| Message ID | 1424103173-17387-1-git-send-email-thomas.daniel@intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Tested-By: PRC QA PRTS (Patch Regression Test System Contact: shuang.he@intel.com)
Task id: 5785
-------------------------------------Summary-------------------------------------
Platform Delta drm-intel-nightly Series Applied
PNV -5 277/277 272/277
ILK 313/313 313/313
SNB -1 309/309 308/309
IVB -1 382/382 381/382
BYT 296/296 296/296
HSW 425/425 425/425
BDW -1 318/318 317/318
-------------------------------------Detailed-------------------------------------
Platform Test drm-intel-nightly Series Applied
*PNV igt_gem_fence_thrash_bo-write-verify-none NRUN(1)PASS(2) FAIL(1)PASS(1)
*PNV igt_gem_fence_thrash_bo-write-verify-x PASS(3) FAIL(1)PASS(1)
*PNV igt_gem_fence_thrash_bo-write-verify-y PASS(3) FAIL(1)PASS(1)
PNV igt_gem_userptr_blits_coherency-sync CRASH(1)PASS(3) CRASH(1)PASS(1)
PNV igt_gem_userptr_blits_coherency-unsync CRASH(1)PASS(2) CRASH(1)PASS(1)
*SNB igt_kms_pipe_crc_basic_read-crc-pipe-A PASS(2) DMESG_WARN(1)PASS(1)
*IVB igt_gem_storedw_batches_loop_normal PASS(3) DMESG_WARN(1)PASS(1)
*BDW igt_gem_gtt_hog PASS(5) DMESG_WARN(1)PASS(1)
Note: You need to pay more attention to line start with '*'
On 16/02/15 16:12, Thomas Daniel wrote: > Work was getting left behind in LRC contexts during reset. This causes a hang > if the GPU is reset when HEAD==TAIL because the context's ringbuffer head and > tail don't get reset and retiring a request doesn't alter them, so the ring > still appears full. > > Added a function intel_lr_context_reset() to reset head and tail on a LRC and > its ringbuffer. > > Call intel_lr_context_reset() for each context in i915_gem_context_reset() when > in execlists mode. > > Testcase: igt/pm_rps --run-subtest reset #bdw > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=88096 > Signed-off-by: Thomas Daniel <thomas.daniel@intel.com> > --- > drivers/gpu/drm/i915/i915_gem_context.c | 12 +++++++---- > drivers/gpu/drm/i915/intel_lrc.c | 34 +++++++++++++++++++++++++++++++ > drivers/gpu/drm/i915/intel_lrc.h | 2 ++ > 3 files changed, 44 insertions(+), 4 deletions(-) A couple of minor points below, but not ones that require changes, so: Reviewed-by: Dave Gordon <david.s.gordon@intel.com> > diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c > index 8603bf4..70346b0 100644 > --- a/drivers/gpu/drm/i915/i915_gem_context.c > +++ b/drivers/gpu/drm/i915/i915_gem_context.c > @@ -296,11 +296,15 @@ void i915_gem_context_reset(struct drm_device *dev) > struct drm_i915_private *dev_priv = dev->dev_private; > int i; > > - /* In execlists mode we will unreference the context when the execlist > - * queue is cleared and the requests destroyed. > - */ > - if (i915.enable_execlists) > + if (i915.enable_execlists) { > + struct intel_context *ctx; > + > + list_for_each_entry(ctx, &dev_priv->context_list, link) { > + intel_lr_context_reset(dev, ctx); > + } > + > return; > + } > > for (i = 0; i < I915_NUM_RINGS; i++) { > struct intel_engine_cs *ring = &dev_priv->ring[i]; > diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c > index aafcef3..1946bb9 100644 > --- a/drivers/gpu/drm/i915/intel_lrc.c > +++ b/drivers/gpu/drm/i915/intel_lrc.c > @@ -1950,3 +1950,37 @@ error_unpin_ctx: > drm_gem_object_unreference(&ctx_obj->base); > return ret; > } > + > +void intel_lr_context_reset(struct drm_device *dev, > + struct intel_context *ctx) > +{ > + struct drm_i915_private *dev_priv = dev->dev_private; > + struct intel_engine_cs *ring; > + int i; > + > + for_each_ring(ring, dev_priv, i) { > + struct drm_i915_gem_object *ctx_obj = > + ctx->engine[ring->id].state; > + if (ctx_obj) { > + struct intel_ringbuffer *ringbuf = > + ctx->engine[ring->id].ringbuf; > + uint32_t *reg_state; > + struct page *page; > + > + if (i915_gem_object_get_pages(ctx_obj)) { > + WARN(1, "Failed get_pages for context obj\n"); > + continue; > + } This could be folded into a single "if (WARN_ON(...)) continue;" > + page = i915_gem_object_get_page(ctx_obj, 1); Isn't it a pity that we have i915_gem_object_get_page() and i915_gem_object_get_pages() which look so similar but do completely different things :( > + reg_state = kmap_atomic(page); > + > + reg_state[CTX_RING_HEAD+1] = 0; > + reg_state[CTX_RING_TAIL+1] = 0; > + > + kunmap_atomic(reg_state); > + > + ringbuf->head = 0; > + ringbuf->tail = 0; > + } > + } > +} > diff --git a/drivers/gpu/drm/i915/intel_lrc.h b/drivers/gpu/drm/i915/intel_lrc.h > index f635735..5dd0eca 100644 > --- a/drivers/gpu/drm/i915/intel_lrc.h > +++ b/drivers/gpu/drm/i915/intel_lrc.h > @@ -73,6 +73,8 @@ int intel_lr_context_deferred_create(struct intel_context *ctx, > struct intel_engine_cs *ring); > void intel_lr_context_unpin(struct intel_engine_cs *ring, > struct intel_context *ctx); > +void intel_lr_context_reset(struct drm_device *dev, > + struct intel_context *ctx); > > /* Execlists */ > int intel_sanitize_enable_execlists(struct drm_device *dev, int enable_execlists); >
On Tue, Feb 17, 2015 at 11:34:31AM +0000, Dave Gordon wrote: > On 16/02/15 16:12, Thomas Daniel wrote: > > + if (i915_gem_object_get_pages(ctx_obj)) { > > + WARN(1, "Failed get_pages for context obj\n"); > > + continue; > > + } > > This could be folded into a single "if (WARN_ON(...)) continue;" It's also a bit bad style to not handle this case at all. But meh ;-) > > + page = i915_gem_object_get_page(ctx_obj, 1); > > Isn't it a pity that we have i915_gem_object_get_page() and > i915_gem_object_get_pages() which look so similar but do completely > different things :( lookup_page might be a more suitable name for this one indeed ... patch highly welcome. -Daniel
On Mon, Feb 16, 2015 at 04:12:53PM +0000, Thomas Daniel wrote: > Work was getting left behind in LRC contexts during reset. This causes a hang > if the GPU is reset when HEAD==TAIL because the context's ringbuffer head and > tail don't get reset and retiring a request doesn't alter them, so the ring > still appears full. > > Added a function intel_lr_context_reset() to reset head and tail on a LRC and > its ringbuffer. > > Call intel_lr_context_reset() for each context in i915_gem_context_reset() when > in execlists mode. > > Testcase: igt/pm_rps --run-subtest reset #bdw > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=88096 > Signed-off-by: Thomas Daniel <thomas.daniel@intel.com> > --- > drivers/gpu/drm/i915/i915_gem_context.c | 12 +++++++---- > drivers/gpu/drm/i915/intel_lrc.c | 34 +++++++++++++++++++++++++++++++ > drivers/gpu/drm/i915/intel_lrc.h | 2 ++ > 3 files changed, 44 insertions(+), 4 deletions(-) > > diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c > index 8603bf4..70346b0 100644 > --- a/drivers/gpu/drm/i915/i915_gem_context.c > +++ b/drivers/gpu/drm/i915/i915_gem_context.c > @@ -296,11 +296,15 @@ void i915_gem_context_reset(struct drm_device *dev) > struct drm_i915_private *dev_priv = dev->dev_private; > int i; > > - /* In execlists mode we will unreference the context when the execlist > - * queue is cleared and the requests destroyed. > - */ > - if (i915.enable_execlists) > + if (i915.enable_execlists) { > + struct intel_context *ctx; > + > + list_for_each_entry(ctx, &dev_priv->context_list, link) { > + intel_lr_context_reset(dev, ctx); > + } > + > return; > + } I'm somewhat voting for a vfunc for engine state reset since hiding the lrc callback by going through legacy code doesn't look too pretty. Follow-up patch if you're bored would be neat ;-) > > for (i = 0; i < I915_NUM_RINGS; i++) { > struct intel_engine_cs *ring = &dev_priv->ring[i]; > diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c > index aafcef3..1946bb9 100644 > --- a/drivers/gpu/drm/i915/intel_lrc.c > +++ b/drivers/gpu/drm/i915/intel_lrc.c > @@ -1950,3 +1950,37 @@ error_unpin_ctx: > drm_gem_object_unreference(&ctx_obj->base); > return ret; > } > + > +void intel_lr_context_reset(struct drm_device *dev, > + struct intel_context *ctx) > +{ > + struct drm_i915_private *dev_priv = dev->dev_private; > + struct intel_engine_cs *ring; > + int i; > + > + for_each_ring(ring, dev_priv, i) { > + struct drm_i915_gem_object *ctx_obj = > + ctx->engine[ring->id].state; > + if (ctx_obj) { The common patter is if (!ctx_obj) continue; which allows you to fold out one indent level. I couldn't resist and done this while applying ... Queued for -next, thanks for the patch. -Daniel
diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c index 8603bf4..70346b0 100644 --- a/drivers/gpu/drm/i915/i915_gem_context.c +++ b/drivers/gpu/drm/i915/i915_gem_context.c @@ -296,11 +296,15 @@ void i915_gem_context_reset(struct drm_device *dev) struct drm_i915_private *dev_priv = dev->dev_private; int i; - /* In execlists mode we will unreference the context when the execlist - * queue is cleared and the requests destroyed. - */ - if (i915.enable_execlists) + if (i915.enable_execlists) { + struct intel_context *ctx; + + list_for_each_entry(ctx, &dev_priv->context_list, link) { + intel_lr_context_reset(dev, ctx); + } + return; + } for (i = 0; i < I915_NUM_RINGS; i++) { struct intel_engine_cs *ring = &dev_priv->ring[i]; diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index aafcef3..1946bb9 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -1950,3 +1950,37 @@ error_unpin_ctx: drm_gem_object_unreference(&ctx_obj->base); return ret; } + +void intel_lr_context_reset(struct drm_device *dev, + struct intel_context *ctx) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_engine_cs *ring; + int i; + + for_each_ring(ring, dev_priv, i) { + struct drm_i915_gem_object *ctx_obj = + ctx->engine[ring->id].state; + if (ctx_obj) { + struct intel_ringbuffer *ringbuf = + ctx->engine[ring->id].ringbuf; + uint32_t *reg_state; + struct page *page; + + if (i915_gem_object_get_pages(ctx_obj)) { + WARN(1, "Failed get_pages for context obj\n"); + continue; + } + page = i915_gem_object_get_page(ctx_obj, 1); + reg_state = kmap_atomic(page); + + reg_state[CTX_RING_HEAD+1] = 0; + reg_state[CTX_RING_TAIL+1] = 0; + + kunmap_atomic(reg_state); + + ringbuf->head = 0; + ringbuf->tail = 0; + } + } +} diff --git a/drivers/gpu/drm/i915/intel_lrc.h b/drivers/gpu/drm/i915/intel_lrc.h index f635735..5dd0eca 100644 --- a/drivers/gpu/drm/i915/intel_lrc.h +++ b/drivers/gpu/drm/i915/intel_lrc.h @@ -73,6 +73,8 @@ int intel_lr_context_deferred_create(struct intel_context *ctx, struct intel_engine_cs *ring); void intel_lr_context_unpin(struct intel_engine_cs *ring, struct intel_context *ctx); +void intel_lr_context_reset(struct drm_device *dev, + struct intel_context *ctx); /* Execlists */ int intel_sanitize_enable_execlists(struct drm_device *dev, int enable_execlists);
Work was getting left behind in LRC contexts during reset. This causes a hang if the GPU is reset when HEAD==TAIL because the context's ringbuffer head and tail don't get reset and retiring a request doesn't alter them, so the ring still appears full. Added a function intel_lr_context_reset() to reset head and tail on a LRC and its ringbuffer. Call intel_lr_context_reset() for each context in i915_gem_context_reset() when in execlists mode. Testcase: igt/pm_rps --run-subtest reset #bdw Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=88096 Signed-off-by: Thomas Daniel <thomas.daniel@intel.com> --- drivers/gpu/drm/i915/i915_gem_context.c | 12 +++++++---- drivers/gpu/drm/i915/intel_lrc.c | 34 +++++++++++++++++++++++++++++++ drivers/gpu/drm/i915/intel_lrc.h | 2 ++ 3 files changed, 44 insertions(+), 4 deletions(-)