 @@ static void amdgpu_mn_invalidate_range_start(struct mmu_notifier *mn, amdgpu_bo_unreserve(bo); } } + + mutex_unlock(&rmn->lock); } static const struct mmu_notifier_ops amdgpu_mn_ops = { @@ -196,6 +201,7 @@ static struct amdgpu_mn *amdgpu_mn_get(struct amdgpu_device *adev) rmn->adev = adev; rmn->mm = mm; rmn->mn.ops = &amdgpu_mn_ops; + mutex_init(&rmn->lock); rmn->objects = RB_ROOT; r = __mmu_notifier_register(&rmn->mn, mm); @@ -242,7 +248,7 @@ int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr) INIT_LIST_HEAD(&bos); - down_write(&rmn->mm->mmap_sem); + mutex_lock(&rmn->lock); while ((it = interval_tree_iter_first(&rmn->objects, addr, end))) { kfree(node); @@ -256,7 +262,7 @@ int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr) if (!node) { node = kmalloc(sizeof(struct amdgpu_mn_node), GFP_KERNEL); if (!node) { - up_write(&rmn->mm->mmap_sem); + mutex_unlock(&rmn->lock); return -ENOMEM; } } @@ -271,7 +277,7 @@ int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr) interval_tree_insert(&node->it, &rmn->objects); - up_write(&rmn->mm->mmap_sem); + mutex_unlock(&rmn->lock); return 0; } @@ -297,7 +303,7 @@ void amdgpu_mn_unregister(struct amdgpu_bo *bo) return; } - down_write(&rmn->mm->mmap_sem); + mutex_lock(&rmn->lock); /* save the next list entry for later */ head = bo->mn_list.next; @@ -312,6 +318,6 @@ void amdgpu_mn_unregister(struct amdgpu_bo *bo) kfree(node); } - up_write(&rmn->mm->mmap_sem); + mutex_unlock(&rmn->lock); mutex_unlock(&adev->mn_lock); } -- cgit v1.1 From ae20f12d2de6629ee6f679ccf22f9b7b209c464d Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Fri, 18 Mar 2016 19:29:52 +0100 Subject: drm/amdgpu: add invalidate_page callback for userptrs MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Otherwise we can run into problems with the writeback code. Signed-off-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c | 98 +++++++++++++++++++++++++--------- 1 file changed, 72 insertions(+), 26 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c index c47f2222..9f4a45c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c @@ -106,6 +106,76 @@ static void amdgpu_mn_release(struct mmu_notifier *mn, } /** + * amdgpu_mn_invalidate_node - unmap all BOs of a node + * + * @node: the node with the BOs to unmap + * + * We block for all BOs and unmap them by move them + * into system domain again. + */ +static void amdgpu_mn_invalidate_node(struct amdgpu_mn_node *node, + unsigned long start, + unsigned long end) +{ + struct amdgpu_bo *bo; + long r; + + list_for_each_entry(bo, &node->bos, mn_list) { + + if (!amdgpu_ttm_tt_affect_userptr(bo->tbo.ttm, start, end)) + continue; + + r = amdgpu_bo_reserve(bo, true); + if (r) { + DRM_ERROR("(%ld) failed to reserve user bo\n", r); + continue; + } + + r = reservation_object_wait_timeout_rcu(bo->tbo.resv, + true, false, MAX_SCHEDULE_TIMEOUT); + if (r <= 0) + DRM_ERROR("(%ld) failed to wait for user bo\n", r); + + amdgpu_ttm_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU); + r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false); + if (r) + DRM_ERROR("(%ld) failed to validate user bo\n", r); + + amdgpu_bo_unreserve(bo); + } +} + +/** + * amdgpu_mn_invalidate_page - callback to notify about mm change + * + * @mn: our notifier + * @mn: the mm this callback is about + * @address: address of invalidate page + * + * Invalidation of a single page. Blocks for all BOs mapping it + * and unmap them by move them into system domain again. + */ +static void amdgpu_mn_invalidate_page(struct mmu_notifier *mn, + struct mm_struct *mm, + unsigned long address) +{ + struct amdgpu_mn *rmn = container_of(mn, struct amdgpu_mn, mn); + struct interval_tree_node *it; + + mutex_lock(&rmn->lock); + + it = interval_tree_iter_first(&rmn->objects, address, address); + if (it) { + struct amdgpu_mn_node *node; + + node = container_of(it, struct amdgpu_mn_node, it); + amdgpu_mn_invalidate_node(node, address, address); + } + + mutex_unlock(&rmn->lock); +} + +/** * amdgpu_mn_invalidate_range_start - callback to notify about mm change * * @mn: our notifier @@ -132,36 +202,11 @@ static void amdgpu_mn_invalidate_range_start(struct mmu_notifier *mn, it = interval_tree_iter_first(&rmn->objects, start, end); while (it) { struct amdgpu_mn_node *node; - struct amdgpu_bo *bo; - long r; node = container_of(it, struct amdgpu_mn_node, it); it = interval_tree_iter_next(it, start, end); - list_for_each_entry(bo, &node->bos, mn_list) { - - if (!amdgpu_ttm_tt_affect_userptr(bo->tbo.ttm, start, - end)) - continue; - - r = amdgpu_bo_reserve(bo, true); - if (r) { - DRM_ERROR("(%ld) failed to reserve user bo\n", r); - continue; - } - - r = reservation_object_wait_timeout_rcu(bo->tbo.resv, - true, false, MAX_SCHEDULE_TIMEOUT); - if (r <= 0) - DRM_ERROR("(%ld) failed to wait for user bo\n", r); - - amdgpu_ttm_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU); - r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false); - if (r) - DRM_ERROR("(%ld) failed to validate user bo\n", r); - - amdgpu_bo_unreserve(bo); - } + amdgpu_mn_invalidate_node(node, start, end); } mutex_unlock(&rmn->lock); @@ -169,6 +214,7 @@ static void amdgpu_mn_invalidate_range_start(struct mmu_notifier *mn, static const struct mmu_notifier_ops amdgpu_mn_ops = { .release = amdgpu_mn_release, + .invalidate_page = amdgpu_mn_invalidate_page, .invalidate_range_start = amdgpu_mn_invalidate_range_start, }; -- cgit v1.1 From b36f7d26a7fdc0b07b1217368ee09bb8560269f8 Mon Sep 17 00:00:00 2001 From: Dave Airlie Date: Tue, 22 Mar 2016 09:38:18 +1000 Subject: drm/radeon/mst: fix regression in lane/link handling. The function this used changed in 092c96a8ab9d1bd60ada2ed385cc364ce084180e drm/radeon: fix dp link rate selection (v2) However for MST we should just always train to the max link/rate. Though we probably need to limit this for future hw, in theory radeon won't support it. This fixes my 30" monitor with MST enabled. Cc: stable@vger.kernel.org # v4.4 Signed-off-by: Dave Airlie Signed-off-by: Alex Deucher --- drivers/gpu/drm/radeon/radeon_dp_mst.c | 12 ++---------- 1 file changed, 2 insertions(+), 10 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon_dp_mst.c b/drivers/gpu/drm/radeon/radeon_dp_mst.c index df7a171..9d210bb 100644 --- a/drivers/gpu/drm/radeon/radeon_dp_mst.c +++ b/drivers/gpu/drm/radeon/radeon_dp_mst.c @@ -525,17 +525,9 @@ static bool radeon_mst_mode_fixup(struct drm_encoder *encoder, drm_mode_set_crtcinfo(adjusted_mode, 0); { struct radeon_connector_atom_dig *dig_connector; - int ret; - dig_connector = mst_enc->connector->con_priv; - ret = radeon_dp_get_dp_link_config(&mst_enc->connector->base, - dig_connector->dpcd, adjusted_mode->clock, - &dig_connector->dp_lane_count, - &dig_connector->dp_clock); - if (ret) { - dig_connector->dp_lane_count = 0; - dig_connector->dp_clock = 0; - } + dig_connector->dp_lane_count = drm_dp_max_lane_count(dig_connector->dpcd); + dig_connector->dp_clock = drm_dp_max_link_rate(dig_connector->dpcd); DRM_DEBUG_KMS("dig clock %p %d %d\n", dig_connector, dig_connector->dp_lane_count, dig_connector->dp_clock); } -- cgit v1.1 From 1135035d9275ef9aecad23fc715a69ff78904adf Mon Sep 17 00:00:00 2001 From: Dave Airlie Date: Tue, 22 Mar 2016 09:38:19 +1000 Subject: drm/radeon/mst: cleanup code indentation This was all sorts of ugly from when I hacked it up, just clean it up now and remove the extra indents. Signed-off-by: Dave Airlie Signed-off-by: Alex Deucher --- drivers/gpu/drm/radeon/radeon_dp_mst.c | 14 ++++++-------- 1 file changed, 6 insertions(+), 8 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon_dp_mst.c b/drivers/gpu/drm/radeon/radeon_dp_mst.c index 9d210bb..43cffb5 100644 --- a/drivers/gpu/drm/radeon/radeon_dp_mst.c +++ b/drivers/gpu/drm/radeon/radeon_dp_mst.c @@ -510,6 +510,7 @@ static bool radeon_mst_mode_fixup(struct drm_encoder *encoder, { struct radeon_encoder_mst *mst_enc; struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); + struct radeon_connector_atom_dig *dig_connector; int bpp = 24; mst_enc = radeon_encoder->enc_priv; @@ -523,14 +524,11 @@ static bool radeon_mst_mode_fixup(struct drm_encoder *encoder, drm_mode_set_crtcinfo(adjusted_mode, 0); - { - struct radeon_connector_atom_dig *dig_connector; - dig_connector = mst_enc->connector->con_priv; - dig_connector->dp_lane_count = drm_dp_max_lane_count(dig_connector->dpcd); - dig_connector->dp_clock = drm_dp_max_link_rate(dig_connector->dpcd); - DRM_DEBUG_KMS("dig clock %p %d %d\n", dig_connector, - dig_connector->dp_lane_count, dig_connector->dp_clock); - } + dig_connector = mst_enc->connector->con_priv; + dig_connector->dp_lane_count = drm_dp_max_lane_count(dig_connector->dpcd); + dig_connector->dp_clock = drm_dp_max_link_rate(dig_connector->dpcd); + DRM_DEBUG_KMS("dig clock %p %d %d\n", dig_connector, + dig_connector->dp_lane_count, dig_connector->dp_clock); return true; } -- cgit v1.1