All of lore.kernel.org
 help / color / mirror / Atom feed
From: tiffany lin <tiffany.lin@mediatek.com>
To: Hans Verkuil <hverkuil@xs4all.nl>
Cc: Hans Verkuil <hans.verkuil@cisco.com>,
	<daniel.thompson@linaro.org>, "Rob Herring" <robh+dt@kernel.org>,
	Mauro Carvalho Chehab <mchehab@osg.samsung.com>,
	Matthias Brugger <matthias.bgg@gmail.com>,
	Daniel Kurtz <djkurtz@chromium.org>,
	Pawel Osciak <posciak@chromium.org>,
	Eddie Huang <eddie.huang@mediatek.com>,
	Yingjoe Chen <yingjoe.chen@mediatek.com>,
	<devicetree@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
	<linux-arm-kernel@lists.infradead.org>,
	<linux-media@vger.kernel.org>,
	<linux-mediatek@lists.infradead.org>, <PoChun.Lin@mediatek.com>,
	Andrew-CT Chen <andrew-ct.chen@mediatek.com>
Subject: Re: [PATCH v4 5/8] [Media] vcodec: mediatek: Add Mediatek V4L2 Video Encoder Driver
Date: Wed, 17 Feb 2016 16:33:12 +0800	[thread overview]
Message-ID: <1455697992.26202.13.camel@mtksdaap41> (raw)
In-Reply-To: <56C425A0.6080005@xs4all.nl>

Hi Hans,

On Wed, 2016-02-17 at 08:47 +0100, Hans Verkuil wrote:
> On 02/16/2016 07:37 AM, tiffany lin wrote:
> >>> +static int fops_vcodec_open(struct file *file)
> >>> +{
> >>> +	struct video_device *vfd = video_devdata(file);
> >>> +	struct mtk_vcodec_dev *dev = video_drvdata(file);
> >>> +	struct mtk_vcodec_ctx *ctx = NULL;
> >>> +	int ret = 0;
> >>> +
> >>> +	mutex_lock(&dev->dev_mutex);
> >>> +
> >>> +	ctx = devm_kzalloc(&dev->plat_dev->dev, sizeof(*ctx), GFP_KERNEL);
> >>> +	if (!ctx) {
> >>> +		ret = -ENOMEM;
> >>> +		goto err_alloc;
> >>> +	}
> >>> +
> >>> +	if (dev->num_instances >= MTK_VCODEC_MAX_ENCODER_INSTANCES) {
> >>> +		mtk_v4l2_err("Too many open contexts\n");
> >>> +		ret = -EBUSY;
> >>> +		goto err_no_ctx;
> >>
> >> Hmm. I never like it if you can't open a video node because of a reason like this.
> >>
> >> I.e. a simple 'v4l2-ctl -D' (i.e. calling QUERYCAP) should never fail.
> >>
> >> If there are hardware limitation that prevent more than X instances from running at
> >> the same time, then those limitations typically kick in when you start to stream
> >> (or possibly when calling REQBUFS). But before that it should always be possible to
> >> open the device.
> >>
> >> Having this check at open() is an indication of a poor design.
> >>
> >> Is this is a hardware limitation at all?
> >>
> > This is to make sure performance meet requirements, such as bitrate and
> > framerate.
> 
> Is it the driver's job to make enforce this? What if the application only
> deals with low-res video, but wants to encode a lot of those? Or is encoding
> a video off-line?
> 
> The driver generally doesn't know the use-case, so if this is an artificial
> limitation as opposed to a hardware limitation, then I would just drop this.
> 
We got your point, we will remove this artificial limitation in next
version.

best regards,
Tiffany


> Regards,
> 
> 	Hans
> 
> > We got your point. We will remove this and move limitation control to
> > start_streaming or REQBUFS.
> > Appreciated for your suggestion.:)
> > 
> > 
> >>> +	}
> 

WARNING: multiple messages have this Message-ID (diff)
From: tiffany lin <tiffany.lin@mediatek.com>
To: Hans Verkuil <hverkuil@xs4all.nl>
Cc: Hans Verkuil <hans.verkuil@cisco.com>,
	daniel.thompson@linaro.org, Rob Herring <robh+dt@kernel.org>,
	Mauro Carvalho Chehab <mchehab@osg.samsung.com>,
	Matthias Brugger <matthias.bgg@gmail.com>,
	Daniel Kurtz <djkurtz@chromium.org>,
	Pawel Osciak <posciak@chromium.org>,
	Eddie Huang <eddie.huang@mediatek.com>,
	Yingjoe Chen <yingjoe.chen@mediatek.com>,
	devicetree@vger.kernel.org, linux-kernel@vger.kernel.org,
	linux-arm-kernel@lists.infradead.org,
	linux-media@vger.kernel.org, linux-mediatek@lists.infradead.org,
	PoChun.Lin@mediatek.com,
	Andrew-CT Chen <andrew-ct.chen@mediatek.com>
Subject: Re: [PATCH v4 5/8] [Media] vcodec: mediatek: Add Mediatek V4L2 Video Encoder Driver
Date: Wed, 17 Feb 2016 16:33:12 +0800	[thread overview]
Message-ID: <1455697992.26202.13.camel@mtksdaap41> (raw)
In-Reply-To: <56C425A0.6080005@xs4all.nl>

Hi Hans,

On Wed, 2016-02-17 at 08:47 +0100, Hans Verkuil wrote:
> On 02/16/2016 07:37 AM, tiffany lin wrote:
> >>> +static int fops_vcodec_open(struct file *file)
> >>> +{
> >>> +	struct video_device *vfd = video_devdata(file);
> >>> +	struct mtk_vcodec_dev *dev = video_drvdata(file);
> >>> +	struct mtk_vcodec_ctx *ctx = NULL;
> >>> +	int ret = 0;
> >>> +
> >>> +	mutex_lock(&dev->dev_mutex);
> >>> +
> >>> +	ctx = devm_kzalloc(&dev->plat_dev->dev, sizeof(*ctx), GFP_KERNEL);
> >>> +	if (!ctx) {
> >>> +		ret = -ENOMEM;
> >>> +		goto err_alloc;
> >>> +	}
> >>> +
> >>> +	if (dev->num_instances >= MTK_VCODEC_MAX_ENCODER_INSTANCES) {
> >>> +		mtk_v4l2_err("Too many open contexts\n");
> >>> +		ret = -EBUSY;
> >>> +		goto err_no_ctx;
> >>
> >> Hmm. I never like it if you can't open a video node because of a reason like this.
> >>
> >> I.e. a simple 'v4l2-ctl -D' (i.e. calling QUERYCAP) should never fail.
> >>
> >> If there are hardware limitation that prevent more than X instances from running at
> >> the same time, then those limitations typically kick in when you start to stream
> >> (or possibly when calling REQBUFS). But before that it should always be possible to
> >> open the device.
> >>
> >> Having this check at open() is an indication of a poor design.
> >>
> >> Is this is a hardware limitation at all?
> >>
> > This is to make sure performance meet requirements, such as bitrate and
> > framerate.
> 
> Is it the driver's job to make enforce this? What if the application only
> deals with low-res video, but wants to encode a lot of those? Or is encoding
> a video off-line?
> 
> The driver generally doesn't know the use-case, so if this is an artificial
> limitation as opposed to a hardware limitation, then I would just drop this.
> 
We got your point, we will remove this artificial limitation in next
version.

best regards,
Tiffany


> Regards,
> 
> 	Hans
> 
> > We got your point. We will remove this and move limitation control to
> > start_streaming or REQBUFS.
> > Appreciated for your suggestion.:)
> > 
> > 
> >>> +	}
> 

WARNING: multiple messages have this Message-ID (diff)
From: tiffany.lin@mediatek.com (tiffany lin)
To: linux-arm-kernel@lists.infradead.org
Subject: [PATCH v4 5/8] [Media] vcodec: mediatek: Add Mediatek V4L2 Video Encoder Driver
Date: Wed, 17 Feb 2016 16:33:12 +0800	[thread overview]
Message-ID: <1455697992.26202.13.camel@mtksdaap41> (raw)
In-Reply-To: <56C425A0.6080005@xs4all.nl>

Hi Hans,

On Wed, 2016-02-17 at 08:47 +0100, Hans Verkuil wrote:
> On 02/16/2016 07:37 AM, tiffany lin wrote:
> >>> +static int fops_vcodec_open(struct file *file)
> >>> +{
> >>> +	struct video_device *vfd = video_devdata(file);
> >>> +	struct mtk_vcodec_dev *dev = video_drvdata(file);
> >>> +	struct mtk_vcodec_ctx *ctx = NULL;
> >>> +	int ret = 0;
> >>> +
> >>> +	mutex_lock(&dev->dev_mutex);
> >>> +
> >>> +	ctx = devm_kzalloc(&dev->plat_dev->dev, sizeof(*ctx), GFP_KERNEL);
> >>> +	if (!ctx) {
> >>> +		ret = -ENOMEM;
> >>> +		goto err_alloc;
> >>> +	}
> >>> +
> >>> +	if (dev->num_instances >= MTK_VCODEC_MAX_ENCODER_INSTANCES) {
> >>> +		mtk_v4l2_err("Too many open contexts\n");
> >>> +		ret = -EBUSY;
> >>> +		goto err_no_ctx;
> >>
> >> Hmm. I never like it if you can't open a video node because of a reason like this.
> >>
> >> I.e. a simple 'v4l2-ctl -D' (i.e. calling QUERYCAP) should never fail.
> >>
> >> If there are hardware limitation that prevent more than X instances from running at
> >> the same time, then those limitations typically kick in when you start to stream
> >> (or possibly when calling REQBUFS). But before that it should always be possible to
> >> open the device.
> >>
> >> Having this check at open() is an indication of a poor design.
> >>
> >> Is this is a hardware limitation at all?
> >>
> > This is to make sure performance meet requirements, such as bitrate and
> > framerate.
> 
> Is it the driver's job to make enforce this? What if the application only
> deals with low-res video, but wants to encode a lot of those? Or is encoding
> a video off-line?
> 
> The driver generally doesn't know the use-case, so if this is an artificial
> limitation as opposed to a hardware limitation, then I would just drop this.
> 
We got your point, we will remove this artificial limitation in next
version.

best regards,
Tiffany


> Regards,
> 
> 	Hans
> 
> > We got your point. We will remove this and move limitation control to
> > start_streaming or REQBUFS.
> > Appreciated for your suggestion.:)
> > 
> > 
> >>> +	}
> 

  reply	other threads:[~2016-02-17  8:33 UTC|newest]

Thread overview: 102+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-02-04 11:34 [PATCH v4 0/8] Add MT8173 Video Encoder Driver and VPU Driver Tiffany Lin
2016-02-04 11:34 ` Tiffany Lin
2016-02-04 11:34 ` Tiffany Lin
2016-02-04 11:34 ` [PATCH v4 1/8] dt-bindings: Add a binding for Mediatek Video Processor Tiffany Lin
2016-02-04 11:34   ` Tiffany Lin
2016-02-04 11:34   ` Tiffany Lin
2016-02-04 11:34   ` [PATCH v4 2/8] [media] VPU: mediatek: support Mediatek VPU Tiffany Lin
2016-02-04 11:34     ` Tiffany Lin
2016-02-04 11:34     ` Tiffany Lin
2016-02-04 11:34     ` [PATCH v4 3/8] arm64: dts: mediatek: Add node for Mediatek Video Processor Unit Tiffany Lin
2016-02-04 11:34       ` Tiffany Lin
2016-02-04 11:34       ` Tiffany Lin
2016-02-04 11:34       ` [PATCH v4 4/8] dt-bindings: Add a binding for Mediatek Video Encoder Tiffany Lin
2016-02-04 11:34         ` Tiffany Lin
2016-02-04 11:34         ` Tiffany Lin
2016-02-04 11:35         ` [PATCH v4 5/8] [Media] vcodec: mediatek: Add Mediatek V4L2 Video Encoder Driver Tiffany Lin
2016-02-04 11:35           ` Tiffany Lin
2016-02-04 11:35           ` [PATCH v4 6/8] [media] vcodec: mediatek: Add Mediatek VP8 " Tiffany Lin
2016-02-04 11:35             ` Tiffany Lin
2016-02-04 11:35             ` Tiffany Lin
2016-02-04 11:35             ` [PATCH v4 7/8] [media] vcodec: mediatek: Add Mediatek H264 " Tiffany Lin
2016-02-04 11:35               ` Tiffany Lin
2016-02-04 11:35               ` Tiffany Lin
2016-02-04 11:35               ` [PATCH v4 8/8] arm64: dts: mediatek: Add Video Encoder for MT8173 Tiffany Lin
2016-02-04 11:35                 ` Tiffany Lin
2016-02-04 11:35                 ` Tiffany Lin
2016-02-15 11:33               ` [PATCH v4 7/8] [media] vcodec: mediatek: Add Mediatek H264 Video Encoder Driver Hans Verkuil
2016-02-15 11:33                 ` Hans Verkuil
2016-02-16 11:57                 ` pochun lin
2016-02-16 11:57                   ` pochun lin
2016-02-16 11:57                   ` pochun lin
2016-02-15 11:21           ` [PATCH v4 5/8] [Media] vcodec: mediatek: Add Mediatek V4L2 " Hans Verkuil
2016-02-16  6:37             ` tiffany lin
2016-02-16  6:37               ` tiffany lin
2016-02-16  7:44               ` Hans Verkuil
2016-02-16  7:44                 ` Hans Verkuil
2016-02-16  7:44                 ` Hans Verkuil
2016-02-16 13:20                 ` tiffany lin
2016-02-16 13:20                   ` tiffany lin
2016-02-16 13:20                   ` tiffany lin
2016-02-16 13:48                   ` Hans Verkuil
2016-02-16 13:48                     ` Hans Verkuil
2016-02-17  8:01                     ` tiffany lin
2016-02-17  8:01                       ` tiffany lin
2016-02-17  8:01                       ` tiffany lin
2016-02-17  8:31                       ` Hans Verkuil
2016-02-17  8:31                         ` Hans Verkuil
2016-02-17  9:23                         ` tiffany lin
2016-02-17  9:23                           ` tiffany lin
2016-02-17  9:23                           ` tiffany lin
2016-02-20  9:11                 ` tiffany lin
2016-02-20  9:11                   ` tiffany lin
2016-02-20  9:11                   ` tiffany lin
2016-02-20  9:18                   ` Hans Verkuil
2016-02-20  9:18                     ` Hans Verkuil
2016-02-20  9:18                     ` Hans Verkuil
2016-02-22 15:19                     ` tiffany lin
2016-02-22 15:19                       ` tiffany lin
2016-02-22 15:19                       ` tiffany lin
2016-02-17  7:47               ` Hans Verkuil
2016-02-17  7:47                 ` Hans Verkuil
2016-02-17  8:33                 ` tiffany lin [this message]
2016-02-17  8:33                   ` tiffany lin
2016-02-17  8:33                   ` tiffany lin
2016-02-23  5:46           ` Wu-Cheng Li (李務誠)
2016-02-24  8:26             ` tiffany lin
2016-02-24  8:26               ` tiffany lin
2016-02-08 18:56         ` [PATCH v4 4/8] dt-bindings: Add a binding for Mediatek Video Encoder Rob Herring
2016-02-08 18:56           ` Rob Herring
2016-02-08 18:56           ` Rob Herring
2016-02-09 11:29         ` Daniel Kurtz
2016-02-09 11:29           ` Daniel Kurtz
2016-02-09 11:29           ` Daniel Kurtz
2016-02-15 10:42           ` Daniel Kurtz
2016-02-15 10:42             ` Daniel Kurtz
2016-02-15 10:42             ` Daniel Kurtz
2016-02-16  2:09             ` tiffany lin
2016-02-16  2:09               ` tiffany lin
2016-02-16  2:09               ` tiffany lin
2016-02-16  2:09               ` tiffany lin
2016-02-15 10:07     ` [PATCH v4 2/8] [media] VPU: mediatek: support Mediatek VPU Hans Verkuil
2016-02-15 10:07       ` Hans Verkuil
2016-02-15 11:20       ` tiffany lin
2016-02-15 11:20         ` tiffany lin
2016-02-15 11:20         ` tiffany lin
2016-02-15 10:13     ` Hans Verkuil
2016-02-15 10:13       ` Hans Verkuil
2016-02-15 11:27       ` tiffany lin
2016-02-15 11:27         ` tiffany lin
2016-02-15 11:27         ` tiffany lin
2016-02-15 13:59     ` Wu-Cheng Li (李務誠)
2016-02-15 13:59       ` Wu-Cheng Li (李務誠)
2016-02-16  9:36       ` andrew-ct chen
2016-02-16  9:36         ` andrew-ct chen
2016-02-16  9:36         ` andrew-ct chen
2016-02-08 18:54   ` [PATCH v4 1/8] dt-bindings: Add a binding for Mediatek Video Processor Rob Herring
2016-02-08 18:54     ` Rob Herring
2016-02-15 12:04 ` [PATCH v4 0/8] Add MT8173 Video Encoder Driver and VPU Driver Hans Verkuil
2016-02-15 12:04   ` Hans Verkuil
2016-02-16  6:46   ` tiffany lin
2016-02-16  6:46     ` tiffany lin
2016-02-16  6:46     ` tiffany lin

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=1455697992.26202.13.camel@mtksdaap41 \
    --to=tiffany.lin@mediatek.com \
    --cc=PoChun.Lin@mediatek.com \
    --cc=andrew-ct.chen@mediatek.com \
    --cc=daniel.thompson@linaro.org \
    --cc=devicetree@vger.kernel.org \
    --cc=djkurtz@chromium.org \
    --cc=eddie.huang@mediatek.com \
    --cc=hans.verkuil@cisco.com \
    --cc=hverkuil@xs4all.nl \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-media@vger.kernel.org \
    --cc=linux-mediatek@lists.infradead.org \
    --cc=matthias.bgg@gmail.com \
    --cc=mchehab@osg.samsung.com \
    --cc=posciak@chromium.org \
    --cc=robh+dt@kernel.org \
    --cc=yingjoe.chen@mediatek.com \
    /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.