All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Rémi Denis-Courmont" <remi@remlab.net>
To: Tomasz Stanislawski <t.stanislaws@samsung.com>
Cc: Mauro Carvalho Chehab <mchehab@redhat.com>,
	Laurent Pinchart <laurent.pinchart@ideasonboard.com>,
	<linux-media@vger.kernel.org>, <dri-devel@lists.freedesktop.org>,
	<airlied@redhat.com>, <m.szyprowski@samsung.com>,
	<kyungmin.park@samsung.com>, <sumit.semwal@ti.com>,
	<daeinki@gmail.com>, <daniel.vetter@ffwll.ch>,
	<robdclark@gmail.com>, <pawel@osciak.com>,
	<linaro-mm-sig@lists.linaro.org>, <hverkuil@xs4all.nl>,
	<subashrp@gmail.com>
Subject: Re: [PATCH v4 02/14] Documentation: media: description of DMABUF importing in V4L2
Date: Fri, 20 Apr 2012 12:56:53 +0200	[thread overview]
Message-ID: <d24e8c6e35352ed5800161713f728591@chewa.net> (raw)
In-Reply-To: <4F912141.8060200@samsung.com>

On Fri, 20 Apr 2012 10:41:37 +0200, Tomasz Stanislawski

<t.stanislaws@samsung.com> wrote:

>> Am I understanding wrong or are you saying that you want to drop

userptr

>> from V4L2 API in long-term? If so, why?

> 

> Dropping userptr is just some brainstorming idea.

> It was found out that userptr is not a good mean

> for buffer exchange between to two devices.



I can believe that. But I am also inclined to believe that DMABUF is

targetted at device-to-device transfer, while USERPTR is targetted at

device-to-user (or user-to-device) transfers. Are you saying applications

should use DMABUF and memory map the buffers? Or would you care to explain

how DMABUF addresses the problem space of USERPTR?



> The USERPTR simplifies userspace code but introduce

> a lot of complexity problems for the kernel drivers

> and frameworks.



It is not only a simplification. In some cases, USERPTR is the only I/O

method that supports zero copy in pretty much any circumstance. When the

user cannot reliably predict the maximum number of required buffers,

predicts a value larger than the device will negotiate, or needs buffers to

outlive STREAMOFF (?), MMAP requires memory copying. USERPTR does not.



Now, I do realize that some devices cannot support USERPTR efficiently,

then they should not support USERPTR. But for those devices that can, it

seems quite a nice performance enhancement.



-- 

Rémi Denis-Courmont

Sent from my collocated server

WARNING: multiple messages have this Message-ID
From: "Rémi Denis-Courmont" <remi@remlab.net>
To: Tomasz Stanislawski <t.stanislaws@samsung.com>
Cc: Mauro Carvalho Chehab <mchehab@redhat.com>,
	Laurent Pinchart <laurent.pinchart@ideasonboard.com>,
	linux-media@vger.kernel.org, dri-devel@lists.freedesktop.org,
	airlied@redhat.com, m.szyprowski@samsung.com,
	kyungmin.park@samsung.com, sumit.semwal@ti.com,
	daeinki@gmail.com, daniel.vetter@ffwll.ch, robdclark@gmail.com,
	pawel@osciak.com, linaro-mm-sig@lists.linaro.org,
	hverkuil@xs4all.nl, subashrp@gmail.com
Subject: Re: [PATCH v4 02/14] Documentation: media: description of DMABUF importing in V4L2
Date: Fri, 20 Apr 2012 12:56:53 +0200	[thread overview]
Message-ID: <d24e8c6e35352ed5800161713f728591@chewa.net> (raw)
In-Reply-To: <4F912141.8060200@samsung.com>

On Fri, 20 Apr 2012 10:41:37 +0200, Tomasz Stanislawski\r
<t.stanislaws@samsung.com> wrote:\r
>> Am I understanding wrong or are you saying that you want to drop\r
userptr\r
>> from V4L2 API in long-term? If so, why?\r
> \r
> Dropping userptr is just some brainstorming idea.\r
> It was found out that userptr is not a good mean\r
> for buffer exchange between to two devices.\r
\r
I can believe that. But I am also inclined to believe that DMABUF is\r
targetted at device-to-device transfer, while USERPTR is targetted at\r
device-to-user (or user-to-device) transfers. Are you saying applications\r
should use DMABUF and memory map the buffers? Or would you care to explain\r
how DMABUF addresses the problem space of USERPTR?\r
\r
> The USERPTR simplifies userspace code but introduce\r
> a lot of complexity problems for the kernel drivers\r
> and frameworks.\r
\r
It is not only a simplification. In some cases, USERPTR is the only I/O\r
method that supports zero copy in pretty much any circumstance. When the\r
user cannot reliably predict the maximum number of required buffers,\r
predicts a value larger than the device will negotiate, or needs buffers to\r
outlive STREAMOFF (?), MMAP requires memory copying. USERPTR does not.\r
\r
Now, I do realize that some devices cannot support USERPTR efficiently,\r
then they should not support USERPTR. But for those devices that can, it\r
seems quite a nice performance enhancement.\r
\r
-- \r
Rémi Denis-Courmont\r
Sent from my collocated server

  reply	other threads:[~2012-04-20 10:57 UTC|newest]

Thread overview: 43+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-04-13 15:47 [PATCH v4 00/14] Integration of videobuf2 with dmabuf Tomasz Stanislawski
2012-04-13 15:47 ` [PATCH v4 01/14] v4l: Add DMABUF as a memory type Tomasz Stanislawski
2012-04-17  0:57   ` Laurent Pinchart
2012-04-13 15:47 ` [PATCH v4 02/14] Documentation: media: description of DMABUF importing in V4L2 Tomasz Stanislawski
2012-04-16 23:25   ` Laurent Pinchart
2012-04-19 14:32     ` Tomasz Stanislawski
2012-04-19 20:36       ` Mauro Carvalho Chehab
2012-04-19 20:37       ` Mauro Carvalho Chehab
2012-04-20  8:41         ` Tomasz Stanislawski
2012-04-20 10:56           ` Rémi Denis-Courmont [this message]
2012-04-20 10:56             ` Rémi Denis-Courmont
2012-04-20 12:25             ` Tomasz Stanislawski
2012-04-20 13:03               ` Rémi Denis-Courmont
2012-04-20 13:03                 ` Rémi Denis-Courmont
2012-04-21 17:10                 ` Laurent Pinchart
2012-04-20 14:48               ` Mauro Carvalho Chehab
2012-04-20 13:36             ` Mauro Carvalho Chehab
2012-04-23  7:50               ` Marek Szyprowski
2012-04-23 14:00                 ` Mauro Carvalho Chehab
2012-04-13 15:47 ` [PATCH v4 03/14] v4l: vb2: add support for shared buffer (dma_buf) Tomasz Stanislawski
2012-04-17  0:57   ` Laurent Pinchart
2012-04-13 15:47 ` [PATCH v4 04/14] v4l: vb: remove warnings about MEMORY_DMABUF Tomasz Stanislawski
2012-04-17  0:57   ` Laurent Pinchart
2012-04-13 15:47 ` [PATCH v4 05/14] v4l: vb2-dma-contig: Shorten vb2_dma_contig prefix to vb2_dc Tomasz Stanislawski
2012-04-13 15:47 ` [PATCH v4 06/14] v4l: vb2-dma-contig: Remove unneeded allocation context structure Tomasz Stanislawski
2012-04-17  0:57   ` Laurent Pinchart
2012-04-13 15:47 ` [PATCH v4 07/14] v4l: vb2-dma-contig: Reorder functions Tomasz Stanislawski
2012-04-13 15:47 ` [PATCH v4 08/14] v4l: vb2-dma-contig: add support for scatterlist in userptr mode Tomasz Stanislawski
2012-04-17  0:43   ` Laurent Pinchart
2012-04-17  0:43     ` Laurent Pinchart
2012-04-20  8:52     ` Tomasz Stanislawski
2012-04-13 15:47 ` [PATCH v4 09/14] v4l: vb2: add prepare/finish callbacks to allocators Tomasz Stanislawski
2012-04-17  0:57   ` Laurent Pinchart
2012-04-13 15:47 ` [PATCH v4 10/14] v4l: vb2-dma-contig: add prepare/finish to dma-contig allocator Tomasz Stanislawski
2012-04-13 15:47 ` [PATCH v4 11/14] v4l: vb2-dma-contig: add support for dma_buf importing Tomasz Stanislawski
2012-04-17  0:57   ` Laurent Pinchart
2012-04-17  0:57     ` Laurent Pinchart
2012-04-19 11:38     ` Tomasz Stanislawski
2012-04-13 15:47 ` [PATCH v4 12/14] v4l: vb2-dma-contig: change map/unmap behaviour for importers Tomasz Stanislawski
2012-04-17  1:03   ` Laurent Pinchart
2012-04-13 15:47 ` [PATCH v4 13/14] v4l: s5p-tv: mixer: support for dmabuf importing Tomasz Stanislawski
2012-04-13 15:47 ` [PATCH v4 14/14] v4l: fimc: " Tomasz Stanislawski
2012-04-20 12:56   ` Sylwester Nawrocki

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=d24e8c6e35352ed5800161713f728591@chewa.net \
    --to=remi@remlab.net \
    --cc=airlied@redhat.com \
    --cc=daeinki@gmail.com \
    --cc=daniel.vetter@ffwll.ch \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=hverkuil@xs4all.nl \
    --cc=kyungmin.park@samsung.com \
    --cc=laurent.pinchart@ideasonboard.com \
    --cc=linaro-mm-sig@lists.linaro.org \
    --cc=linux-media@vger.kernel.org \
    --cc=m.szyprowski@samsung.com \
    --cc=mchehab@redhat.com \
    --cc=pawel@osciak.com \
    --cc=robdclark@gmail.com \
    --cc=subashrp@gmail.com \
    --cc=sumit.semwal@ti.com \
    --cc=t.stanislaws@samsung.com \
    --subject='Re: [PATCH v4 02/14] Documentation: media: description of DMABUF importing in V4L2' \
    /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

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.