All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Christian König" <ckoenig.leichtzumerken@gmail.com>
To: intel-gfx@lists.freedesktop.org, linaro-mm-sig@lists.linaro.org,
	dri-devel@lists.freedesktop.org, chris@chris-wilson.co.uk
Subject: [PATCH 1/8] dma-buf: fix busy wait for new shared fences
Date: Tue,  6 Aug 2019 17:01:27 +0200	[thread overview]
Message-ID: <20190806150134.104222-1-christian.koenig@amd.com> (raw)

When reservation_object_add_shared_fence is replacing an old fence with a new
one we should not drop the old one before the new one is in place.

Otherwise other cores can busy wait for the new one to appear.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/dma-buf/reservation.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

diff --git a/drivers/dma-buf/reservation.c b/drivers/dma-buf/reservation.c
index c71b85c8c159..d59207ca72d2 100644
--- a/drivers/dma-buf/reservation.c
+++ b/drivers/dma-buf/reservation.c
@@ -196,6 +196,7 @@ void reservation_object_add_shared_fence(struct reservation_object *obj,
 					 struct dma_fence *fence)
 {
 	struct reservation_object_list *fobj;
+	struct dma_fence *old;
 	unsigned int i, count;
 
 	dma_fence_get(fence);
@@ -209,18 +210,16 @@ void reservation_object_add_shared_fence(struct reservation_object *obj,
 	write_seqcount_begin(&obj->seq);
 
 	for (i = 0; i < count; ++i) {
-		struct dma_fence *old_fence;
 
-		old_fence = rcu_dereference_protected(fobj->shared[i],
-						      reservation_object_held(obj));
-		if (old_fence->context == fence->context ||
-		    dma_fence_is_signaled(old_fence)) {
-			dma_fence_put(old_fence);
+		old = rcu_dereference_protected(fobj->shared[i],
+						reservation_object_held(obj));
+		if (old->context == fence->context ||
+		    dma_fence_is_signaled(old))
 			goto replace;
-		}
 	}
 
 	BUG_ON(fobj->shared_count >= fobj->shared_max);
+	old = NULL;
 	count++;
 
 replace:
@@ -230,6 +229,7 @@ void reservation_object_add_shared_fence(struct reservation_object *obj,
 
 	write_seqcount_end(&obj->seq);
 	preempt_enable();
+	dma_fence_put(old);
 }
 EXPORT_SYMBOL(reservation_object_add_shared_fence);
 
-- 
2.17.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

             reply	other threads:[~2019-08-06 15:01 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-08-06 15:01 Christian König [this message]
2019-08-06 15:01 ` [PATCH 2/8] dma-buf: fix shared fence list handling in reservation_object_copy_fences Christian König
2019-08-06 19:06   ` Chris Wilson
2019-08-07 10:43     ` Christian König
2019-08-06 15:01 ` [PATCH 3/8] drm/i915: stop using seqcount for fenc pruning Christian König
2019-08-06 19:07   ` Chris Wilson
2019-08-06 15:01 ` [PATCH 4/8] drm/i915: use new reservation_object_fences helper Christian König
2019-08-06 19:09   ` Chris Wilson
2019-08-06 15:01 ` [PATCH 5/8] dma-buf: further relax reservation_object_add_shared_fence Christian König
2019-08-06 19:25   ` Chris Wilson
2019-08-06 15:01 ` [PATCH 6/8] dma-buf: simplify reservation_object_get_fences_rcu a bit Christian König
2019-08-06 19:11   ` Chris Wilson
2019-08-06 15:01 ` [PATCH 7/8] dma-buf: add reservation_object_fences helper Christian König
2019-08-06 19:24   ` Chris Wilson
2019-08-07  9:06     ` Christian König
2019-08-06 15:01 ` [PATCH 8/8] dma-buf: nuke reservation_object seq number Christian König
2019-08-06 19:57   ` Chris Wilson
2019-08-07 12:08     ` Christian König
2019-08-07 12:19       ` Chris Wilson
2019-08-07 13:05         ` Koenig, Christian
2019-08-06 18:57 ` [PATCH 1/8] dma-buf: fix busy wait for new shared fences Chris Wilson
2019-08-09 13:27 ` ✗ Fi.CI.BAT: failure for series starting with [1/8] " Patchwork

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20190806150134.104222-1-christian.koenig@amd.com \
    --to=ckoenig.leichtzumerken@gmail.com \
    --cc=chris@chris-wilson.co.uk \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=linaro-mm-sig@lists.linaro.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.