From: Hans Verkuil <hverkuil@xs4all.nl>
To: Sakari Ailus <sakari.ailus@linux.intel.com>, linux-media@vger.kernel.org
Cc: pawel@osciak.com, m.szyprowski@samsung.com,
kyungmin.park@samsung.com, sumit.semwal@linaro.org,
robdclark@gmail.com, daniel.vetter@ffwll.ch, labbott@redhat.com
Subject: Re: [RFC RESEND 02/11] vb2: Move buffer cache synchronisation to prepare from queue
Date: Fri, 11 Sep 2015 18:11:16 +0200 [thread overview]
Message-ID: <55F2FD24.90506@xs4all.nl> (raw)
In-Reply-To: <1441972234-8643-3-git-send-email-sakari.ailus@linux.intel.com>
On 09/11/2015 01:50 PM, Sakari Ailus wrote:
> The buffer cache should be synchronised in buffer preparation, not when
> the buffer is queued to the device. Fix this.
>
> Mmap buffers do not need cache synchronisation since they are always
> coherent.
>
While I agree with this change, I also think it is incomplete. The memop prepare()
is now done before it is queued to the driver, but the corresponding memop
finish() is only called when it is returned by the driver.
So if the application queues buffers and (without calling STREAMON) destroys them
all (REQBUFS(0)), you should get 'unbalanced' messages in the kernel log.
I'm pretty sure v4l2-compliance tests this scenario, so that would be a good thing
to check.
Regards,
Hans
> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
> ---
> drivers/media/v4l2-core/videobuf2-core.c | 18 +++++++++++-------
> 1 file changed, 11 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/media/v4l2-core/videobuf2-core.c b/drivers/media/v4l2-core/videobuf2-core.c
> index af6a23a..64fce4d 100644
> --- a/drivers/media/v4l2-core/videobuf2-core.c
> +++ b/drivers/media/v4l2-core/videobuf2-core.c
> @@ -1636,10 +1636,6 @@ static void __enqueue_in_driver(struct vb2_buffer *vb)
>
> trace_vb2_buf_queue(q, vb);
>
> - /* sync buffers */
> - for (plane = 0; plane < vb->num_planes; ++plane)
> - call_void_memop(vb, prepare, vb->planes[plane].mem_priv);
> -
> call_void_vb_qop(vb, buf_queue, vb);
> }
>
> @@ -1694,11 +1690,19 @@ static int __buf_prepare(struct vb2_buffer *vb, const struct v4l2_buffer *b)
> ret = -EINVAL;
> }
>
> - if (ret)
> + if (ret) {
> dprintk(1, "buffer preparation failed: %d\n", ret);
> - vb->state = ret ? VB2_BUF_STATE_DEQUEUED : VB2_BUF_STATE_PREPARED;
> + vb->state = VB2_BUF_STATE_DEQUEUED;
> + return ret;
> + }
>
> - return ret;
> + /* sync buffers */
> + for (plane = 0; plane < vb->num_planes; ++plane)
> + call_void_memop(vb, prepare, vb->planes[plane].mem_priv);
> +
> + vb->state = VB2_BUF_STATE_PREPARED;
> +
> + return 0;
> }
>
> static int vb2_queue_or_prepare_buf(struct vb2_queue *q, struct v4l2_buffer *b,
>
next prev parent reply other threads:[~2015-09-11 16:12 UTC|newest]
Thread overview: 38+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-09-11 11:50 [RFC RESEND 00/11] vb2: Handle user cache hints, allow drivers to choose cache coherency Sakari Ailus
2015-09-11 11:50 ` [RFC RESEND 01/11] vb2: Rename confusingly named internal buffer preparation functions Sakari Ailus
2015-09-11 15:57 ` Hans Verkuil
2016-12-15 15:40 ` Laurent Pinchart
2015-09-11 11:50 ` [RFC RESEND 02/11] vb2: Move buffer cache synchronisation to prepare from queue Sakari Ailus
2015-09-11 16:11 ` Hans Verkuil [this message]
2015-09-11 11:50 ` [RFC RESEND 03/11] vb2: Move cache synchronisation from buffer done to dqbuf handler Sakari Ailus
2015-09-11 16:25 ` Hans Verkuil
2015-09-15 7:51 ` Sakari Ailus
2016-12-15 23:47 ` Laurent Pinchart
2015-09-11 11:50 ` [RFC RESEND 04/11] v4l: Unify cache management hint buffer flags Sakari Ailus
2015-09-11 16:26 ` Hans Verkuil
2015-09-11 16:44 ` Hans Verkuil
2016-12-15 20:15 ` Laurent Pinchart
2016-12-17 0:35 ` Sakari Ailus
2015-09-11 11:50 ` [RFC RESEND 05/11] v4l2-core: Don't sync cache for a buffer if so requested Sakari Ailus
2015-09-11 17:12 ` Hans Verkuil
2015-09-15 8:22 ` Sakari Ailus
2015-09-15 9:06 ` Hans Verkuil
2016-12-15 20:37 ` Laurent Pinchart
2016-12-17 0:36 ` Sakari Ailus
2015-09-11 11:50 ` [RFC RESEND 06/11] vb2: Improve struct vb2_mem_ops documentation; alloc and put are for MMAP Sakari Ailus
2015-09-11 17:13 ` Hans Verkuil
2016-12-15 20:50 ` Laurent Pinchart
2015-09-11 11:50 ` [RFC RESEND 07/11] vb2: dma-contig: Remove redundant sgt_base field Sakari Ailus
2015-09-11 17:28 ` Hans Verkuil
2015-09-15 8:26 ` Sakari Ailus
2016-12-15 21:08 ` Laurent Pinchart
2016-12-17 0:40 ` Sakari Ailus
2015-09-11 11:50 ` [RFC RESEND 08/11] vb2: dma-contig: Move vb2_dc_get_base_sgt() up Sakari Ailus
2016-12-15 21:12 ` Laurent Pinchart
2015-09-11 11:50 ` [RFC RESEND 09/11] vb2: dma-contig: Don't warn on failure in obtaining scatterlist Sakari Ailus
2016-12-15 21:13 ` Laurent Pinchart
2015-09-11 11:50 ` [RFC RESEND 10/11] vb2: dma-contig: Let drivers decide DMA attrs of MMAP and USERPTR bufs Sakari Ailus
2016-12-15 21:40 ` Laurent Pinchart
2015-09-11 11:50 ` [RFC RESEND 11/11] vb2: dma-contig: Add WARN_ON_ONCE() to check for potential bugs Sakari Ailus
2016-12-15 21:57 ` Laurent Pinchart
2016-12-17 0:50 ` Sakari Ailus
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=55F2FD24.90506@xs4all.nl \
--to=hverkuil@xs4all.nl \
--cc=daniel.vetter@ffwll.ch \
--cc=kyungmin.park@samsung.com \
--cc=labbott@redhat.com \
--cc=linux-media@vger.kernel.org \
--cc=m.szyprowski@samsung.com \
--cc=pawel@osciak.com \
--cc=robdclark@gmail.com \
--cc=sakari.ailus@linux.intel.com \
--cc=sumit.semwal@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 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).