diff options
Diffstat (limited to 'drivers/gpu/drm/i915/i915_gpu_error.c')
-rw-r--r-- | drivers/gpu/drm/i915/i915_gpu_error.c | 203 |
1 files changed, 125 insertions, 78 deletions
diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c index 89725c9..9d73d22 100644 --- a/drivers/gpu/drm/i915/i915_gpu_error.c +++ b/drivers/gpu/drm/i915/i915_gpu_error.c @@ -332,7 +332,7 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m, const struct i915_error_state_file_priv *error_priv) { struct drm_device *dev = error_priv->dev; - struct drm_i915_private *dev_priv = dev->dev_private; + struct drm_i915_private *dev_priv = to_i915(dev); struct drm_i915_error_state *error = error_priv->error; struct drm_i915_error_object *obj; int i, j, offset, elt; @@ -411,7 +411,7 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m, err_printf(m, "DONE_REG: 0x%08x\n", error->done_reg); } - if (INTEL_INFO(dev)->gen == 7) + if (IS_GEN7(dev)) err_printf(m, "ERR_INT: 0x%08x\n", error->err_int); for (i = 0; i < ARRAY_SIZE(error->ring); i++) @@ -463,6 +463,18 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m, } } + if (error->ring[i].num_waiters) { + err_printf(m, "%s --- %d waiters\n", + dev_priv->engine[i].name, + error->ring[i].num_waiters); + for (j = 0; j < error->ring[i].num_waiters; j++) { + err_printf(m, " seqno 0x%08x for %s [%d]\n", + error->ring[i].waiters[j].seqno, + error->ring[i].waiters[j].comm, + error->ring[i].waiters[j].pid); + } + } + if ((obj = error->ring[i].ringbuffer)) { err_printf(m, "%s --- ringbuffer = 0x%08x\n", dev_priv->engine[i].name, @@ -488,7 +500,7 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m, hws_page[elt+1], hws_page[elt+2], hws_page[elt+3]); - offset += 16; + offset += 16; } } @@ -605,8 +617,9 @@ static void i915_error_state_free(struct kref *error_ref) i915_error_object_free(error->ring[i].ringbuffer); i915_error_object_free(error->ring[i].hws_page); i915_error_object_free(error->ring[i].ctx); - kfree(error->ring[i].requests); i915_error_object_free(error->ring[i].wa_ctx); + kfree(error->ring[i].requests); + kfree(error->ring[i].waiters); } i915_error_object_free(error->semaphore_obj); @@ -824,19 +837,18 @@ static uint32_t i915_error_generate_code(struct drm_i915_private *dev_priv, return error_code; } -static void i915_gem_record_fences(struct drm_device *dev, +static void i915_gem_record_fences(struct drm_i915_private *dev_priv, struct drm_i915_error_state *error) { - struct drm_i915_private *dev_priv = dev->dev_private; int i; - if (IS_GEN3(dev) || IS_GEN2(dev)) { + if (IS_GEN3(dev_priv) || IS_GEN2(dev_priv)) { for (i = 0; i < dev_priv->num_fence_regs; i++) error->fence[i] = I915_READ(FENCE_REG(i)); - } else if (IS_GEN5(dev) || IS_GEN4(dev)) { + } else if (IS_GEN5(dev_priv) || IS_GEN4(dev_priv)) { for (i = 0; i < dev_priv->num_fence_regs; i++) error->fence[i] = I915_READ64(FENCE_REG_965_LO(i)); - } else if (INTEL_INFO(dev)->gen >= 6) { + } else if (INTEL_GEN(dev_priv) >= 6) { for (i = 0; i < dev_priv->num_fence_regs; i++) error->fence[i] = I915_READ64(FENCE_REG_GEN6_LO(i)); } @@ -851,7 +863,7 @@ static void gen8_record_semaphore_state(struct drm_i915_private *dev_priv, struct intel_engine_cs *to; enum intel_engine_id id; - if (!i915_semaphore_is_enabled(dev_priv->dev)) + if (!i915_semaphore_is_enabled(dev_priv)) return; if (!error->semaphore_obj) @@ -893,31 +905,71 @@ static void gen6_record_semaphore_state(struct drm_i915_private *dev_priv, } } -static void i915_record_ring_state(struct drm_device *dev, +static void engine_record_waiters(struct intel_engine_cs *engine, + struct drm_i915_error_ring *ering) +{ + struct intel_breadcrumbs *b = &engine->breadcrumbs; + struct drm_i915_error_waiter *waiter; + struct rb_node *rb; + int count; + + ering->num_waiters = 0; + ering->waiters = NULL; + + spin_lock(&b->lock); + count = 0; + for (rb = rb_first(&b->waiters); rb != NULL; rb = rb_next(rb)) + count++; + spin_unlock(&b->lock); + + waiter = NULL; + if (count) + waiter = kmalloc_array(count, + sizeof(struct drm_i915_error_waiter), + GFP_ATOMIC); + if (!waiter) + return; + + ering->waiters = waiter; + + spin_lock(&b->lock); + for (rb = rb_first(&b->waiters); rb; rb = rb_next(rb)) { + struct intel_wait *w = container_of(rb, typeof(*w), node); + + strcpy(waiter->comm, w->tsk->comm); + waiter->pid = w->tsk->pid; + waiter->seqno = w->seqno; + waiter++; + + if (++ering->num_waiters == count) + break; + } + spin_unlock(&b->lock); +} + +static void i915_record_ring_state(struct drm_i915_private *dev_priv, struct drm_i915_error_state *error, struct intel_engine_cs *engine, struct drm_i915_error_ring *ering) { - struct drm_i915_private *dev_priv = dev->dev_private; - - if (INTEL_INFO(dev)->gen >= 6) { + if (INTEL_GEN(dev_priv) >= 6) { ering->rc_psmi = I915_READ(RING_PSMI_CTL(engine->mmio_base)); ering->fault_reg = I915_READ(RING_FAULT_REG(engine)); - if (INTEL_INFO(dev)->gen >= 8) + if (INTEL_GEN(dev_priv) >= 8) gen8_record_semaphore_state(dev_priv, error, engine, ering); else gen6_record_semaphore_state(dev_priv, engine, ering); } - if (INTEL_INFO(dev)->gen >= 4) { + if (INTEL_GEN(dev_priv) >= 4) { ering->faddr = I915_READ(RING_DMA_FADD(engine->mmio_base)); ering->ipeir = I915_READ(RING_IPEIR(engine->mmio_base)); ering->ipehr = I915_READ(RING_IPEHR(engine->mmio_base)); ering->instdone = I915_READ(RING_INSTDONE(engine->mmio_base)); ering->instps = I915_READ(RING_INSTPS(engine->mmio_base)); ering->bbaddr = I915_READ(RING_BBADDR(engine->mmio_base)); - if (INTEL_INFO(dev)->gen >= 8) { + if (INTEL_GEN(dev_priv) >= 8) { ering->faddr |= (u64) I915_READ(RING_DMA_FADD_UDW(engine->mmio_base)) << 32; ering->bbaddr |= (u64) I915_READ(RING_BBADDR_UDW(engine->mmio_base)) << 32; } @@ -929,20 +981,20 @@ static void i915_record_ring_state(struct drm_device *dev, ering->instdone = I915_READ(GEN2_INSTDONE); } - ering->waiting = waitqueue_active(&engine->irq_queue); + ering->waiting = intel_engine_has_waiter(engine); ering->instpm = I915_READ(RING_INSTPM(engine->mmio_base)); ering->acthd = intel_ring_get_active_head(engine); - ering->seqno = engine->get_seqno(engine); + ering->seqno = intel_engine_get_seqno(engine); ering->last_seqno = engine->last_submitted_seqno; ering->start = I915_READ_START(engine); ering->head = I915_READ_HEAD(engine); ering->tail = I915_READ_TAIL(engine); ering->ctl = I915_READ_CTL(engine); - if (I915_NEED_GFX_HWS(dev)) { + if (I915_NEED_GFX_HWS(dev_priv)) { i915_reg_t mmio; - if (IS_GEN7(dev)) { + if (IS_GEN7(dev_priv)) { switch (engine->id) { default: case RCS: @@ -958,7 +1010,7 @@ static void i915_record_ring_state(struct drm_device *dev, mmio = VEBOX_HWS_PGA_GEN7; break; } - } else if (IS_GEN6(engine->dev)) { + } else if (IS_GEN6(engine->i915)) { mmio = RING_HWS_PGA_GEN6(engine->mmio_base); } else { /* XXX: gen8 returns to sanity */ @@ -971,18 +1023,18 @@ static void i915_record_ring_state(struct drm_device *dev, ering->hangcheck_score = engine->hangcheck.score; ering->hangcheck_action = engine->hangcheck.action; - if (USES_PPGTT(dev)) { + if (USES_PPGTT(dev_priv)) { int i; ering->vm_info.gfx_mode = I915_READ(RING_MODE_GEN7(engine)); - if (IS_GEN6(dev)) + if (IS_GEN6(dev_priv)) ering->vm_info.pp_dir_base = I915_READ(RING_PP_DIR_BASE_READ(engine)); - else if (IS_GEN7(dev)) + else if (IS_GEN7(dev_priv)) ering->vm_info.pp_dir_base = I915_READ(RING_PP_DIR_BASE(engine)); - else if (INTEL_INFO(dev)->gen >= 8) + else if (INTEL_GEN(dev_priv) >= 8) for (i = 0; i < 4; i++) { ering->vm_info.pdp[i] = I915_READ(GEN8_RING_PDP_UDW(engine, i)); @@ -998,7 +1050,7 @@ static void i915_gem_record_active_context(struct intel_engine_cs *engine, struct drm_i915_error_state *error, struct drm_i915_error_ring *ering) { - struct drm_i915_private *dev_priv = engine->dev->dev_private; + struct drm_i915_private *dev_priv = engine->i915; struct drm_i915_gem_object *obj; /* Currently render ring is the only HW context user */ @@ -1016,34 +1068,33 @@ static void i915_gem_record_active_context(struct intel_engine_cs *engine, } } -static void i915_gem_record_rings(struct drm_device *dev, +static void i915_gem_record_rings(struct drm_i915_private *dev_priv, struct drm_i915_error_state *error) { - struct drm_i915_private *dev_priv = to_i915(dev); struct i915_ggtt *ggtt = &dev_priv->ggtt; struct drm_i915_gem_request *request; int i, count; for (i = 0; i < I915_NUM_ENGINES; i++) { struct intel_engine_cs *engine = &dev_priv->engine[i]; - struct intel_ringbuffer *rbuf; error->ring[i].pid = -1; - if (engine->dev == NULL) + if (!intel_engine_initialized(engine)) continue; error->ring[i].valid = true; - i915_record_ring_state(dev, error, engine, &error->ring[i]); + i915_record_ring_state(dev_priv, error, engine, &error->ring[i]); + engine_record_waiters(engine, &error->ring[i]); request = i915_gem_find_active_request(engine); if (request) { struct i915_address_space *vm; + struct intel_ringbuffer *rb; - vm = request->ctx && request->ctx->ppgtt ? - &request->ctx->ppgtt->base : - &ggtt->base; + vm = request->ctx->ppgtt ? + &request->ctx->ppgtt->base : &ggtt->base; /* We need to copy these to an anonymous buffer * as the simplest method to avoid being overwritten @@ -1070,26 +1121,17 @@ static void i915_gem_record_rings(struct drm_device *dev, } rcu_read_unlock(); } - } - if (i915.enable_execlists) { - /* TODO: This is only a small fix to keep basic error - * capture working, but we need to add more information - * for it to be useful (e.g. dump the context being - * executed). - */ - if (request) - rbuf = request->ctx->engine[engine->id].ringbuf; - else - rbuf = dev_priv->kernel_context->engine[engine->id].ringbuf; - } else - rbuf = engine->buffer; - - error->ring[i].cpu_ring_head = rbuf->head; - error->ring[i].cpu_ring_tail = rbuf->tail; + error->simulated |= + request->ctx->flags & CONTEXT_NO_ERROR_CAPTURE; - error->ring[i].ringbuffer = - i915_error_ggtt_object_create(dev_priv, rbuf->obj); + rb = request->ringbuf; + error->ring[i].cpu_ring_head = rb->head; + error->ring[i].cpu_ring_tail = rb->tail; + error->ring[i].ringbuffer = + i915_error_ggtt_object_create(dev_priv, + rb->obj); + } error->ring[i].hws_page = i915_error_ggtt_object_create(dev_priv, @@ -1234,7 +1276,7 @@ static void i915_gem_capture_buffers(struct drm_i915_private *dev_priv, static void i915_capture_reg_state(struct drm_i915_private *dev_priv, struct drm_i915_error_state *error) { - struct drm_device *dev = dev_priv->dev; + struct drm_device *dev = &dev_priv->drm; int i; /* General organization @@ -1301,15 +1343,14 @@ static void i915_capture_reg_state(struct drm_i915_private *dev_priv, error->eir = I915_READ(EIR); error->pgtbl_er = I915_READ(PGTBL_ER); - i915_get_extra_instdone(dev, error->extra_instdone); + i915_get_extra_instdone(dev_priv, error->extra_instdone); } -static void i915_error_capture_msg(struct drm_device *dev, +static void i915_error_capture_msg(struct drm_i915_private *dev_priv, struct drm_i915_error_state *error, u32 engine_mask, const char *error_msg) { - struct drm_i915_private *dev_priv = dev->dev_private; u32 ecode; int ring_id = -1, len; @@ -1317,7 +1358,7 @@ static void i915_error_capture_msg(struct drm_device *dev, len = scnprintf(error->error_msg, sizeof(error->error_msg), "GPU HANG: ecode %d:%d:0x%08x", - INTEL_INFO(dev)->gen, ring_id, ecode); + INTEL_GEN(dev_priv), ring_id, ecode); if (ring_id != -1 && error->ring[ring_id].pid != -1) len += scnprintf(error->error_msg + len, @@ -1352,14 +1393,17 @@ static void i915_capture_gen_state(struct drm_i915_private *dev_priv, * out a structure which becomes available in debugfs for user level tools * to pick up. */ -void i915_capture_error_state(struct drm_device *dev, u32 engine_mask, +void i915_capture_error_state(struct drm_i915_private *dev_priv, + u32 engine_mask, const char *error_msg) { static bool warned; - struct drm_i915_private *dev_priv = dev->dev_private; struct drm_i915_error_state *error; unsigned long flags; + if (READ_ONCE(dev_priv->gpu_error.first_error)) + return; + /* Account for pipe specific data like PIPE*STAT */ error = kzalloc(sizeof(*error), GFP_ATOMIC); if (!error) { @@ -1372,23 +1416,25 @@ void i915_capture_error_state(struct drm_device *dev, u32 engine_mask, i915_capture_gen_state(dev_priv, error); i915_capture_reg_state(dev_priv, error); i915_gem_capture_buffers(dev_priv, error); - i915_gem_record_fences(dev, error); - i915_gem_record_rings(dev, error); + i915_gem_record_fences(dev_priv, error); + i915_gem_record_rings(dev_priv, error); do_gettimeofday(&error->time); - error->overlay = intel_overlay_capture_error_state(dev); - error->display = intel_display_capture_error_state(dev); + error->overlay = intel_overlay_capture_error_state(dev_priv); + error->display = intel_display_capture_error_state(dev_priv); - i915_error_capture_msg(dev, error, engine_mask, error_msg); + i915_error_capture_msg(dev_priv, error, engine_mask, error_msg); DRM_INFO("%s\n", error->error_msg); - spin_lock_irqsave(&dev_priv->gpu_error.lock, flags); - if (dev_priv->gpu_error.first_error == NULL) { - dev_priv->gpu_error.first_error = error; - error = NULL; + if (!error->simulated) { + spin_lock_irqsave(&dev_priv->gpu_error.lock, flags); + if (!dev_priv->gpu_error.first_error) { + dev_priv->gpu_error.first_error = error; + error = NULL; + } + spin_unlock_irqrestore(&dev_priv->gpu_error.lock, flags); } - spin_unlock_irqrestore(&dev_priv->gpu_error.lock, flags); if (error) { i915_error_state_free(&error->ref); @@ -1400,7 +1446,8 @@ void i915_capture_error_state(struct drm_device *dev, u32 engine_mask, DRM_INFO("Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel\n"); DRM_INFO("drm/i915 developers can then reassign to the right component if it's not a kernel issue.\n"); DRM_INFO("The gpu crash dump is required to analyze gpu hangs, so please always attach it.\n"); - DRM_INFO("GPU crash dump saved to /sys/class/drm/card%d/error\n", dev->primary->index); + DRM_INFO("GPU crash dump saved to /sys/class/drm/card%d/error\n", + dev_priv->drm.primary->index); warned = true; } } @@ -1408,7 +1455,7 @@ void i915_capture_error_state(struct drm_device *dev, u32 engine_mask, void i915_error_state_get(struct drm_device *dev, struct i915_error_state_file_priv *error_priv) { - struct drm_i915_private *dev_priv = dev->dev_private; + struct drm_i915_private *dev_priv = to_i915(dev); spin_lock_irq(&dev_priv->gpu_error.lock); error_priv->error = dev_priv->gpu_error.first_error; @@ -1426,7 +1473,7 @@ void i915_error_state_put(struct i915_error_state_file_priv *error_priv) void i915_destroy_error_state(struct drm_device *dev) { - struct drm_i915_private *dev_priv = dev->dev_private; + struct drm_i915_private *dev_priv = to_i915(dev); struct drm_i915_error_state *error; spin_lock_irq(&dev_priv->gpu_error.lock); @@ -1450,17 +1497,17 @@ const char *i915_cache_level_str(struct drm_i915_private *i915, int type) } /* NB: please notice the memset */ -void i915_get_extra_instdone(struct drm_device *dev, uint32_t *instdone) +void i915_get_extra_instdone(struct drm_i915_private *dev_priv, + uint32_t *instdone) { - struct drm_i915_private *dev_priv = dev->dev_private; memset(instdone, 0, sizeof(*instdone) * I915_NUM_INSTDONE_REG); - if (IS_GEN2(dev) || IS_GEN3(dev)) + if (IS_GEN2(dev_priv) || IS_GEN3(dev_priv)) instdone[0] = I915_READ(GEN2_INSTDONE); - else if (IS_GEN4(dev) || IS_GEN5(dev) || IS_GEN6(dev)) { + else if (IS_GEN4(dev_priv) || IS_GEN5(dev_priv) || IS_GEN6(dev_priv)) { instdone[0] = I915_READ(RING_INSTDONE(RENDER_RING_BASE)); instdone[1] = I915_READ(GEN4_INSTDONE1); - } else if (INTEL_INFO(dev)->gen >= 7) { + } else if (INTEL_GEN(dev_priv) >= 7) { instdone[0] = I915_READ(RING_INSTDONE(RENDER_RING_BASE)); instdone[1] = I915_READ(GEN7_SC_INSTDONE); instdone[2] = I915_READ(GEN7_SAMPLER_INSTDONE); |