From: Daniel Vetter <daniel@ffwll.ch>
To: Chris Wilson <chris@chris-wilson.co.uk>
Cc: linaro-mm-sig@lists.linaro.org, intel-gfx@lists.freedesktop.org,
dri-devel@lists.freedesktop.org, linux-media@vger.kernel.org
Subject: Re: [Intel-gfx] [PATCH 11/11] dma-buf: Do a fast lockless check for poll with timeout=0
Date: Fri, 23 Sep 2016 15:50:44 +0200 [thread overview]
Message-ID: <20160923135044.GM3988@dvetter-linux.ger.corp.intel.com> (raw)
In-Reply-To: <20160829070834.22296-11-chris@chris-wilson.co.uk>
On Mon, Aug 29, 2016 at 08:08:34AM +0100, Chris Wilson wrote:
> Currently we install a callback for performing poll on a dma-buf,
> irrespective of the timeout. This involves taking a spinlock, as well as
> unnecessary work, and greatly reduces scaling of poll(.timeout=0) across
> multiple threads.
>
> We can query whether the poll will block prior to installing the
> callback to make the busy-query fast.
>
> Single thread: 60% faster
> 8 threads on 4 (+4 HT) cores: 600% faster
>
> Still not quite the perfect scaling we get with a native busy ioctl, but
> poll(dmabuf) is faster due to the quicker lookup of the object and
> avoiding drm_ioctl().
>
> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> Cc: Sumit Semwal <sumit.semwal@linaro.org>
> Cc: linux-media@vger.kernel.org
> Cc: dri-devel@lists.freedesktop.org
> Cc: linaro-mm-sig@lists.linaro.org
> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Need to strike the r-b here, since Christian König pointed out that
objects won't magically switch signalling on.
-Daniel
> ---
> drivers/dma-buf/dma-buf.c | 12 ++++++++++++
> 1 file changed, 12 insertions(+)
>
> diff --git a/drivers/dma-buf/dma-buf.c b/drivers/dma-buf/dma-buf.c
> index cf04d249a6a4..c7a7bc579941 100644
> --- a/drivers/dma-buf/dma-buf.c
> +++ b/drivers/dma-buf/dma-buf.c
> @@ -156,6 +156,18 @@ static unsigned int dma_buf_poll(struct file *file, poll_table *poll)
> if (!events)
> return 0;
>
> + if (poll_does_not_wait(poll)) {
> + if (events & POLLOUT &&
> + !reservation_object_test_signaled_rcu(resv, true))
> + events &= ~(POLLOUT | POLLIN);
> +
> + if (events & POLLIN &&
> + !reservation_object_test_signaled_rcu(resv, false))
> + events &= ~POLLIN;
> +
> + return events;
> + }
> +
> retry:
> seq = read_seqcount_begin(&resv->seq);
> rcu_read_lock();
> --
> 2.9.3
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
next prev parent reply other threads:[~2016-09-23 13:50 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-08-29 7:08 [PATCH 01/11] drm/amdgpu: Remove call to reservation_object_test_signaled_rcu before wait Chris Wilson
2016-08-29 7:08 ` [PATCH 02/11] drm/etnaviv: Remove manual " Chris Wilson
2016-09-23 12:55 ` Daniel Vetter
2016-10-05 16:15 ` Sumit Semwal
2016-10-10 13:17 ` Lucas Stach
2016-08-29 7:08 ` [PATCH 03/11] drm/msm: Remove " Chris Wilson
2016-09-23 12:55 ` Daniel Vetter
2016-09-23 13:07 ` [Intel-gfx] " Rob Clark
2016-08-29 7:08 ` [PATCH 04/11] drm/nouveau: " Chris Wilson
2016-09-23 12:55 ` Daniel Vetter
2016-10-05 16:05 ` Sumit Semwal
2016-08-29 7:08 ` [PATCH 05/11] drm/vmwgfx: " Chris Wilson
2016-09-23 12:56 ` Daniel Vetter
2016-10-05 16:11 ` [Intel-gfx] " Sumit Semwal
2016-10-05 17:03 ` Sinclair Yeh
2016-08-29 7:08 ` [PATCH 06/11] dma-buf: Introduce fence_get_rcu_safe() Chris Wilson
2016-09-23 12:59 ` Daniel Vetter
2016-09-23 13:34 ` Markus Heiser
2016-08-29 7:08 ` [PATCH 07/11] dma-buf: Restart reservation_object_get_fences_rcu() after writes Chris Wilson
2016-09-23 13:03 ` Daniel Vetter
2016-08-29 7:08 ` [PATCH 08/11] dma-buf: Restart reservation_object_wait_timeout_rcu() " Chris Wilson
2016-09-23 13:18 ` Daniel Vetter
2016-08-29 7:08 ` [PATCH 09/11] dma-buf: Restart reservation_object_test_signaled_rcu() " Chris Wilson
2016-09-23 13:43 ` Daniel Vetter
2016-08-29 7:08 ` [PATCH 10/11] dma-buf: Use seqlock to close RCU race in test_signaled_single Chris Wilson
2016-09-23 13:49 ` [Linaro-mm-sig] " Daniel Vetter
2016-09-23 14:02 ` Chris Wilson
2016-09-25 20:43 ` Daniel Vetter
2016-08-29 7:08 ` [PATCH 11/11] dma-buf: Do a fast lockless check for poll with timeout=0 Chris Wilson
2016-08-29 18:16 ` [PATCH] dma-buf/sync-file: Avoid enable fence signaling if poll(.timeout=0) Chris Wilson
2016-08-29 18:26 ` Gustavo Padovan
2016-09-13 14:46 ` Sumit Semwal
2016-09-15 0:00 ` Rafael Antognolli
2016-09-21 7:26 ` Gustavo Padovan
2016-09-21 11:08 ` Chris Wilson
2016-09-23 13:50 ` Daniel Vetter [this message]
2016-09-23 14:15 ` [Intel-gfx] [PATCH 11/11] dma-buf: Do a fast lockless check for poll with timeout=0 Chris Wilson
2016-09-23 15:06 ` Chris Wilson
2016-09-23 15:20 ` [Intel-gfx] " Chris Wilson
2016-09-23 17:59 ` Christian König
2016-09-25 20:44 ` Daniel Vetter
2016-08-29 8:20 ` [PATCH 01/11] drm/amdgpu: Remove call to reservation_object_test_signaled_rcu before wait Christian König
2016-09-23 12:54 ` Daniel Vetter
2016-10-05 16:03 ` Sumit Semwal
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=20160923135044.GM3988@dvetter-linux.ger.corp.intel.com \
--to=daniel@ffwll.ch \
--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 \
--cc=linux-media@vger.kernel.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).