Re: [PATCH] Support for write-only controls

Re: [PATCH] Support for write-only controls

[Brandon Philips]

On 01:01 Tue 18 Dec 2007, Martin Rubli wrote:
> Thanks a lot for all your feedback and the constructive discussion and sorry 
> for the delay while I was without Internet on the weekend. I'll try to 
> summarize what we have so far:
>
> Write-only controls:
>
> It seems, everybody likes EACCES. Michael, maybe we could get some feedback 
> from you on this? It would be nice to change the spec, so that EACCES also 
> becomes the error for writing read-only controls--it seems appropriate. But 
> if for some reason we can't change that we should probably make the 
> write-only controls consistent and return EINVAL as well.
>
> Unusable controls due to device communication error:
>
> The easiest solution seems to be to set the V4L2_CTRL_FLAG_DISABLED flag as 
> was suggested. The documentation currently says "permanently disabled and 
> should be ignored by the application" which I think is exactly what applies 
> to the situation. The V4L2_CTRL_FLAG_NEXT_CTRL would still be respected by 
> drivers supporting the extended control enumeration, so no need to the spec 
> is required. But I would still add a short paragraph about the first part as 
> a guide for future implementations and a witness of this thread. ;-)
>
> As soon as everyone agrees on this, I will propose new patches. Let me know 
> what you think ...

Did you ever send out patches on this?  I can't seem to find them.

Cheers,

	Brandon

--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

Re: [PATCH] Support for write-only controls

[Brandon Philips]

On 15:12 Wed 13 Feb 2008, Brandon Philips wrote:
> On 01:01 Tue 18 Dec 2007, Martin Rubli wrote:
> > Thanks a lot for all your feedback and the constructive discussion and sorry 
> > for the delay while I was without Internet on the weekend. I'll try to 
> > summarize what we have so far:
> >
> > Write-only controls:
> >
> > It seems, everybody likes EACCES. Michael, maybe we could get some feedback 
> > from you on this? It would be nice to change the spec, so that EACCES also 
> > becomes the error for writing read-only controls--it seems appropriate. But 
> > if for some reason we can't change that we should probably make the 
> > write-only controls consistent and return EINVAL as well.
> >
> > Unusable controls due to device communication error:
> >
> > The easiest solution seems to be to set the V4L2_CTRL_FLAG_DISABLED flag as 
> > was suggested. The documentation currently says "permanently disabled and 
> > should be ignored by the application" which I think is exactly what applies 
> > to the situation. The V4L2_CTRL_FLAG_NEXT_CTRL would still be respected by 
> > drivers supporting the extended control enumeration, so no need to the spec 
> > is required. But I would still add a short paragraph about the first part as 
> > a guide for future implementations and a witness of this thread. ;-)
> >
> > As soon as everyone agrees on this, I will propose new patches. Let me know 
> > what you think ...
> 

Ping.  I never saw patches come across for this.

Thanks,

	Brandon

--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

RE: [PATCH] Support for write-only controls

[Martin Rubli]

[-- Attachment #1: Type: text/plain, Size: 1952 bytes --]

Hi Brandon,

> -----Original Message-----
> From: Brandon Philips [mailto:brandon@ifup.org]
> Sent: Tuesday, April 15, 2008 08:44
> To: Martin Rubli; linux1@rubli.info; Laurent Pinchart
> Cc: Linux and Kernel Video
> Subject: Re: [PATCH] Support for write-only controls
> 
> On 15:12 Wed 13 Feb 2008, Brandon Philips wrote:
> > On 01:01 Tue 18 Dec 2007, Martin Rubli wrote:
> > > Thanks a lot for all your feedback and the constructive discussion and
> sorry
> > > for the delay while I was without Internet on the weekend. I'll try to
> > > summarize what we have so far:
> > >
> > > Write-only controls:
> > >
> > > It seems, everybody likes EACCES. Michael, maybe we could get some
> feedback
> > > from you on this? It would be nice to change the spec, so that EACCES
> also
> > > becomes the error for writing read-only controls--it seems
appropriate.
> But
> > > if for some reason we can't change that we should probably make the
> > > write-only controls consistent and return EINVAL as well.
> > >
> > > Unusable controls due to device communication error:
> > >
> > > The easiest solution seems to be to set the V4L2_CTRL_FLAG_DISABLED
> flag as
> > > was suggested. The documentation currently says "permanently disabled
> and
> > > should be ignored by the application" which I think is exactly what
> applies
> > > to the situation. The V4L2_CTRL_FLAG_NEXT_CTRL would still be
> respected by
> > > drivers supporting the extended control enumeration, so no need to the
> spec
> > > is required. But I would still add a short paragraph about the first
part as
> > > a guide for future implementations and a witness of this thread. ;-)
> > >
> > > As soon as everyone agrees on this, I will propose new patches. Let me
> know
> > > what you think ...
> >
> 
> Ping.  I never saw patches come across for this.

Attached the patch for videodev2.h and the documentation sources.

Signed-off-by: Martin Rubli <linux1 rubli info>

Cheers,
Martin

[-- Attachment #2: v4l2_ctrl_flag_write_only.patch --]
[-- Type: application/octet-stream, Size: 578 bytes --]

diff -Naur v4l-dvb-5de66b04c598/linux/include/linux/videodev2.h v4l-dvb-new/linux/include/linux/videodev2.h
--- v4l-dvb-5de66b04c598/linux/include/linux/videodev2.h	2008-04-16 06:22:11.000000000 +0800
+++ v4l-dvb-new/linux/include/linux/videodev2.h	2008-04-16 17:37:38.000000000 +0800
@@ -831,6 +831,7 @@
 #define V4L2_CTRL_FLAG_UPDATE 		0x0008
 #define V4L2_CTRL_FLAG_INACTIVE 	0x0010
 #define V4L2_CTRL_FLAG_SLIDER 		0x0020
+#define V4L2_CTRL_FLAG_WRITE_ONLY 	0x0040
 
 /*  Query flag, to be ORed with the control ID */
 #define V4L2_CTRL_FLAG_NEXT_CTRL	0x80000000

[-- Attachment #3: v4l2_ctrl_flag_write_only-doc.patch --]
[-- Type: application/octet-stream, Size: 960 bytes --]

diff -Naur v4l2spec-0.24/vidioc-queryctrl.sgml v4l2spec-0.24a/vidioc-queryctrl.sgml
--- v4l2spec-0.24/vidioc-queryctrl.sgml	2008-03-06 23:42:13.000000000 +0800
+++ v4l2spec-0.24a/vidioc-queryctrl.sgml	2008-04-16 17:27:36.000000000 +0800
@@ -361,6 +361,15 @@
             <entry>A hint that this control is best represented as a
 slider-like element in a user interface.</entry>
           </row>
+          <row>
+            <entry><constant>V4L2_CTRL_FLAG_WRITE_ONLY</constant></entry>
+            <entry>0x0040</entry>
+            <entry>This control is permanently writable only. Any
+attempt to read the control will result in an EACCES error code. This
+flag is typically present for relative controls or action controls where
+writing a value will cause the device to carry out a given action
+(e. g. motor control) but no meaningful value can be returned.</entry>
+          </row>
         </tbody>
       </tgroup>
     </table>

[-- Attachment #4: Type: text/plain, Size: 164 bytes --]

--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

Re: [PATCH] Support for write-only controls

[Mauro Carvalho Chehab]

On Thu, 17 Apr 2008 00:41:43 +0200
Laurent Pinchart <laurent.pinchart@skynet.be> wrote:

> Hi Mauro,
> 
> On Tuesday 15 April 2008, Mauro Carvalho Chehab wrote:
> > On Mon, 14 Apr 2008 17:44:16 -0700
> >
> > Brandon Philips <brandon@ifup.org> wrote:
> > > Ping.  I never saw patches come across for this.
> >
> > Brandon, Could you please add this on one of your trees, together with
> > those pending V4L2 API patches for UVC? I want to merge those changes
> > together with the in-kernel driver that firstly requires such changes.
> >
> > Btw, Laurent, are you ready for uvc inclusion? The window for the next
> > kernel is about to open.
> 
> The driver is constantly evolving, but I don't see any showstopper.
> 
> How exactly should we proceed ? Should I submit a patch based on git 2.6.25 ? 

The better is to generate against -hg, but against 2.6.25 should also be fine.

> What about future patches ?

You may submit to me as they are tested. I prefer if you use hg. Otherwise,
please submit via email.

> Do you have any opinion regarding keeping or closing the linux-uvc-devel 
> mailing list ?

It is up to you. There are some drivers already with specific lists, like ivtv
and pvrusb2. If you prefer, you may also use the std V4L ML.

Cheers,
Mauro

--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

Re: [PATCH] Support for write-only controls

[Laurent Pinchart]

Hi Mauro,

On Tuesday 15 April 2008, Mauro Carvalho Chehab wrote:
> On Mon, 14 Apr 2008 17:44:16 -0700
>
> Brandon Philips <brandon@ifup.org> wrote:
> > Ping.  I never saw patches come across for this.
>
> Brandon, Could you please add this on one of your trees, together with
> those pending V4L2 API patches for UVC? I want to merge those changes
> together with the in-kernel driver that firstly requires such changes.
>
> Btw, Laurent, are you ready for uvc inclusion? The window for the next
> kernel is about to open.

The driver is constantly evolving, but I don't see any showstopper.

How exactly should we proceed ? Should I submit a patch based on git 2.6.25 ? 
What about future patches ?

Do you have any opinion regarding keeping or closing the linux-uvc-devel 
mailing list ?

Laurent Pinchart

--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

[PATCH] v4l2-spec: write-only controls

['Brandon Philips']

Hello Michael-

Please fold this new control flag into the v4l2spec.

Thanks,

	Brandon

diff -Naur v4l2spec-0.24/vidioc-queryctrl.sgml v4l2spec-0.24a/vidioc-queryctrl.sgml
--- v4l2spec-0.24/vidioc-queryctrl.sgml	2008-03-06 23:42:13.000000000 +0800
+++ v4l2spec-0.24a/vidioc-queryctrl.sgml	2008-04-16 17:27:36.000000000 +0800
@@ -361,6 +361,15 @@
             <entry>A hint that this control is best represented as a
 slider-like element in a user interface.</entry>
           </row>
+          <row>
+            <entry><constant>V4L2_CTRL_FLAG_WRITE_ONLY</constant></entry>
+            <entry>0x0040</entry>
+            <entry>This control is permanently writable only. Any
+attempt to read the control will result in an EACCES error code. This
+flag is typically present for relative controls or action controls where
+writing a value will cause the device to carry out a given action
+(e. g. motor control) but no meaningful value can be returned.</entry>
+          </row>
         </tbody>
       </tgroup>
     </table>

--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

Re: [PATCH] Support for write-only controls

[Brandon Philips]

On 00:19 Tue 15 Apr 2008, Mauro Carvalho Chehab wrote:
> Brandon, Could you please add this on one of your trees, together with those
> pending V4L2 API patches for UVC? I want to merge those changes together with the
> in-kernel driver that firstly requires such changes.

I have a tree with the change sets.  Please don't pull from the tip
though: hg pull -r 4ca1ed646f89 http://ifup.org/hg/v4l-uvc

The tip of that tree has UVC and all of the Kconfig/Makefile bits too.

The patch set for the tree: http://ifup.org/hg/uvc-v4l-patches

If Laurent wants to add his sign off to that last patch (based on r204)
we can commit that too :D

Cheers,

	Brandon

--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

Re: [PATCH] Support for write-only controls

[Laurent Pinchart]

On Tuesday 22 April 2008 04:22, Brandon Philips wrote:
> On 00:19 Tue 15 Apr 2008, Mauro Carvalho Chehab wrote:
> > Brandon, Could you please add this on one of your trees, together with
> > those pending V4L2 API patches for UVC? I want to merge those changes
> > together with the in-kernel driver that firstly requires such changes.
> 
> I have a tree with the change sets.  Please don't pull from the tip
> though: hg pull -r 4ca1ed646f89 http://ifup.org/hg/v4l-uvc
> 
> The tip of that tree has UVC and all of the Kconfig/Makefile bits too.
> 
> The patch set for the tree: http://ifup.org/hg/uvc-v4l-patches
> 
> If Laurent wants to add his sign off to that last patch (based on r204)
> we can commit that too :D

I want the driver to be properly reviewed on both video4linux-list and 
linux-usb. I will post a patch on both mailing lists today.

Thanks for taking care of the Kconfig/Makefile bits. Could you send them to me 
so that I can include them in the patch to be reviewed ?

Cheers,

Laurent Pinchart

--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list

Re: [PATCH] Support for write-only controls

[Brandon Philips]

On 10:40 Tue 22 Apr 2008, Laurent Pinchart wrote:
> On Tuesday 22 April 2008 04:22, Brandon Philips wrote:
> > On 00:19 Tue 15 Apr 2008, Mauro Carvalho Chehab wrote:
> > > Brandon, Could you please add this on one of your trees, together with
> > > those pending V4L2 API patches for UVC? I want to merge those changes
> > > together with the in-kernel driver that firstly requires such changes.
> > 
> > I have a tree with the change sets.  Please don't pull from the tip
> > though: hg pull -r 4ca1ed646f89 http://ifup.org/hg/v4l-uvc
> > 
> > The tip of that tree has UVC and all of the Kconfig/Makefile bits too.
> > 
> > The patch set for the tree: http://ifup.org/hg/uvc-v4l-patches
> > 
> > If Laurent wants to add his sign off to that last patch (based on r204)
> > we can commit that too :D
> 
> I want the driver to be properly reviewed on both video4linux-list and 
> linux-usb. I will post a patch on both mailing lists today.

Certainly :D

I put the UVC patch on my tree just to ensure I got all the right bits
together ;)

> Thanks for taking care of the Kconfig/Makefile bits. Could you send
> them to me so that I can include them in the patch to be reviewed ?

Signed-off-by: Brandon Philips <bphilips@suse.de>

--- a/linux/drivers/media/video/Kconfig
+++ b/linux/drivers/media/video/Kconfig
@@ -741,6 +741,14 @@ menuconfig V4L_USB_DRIVERS
 
 if V4L_USB_DRIVERS && USB
 
+config USB_UVC
+	tristate "USB Video Class (UVC)"
+	---help---
+	  Support for the USB Video Class (UVC).  Currently only video
+	  input devices, such as webcams, are supported.
+
+	  For more information see: <http://linux-uvc.berlios.de/>
+
 source "drivers/media/video/pvrusb2/Kconfig"
 
 source "drivers/media/video/em28xx/Kconfig"
--- a/linux/drivers/media/video/Makefile
+++ b/linux/drivers/media/video/Makefile
@@ -144,6 +144,8 @@ obj-$(CONFIG_SOC_CAMERA_MT9M001)	+= mt9m
 obj-$(CONFIG_SOC_CAMERA_MT9M001)	+= mt9m001.o
 obj-$(CONFIG_SOC_CAMERA_MT9V022)	+= mt9v022.o
 
+obj-$(CONFIG_USB_UVC) += uvc/
+
 obj-$(CONFIG_VIDEO_AU0828) += au0828/
 
 EXTRA_CFLAGS += -Idrivers/media/dvb/dvb-core
--- /dev/null
+++ b/linux/drivers/media/video/uvc/Makefile
@@ -0,0 +1,3 @@
+uvcvideo-objs  := uvc_driver.o uvc_queue.o uvc_v4l2.o uvc_video.o uvc_ctrl.o\
+			uvc_status.o uvc_isight.o
+obj-$(CONFIG_USB_UVC) := uvcvideo.o

--
video4linux-list mailing list
Unsubscribe mailto:video4linux-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/video4linux-list