[PATCH 0/1] media: uvcvideo: Support devices that report an OT as an entity source

Hi Mauro, et al.,

Fedora has been receiving quite a lot of bug reports (about 1 per day)
for a while now about the WARN_ON recently added by the:
"media: mc-entity.c: use WARN_ON, validate link pads" commit
triggering with uvcvideo (causing backtraces in dmesg).

I've been acting as a go-between for the Fedora bug reporters and
Laurent.

Laurent has written the fix which I'm posting here; I've provided
a Fedora kernel to the users with the fix added and they have confirmed
that this fixes the WARN_ON triggering.

Since Laurent is a bit short on time ATM I'm submitting his patch
upstream now that it has been tested.

Regards,

Hans


Laurent Pinchart (1):
  media: uvcvideo: Support devices that report an OT as an entity source

 drivers/media/usb/uvc/uvc_driver.c | 32 ++++++++++++++++++++++++++++++
 1 file changed, 32 insertions(+)

-- 
2.30.1

From: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

Some devices reference an output terminal as the source of extension
units. This is incorrect, as output terminals only have an input pin,
and thus can't be connected to any entity in the forward direction. The
resulting topology would cause issues when registering the media
controller graph. To avoid this problem, connect the extension unit to
the source of the output terminal instead.

While at it, and while no device has been reported to be affected by
this issue, also handle forward scans where two output terminals would
be connected together, and skip the terminals found through such an
invalid connection.

Reported-and-tested-by: John Nealy <redacted>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans de Goede <redacted>
---
 drivers/media/usb/uvc/uvc_driver.c | 32 ++++++++++++++++++++++++++++++
 1 file changed, 32 insertions(+)

diff --git a/drivers/media/usb/uvc/uvc_driver.c b/drivers/media/usb/uvc/uvc_driver.c
index 30ef2a3110f7..8df58f04dac6 100644
--- a/drivers/media/usb/uvc/uvc_driver.c
+++ b/drivers/media/usb/uvc/uvc_driver.c

@@ -1716,6 +1716,31 @@ static int uvc_scan_chain_forward(struct uvc_video_chain *chain,
 				return -EINVAL;
 			}
 
+			/*
+			 * Some devices reference an output terminal as the
+			 * source of extension units. This is incorrect, as
+			 * output terminals only have an input pin, and thus
+			 * can't be connected to any entity in the forward
+			 * direction. The resulting topology would cause issues
+			 * when registering the media controller graph. To
+			 * avoid this problem, connect the extension unit to
+			 * the source of the output terminal instead.
+			 */
+			if (UVC_ENTITY_IS_OTERM(entity)) {
+				struct uvc_entity *source;
+
+				source = uvc_entity_by_id(chain->dev,
+							  entity->baSourceID[0]);
+				if (!source) {
+					uvc_dbg(chain->dev, DESCR,
+						"Can't connect extension unit %u in chain\n",
+						forward->id);
+					break;
+				}
+
+				forward->baSourceID[0] = source->id;
+			}
+
 			list_add_tail(&forward->chain, &chain->entities);
 			if (!found)
 				uvc_dbg_cont(PROBE, " (->");

@@ -1735,6 +1760,13 @@ static int uvc_scan_chain_forward(struct uvc_video_chain *chain,
 				return -EINVAL;
 			}
 
+			if (UVC_ENTITY_IS_OTERM(entity)) {
+				uvc_dbg(chain->dev, DESCR,
+					"Unsupported connection between output terminals %u and %u\n",
+					entity->id, forward->id);
+				break;
+			}
+
 			list_add_tail(&forward->chain, &chain->entities);
 			if (!found)
 				uvc_dbg_cont(PROBE, " (->");

-- 
2.30.1

Hi All,

On 3/8/21 11:31 AM, Hans de Goede wrote:

From: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

Some devices reference an output terminal as the source of extension
units. This is incorrect, as output terminals only have an input pin,
and thus can't be connected to any entity in the forward direction. The
resulting topology would cause issues when registering the media
controller graph. To avoid this problem, connect the extension unit to
the source of the output terminal instead.

While at it, and while no device has been reported to be affected by
this issue, also handle forward scans where two output terminals would
be connected together, and skip the terminals found through such an
invalid connection.

Reported-and-tested-by: John Nealy <redacted>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans de Goede <redacted>

Ping? This is a bug-fix which fixes a WARN triggering, causing many
users to see a backtrace in their kernel-logs and reporting bugs about this:

https://bugzilla.redhat.com/buglist.cgi?quicksearch=mc-entity.c

Currently shows 12 open bugs for this and this is not counting all the
ones which have already been triaged and matched as dups.

As such it would be really good if we can get this merged and on
its way to 5.12-rc# as a fix for 5.12 (and to be backported to the
stable series).

Regards,

Hans

Hi Hans,

On Wed, Mar 31, 2021 at 12:38:07PM +0200, Hans de Goede wrote:

On 3/8/21 11:31 AM, Hans de Goede wrote:

quoted

From: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

Some devices reference an output terminal as the source of extension
units. This is incorrect, as output terminals only have an input pin,
and thus can't be connected to any entity in the forward direction. The
resulting topology would cause issues when registering the media
controller graph. To avoid this problem, connect the extension unit to
the source of the output terminal instead.

While at it, and while no device has been reported to be affected by
this issue, also handle forward scans where two output terminals would
be connected together, and skip the terminals found through such an
invalid connection.

Reported-and-tested-by: John Nealy <redacted>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans de Goede <redacted>

Ping? This is a bug-fix which fixes a WARN triggering, causing many
users to see a backtrace in their kernel-logs and reporting bugs about this:

https://bugzilla.redhat.com/buglist.cgi?quicksearch=mc-entity.c

Currently shows 12 open bugs for this and this is not counting all the
ones which have already been triaged and matched as dups.

As such it would be really good if we can get this merged and on
its way to 5.12-rc# as a fix for 5.12 (and to be backported to the
stable series).

The patch is included in "[GIT PULL FOR v5.13] Miscellaneous changes". I
have no personal issue with it being merged in v5.12-rc, but technically
it's not a regression fix, is it ? I'll let Mauro decide what works best
for him.

quoted

---
 drivers/media/usb/uvc/uvc_driver.c | 32 ++++++++++++++++++++++++++++++
 1 file changed, 32 insertions(+)

diff --git a/drivers/media/usb/uvc/uvc_driver.c b/drivers/media/usb/uvc/uvc_driver.c
index 30ef2a3110f7..8df58f04dac6 100644
--- a/drivers/media/usb/uvc/uvc_driver.c
+++ b/drivers/media/usb/uvc/uvc_driver.c

@@ -1716,6 +1716,31 @@ static int uvc_scan_chain_forward(struct uvc_video_chain *chain,
 				return -EINVAL;
 			}
 
+			/*
+			 * Some devices reference an output terminal as the
+			 * source of extension units. This is incorrect, as
+			 * output terminals only have an input pin, and thus
+			 * can't be connected to any entity in the forward
+			 * direction. The resulting topology would cause issues
+			 * when registering the media controller graph. To
+			 * avoid this problem, connect the extension unit to
+			 * the source of the output terminal instead.
+			 */
+			if (UVC_ENTITY_IS_OTERM(entity)) {
+				struct uvc_entity *source;
+
+				source = uvc_entity_by_id(chain->dev,
+							  entity->baSourceID[0]);
+				if (!source) {
+					uvc_dbg(chain->dev, DESCR,
+						"Can't connect extension unit %u in chain\n",
+						forward->id);
+					break;
+				}
+
+				forward->baSourceID[0] = source->id;
+			}
+
 			list_add_tail(&forward->chain, &chain->entities);
 			if (!found)
 				uvc_dbg_cont(PROBE, " (->");

@@ -1735,6 +1760,13 @@ static int uvc_scan_chain_forward(struct uvc_video_chain *chain,
 				return -EINVAL;
 			}
 
+			if (UVC_ENTITY_IS_OTERM(entity)) {
+				uvc_dbg(chain->dev, DESCR,
+					"Unsupported connection between output terminals %u and %u\n",
+					entity->id, forward->id);
+				break;
+			}
+
 			list_add_tail(&forward->chain, &chain->entities);
 			if (!found)
 				uvc_dbg_cont(PROBE, " (->");

-- 
Regards,

Laurent Pinchart

Hi,

On 3/31/21 1:03 PM, Laurent Pinchart wrote:

Hi Hans,

On Wed, Mar 31, 2021 at 12:38:07PM +0200, Hans de Goede wrote:

quoted

On 3/8/21 11:31 AM, Hans de Goede wrote:

quoted

From: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

Some devices reference an output terminal as the source of extension
units. This is incorrect, as output terminals only have an input pin,
and thus can't be connected to any entity in the forward direction. The
resulting topology would cause issues when registering the media
controller graph. To avoid this problem, connect the extension unit to
the source of the output terminal instead.

While at it, and while no device has been reported to be affected by
this issue, also handle forward scans where two output terminals would
be connected together, and skip the terminals found through such an
invalid connection.

Reported-and-tested-by: John Nealy <redacted>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans de Goede <redacted>

Ping? This is a bug-fix which fixes a WARN triggering, causing many
users to see a backtrace in their kernel-logs and reporting bugs about this:

https://bugzilla.redhat.com/buglist.cgi?quicksearch=mc-entity.c

Currently shows 12 open bugs for this and this is not counting all the
ones which have already been triaged and matched as dups.

As such it would be really good if we can get this merged and on
its way to 5.12-rc# as a fix for 5.12 (and to be backported to the
stable series).

The patch is included in "[GIT PULL FOR v5.13] Miscellaneous changes".

Ah I missed that.

I have no personal issue with it being merged in v5.12-rc, but technically
it's not a regression fix, is it ? I'll let Mauro decide what works best
for him.

It is true that this is not a regression-fix but it is a bug-fix and for
a bug which users are actively hitting.

Regards,

Hans

quoted

quoted

---
 drivers/media/usb/uvc/uvc_driver.c | 32 ++++++++++++++++++++++++++++++
 1 file changed, 32 insertions(+)

diff --git a/drivers/media/usb/uvc/uvc_driver.c b/drivers/media/usb/uvc/uvc_driver.c
index 30ef2a3110f7..8df58f04dac6 100644
--- a/drivers/media/usb/uvc/uvc_driver.c
+++ b/drivers/media/usb/uvc/uvc_driver.c

@@ -1716,6 +1716,31 @@ static int uvc_scan_chain_forward(struct uvc_video_chain *chain,
 				return -EINVAL;
 			}
 
+			/*
+			 * Some devices reference an output terminal as the
+			 * source of extension units. This is incorrect, as
+			 * output terminals only have an input pin, and thus
+			 * can't be connected to any entity in the forward
+			 * direction. The resulting topology would cause issues
+			 * when registering the media controller graph. To
+			 * avoid this problem, connect the extension unit to
+			 * the source of the output terminal instead.
+			 */
+			if (UVC_ENTITY_IS_OTERM(entity)) {
+				struct uvc_entity *source;
+
+				source = uvc_entity_by_id(chain->dev,
+							  entity->baSourceID[0]);
+				if (!source) {
+					uvc_dbg(chain->dev, DESCR,
+						"Can't connect extension unit %u in chain\n",
+						forward->id);
+					break;
+				}
+
+				forward->baSourceID[0] = source->id;
+			}
+
 			list_add_tail(&forward->chain, &chain->entities);
 			if (!found)
 				uvc_dbg_cont(PROBE, " (->");

@@ -1735,6 +1760,13 @@ static int uvc_scan_chain_forward(struct uvc_video_chain *chain,
 				return -EINVAL;
 			}
 
+			if (UVC_ENTITY_IS_OTERM(entity)) {
+				uvc_dbg(chain->dev, DESCR,
+					"Unsupported connection between output terminals %u and %u\n",
+					entity->id, forward->id);
+				break;
+			}
+
 			list_add_tail(&forward->chain, &chain->entities);
 			if (!found)
 				uvc_dbg_cont(PROBE, " (->");