[Intel-gfx] [PATCH v5 08/10] drm/i915/guc: Plumb GuC-capture into gpu_coredump
Umesh Nerlige Ramappa
umesh.nerlige.ramappa at intel.com
Fri Feb 11 02:11:36 UTC 2022
On Wed, Jan 26, 2022 at 02:48:20AM -0800, Alan Previn wrote:
>Add a flags parameter through all of the coredump creation
>functions. Add a bitmask flag to indicate if the top
>level gpu_coredump event is triggered in response to
>a GuC context reset notification.
>
>Using that flag, ensure all coredump functions that
>read or print mmio-register values related to work submission
>or command-streamer engines are skipped and replaced with
>a calls guc-capture module equivalent functions to retrieve
>or print the register dump.
>
>While here, split out display related register reading
>and printing into its own function that is called agnostic
>to whether GuC had triggered the reset.
>
>For now, introduce an empty printing function that can
>filled in on a subsequent patch just to handle formatting.
>
>Signed-off-by: Alan Previn <alan.previn.teres.alexis at intel.com>
>---
> .../drm/i915/gt/intel_execlists_submission.c | 4 +-
> drivers/gpu/drm/i915/gt/intel_reset.c | 2 +-
> .../gpu/drm/i915/gt/uc/intel_guc_capture.c | 78 ++++++
> .../gpu/drm/i915/gt/uc/intel_guc_capture.h | 11 +-
> .../gpu/drm/i915/gt/uc/intel_guc_submission.c | 2 +-
> drivers/gpu/drm/i915/i915_debugfs.c | 2 +-
> drivers/gpu/drm/i915/i915_gpu_error.c | 263 ++++++++++++------
> drivers/gpu/drm/i915/i915_gpu_error.h | 31 ++-
> 8 files changed, 295 insertions(+), 98 deletions(-)
>
>diff --git a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
>index 960a9aaf4f3a..c8bb43863461 100644
>--- a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
>+++ b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
>@@ -2230,11 +2230,11 @@ static struct execlists_capture *capture_regs(struct intel_engine_cs *engine)
> if (!cap->error)
> goto err_cap;
>
>- cap->error->gt = intel_gt_coredump_alloc(engine->gt, gfp);
>+ cap->error->gt = intel_gt_coredump_alloc(engine->gt, gfp, CORE_DUMP_FLAG_NONE);
> if (!cap->error->gt)
> goto err_gpu;
>
>- cap->error->gt->engine = intel_engine_coredump_alloc(engine, gfp);
>+ cap->error->gt->engine = intel_engine_coredump_alloc(engine, gfp, CORE_DUMP_FLAG_NONE);
> if (!cap->error->gt->engine)
> goto err_gt;
>
>diff --git a/drivers/gpu/drm/i915/gt/intel_reset.c b/drivers/gpu/drm/i915/gt/intel_reset.c
>index 6f2821cca409..c1dc3f8b1108 100644
>--- a/drivers/gpu/drm/i915/gt/intel_reset.c
>+++ b/drivers/gpu/drm/i915/gt/intel_reset.c
>@@ -1305,7 +1305,7 @@ void intel_gt_handle_error(struct intel_gt *gt,
> engine_mask &= gt->info.engine_mask;
>
> if (flags & I915_ERROR_CAPTURE) {
>- i915_capture_error_state(gt, engine_mask);
>+ i915_capture_error_state(gt, engine_mask, CORE_DUMP_FLAG_NONE);
> intel_gt_clear_error_registers(gt, engine_mask);
> }
>
>diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_capture.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_capture.c
>index 0b6d743712a6..2f5dc413dddc 100644
>--- a/drivers/gpu/drm/i915/gt/uc/intel_guc_capture.c
>+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_capture.c
>@@ -728,6 +728,17 @@ int intel_guc_capture_output_min_size_est(struct intel_guc *guc)
> * instance). This node id added to a linked list stored in
> * guc->capture->priv for matchup and printout when triggered by
> * i915_gpu_coredump and err_print_gt (via error capture sysfs) later.
>+ *
>+ * GUC --> notify context reset:
>+ * -----------------------------
>+ * --> G2H CONTEXT RESET
>+ * L--> guc_handle_context_reset --> i915_capture_error_state
>+ * L--> i915_gpu_coredump(..IS_GUC_CAPTURE) --> gt_record_engines
>+ * --> capture_engine(..IS_GUC_CAPTURE)
>+ * L--> detach C from internal linked list and add into
>+ * intel_engine_coredump struct (if the context and
>+ * engine of the event notification matches a node
>+ * in the link list)
> */
>
> static int guc_capture_buf_cnt(struct __guc_capture_bufstate *buf)
>@@ -1203,6 +1214,73 @@ static void __guc_capture_store_snapshot_work(struct intel_guc *guc)
> mutex_unlock(&guc->log_state[GUC_CAPTURE_LOG_BUFFER].lock);
> }
>
>+#if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)
>+
>+int intel_guc_capture_print_engine_node(struct drm_i915_error_state_buf *ebuf,
>+ const struct intel_engine_coredump *ee)
>+{
>+ return 0;
What goes here? Also return value is not used.
>+}
>+
>+#endif //CONFIG_DRM_I915_DEBUG_GUC
>+
>+void intel_guc_capture_free_node(struct intel_engine_coredump *ee)
>+{
>+ int i;
>+
>+ if (!ee)
>+ return;
>+ if (!ee->guc_capture_node)
>+ return;
if (!ee || !ee->guc_capture_node) ^
>+ for (i = GUC_CAPTURE_LIST_TYPE_GLOBAL; i < GUC_CAPTURE_LIST_TYPE_MAX; ++i) {
>+ if (ee->guc_capture_node->reginfo[i].regs)
>+ kfree(ee->guc_capture_node->reginfo[i].regs);
>+ }
>+ kfree(ee->guc_capture_node);
>+ ee->guc_capture_node = NULL;
>+}
>+
>+void intel_guc_capture_get_matching_node(struct intel_gt *gt,
>+ struct intel_engine_coredump *ee,
>+ struct intel_context *ce)
>+{
>+ struct intel_guc *guc;
>+ struct drm_i915_private *i915;
>+
>+ if (!gt || !ee || !ce)
>+ return;
>+
>+ i915 = gt->i915;
>+ guc = >->uc.guc;
>+ if (!guc->capture.priv)
>+ return;
>+
>+ GEM_BUG_ON(ee->guc_capture_node);
>+ /*
>+ * Look for a matching GuC reported error capture node from
>+ * the internal output link-list based on lrca, guc-id and engine
>+ * identification.
>+ */
>+ if (!list_empty(&guc->capture.priv->outlist)) {
You would just not execute anything in the loop below if empty, so no
need to check for empty.
>+ struct __guc_capture_parsed_output *n, *ntmp;
>+
>+ list_for_each_entry_safe(n, ntmp, &guc->capture.priv->outlist, link) {
>+ if (n->eng_inst == GUC_ID_TO_ENGINE_INSTANCE(ee->engine->guc_id) &&
>+ n->eng_class == GUC_ID_TO_ENGINE_CLASS(ee->engine->guc_id) &&
>+ n->guc_id == ce->guc_id.id &&
>+ (n->lrca & CTX_GTT_ADDRESS_MASK) ==
>+ (ce->lrc.lrca & CTX_GTT_ADDRESS_MASK)) {
>+ list_del(&n->link);
>+ --guc->capture.priv->listcount;
>+ ee->guc_capture_node = n;
>+ ee->capture = &guc->capture;
>+ return;
>+ }
>+ }
>+ }
>+ drm_warn(&i915->drm, "GuC capture can't match ee to node\n");
>+}
lgtm. In general the implementation differences between submission
backends (guc vs execlists) need to be vfuncs so we can avoid having to
use the flags and conditions. Regardless, with the above comments
addressed, this is:
Reviewed-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa at intel.com>
Umesh
More information about the Intel-gfx
mailing list