| Message ID | 1405674316-3832-1-git-send-email-rodrigo.vivi@intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Fri, Jul 18, 2014 at 02:05:16AM -0700, Rodrigo Vivi wrote: > semaphore _sync_seqno, _seqno and _mbox are smaller than number of rings. > This optimization is to remove the ring itself from the list and the logic to do that > is at intel_ring_sync_index as below: > > /* > * rcs -> 0 = vcs, 1 = bcs, 2 = vecs, 3 = vcs2; > * vcs -> 0 = bcs, 1 = vecs, 2 = vcs2, 3 = rcs; > * bcs -> 0 = vecs, 1 = vcs2. 2 = rcs, 3 = vcs; > * vecs -> 0 = vcs2, 1 = rcs, 2 = vcs, 3 = bcs; > * vcs2 -> 0 = rcs, 1 = vcs, 2 = bcs, 3 = vecs; > */ > > v2: Skip when from == to (Damien). > v3: avoid computing idx when from == to (Damien). > use ring == to instead of ring->id == to->id (Damien). > use continue instead of return (Rodrigo). > > Cc: Damien Lespiau <damien.lespiau@intel.com> > Cc: Ben Widawsky <benjamin.widawsky@intel.com> > Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> I guess there are still some details that look weird: - idx's scope could be reduced to the loop - there's still some code that is executed on the skipped iteration that doesn't need to be. But I believe we shouldn't overflow now at least, so, whether you can be bothered with those last 2 comments or not: Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c index 9faebbc..1efcf1f 100644 --- a/drivers/gpu/drm/i915/i915_gpu_error.c +++ b/drivers/gpu/drm/i915/i915_gpu_error.c @@ -764,8 +764,8 @@ static void gen8_record_semaphore_state(struct drm_i915_private *dev_priv, struct intel_engine_cs *ring, struct drm_i915_error_ring *ering) { - struct intel_engine_cs *useless; - int i; + struct intel_engine_cs *to; + int i, idx; if (!i915_semaphore_is_enabled(dev_priv->dev)) return; @@ -776,13 +776,18 @@ static void gen8_record_semaphore_state(struct drm_i915_private *dev_priv, dev_priv->semaphore_obj, &dev_priv->gtt.base); - for_each_ring(useless, dev_priv, i) { + for_each_ring(to, dev_priv, i) { u16 signal_offset = (GEN8_SIGNAL_OFFSET(ring, i) & PAGE_MASK) / 4; u32 *tmp = error->semaphore_obj->pages[0]; - ering->semaphore_mboxes[i] = tmp[signal_offset]; - ering->semaphore_seqno[i] = ring->semaphore.sync_seqno[i]; + if (ring == to) + continue; + + idx = intel_ring_sync_index(ring, to); + + ering->semaphore_mboxes[idx] = tmp[signal_offset]; + ering->semaphore_seqno[idx] = ring->semaphore.sync_seqno[idx]; } }
semaphore _sync_seqno, _seqno and _mbox are smaller than number of rings. This optimization is to remove the ring itself from the list and the logic to do that is at intel_ring_sync_index as below: /* * rcs -> 0 = vcs, 1 = bcs, 2 = vecs, 3 = vcs2; * vcs -> 0 = bcs, 1 = vecs, 2 = vcs2, 3 = rcs; * bcs -> 0 = vecs, 1 = vcs2. 2 = rcs, 3 = vcs; * vecs -> 0 = vcs2, 1 = rcs, 2 = vcs, 3 = bcs; * vcs2 -> 0 = rcs, 1 = vcs, 2 = bcs, 3 = vecs; */ v2: Skip when from == to (Damien). v3: avoid computing idx when from == to (Damien). use ring == to instead of ring->id == to->id (Damien). use continue instead of return (Rodrigo). Cc: Damien Lespiau <damien.lespiau@intel.com> Cc: Ben Widawsky <benjamin.widawsky@intel.com> Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> --- drivers/gpu/drm/i915/i915_gpu_error.c | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-)