[PATCH] media: uvcvideo: Fix deferred probing error

[PATCH] media: uvcvideo: Fix deferred probing error

[Ricardo Ribalda]

uvc_gpio_parse() can return -EPROBE_DEFER when the GPIOs it depends on
have not yet been probed. This return code should be propagated to the
caller of uvc_probe() to ensure that probing is retried when the required
GPIOs become available.

Currently, this error code is incorrectly converted to -ENODEV,
causing some internal cameras to be ignored.

This commit fixes this issue by propagating the -EPROBE_DEFER error.

Cc: stable@vger.kernel.org
Fixes: 2886477ff987 ("media: uvcvideo: Implement UVC_EXT_GPIO_UNIT")
Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
---
 drivers/media/usb/uvc/uvc_driver.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/drivers/media/usb/uvc/uvc_driver.c b/drivers/media/usb/uvc/uvc_driver.c
index a10d4f4d9f95..73a7f23b616c 100644
--- a/drivers/media/usb/uvc/uvc_driver.c
+++ b/drivers/media/usb/uvc/uvc_driver.c
@@ -2253,9 +2253,10 @@ static int uvc_probe(struct usb_interface *intf,
 	}
 
 	/* Parse the associated GPIOs. */
-	if (uvc_gpio_parse(dev) < 0) {
+	ret = uvc_gpio_parse(dev);
+	if (ret < 0) {
 		uvc_dbg(dev, PROBE, "Unable to parse UVC GPIOs\n");
-		goto error;
+		goto error_retcode;
 	}
 
 	dev_info(&dev->udev->dev, "Found UVC %u.%02x device %s (%04x:%04x)\n",
@@ -2328,9 +2329,11 @@ static int uvc_probe(struct usb_interface *intf,
 	return 0;
 
 error:
+	ret = -ENODEV;
+error_retcode:
 	uvc_unregister_video(dev);
 	kref_put(&dev->ref, uvc_delete);
-	return -ENODEV;
+	return ret;
 }
 
 static void uvc_disconnect(struct usb_interface *intf)

---
base-commit: c4b7779abc6633677e6edb79e2809f4f61fde157
change-id: 20250129-uvc-eprobedefer-b5ebb4db63cc

Best regards,
-- 
Ricardo Ribalda <ribalda@chromium.org>

Re: [PATCH] media: uvcvideo: Fix deferred probing error

[Doug Anderson]

Hi,

On Wed, Jan 29, 2025 at 4:40 AM Ricardo Ribalda <ribalda@chromium.org> wrote:
>
> uvc_gpio_parse() can return -EPROBE_DEFER when the GPIOs it depends on
> have not yet been probed. This return code should be propagated to the
> caller of uvc_probe() to ensure that probing is retried when the required
> GPIOs become available.
>
> Currently, this error code is incorrectly converted to -ENODEV,
> causing some internal cameras to be ignored.
>
> This commit fixes this issue by propagating the -EPROBE_DEFER error.
>
> Cc: stable@vger.kernel.org
> Fixes: 2886477ff987 ("media: uvcvideo: Implement UVC_EXT_GPIO_UNIT")
> Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
> ---
>  drivers/media/usb/uvc/uvc_driver.c | 9 ++++++---
>  1 file changed, 6 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/media/usb/uvc/uvc_driver.c b/drivers/media/usb/uvc/uvc_driver.c
> index a10d4f4d9f95..73a7f23b616c 100644
> --- a/drivers/media/usb/uvc/uvc_driver.c
> +++ b/drivers/media/usb/uvc/uvc_driver.c
> @@ -2253,9 +2253,10 @@ static int uvc_probe(struct usb_interface *intf,
>         }
>
>         /* Parse the associated GPIOs. */
> -       if (uvc_gpio_parse(dev) < 0) {
> +       ret = uvc_gpio_parse(dev);
> +       if (ret < 0) {
>                 uvc_dbg(dev, PROBE, "Unable to parse UVC GPIOs\n");
> -               goto error;
> +               goto error_retcode;

FWIW, since you're specifically considering the -EPROBE_DEFER case,
it's probably worthwhile to make sure that dev_err_probe() is called.
That eventually calls device_set_deferred_probe_reason() which can be
helpful for tracking down problems.

It looks like uvc_gpio_parse() already calls this if gpiod_to_irq()
returns an error code probably you also want to make sure that
dev_err_probe() also gets called in the case where
devm_gpiod_get_optional() returns -EPROBE_DEFER. In that case,
potentially one could also get rid of the uvc_dbg() above.

In any case, IMO even without those changes your patch is still worth
landing. ...and maybe my suggestion should be in a separate follow-on
patch anyway. Thus:

Reviewed-by: Douglas Anderson <dianders@chromium.org>

Re: [PATCH] media: uvcvideo: Fix deferred probing error

[Laurent Pinchart]

Hi Ricardo,

Thank you for the patch.

On Wed, Jan 29, 2025 at 12:39:46PM +0000, Ricardo Ribalda wrote:
> uvc_gpio_parse() can return -EPROBE_DEFER when the GPIOs it depends on
> have not yet been probed. This return code should be propagated to the
> caller of uvc_probe() to ensure that probing is retried when the required
> GPIOs become available.
> 
> Currently, this error code is incorrectly converted to -ENODEV,
> causing some internal cameras to be ignored.
> 
> This commit fixes this issue by propagating the -EPROBE_DEFER error.
> 
> Cc: stable@vger.kernel.org
> Fixes: 2886477ff987 ("media: uvcvideo: Implement UVC_EXT_GPIO_UNIT")
> Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
> ---
>  drivers/media/usb/uvc/uvc_driver.c | 9 ++++++---
>  1 file changed, 6 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/media/usb/uvc/uvc_driver.c b/drivers/media/usb/uvc/uvc_driver.c
> index a10d4f4d9f95..73a7f23b616c 100644
> --- a/drivers/media/usb/uvc/uvc_driver.c
> +++ b/drivers/media/usb/uvc/uvc_driver.c
> @@ -2253,9 +2253,10 @@ static int uvc_probe(struct usb_interface *intf,
>  	}
>  
>  	/* Parse the associated GPIOs. */
> -	if (uvc_gpio_parse(dev) < 0) {
> +	ret = uvc_gpio_parse(dev);
> +	if (ret < 0) {
>  		uvc_dbg(dev, PROBE, "Unable to parse UVC GPIOs\n");
> -		goto error;
> +		goto error_retcode;
>  	}
>  
>  	dev_info(&dev->udev->dev, "Found UVC %u.%02x device %s (%04x:%04x)\n",
> @@ -2328,9 +2329,11 @@ static int uvc_probe(struct usb_interface *intf,
>  	return 0;
>  
>  error:
> +	ret = -ENODEV;
> +error_retcode:

This isn't very nice. Could we instead also propagate error codes from
other locations in the uvc_probe() function ? If you want to minimize
changes, you can initialize ret to -ENODEV, and turn the (ret < 0) check
for uvc_gpio_parse() to a (ret) check.

>  	uvc_unregister_video(dev);
>  	kref_put(&dev->ref, uvc_delete);
> -	return -ENODEV;
> +	return ret;
>  }
>  
>  static void uvc_disconnect(struct usb_interface *intf)
> 
> ---
> base-commit: c4b7779abc6633677e6edb79e2809f4f61fde157
> change-id: 20250129-uvc-eprobedefer-b5ebb4db63cc

-- 
Regards,

Laurent Pinchart

Re: [PATCH] media: uvcvideo: Fix deferred probing error

[Ricardo Ribalda]

On Sun, 23 Feb 2025 at 15:36, Laurent Pinchart
<laurent.pinchart@ideasonboard.com> wrote:
>
> Hi Ricardo,
>
> Thank you for the patch.
>
> On Wed, Jan 29, 2025 at 12:39:46PM +0000, Ricardo Ribalda wrote:
> > uvc_gpio_parse() can return -EPROBE_DEFER when the GPIOs it depends on
> > have not yet been probed. This return code should be propagated to the
> > caller of uvc_probe() to ensure that probing is retried when the required
> > GPIOs become available.
> >
> > Currently, this error code is incorrectly converted to -ENODEV,
> > causing some internal cameras to be ignored.
> >
> > This commit fixes this issue by propagating the -EPROBE_DEFER error.
> >
> > Cc: stable@vger.kernel.org
> > Fixes: 2886477ff987 ("media: uvcvideo: Implement UVC_EXT_GPIO_UNIT")
> > Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
> > ---
> >  drivers/media/usb/uvc/uvc_driver.c | 9 ++++++---
> >  1 file changed, 6 insertions(+), 3 deletions(-)
> >
> > diff --git a/drivers/media/usb/uvc/uvc_driver.c b/drivers/media/usb/uvc/uvc_driver.c
> > index a10d4f4d9f95..73a7f23b616c 100644
> > --- a/drivers/media/usb/uvc/uvc_driver.c
> > +++ b/drivers/media/usb/uvc/uvc_driver.c
> > @@ -2253,9 +2253,10 @@ static int uvc_probe(struct usb_interface *intf,
> >       }
> >
> >       /* Parse the associated GPIOs. */
> > -     if (uvc_gpio_parse(dev) < 0) {
> > +     ret = uvc_gpio_parse(dev);
> > +     if (ret < 0) {
> >               uvc_dbg(dev, PROBE, "Unable to parse UVC GPIOs\n");
> > -             goto error;
> > +             goto error_retcode;
> >       }
> >
> >       dev_info(&dev->udev->dev, "Found UVC %u.%02x device %s (%04x:%04x)\n",
> > @@ -2328,9 +2329,11 @@ static int uvc_probe(struct usb_interface *intf,
> >       return 0;
> >
> >  error:
> > +     ret = -ENODEV;
> > +error_retcode:
>
> This isn't very nice. Could we instead also propagate error codes from
> other locations in the uvc_probe() function ? If you want to minimize
> changes, you can initialize ret to -ENODEV, and turn the (ret < 0) check
> for uvc_gpio_parse() to a (ret) check.

Not very nice, but easy to backport to stables. What about a follow-up
change like this:

index c93abe2367aa..8c67feca1688 100644
--- a/drivers/media/usb/uvc/uvc_driver.c
+++ b/drivers/media/usb/uvc/uvc_driver.c
@@ -2261,7 +2261,7 @@ static int uvc_probe(struct usb_interface *intf,
        ret = uvc_gpio_parse(dev);
        if (ret < 0) {
                uvc_dbg(dev, PROBE, "Unable to parse UVC GPIOs\n");
-               goto error_retcode;
+               goto error;
        }

        dev_info(&dev->udev->dev, "Found UVC %u.%02x device %s (%04x:%04x)\n",
@@ -2285,24 +2285,32 @@ static int uvc_probe(struct usb_interface *intf,
        }

        /* Register the V4L2 device. */
-       if (v4l2_device_register(&intf->dev, &dev->vdev) < 0)
+       ret = v4l2_device_register(&intf->dev, &dev->vdev);
+       if (ret < 0)
                goto error;

        /* Scan the device for video chains. */
-       if (uvc_scan_device(dev) < 0)
+       if (uvc_scan_device(dev) < 0) {
+               ret = -ENODEV;
                goto error;
+       }

        /* Initialize controls. */
-       if (uvc_ctrl_init_device(dev) < 0)
+       if (uvc_ctrl_init_device(dev) < 0) {
+               ret = -ENODEV;
                goto error;
+       }

        /* Register video device nodes. */
-       if (uvc_register_chains(dev) < 0)
+       if (uvc_register_chains(dev) < 0) {
+               ret = -ENODEV;
                goto error;
+       }

 #ifdef CONFIG_MEDIA_CONTROLLER
        /* Register the media device node */
-       if (media_device_register(&dev->mdev) < 0)
+       ret = media_device_register(&dev->mdev);
+       if (ret < 0)
                goto error;
 #endif
        /* Save our data pointer in the interface data. */
@@ -2334,8 +2342,6 @@ static int uvc_probe(struct usb_interface *intf,
        return 0;

 error:
-       ret = -ENODEV;
-error_retcode:
        uvc_unregister_video(dev);
        kref_put(&dev->ref, uvc_delete);
        return ret;


>
> >       uvc_unregister_video(dev);
> >       kref_put(&dev->ref, uvc_delete);
> > -     return -ENODEV;
> > +     return ret;
> >  }
> >
> >  static void uvc_disconnect(struct usb_interface *intf)
> >
> > ---
> > base-commit: c4b7779abc6633677e6edb79e2809f4f61fde157
> > change-id: 20250129-uvc-eprobedefer-b5ebb4db63cc
>
> --
> Regards,
>
> Laurent Pinchart

Let me know what do you think so I can send a v2 with the change
proposed by Doug.

Regards!

-- 
Ricardo Ribalda

Re: [PATCH] media: uvcvideo: Fix deferred probing error

[Laurent Pinchart]

On Sun, Feb 23, 2025 at 08:32:24PM +0100, Ricardo Ribalda wrote:
> On Sun, 23 Feb 2025 at 15:36, Laurent Pinchart wrote:
> > On Wed, Jan 29, 2025 at 12:39:46PM +0000, Ricardo Ribalda wrote:
> > > uvc_gpio_parse() can return -EPROBE_DEFER when the GPIOs it depends on
> > > have not yet been probed. This return code should be propagated to the
> > > caller of uvc_probe() to ensure that probing is retried when the required
> > > GPIOs become available.
> > >
> > > Currently, this error code is incorrectly converted to -ENODEV,
> > > causing some internal cameras to be ignored.
> > >
> > > This commit fixes this issue by propagating the -EPROBE_DEFER error.
> > >
> > > Cc: stable@vger.kernel.org
> > > Fixes: 2886477ff987 ("media: uvcvideo: Implement UVC_EXT_GPIO_UNIT")
> > > Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
> > > ---
> > >  drivers/media/usb/uvc/uvc_driver.c | 9 ++++++---
> > >  1 file changed, 6 insertions(+), 3 deletions(-)
> > >
> > > diff --git a/drivers/media/usb/uvc/uvc_driver.c b/drivers/media/usb/uvc/uvc_driver.c
> > > index a10d4f4d9f95..73a7f23b616c 100644
> > > --- a/drivers/media/usb/uvc/uvc_driver.c
> > > +++ b/drivers/media/usb/uvc/uvc_driver.c
> > > @@ -2253,9 +2253,10 @@ static int uvc_probe(struct usb_interface *intf,
> > >       }
> > >
> > >       /* Parse the associated GPIOs. */
> > > -     if (uvc_gpio_parse(dev) < 0) {
> > > +     ret = uvc_gpio_parse(dev);
> > > +     if (ret < 0) {
> > >               uvc_dbg(dev, PROBE, "Unable to parse UVC GPIOs\n");
> > > -             goto error;
> > > +             goto error_retcode;
> > >       }
> > >
> > >       dev_info(&dev->udev->dev, "Found UVC %u.%02x device %s (%04x:%04x)\n",
> > > @@ -2328,9 +2329,11 @@ static int uvc_probe(struct usb_interface *intf,
> > >       return 0;
> > >
> > >  error:
> > > +     ret = -ENODEV;
> > > +error_retcode:
> >
> > This isn't very nice. Could we instead also propagate error codes from
> > other locations in the uvc_probe() function ? If you want to minimize
> > changes, you can initialize ret to -ENODEV, and turn the (ret < 0) check
> > for uvc_gpio_parse() to a (ret) check.
> 
> Not very nice, but easy to backport to stables. What about a follow-up
> change like this:

How about the nicer fix for mainline, and a simpler to backport one for
stable ? The main focus should be on mainline, not backports.

> index c93abe2367aa..8c67feca1688 100644
> --- a/drivers/media/usb/uvc/uvc_driver.c
> +++ b/drivers/media/usb/uvc/uvc_driver.c
> @@ -2261,7 +2261,7 @@ static int uvc_probe(struct usb_interface *intf,
>         ret = uvc_gpio_parse(dev);
>         if (ret < 0) {
>                 uvc_dbg(dev, PROBE, "Unable to parse UVC GPIOs\n");
> -               goto error_retcode;
> +               goto error;
>         }
> 
>         dev_info(&dev->udev->dev, "Found UVC %u.%02x device %s (%04x:%04x)\n",
> @@ -2285,24 +2285,32 @@ static int uvc_probe(struct usb_interface *intf,
>         }
> 
>         /* Register the V4L2 device. */
> -       if (v4l2_device_register(&intf->dev, &dev->vdev) < 0)
> +       ret = v4l2_device_register(&intf->dev, &dev->vdev);
> +       if (ret < 0)
>                 goto error;
> 
>         /* Scan the device for video chains. */
> -       if (uvc_scan_device(dev) < 0)
> +       if (uvc_scan_device(dev) < 0) {
> +               ret = -ENODEV;
>                 goto error;
> +       }
> 
>         /* Initialize controls. */
> -       if (uvc_ctrl_init_device(dev) < 0)
> +       if (uvc_ctrl_init_device(dev) < 0) {
> +               ret = -ENODEV;
>                 goto error;
> +       }
> 
>         /* Register video device nodes. */
> -       if (uvc_register_chains(dev) < 0)
> +       if (uvc_register_chains(dev) < 0) {
> +               ret = -ENODEV;
>                 goto error;
> +       }
> 
>  #ifdef CONFIG_MEDIA_CONTROLLER
>         /* Register the media device node */
> -       if (media_device_register(&dev->mdev) < 0)
> +       ret = media_device_register(&dev->mdev);
> +       if (ret < 0)
>                 goto error;
>  #endif
>         /* Save our data pointer in the interface data. */
> @@ -2334,8 +2342,6 @@ static int uvc_probe(struct usb_interface *intf,
>         return 0;
> 
>  error:
> -       ret = -ENODEV;
> -error_retcode:
>         uvc_unregister_video(dev);
>         kref_put(&dev->ref, uvc_delete);
>         return ret;
> 
> > >       uvc_unregister_video(dev);
> > >       kref_put(&dev->ref, uvc_delete);
> > > -     return -ENODEV;
> > > +     return ret;
> > >  }
> > >
> > >  static void uvc_disconnect(struct usb_interface *intf)
> > >
> > > ---
> > > base-commit: c4b7779abc6633677e6edb79e2809f4f61fde157
> > > change-id: 20250129-uvc-eprobedefer-b5ebb4db63cc
> >
> > --
> > Regards,
> >
> > Laurent Pinchart
> 
> Let me know what do you think so I can send a v2 with the change
> proposed by Doug.
> 
> Regards!

-- 
Regards,

Laurent Pinchart