From: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
To: Ricardo Ribalda <ribalda@chromium.org>
Cc: Hans de Goede <hansg@kernel.org>,
Mauro Carvalho Chehab <mchehab@kernel.org>,
linux-media@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] media: uvcvideo: Create a specific id namespace for output entities
Date: Wed, 22 Oct 2025 15:08:49 +0300 [thread overview]
Message-ID: <20251022120849.GD727@pendragon.ideasonboard.com> (raw)
In-Reply-To: <20251022-uvc-grandstream-laurent-v1-1-0925738a3484@chromium.org>
On Wed, Oct 22, 2025 at 11:55:16AM +0000, Ricardo Ribalda wrote:
> Nothing can be connected from an output entity. Which means that no
s/output entity/output terminal. Same below.
Did you mean s/from an/to an/ ?
> other entity can reference an output entity as baSourceId.
>
Some output terminals have controls, so we need to preserve their ID.
That's why my proposal only set the UVC_TERM_OUTPUT bit for the
*streaming* output terminals, not for all output terminals.
> Use this fact to move all the output entities to a different namespace
> id.
>
> The output entities are usually named after the dev_name() of the usb
> device, so there should not be any uAPI change from this change.
>
> Although with this change we can handle some id collisions
> automagically, change the logic of uvc_alloc_new_entity() to keep
> showing a warning when a camera has invalid descriptors. Hopefully this
> message will help vendors fix their invalid descriptors.
>
> Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
> ---
> Hi, this patch fixes support for some devices with invalid USB
> descriptor.
>
> It is orthogonal to:
> https://lore.kernel.org/linux-media/20251021184213.GC19043@pendragon.ideasonboard.com/T/#t
>
> Some devices will be fixed by the other patch, other devices will be
> fixed by this. In my opinion is worth to land both patches.
>
> Tested with GRANDSTREAM GUV3100 in a 6.6 kernel.
> ---
> drivers/media/usb/uvc/uvc_driver.c | 23 +++++++++++++++++++----
> 1 file changed, 19 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/media/usb/uvc/uvc_driver.c b/drivers/media/usb/uvc/uvc_driver.c
> index fb6afb8e84f00961f86fd8f840fba48d706d7a9a..40f8ae0df89e104992f5d55af3d3539dea3d146e 100644
> --- a/drivers/media/usb/uvc/uvc_driver.c
> +++ b/drivers/media/usb/uvc/uvc_driver.c
> @@ -165,10 +165,14 @@ static struct uvc_entity *uvc_entity_by_reference(struct uvc_device *dev,
> return NULL;
> }
>
> +#define ENTITY_HARDWARE_ID(id) ((id) & ~UVC_TERM_OUTPUT)
This needs a UVC_ prefix, and should probably go to uvcvideo.h. You can
also & 0xff, as the UVC descriptors store IDs in 8-bit fields.
> +
> static struct uvc_streaming *uvc_stream_by_id(struct uvc_device *dev, int id)
> {
> struct uvc_streaming *stream;
>
> + id = ENTITY_HARDWARE_ID(id);
> +
> list_for_each_entry(stream, &dev->streams, list) {
> if (stream->header.bTerminalLink == id)
> return stream;
> @@ -810,10 +814,12 @@ static struct uvc_entity *uvc_alloc_new_entity(struct uvc_device *dev, u16 type,
> }
>
> /* Per UVC 1.1+ spec 3.7.2, the ID is unique. */
> - if (uvc_entity_by_id(dev, id)) {
> - dev_err(&dev->intf->dev, "Found multiple Units with ID %u\n", id);
> + if (uvc_entity_by_id(dev, ENTITY_HARDWARE_ID(id)))
> + dev_err(&dev->intf->dev, "Found multiple Units with ID %u\n",
> + ENTITY_HARDWARE_ID(id));
It's not an error anymore if there's no collision of the full 16-bit ID,
right ? Should it be demoted to a dev_warn() ?
> +
> + if (uvc_entity_by_id(dev, id))
> id = UVC_INVALID_ENTITY_ID;
> - }
>
> extra_size = roundup(extra_size, sizeof(*entity->pads));
> if (num_pads)
> @@ -969,6 +975,7 @@ static int uvc_parse_standard_control(struct uvc_device *dev,
> struct usb_host_interface *alts = dev->intf->cur_altsetting;
> unsigned int i, n, p, len;
> const char *type_name;
> + unsigned int id;
> u16 type;
>
> switch (buffer[2]) {
> @@ -1107,8 +1114,16 @@ static int uvc_parse_standard_control(struct uvc_device *dev,
> return 0;
> }
>
> + /*
> + * Nothing can be connected from an output terminal. To avoid
> + * entity-id's collisions in devices with invalid USB
> + * descriptors, move the output terminal id to its own
> + * namespace.
> + */
> + id = buffer[3] | UVC_TERM_OUTPUT;
> +
> term = uvc_alloc_new_entity(dev, type | UVC_TERM_OUTPUT,
> - buffer[3], 1, 0);
> + id, 1, 0);
> if (IS_ERR(term))
> return PTR_ERR(term);
>
>
> ---
> base-commit: ea299a2164262ff787c9d33f46049acccd120672
> change-id: 20251022-uvc-grandstream-laurent-3f9abb8a0d5b
--
Regards,
Laurent Pinchart
next prev parent reply other threads:[~2025-10-22 12:09 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-10-22 11:55 [PATCH] media: uvcvideo: Create a specific id namespace for output entities Ricardo Ribalda
2025-10-22 12:08 ` Laurent Pinchart [this message]
2025-10-22 12:29 ` Ricardo Ribalda
2025-10-22 12:49 ` Laurent Pinchart
2025-10-22 13:08 ` Ricardo Ribalda
2025-10-22 13:12 ` Laurent Pinchart
2025-10-22 13:14 ` Ricardo Ribalda
2025-10-23 11:24 ` Laurent Pinchart
2025-10-23 11:47 ` Ricardo Ribalda
2025-10-24 10:54 ` Laurent Pinchart
2025-10-24 11:00 ` Ricardo Ribalda
2025-10-27 2:46 ` Laurent Pinchart
2025-10-27 2:47 ` Laurent Pinchart
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=20251022120849.GD727@pendragon.ideasonboard.com \
--to=laurent.pinchart@ideasonboard.com \
--cc=hansg@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-media@vger.kernel.org \
--cc=mchehab@kernel.org \
--cc=ribalda@chromium.org \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox