* Few TMZ BO move patches
@ 2020-03-02 12:17 Christian König
2020-03-02 12:17 ` [PATCH 1/3] drm/amdgpu: also add the TMZ flag to GART Christian König
` (4 more replies)
0 siblings, 5 replies; 8+ messages in thread
From: Christian König @ 2020-03-02 12:17 UTC (permalink / raw)
To: amd-gfx
Because of a shift in priorities I won't work on TMZ this week.
So attached are a few smaller patches I already prepared, but the bounce copy for system eviction is still missing.
Patches are totally untested, but I anybody find them useful feel free to test and review them.
Regards,
Christian.
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
^ permalink raw reply [flat|nested] 8+ messages in thread* [PATCH 1/3] drm/amdgpu: also add the TMZ flag to GART 2020-03-02 12:17 Few TMZ BO move patches Christian König @ 2020-03-02 12:17 ` Christian König 2020-03-02 12:17 ` [PATCH 2/3] drm/amdgpu: add TMZ handling to amdgpu_move_blit Christian König ` (3 subsequent siblings) 4 siblings, 0 replies; 8+ messages in thread From: Christian König @ 2020-03-02 12:17 UTC (permalink / raw) To: amd-gfx This is necessary for TMZ handling during buffer moves and scanout. Signed-off-by: Christian König <christian.koenig@amd.com> --- drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c index 660867cf2597..18f6caa7b6a2 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -1033,6 +1033,9 @@ int amdgpu_ttm_gart_bind(struct amdgpu_device *adev, struct amdgpu_ttm_tt *gtt = (void *)ttm; int r; + if (amdgpu_bo_encrypted(abo)) + flags |= AMDGPU_PTE_TMZ; + if (abo->flags & AMDGPU_GEM_CREATE_MQD_GFX9) { uint64_t page_idx = 1; -- 2.17.1 _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx ^ permalink raw reply related [flat|nested] 8+ messages in thread
* [PATCH 2/3] drm/amdgpu: add TMZ handling to amdgpu_move_blit 2020-03-02 12:17 Few TMZ BO move patches Christian König 2020-03-02 12:17 ` [PATCH 1/3] drm/amdgpu: also add the TMZ flag to GART Christian König @ 2020-03-02 12:17 ` Christian König 2020-03-02 12:17 ` [PATCH 3/3] drm/amdgpu: stop evicting encrypted BOs to swap Christian König ` (2 subsequent siblings) 4 siblings, 0 replies; 8+ messages in thread From: Christian König @ 2020-03-02 12:17 UTC (permalink / raw) To: amd-gfx This way we should be at least able to move buffers from VRAM to GTT. Signed-off-by: Christian König <christian.koenig@amd.com> --- drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 28 +++++++++++++++++-------- drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h | 2 +- 2 files changed, 20 insertions(+), 10 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c index 18f6caa7b6a2..25d3b93f446e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -65,7 +65,7 @@ static int amdgpu_map_buffer(struct ttm_buffer_object *bo, struct ttm_mem_reg *mem, unsigned num_pages, uint64_t offset, unsigned window, - struct amdgpu_ring *ring, + struct amdgpu_ring *ring, bool tmz, uint64_t *addr); static int amdgpu_invalidate_caches(struct ttm_bo_device *bdev, uint32_t flags) @@ -295,17 +295,23 @@ static struct drm_mm_node *amdgpu_find_mm_node(struct ttm_mem_reg *mem, /** * amdgpu_copy_ttm_mem_to_mem - Helper function for copy + * @adev: amdgpu device + * @src: buffer/address where to read from + * @dst: buffer/address where to write to + * @size: number of bytes to copy + * @tmz: if a secure copy should be used + * @resv: resv object to sync to + * @f: Returns the last fence if multiple jobs are submitted. * * The function copies @size bytes from {src->mem + src->offset} to * {dst->mem + dst->offset}. src->bo and dst->bo could be same BO for a * move and different for a BO to BO copy. * - * @f: Returns the last fence if multiple jobs are submitted. */ int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev, struct amdgpu_copy_mem *src, struct amdgpu_copy_mem *dst, - uint64_t size, + uint64_t size, bool tmz, struct dma_resv *resv, struct dma_fence **f) { @@ -357,7 +363,7 @@ int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev, if (src->mem->start == AMDGPU_BO_INVALID_OFFSET) { r = amdgpu_map_buffer(src->bo, src->mem, PFN_UP(cur_size + src_page_offset), - src_node_start, 0, ring, + src_node_start, 0, ring, tmz, &from); if (r) goto error; @@ -370,7 +376,7 @@ int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev, if (dst->mem->start == AMDGPU_BO_INVALID_OFFSET) { r = amdgpu_map_buffer(dst->bo, dst->mem, PFN_UP(cur_size + dst_page_offset), - dst_node_start, 1, ring, + dst_node_start, 1, ring, tmz, &to); if (r) goto error; @@ -378,7 +384,7 @@ int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev, } r = amdgpu_copy_buffer(ring, from, to, cur_size, - resv, &next, false, true, false); + resv, &next, false, true, tmz); if (r) goto error; @@ -430,6 +436,7 @@ static int amdgpu_move_blit(struct ttm_buffer_object *bo, struct ttm_mem_reg *old_mem) { struct amdgpu_device *adev = amdgpu_ttm_adev(bo->bdev); + struct amdgpu_bo *abo = ttm_to_amdgpu_bo(bo); struct amdgpu_copy_mem src, dst; struct dma_fence *fence = NULL; int r; @@ -443,14 +450,14 @@ static int amdgpu_move_blit(struct ttm_buffer_object *bo, r = amdgpu_ttm_copy_mem_to_mem(adev, &src, &dst, new_mem->num_pages << PAGE_SHIFT, + amdgpu_bo_encrypted(abo), bo->base.resv, &fence); if (r) goto error; /* clear the space being freed */ if (old_mem->mem_type == TTM_PL_VRAM && - (ttm_to_amdgpu_bo(bo)->flags & - AMDGPU_GEM_CREATE_VRAM_WIPE_ON_RELEASE)) { + (abo->flags & AMDGPU_GEM_CREATE_VRAM_WIPE_ON_RELEASE)) { struct dma_fence *wipe_fence = NULL; r = amdgpu_fill_buffer(ttm_to_amdgpu_bo(bo), AMDGPU_POISON, @@ -2023,7 +2030,7 @@ int amdgpu_mmap(struct file *filp, struct vm_area_struct *vma) static int amdgpu_map_buffer(struct ttm_buffer_object *bo, struct ttm_mem_reg *mem, unsigned num_pages, uint64_t offset, unsigned window, - struct amdgpu_ring *ring, + struct amdgpu_ring *ring, bool tmz, uint64_t *addr) { struct amdgpu_ttm_tt *gtt = (void *)bo->ttm; @@ -2064,6 +2071,9 @@ static int amdgpu_map_buffer(struct ttm_buffer_object *bo, dma_address = >t->ttm.dma_address[offset >> PAGE_SHIFT]; flags = amdgpu_ttm_tt_pte_flags(adev, ttm, mem); + if (tmz) + flags |= AMDGPU_PTE_TMZ; + r = amdgpu_gart_map(adev, 0, num_pages, dma_address, flags, &job->ibs[0].ptr[num_dw]); if (r) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h index 2c4ad5b589d0..6001ab9bd932 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h @@ -91,7 +91,7 @@ int amdgpu_copy_buffer(struct amdgpu_ring *ring, uint64_t src_offset, int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev, struct amdgpu_copy_mem *src, struct amdgpu_copy_mem *dst, - uint64_t size, + uint64_t size, bool tmz, struct dma_resv *resv, struct dma_fence **f); int amdgpu_fill_buffer(struct amdgpu_bo *bo, -- 2.17.1 _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx ^ permalink raw reply related [flat|nested] 8+ messages in thread
* [PATCH 3/3] drm/amdgpu: stop evicting encrypted BOs to swap 2020-03-02 12:17 Few TMZ BO move patches Christian König 2020-03-02 12:17 ` [PATCH 1/3] drm/amdgpu: also add the TMZ flag to GART Christian König 2020-03-02 12:17 ` [PATCH 2/3] drm/amdgpu: add TMZ handling to amdgpu_move_blit Christian König @ 2020-03-02 12:17 ` Christian König 2020-03-02 21:37 ` Alex Deucher 2020-03-02 13:10 ` Few TMZ BO move patches Pierre-Eric Pelloux-Prayer 2020-03-02 21:47 ` Luben Tuikov 4 siblings, 1 reply; 8+ messages in thread From: Christian König @ 2020-03-02 12:17 UTC (permalink / raw) To: amd-gfx Swapping out encrypted BOs doesn't work because they can't change their physical location without going through a bounce copy. As a workaround disable evicting encrypted BOs to the system domain for now. Signed-off-by: Christian König <christian.koenig@amd.com> --- drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c index 25d3b93f446e..96d97523da65 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -1552,6 +1552,9 @@ static bool amdgpu_ttm_bo_eviction_valuable(struct ttm_buffer_object *bo, switch (bo->mem.mem_type) { case TTM_PL_TT: + if (amdgpu_bo_is_amdgpu_bo(bo) && + amdgpu_bo_encrypted(ttm_to_amdgpu_bo(bo))) + return false; return true; case TTM_PL_VRAM: -- 2.17.1 _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx ^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: [PATCH 3/3] drm/amdgpu: stop evicting encrypted BOs to swap 2020-03-02 12:17 ` [PATCH 3/3] drm/amdgpu: stop evicting encrypted BOs to swap Christian König @ 2020-03-02 21:37 ` Alex Deucher 0 siblings, 0 replies; 8+ messages in thread From: Alex Deucher @ 2020-03-02 21:37 UTC (permalink / raw) To: Christian König; +Cc: amd-gfx list On Mon, Mar 2, 2020 at 7:18 AM Christian König <ckoenig.leichtzumerken@gmail.com> wrote: > > Swapping out encrypted BOs doesn't work because they can't change > their physical location without going through a bounce copy. > > As a workaround disable evicting encrypted BOs to the system > domain for now. > > Signed-off-by: Christian König <christian.koenig@amd.com> Series is: Reviewed-by: Alex Deucher <alexander.deucher@amd.com> > --- > drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 3 +++ > 1 file changed, 3 insertions(+) > > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c > index 25d3b93f446e..96d97523da65 100644 > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c > @@ -1552,6 +1552,9 @@ static bool amdgpu_ttm_bo_eviction_valuable(struct ttm_buffer_object *bo, > > switch (bo->mem.mem_type) { > case TTM_PL_TT: > + if (amdgpu_bo_is_amdgpu_bo(bo) && > + amdgpu_bo_encrypted(ttm_to_amdgpu_bo(bo))) > + return false; > return true; > > case TTM_PL_VRAM: > -- > 2.17.1 > > _______________________________________________ > amd-gfx mailing list > amd-gfx@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/amd-gfx _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: Few TMZ BO move patches 2020-03-02 12:17 Few TMZ BO move patches Christian König ` (2 preceding siblings ...) 2020-03-02 12:17 ` [PATCH 3/3] drm/amdgpu: stop evicting encrypted BOs to swap Christian König @ 2020-03-02 13:10 ` Pierre-Eric Pelloux-Prayer 2020-03-02 21:47 ` Luben Tuikov 4 siblings, 0 replies; 8+ messages in thread From: Pierre-Eric Pelloux-Prayer @ 2020-03-02 13:10 UTC (permalink / raw) To: Christian König, amd-gfx Hi Christian, The 3 patches are: Tested-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com> Regards, Pierre-Eric On 02/03/2020 13:17, Christian König wrote: > Because of a shift in priorities I won't work on TMZ this week. > > So attached are a few smaller patches I already prepared, but the bounce copy for system eviction is still missing. > > Patches are totally untested, but I anybody find them useful feel free to test and review them. > > Regards, > Christian. > > > _______________________________________________ > amd-gfx mailing list > amd-gfx@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/amd-gfx > _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: Few TMZ BO move patches 2020-03-02 12:17 Few TMZ BO move patches Christian König ` (3 preceding siblings ...) 2020-03-02 13:10 ` Few TMZ BO move patches Pierre-Eric Pelloux-Prayer @ 2020-03-02 21:47 ` Luben Tuikov 2020-03-04 12:41 ` Christian König 4 siblings, 1 reply; 8+ messages in thread From: Luben Tuikov @ 2020-03-02 21:47 UTC (permalink / raw) To: Christian König, amd-gfx Thanks Christian. I'll take a look and play with it. Regards, Luben On 2020-03-02 7:17 a.m., Christian König wrote: > Because of a shift in priorities I won't work on TMZ this week. > > So attached are a few smaller patches I already prepared, but the bounce copy for system eviction is still missing. > > Patches are totally untested, but I anybody find them useful feel free to test and review them. > > Regards, > Christian. > > > _______________________________________________ > amd-gfx mailing list > amd-gfx@lists.freedesktop.org > https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.freedesktop.org%2Fmailman%2Flistinfo%2Famd-gfx&data=02%7C01%7Cluben.tuikov%40amd.com%7Cbb0a04fde1ea4b7960d408d7bea3c2d0%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637187482887561950&sdata=F1KJ%2BX%2F04Kkj7Wzw4bsc1TXLReLNVkrcw9R5ohbgvOQ%3D&reserved=0 > _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: Few TMZ BO move patches 2020-03-02 21:47 ` Luben Tuikov @ 2020-03-04 12:41 ` Christian König 0 siblings, 0 replies; 8+ messages in thread From: Christian König @ 2020-03-04 12:41 UTC (permalink / raw) To: Luben Tuikov, amd-gfx Had to commit that today since I need to work on a different project. If you find anything let me know and we can fix it in a follow up patch. Christian. Am 02.03.20 um 22:47 schrieb Luben Tuikov: > Thanks Christian. I'll take a look and play with it. > > Regards, > Luben > > On 2020-03-02 7:17 a.m., Christian König wrote: >> Because of a shift in priorities I won't work on TMZ this week. >> >> So attached are a few smaller patches I already prepared, but the bounce copy for system eviction is still missing. >> >> Patches are totally untested, but I anybody find them useful feel free to test and review them. >> >> Regards, >> Christian. >> >> >> _______________________________________________ >> amd-gfx mailing list >> amd-gfx@lists.freedesktop.org >> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.freedesktop.org%2Fmailman%2Flistinfo%2Famd-gfx&data=02%7C01%7Cluben.tuikov%40amd.com%7Cbb0a04fde1ea4b7960d408d7bea3c2d0%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637187482887561950&sdata=F1KJ%2BX%2F04Kkj7Wzw4bsc1TXLReLNVkrcw9R5ohbgvOQ%3D&reserved=0 >> _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx ^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2020-03-04 12:42 UTC | newest] Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2020-03-02 12:17 Few TMZ BO move patches Christian König 2020-03-02 12:17 ` [PATCH 1/3] drm/amdgpu: also add the TMZ flag to GART Christian König 2020-03-02 12:17 ` [PATCH 2/3] drm/amdgpu: add TMZ handling to amdgpu_move_blit Christian König 2020-03-02 12:17 ` [PATCH 3/3] drm/amdgpu: stop evicting encrypted BOs to swap Christian König 2020-03-02 21:37 ` Alex Deucher 2020-03-02 13:10 ` Few TMZ BO move patches Pierre-Eric Pelloux-Prayer 2020-03-02 21:47 ` Luben Tuikov 2020-03-04 12:41 ` Christian König
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.