* [PATCH] media: uvcvideo: add capture quirks for 1e4e:7102
@ 2026-09-17 21:29 Henry Paradiz
2026-09-18 9:24 ` Ricardo Ribalda
0 siblings, 1 reply; 2+ messages in thread
From: Henry Paradiz @ 2026-09-17 21:29 UTC (permalink / raw)
To: Laurent Pinchart, Hans de Goede; +Cc: linux-media, linux-kernel
The eEver Live Streaming USB Device HDMI capture card (1e4e:7102,
bcdDevice 1.00) exposes two problems when HDMI is connected but the
input signal does not lock.
While idle, the device disconnects shortly after entering USB runtime
suspend and repeatedly re-enumerates. Keeping it runtime-active stops
this loop. Enable the existing UVC_QUIRK_DISABLE_AUTOSUSPEND for this
USB ID to avoid the repeated USB, UVC and USB-audio discovery messages.
Normal USB disconnect handling is preserved.
On each enumeration, VIDIOC_TRY_FMT takes about 920 ms per call because
uvcvideo negotiates with the device through UVC probe control transfers.
PipeWire queries nine MJPEG frame sizes for colorimetry on its main loop,
blocking unrelated audio mute and volume requests for roughly eight
seconds while audio playback continues on a separate thread.
Add UVC_QUIRK_SKIP_TRY_FMT_PROBE to answer this device's TRY_FMT calls
using the parsed format and frame descriptors. The advertised frame
buffer sizes match the observed probe results for all nine resolutions.
S_FMT continues to negotiate with the hardware, and other devices retain
their existing TRY_FMT behavior.
Both quirks are enabled by one device-specific entry. The card can still
provide diagnostic frames without an HDMI signal; HDMI signal recovery
and capture configuration continue through the normal driver paths.
Signed-off-by: Henry Paradiz <henry.paradiz@gmail.com>
---
drivers/media/usb/uvc/uvc_driver.c | 4 ++++
drivers/media/usb/uvc/uvc_v4l2.c | 19 ++++++++++++-------
drivers/media/usb/uvc/uvcvideo.h | 1 +
3 files changed, 17 insertions(+), 7 deletions(-)
diff --git a/drivers/media/usb/uvc/uvc_driver.c b/drivers/media/usb/uvc/uvc_driver.c
--- a/drivers/media/usb/uvc/uvc_driver.c
+++ b/drivers/media/usb/uvc/uvc_driver.c
@@ -3055,6 +3055,10 @@
.bInterfaceSubClass = 1,
.bInterfaceProtocol = 0,
.driver_info = (kernel_ulong_t)&uvc_quirk_probe_def },
+ /* eEver Live Streaming USB Device HDMI capture */
+ { USB_DEVICE_AND_INTERFACE_INFO(0x1e4e, 0x7102, USB_CLASS_VIDEO, 1, 0),
+ .driver_info = UVC_INFO_QUIRK(UVC_QUIRK_SKIP_TRY_FMT_PROBE
+ | UVC_QUIRK_DISABLE_AUTOSUSPEND) },
/* The Imaging Source USB CCD cameras */
{ .match_flags = USB_DEVICE_ID_MATCH_DEVICE
| USB_DEVICE_ID_MATCH_INT_INFO,
diff --git a/drivers/media/usb/uvc/uvc_v4l2.c b/drivers/media/usb/uvc/uvc_v4l2.c
--- a/drivers/media/usb/uvc/uvc_v4l2.c
+++ b/drivers/media/usb/uvc/uvc_v4l2.c
@@ -241,7 +241,7 @@
static int uvc_v4l2_try_format(struct uvc_streaming *stream,
struct v4l2_format *fmt, struct uvc_streaming_control *probe,
const struct uvc_format **uvc_format,
- const struct uvc_frame **uvc_frame)
+ const struct uvc_frame **uvc_frame, bool probe_device)
{
const struct uvc_format *format = NULL;
const struct uvc_frame *frame = NULL;
@@ -336,10 +336,14 @@
probe->dwMaxVideoFrameSize =
stream->ctrl.dwMaxVideoFrameSize;
- /* Probe the device. */
- ret = uvc_probe_video(stream, probe);
- if (ret < 0)
- return ret;
+ /* Some devices stall on probes during idle format enumeration. */
+ if (probe_device) {
+ ret = uvc_probe_video(stream, probe);
+ if (ret < 0)
+ return ret;
+ } else {
+ probe->dwMaxVideoFrameSize = frame->dwMaxVideoFrameBufferSize;
+ }
/*
* After the probe, update fmt with the values returned from
@@ -432,7 +436,7 @@
if (fmt->type != stream->type)
return -EINVAL;
- ret = uvc_v4l2_try_format(stream, fmt, &probe, &format, &frame);
+ ret = uvc_v4l2_try_format(stream, fmt, &probe, &format, &frame, true);
if (ret < 0)
return ret;
@@ -649,7 +653,8 @@
struct uvc_streaming *stream = handle->stream;
struct uvc_streaming_control probe;
- return uvc_v4l2_try_format(stream, fmt, &probe, NULL, NULL);
+ return uvc_v4l2_try_format(stream, fmt, &probe, NULL, NULL,
+ !(stream->dev->quirks & UVC_QUIRK_SKIP_TRY_FMT_PROBE));
}
static int uvc_ioctl_enum_input(struct file *file, void *priv,
diff --git a/drivers/media/usb/uvc/uvcvideo.h b/drivers/media/usb/uvc/uvcvideo.h
--- a/drivers/media/usb/uvc/uvcvideo.h
+++ b/drivers/media/usb/uvc/uvcvideo.h
@@ -81,6 +81,7 @@
#define UVC_QUIRK_INVALID_DEVICE_SOF 0x00010000
#define UVC_QUIRK_MJPEG_NO_EOF 0x00020000
#define UVC_QUIRK_MSXU_META 0x00040000
+#define UVC_QUIRK_SKIP_TRY_FMT_PROBE 0x00080000
/* Format flags */
#define UVC_FMT_FLAG_COMPRESSED 0x00000001
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [PATCH] media: uvcvideo: add capture quirks for 1e4e:7102
2026-09-17 21:29 [PATCH] media: uvcvideo: add capture quirks for 1e4e:7102 Henry Paradiz
@ 2026-09-18 9:24 ` Ricardo Ribalda
0 siblings, 0 replies; 2+ messages in thread
From: Ricardo Ribalda @ 2026-09-18 9:24 UTC (permalink / raw)
To: Henry Paradiz; +Cc: Laurent Pinchart, Hans de Goede, linux-media, linux-kernel
Hi Henry
On Fri, 18 Sept 2026 at 02:35, Henry Paradiz <henry.paradiz@gmail.com> wrote:
>
> The eEver Live Streaming USB Device HDMI capture card (1e4e:7102,
> bcdDevice 1.00) exposes two problems when HDMI is connected but the
> input signal does not lock.
Could you please share the output of lsusb -v -d 1e4e:7102
>
> While idle, the device disconnects shortly after entering USB runtime
> suspend and repeatedly re-enumerates. Keeping it runtime-active stops
> this loop. Enable the existing UVC_QUIRK_DISABLE_AUTOSUSPEND for this
> USB ID to avoid the repeated USB, UVC and USB-audio discovery messages.
> Normal USB disconnect handling is preserved.
>
> On each enumeration, VIDIOC_TRY_FMT takes about 920 ms per call because
> uvcvideo negotiates with the device through UVC probe control transfers.
> PipeWire queries nine MJPEG frame sizes for colorimetry on its main loop,
> blocking unrelated audio mute and volume requests for roughly eight
> seconds while audio playback continues on a separate thread.
Isn't this a userspace or device issue?
If VIDIOC_TRY_FMT stalls the device, shouldn't you simply not call
VIDIOC_TRY_FMT?
This looks more like a hack than a quirk.
>
> Add UVC_QUIRK_SKIP_TRY_FMT_PROBE to answer this device's TRY_FMT calls
> using the parsed format and frame descriptors. The advertised frame
> buffer sizes match the observed probe results for all nine resolutions.
> S_FMT continues to negotiate with the hardware, and other devices retain
> their existing TRY_FMT behavior.
>
> Both quirks are enabled by one device-specific entry. The card can still
> provide diagnostic frames without an HDMI signal; HDMI signal recovery
> and capture configuration continue through the normal driver paths.
>
> Signed-off-by: Henry Paradiz <henry.paradiz@gmail.com>
> ---
> drivers/media/usb/uvc/uvc_driver.c | 4 ++++
> drivers/media/usb/uvc/uvc_v4l2.c | 19 ++++++++++++-------
> drivers/media/usb/uvc/uvcvideo.h | 1 +
> 3 files changed, 17 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/media/usb/uvc/uvc_driver.c b/drivers/media/usb/uvc/uvc_driver.c
> --- a/drivers/media/usb/uvc/uvc_driver.c
> +++ b/drivers/media/usb/uvc/uvc_driver.c
> @@ -3055,6 +3055,10 @@
> .bInterfaceSubClass = 1,
> .bInterfaceProtocol = 0,
> .driver_info = (kernel_ulong_t)&uvc_quirk_probe_def },
> + /* eEver Live Streaming USB Device HDMI capture */
> + { USB_DEVICE_AND_INTERFACE_INFO(0x1e4e, 0x7102, USB_CLASS_VIDEO, 1, 0),
Please keep the list sorted by vid:pid
> + .driver_info = UVC_INFO_QUIRK(UVC_QUIRK_SKIP_TRY_FMT_PROBE
> + | UVC_QUIRK_DISABLE_AUTOSUSPEND) },
> /* The Imaging Source USB CCD cameras */
> { .match_flags = USB_DEVICE_ID_MATCH_DEVICE
> | USB_DEVICE_ID_MATCH_INT_INFO,
> diff --git a/drivers/media/usb/uvc/uvc_v4l2.c b/drivers/media/usb/uvc/uvc_v4l2.c
> --- a/drivers/media/usb/uvc/uvc_v4l2.c
> +++ b/drivers/media/usb/uvc/uvc_v4l2.c
> @@ -241,7 +241,7 @@
> static int uvc_v4l2_try_format(struct uvc_streaming *stream,
> struct v4l2_format *fmt, struct uvc_streaming_control *probe,
> const struct uvc_format **uvc_format,
> - const struct uvc_frame **uvc_frame)
> + const struct uvc_frame **uvc_frame, bool probe_device)
> {
> const struct uvc_format *format = NULL;
> const struct uvc_frame *frame = NULL;
> @@ -336,10 +336,14 @@
> probe->dwMaxVideoFrameSize =
> stream->ctrl.dwMaxVideoFrameSize;
>
> - /* Probe the device. */
> - ret = uvc_probe_video(stream, probe);
> - if (ret < 0)
> - return ret;
> + /* Some devices stall on probes during idle format enumeration. */
> + if (probe_device) {
> + ret = uvc_probe_video(stream, probe);
> + if (ret < 0)
> + return ret;
> + } else {
> + probe->dwMaxVideoFrameSize = frame->dwMaxVideoFrameBufferSize;
> + }
>
> /*
> * After the probe, update fmt with the values returned from
> @@ -432,7 +436,7 @@
> if (fmt->type != stream->type)
> return -EINVAL;
>
> - ret = uvc_v4l2_try_format(stream, fmt, &probe, &format, &frame);
> + ret = uvc_v4l2_try_format(stream, fmt, &probe, &format, &frame, true);
> if (ret < 0)
> return ret;
>
> @@ -649,7 +653,8 @@
> struct uvc_streaming *stream = handle->stream;
> struct uvc_streaming_control probe;
>
> - return uvc_v4l2_try_format(stream, fmt, &probe, NULL, NULL);
> + return uvc_v4l2_try_format(stream, fmt, &probe, NULL, NULL,
> + !(stream->dev->quirks & UVC_QUIRK_SKIP_TRY_FMT_PROBE));
> }
>
> static int uvc_ioctl_enum_input(struct file *file, void *priv,
> diff --git a/drivers/media/usb/uvc/uvcvideo.h b/drivers/media/usb/uvc/uvcvideo.h
> --- a/drivers/media/usb/uvc/uvcvideo.h
> +++ b/drivers/media/usb/uvc/uvcvideo.h
> @@ -81,6 +81,7 @@
> #define UVC_QUIRK_INVALID_DEVICE_SOF 0x00010000
> #define UVC_QUIRK_MJPEG_NO_EOF 0x00020000
> #define UVC_QUIRK_MSXU_META 0x00040000
> +#define UVC_QUIRK_SKIP_TRY_FMT_PROBE 0x00080000
>
> /* Format flags */
> #define UVC_FMT_FLAG_COMPRESSED 0x00000001
>
--
Ricardo Ribalda
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2026-09-18 9:24 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-09-17 21:29 [PATCH] media: uvcvideo: add capture quirks for 1e4e:7102 Henry Paradiz
2026-09-18 9:24 ` Ricardo Ribalda
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®