drm/i915/guc: don't spinwait if the GuC's workqueue is full

Rather than wait to see whether more space becomes available in the GuC
submission workqueue, we can just return -EAGAIN and let the caller try
again in a little while. This gets rid of an uninterruptable sleep in
the polling code :)

We'll also add a counter to the GuC client statistics, to see how often
we find the WQ full.

v2:
    Flag the likely() code path (Tvtrko Ursulin).

v4:
    Add/update comments about failure counters (Tvtrko Ursulin).

Signed-off-by: Dave Gordon <david.s.gordon@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 4c6b48dbd6e2f62c6db0cc80c435e031b5986d42..e606c6acef0fa262b688b6eeafca4ad60a987936 100644 (file)
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -2515,6 +2515,7 @@ static void i915_guc_client_info(struct seq_file *m,
        seq_printf(m, "\tWQ size %d, offset: 0x%x, tail %d\n",
                client->wq_size, client->wq_offset, client->wq_tail);
 
+       seq_printf(m, "\tWork queue full: %u\n", client->no_wq_space);
        seq_printf(m, "\tFailed to queue: %u\n", client->q_fail);
        seq_printf(m, "\tFailed doorbell: %u\n", client->b_fail);
        seq_printf(m, "\tLast submission result: %d\n", client->retcode);

diff --git a/drivers/gpu/drm/i915/i915_guc_submission.c b/drivers/gpu/drm/i915/i915_guc_submission.c
index 87cb7396ff93b9117fea521da003ef35c8f569f1..85b2b89eeb3a67ca9aae518145faae1676ac4561 100644 (file)
--- a/drivers/gpu/drm/i915/i915_guc_submission.c
+++ b/drivers/gpu/drm/i915/i915_guc_submission.c
@@ -468,26 +468,22 @@ static void guc_fini_ctx_desc(struct intel_guc *guc,
  */
 int i915_guc_wq_check_space(struct drm_i915_gem_request *request)
 {
-       const size_t size = sizeof(struct guc_wq_item);
+       const size_t wqi_size = sizeof(struct guc_wq_item);
        struct i915_guc_client *gc = request->i915->guc.execbuf_client;
        struct guc_process_desc *desc;
-       int ret = -ETIMEDOUT, timeout_counter = 200;
+       u32 freespace;
 
        GEM_BUG_ON(gc == NULL);
 
        desc = gc->client_base + gc->proc_desc_offset;
 
-       while (timeout_counter-- > 0) {
-               if (CIRC_SPACE(gc->wq_tail, desc->head, gc->wq_size) >= size) {
-                       ret = 0;
-                       break;
-               }
+       freespace = CIRC_SPACE(gc->wq_tail, desc->head, gc->wq_size);
+       if (likely(freespace >= wqi_size))
+               return 0;
 
-               if (timeout_counter)
-                       usleep_range(1000, 2000);
-       };
+       gc->no_wq_space += 1;
 
-       return ret;
+       return -EAGAIN;
 }
 
 static int guc_add_workqueue_item(struct i915_guc_client *gc,

diff --git a/drivers/gpu/drm/i915/intel_guc.h b/drivers/gpu/drm/i915/intel_guc.h
index 91f315cf073b1566eb128b596785bb1233507e1b..380a743f830f459e44ca454a4f0c7dd08121c46e 100644 (file)
--- a/drivers/gpu/drm/i915/intel_guc.h
+++ b/drivers/gpu/drm/i915/intel_guc.h
@@ -48,9 +48,18 @@ struct drm_i915_gem_request;
  * queue (a circular array of work items), again described in the process
  * descriptor. Work queue pages are mapped momentarily as required.
  *
- * Finally, we also keep a few statistics here, including the number of
- * submissions to each engine, and a record of the last submission failure
- * (if any).
+ * We also keep a few statistics on failures. Ideally, these should all
+ * be zero!
+ *   no_wq_space: times that the submission pre-check found no space was
+ *                available in the work queue (note, the queue is shared,
+ *                not per-engine). It is OK for this to be nonzero, but
+ *                it should not be huge!
+ *   q_fail: failed to enqueue a work item. This should never happen,
+ *           because we check for space beforehand.
+ *   b_fail: failed to ring the doorbell. This should never happen, unless
+ *           somehow the hardware misbehaves, or maybe if the GuC firmware
+ *           crashes? We probably need to reset the GPU to recover.
+ *   retcode: errno from last guc_submit()
  */
 struct i915_guc_client {
        struct drm_i915_gem_object *client_obj;
@@ -71,12 +80,13 @@ struct i915_guc_client {
        uint32_t wq_tail;
        uint32_t unused;                /* Was 'wq_head'                */
 
-       /* GuC submission statistics & status */
-       uint64_t submissions[GUC_MAX_ENGINES_NUM];
+       uint32_t no_wq_space;
        uint32_t q_fail;
        uint32_t b_fail;
        int retcode;
-       int spare;                      /* pad to 32 DWords             */
+
+       /* Per-engine counts of GuC submissions */
+       uint64_t submissions[GUC_MAX_ENGINES_NUM];
 };
 
 enum intel_guc_fw_status {