diff options
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c')
| -rw-r--r-- | drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c | 56 |
1 files changed, 27 insertions, 29 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c index 3a7bab87b5d8..63f62c670df5 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c @@ -129,6 +129,7 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned int num_ibs, struct amdgpu_ib *ib = &ibs[0]; struct dma_fence *tmp = NULL; struct amdgpu_fence *af; + struct amdgpu_fence *vm_af; bool need_ctx_switch; struct amdgpu_vm *vm; uint64_t fence_ctx; @@ -215,25 +216,28 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned int num_ibs, dma_fence_put(tmp); } - if ((ib->flags & AMDGPU_IB_FLAG_EMIT_MEM_SYNC) && ring->funcs->emit_mem_sync) - ring->funcs->emit_mem_sync(ring); + if (job) { + vm_af = job->hw_vm_fence; + /* VM sequence */ + vm_af->ib_wptr = ring->wptr; + amdgpu_vm_flush(ring, job, need_pipe_sync); + vm_af->ib_dw_size = + amdgpu_ring_get_dw_distance(ring, vm_af->ib_wptr, ring->wptr); + } - if (ring->funcs->emit_wave_limit && - ring->hw_prio == AMDGPU_GFX_PIPE_PRIO_HIGH) - ring->funcs->emit_wave_limit(ring, true); + /* IB sequence */ + af->ib_wptr = ring->wptr; + amdgpu_ring_ib_begin(ring); if (ring->funcs->insert_start) ring->funcs->insert_start(ring); - if (job) { - r = amdgpu_vm_flush(ring, job, need_pipe_sync); - if (r) { - amdgpu_ring_undo(ring); - goto free_fence; - } - } + if ((ib->flags & AMDGPU_IB_FLAG_EMIT_MEM_SYNC) && ring->funcs->emit_mem_sync) + ring->funcs->emit_mem_sync(ring); - amdgpu_ring_ib_begin(ring); + if (ring->funcs->emit_wave_limit && + ring->hw_prio == AMDGPU_GFX_PIPE_PRIO_HIGH) + ring->funcs->emit_wave_limit(ring, true); if (ring->funcs->emit_gfx_shadow && adev->gfx.cp_gfx_shadow) amdgpu_ring_emit_gfx_shadow(ring, shadow_va, csa_va, gds_va, @@ -243,6 +247,9 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned int num_ibs, cond_exec = amdgpu_ring_init_cond_exec(ring, ring->cond_exe_gpu_addr); + /* Skip the IB for guilty contexts */ + af->skip_ib_dw_start_offset = + amdgpu_ring_get_dw_distance(ring, af->ib_wptr, ring->wptr); amdgpu_device_flush_hdp(adev, ring); if (need_ctx_switch) @@ -281,6 +288,9 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned int num_ibs, amdgpu_ring_emit_frame_cntl(ring, false, secure); amdgpu_device_invalidate_hdp(adev, ring); + /* Skip the IB for guilty contexts */ + af->skip_ib_dw_end_offset = + amdgpu_ring_get_dw_distance(ring, af->ib_wptr, ring->wptr); if (ib->flags & AMDGPU_IB_FLAG_TC_WB_NOT_INVALIDATE) fence_flags |= AMDGPU_FENCE_FLAG_TC_WB_ONLY; @@ -297,14 +307,7 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned int num_ibs, amdgpu_ring_init_cond_exec(ring, ring->cond_exe_gpu_addr); } - r = amdgpu_fence_emit(ring, af, fence_flags); - if (r) { - dev_err(adev->dev, "failed to emit fence (%d)\n", r); - if (job && job->vmid) - amdgpu_vmid_reset(adev, ring->vm_hub, job->vmid); - amdgpu_ring_undo(ring); - goto free_fence; - } + amdgpu_fence_emit(ring, af, fence_flags); *f = &af->base; /* get a ref for the job */ if (job) @@ -323,15 +326,10 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned int num_ibs, ring->hw_prio == AMDGPU_GFX_PIPE_PRIO_HIGH) ring->funcs->emit_wave_limit(ring, false); - /* Save the wptr associated with this fence. - * This must be last for resets to work properly - * as we need to save the wptr associated with this - * fence so we know what rings contents to backup - * after we reset the queue. - */ - amdgpu_fence_save_wptr(af); - amdgpu_ring_ib_end(ring); + + af->ib_dw_size = amdgpu_ring_get_dw_distance(ring, af->ib_wptr, ring->wptr); + amdgpu_ring_commit(ring); return 0; |