From: Thomas Hellström thomas.hellstrom@intel.com
We want to be able to blit from within a ww transaction, so add blit functions that are able to do that. Also take care to unlock the blit batch-buffer after use so it isn't recycled locked.
Signed-off-by: Thomas Hellström thomas.hellstrom@intel.com Cc: Matthew Auld matthew.auld@intel.com --- .../gpu/drm/i915/gem/i915_gem_object_blt.c | 91 +++++++++++++------ .../gpu/drm/i915/gem/i915_gem_object_blt.h | 10 ++ 2 files changed, 72 insertions(+), 29 deletions(-)
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object_blt.c b/drivers/gpu/drm/i915/gem/i915_gem_object_blt.c index e0b873c3f46a..b41b076f6864 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_object_blt.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_object_blt.c @@ -145,11 +145,11 @@ move_obj_to_gpu(struct drm_i915_gem_object *obj, return i915_request_await_object(rq, obj, write); }
-int i915_gem_object_fill_blt(struct drm_i915_gem_object *obj, - struct intel_context *ce, - u32 value) +int i915_gem_object_ww_fill_blt(struct drm_i915_gem_object *obj, + struct i915_gem_ww_ctx *ww, + struct intel_context *ce, + u32 value) { - struct i915_gem_ww_ctx ww; struct i915_request *rq; struct i915_vma *batch; struct i915_vma *vma; @@ -159,22 +159,16 @@ int i915_gem_object_fill_blt(struct drm_i915_gem_object *obj, if (IS_ERR(vma)) return PTR_ERR(vma);
- i915_gem_ww_ctx_init(&ww, true); intel_engine_pm_get(ce->engine); -retry: - err = i915_gem_object_lock(obj, &ww); + err = intel_context_pin_ww(ce, ww); if (err) goto out;
- err = intel_context_pin_ww(ce, &ww); - if (err) - goto out; - - err = i915_vma_pin_ww(vma, &ww, 0, 0, PIN_USER); + err = i915_vma_pin_ww(vma, ww, 0, 0, PIN_USER); if (err) goto out_ctx;
- batch = intel_emit_vma_fill_blt(ce, vma, &ww, value); + batch = intel_emit_vma_fill_blt(ce, vma, ww, value); if (IS_ERR(batch)) { err = PTR_ERR(batch); goto out_vma; @@ -210,22 +204,43 @@ int i915_gem_object_fill_blt(struct drm_i915_gem_object *obj,
i915_request_add(rq); out_batch: + i915_gem_ww_unlock_single(batch->obj); intel_emit_vma_release(ce, batch); out_vma: i915_vma_unpin(vma); out_ctx: intel_context_unpin(ce); out: + intel_engine_pm_put(ce->engine); + return err; +} + +int i915_gem_object_fill_blt(struct drm_i915_gem_object *obj, + struct intel_context *ce, + u32 value) +{ + struct i915_gem_ww_ctx ww; + int err; + + i915_gem_ww_ctx_init(&ww, true); +retry: + err = i915_gem_object_lock(obj, &ww); + if (err) + goto out_err; + + err = i915_gem_object_ww_fill_blt(obj, &ww, ce, value); +out_err: if (err == -EDEADLK) { err = i915_gem_ww_ctx_backoff(&ww); if (!err) goto retry; } i915_gem_ww_ctx_fini(&ww); - intel_engine_pm_put(ce->engine); + return err; }
+ /* Wa_1209644611:icl,ehl */ static bool wa_1209644611_applies(struct drm_i915_private *i915, u32 size) { @@ -354,13 +369,13 @@ struct i915_vma *intel_emit_vma_copy_blt(struct intel_context *ce, return ERR_PTR(err); }
-int i915_gem_object_copy_blt(struct drm_i915_gem_object *src, - struct drm_i915_gem_object *dst, - struct intel_context *ce) +int i915_gem_object_ww_copy_blt(struct drm_i915_gem_object *src, + struct drm_i915_gem_object *dst, + struct i915_gem_ww_ctx *ww, + struct intel_context *ce) { struct i915_address_space *vm = ce->vm; struct i915_vma *vma[2], *batch; - struct i915_gem_ww_ctx ww; struct i915_request *rq; int err, i;
@@ -372,26 +387,20 @@ int i915_gem_object_copy_blt(struct drm_i915_gem_object *src, if (IS_ERR(vma[1])) return PTR_ERR(vma[1]);
- i915_gem_ww_ctx_init(&ww, true); intel_engine_pm_get(ce->engine); -retry: - err = i915_gem_object_lock(src, &ww); - if (!err) - err = i915_gem_object_lock(dst, &ww); - if (!err) - err = intel_context_pin_ww(ce, &ww); + err = intel_context_pin_ww(ce, ww); if (err) goto out;
- err = i915_vma_pin_ww(vma[0], &ww, 0, 0, PIN_USER); + err = i915_vma_pin_ww(vma[0], ww, 0, 0, PIN_USER); if (err) goto out_ctx;
- err = i915_vma_pin_ww(vma[1], &ww, 0, 0, PIN_USER); + err = i915_vma_pin_ww(vma[1], ww, 0, 0, PIN_USER); if (unlikely(err)) goto out_unpin_src;
- batch = intel_emit_vma_copy_blt(ce, &ww, vma[0], vma[1]); + batch = intel_emit_vma_copy_blt(ce, ww, vma[0], vma[1]); if (IS_ERR(batch)) { err = PTR_ERR(batch); goto out_unpin_dst; @@ -437,6 +446,7 @@ int i915_gem_object_copy_blt(struct drm_i915_gem_object *src,
i915_request_add(rq); out_batch: + i915_gem_ww_unlock_single(batch->obj); intel_emit_vma_release(ce, batch); out_unpin_dst: i915_vma_unpin(vma[1]); @@ -445,13 +455,36 @@ int i915_gem_object_copy_blt(struct drm_i915_gem_object *src, out_ctx: intel_context_unpin(ce); out: + intel_engine_pm_put(ce->engine); + return err; +} + +int i915_gem_object_copy_blt(struct drm_i915_gem_object *src, + struct drm_i915_gem_object *dst, + struct intel_context *ce) +{ + struct i915_gem_ww_ctx ww; + int err; + + i915_gem_ww_ctx_init(&ww, true); +retry: + err = i915_gem_object_lock(src, &ww); + if (err) + goto out_err; + + err = i915_gem_object_lock(dst, &ww); + if (err) + goto out_err; + + err = i915_gem_object_ww_copy_blt(src, dst, &ww, ce); +out_err: if (err == -EDEADLK) { err = i915_gem_ww_ctx_backoff(&ww); if (!err) goto retry; } i915_gem_ww_ctx_fini(&ww); - intel_engine_pm_put(ce->engine); + return err; }
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object_blt.h b/drivers/gpu/drm/i915/gem/i915_gem_object_blt.h index 2409fdcccf0e..da3d66abde64 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_object_blt.h +++ b/drivers/gpu/drm/i915/gem/i915_gem_object_blt.h @@ -36,4 +36,14 @@ int i915_gem_object_copy_blt(struct drm_i915_gem_object *src, struct drm_i915_gem_object *dst, struct intel_context *ce);
+int i915_gem_object_ww_fill_blt(struct drm_i915_gem_object *obj, + struct i915_gem_ww_ctx *ww, + struct intel_context *ce, + u32 value); + +int i915_gem_object_ww_copy_blt(struct drm_i915_gem_object *src, + struct drm_i915_gem_object *dst, + struct i915_gem_ww_ctx *ww, + struct intel_context *ce); + #endif