2007-12-06 22:36:44

by Gregor Jasny

[permalink] [raw]
Subject: [PATCH] vivi driver works only as first device

From: Gregor Jasny <[email protected]>

When the vivi driver allocates a video device, video_register_device() stores the
allocated device minor inside the vivi structure. But when the device node is opened,
the file minor number is compared to the minor in the device list. So this patch
copies the allocated minor in the device list, too.

Without this patch it is impossible to use the vivi driver mith minors greater than zero.

Signed-off-by: Gregor Jasny <[email protected]>
---
diff --git a/drivers/media/video/vivi.c b/drivers/media/video/vivi.c
index ee73dc7..f4e533a 100644
--- a/drivers/media/video/vivi.c
+++ b/drivers/media/video/vivi.c
@@ -1175,6 +1175,9 @@ static int __init vivi_init(void)

ret = video_register_device(&vivi, VFL_TYPE_GRABBER, video_nr);
printk(KERN_INFO "Video Technology Magazine Virtual Video Capture Board (Load status: %d)\n", ret);
+
+ dev->vfd.minor = vivi.minor;
+
return ret;
}


2007-12-10 07:09:50

by Mauro Carvalho Chehab

[permalink] [raw]
Subject: Re: [PATCH] vivi driver works only as first device

Hi Gregor,
Em Qui, 2007-12-06 às 23:06 +0100, Gregor Jasny escreveu:
> From: Gregor Jasny <[email protected]>
>
> When the vivi driver allocates a video device, video_register_device() stores the
> allocated device minor inside the vivi structure. But when the device node is opened,
> the file minor number is compared to the minor in the device list. So this patch
> copies the allocated minor in the device list, too.

Thanks for the report. Instead of applying your patch, I decided to
better analyze the issue, fixing it with the proper solution. The issue
is that vivi_register changes iminor, but this change were not properly
returned to the driver.

--
Cheers,
Mauro

2008-01-06 22:47:27

by Gregor Jasny

[permalink] [raw]
Subject: Re: [PATCH] vivi driver works only as first device

On Dec 10, 2007 8:08 AM, Mauro Carvalho Chehab <[email protected]> wrote:
> Thanks for the report. Instead of applying your patch, I decided to
> better analyze the issue, fixing it with the proper solution. The issue
> is that vivi_register changes iminor, but this change were not properly
> returned to the driver.

Any chance to push this into 2.6.24?

Gregor

2008-01-07 09:46:42

by Mauro Carvalho Chehab

[permalink] [raw]
Subject: Re: [PATCH] vivi driver works only as first device

On Sun, 6 Jan 2008 23:47:19 +0100
"Gregor Jasny" <[email protected]> wrote:

> On Dec 10, 2007 8:08 AM, Mauro Carvalho Chehab <[email protected]> wrote:
> > Thanks for the report. Instead of applying your patch, I decided to
> > better analyze the issue, fixing it with the proper solution. The issue
> > is that vivi_register changes iminor, but this change were not properly
> > returned to the driver.
>
> Any chance to push this into 2.6.24?

I've scheduled this patch for inclusion at 2.6.25-rc.

Cheers,
Mauro