#
2a3b7af1 |
|
13-Feb-2024 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: mark incomplete frames with UVC_STREAM_ERR If an frame was transmitted incomplete to the host, we set the UVC_STREAM_ERR bit in the header for the last request that is going to be queued. This way the host will know that it should drop the frame instead of trying to display the corrupted content. Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20240214-uvc-error-tag-v1-2-37659a3877fe@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
ab9ffa51 |
|
13-Feb-2024 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: dont drop frames if zero length packages are late If the request that was missed was zero bytes long, it is likely that the overall transferred frame was not affected. So don't flag the frame incomplete in that case. Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20240214-uvc-error-tag-v1-1-37659a3877fe@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
58da4e57 |
|
13-Feb-2024 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: rework complete handler We refactor the complete handler since the return path with the locking are really difficult to follow. Just simplify the function by switching the logic return it on an disabled endpoint early. This way the second level of indentation can be removed. Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20240214-uvc-gadget-cleanup-v1-3-de6d78780459@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
5e7ea65d |
|
13-Feb-2024 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: refactor the check for a valid buffer in the pump worker By toggling the condition check for a valid buffer, the else path can be completely avoided. Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20240214-uvc-gadget-cleanup-v1-2-de6d78780459@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
ed6cbac8 |
|
13-Feb-2024 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: drop unnecessary check for always set req The pump function is running in an while(1) loop. The only case this loop will be escaped is the two breaks. In both cases the req is valid. Therefor the check for an not set req can be dropped and setting the req to NULL does also has never any effect. Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20240214-uvc-gadget-cleanup-v1-1-de6d78780459@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
9866dc43 |
|
04-Jan-2024 |
Avichal Rakesh <arakesh@google.com> |
usb: gadget: uvc: Remove nested locking When handling error status from uvcg_video_usb_req_queue, uvc_video_complete currently calls uvcg_queue_cancel with video->req_lock held. uvcg_queue_cancel internally locks queue->irqlock, which nests queue->irqlock inside video->req_lock. This isn't a functional bug at the moment, but does open up possibilities for ABBA deadlocks in the future. This patch fixes the accidental nesting by dropping video->req_lock before calling uvcg_queue_cancel. Fixes: 6acba0345b68 ("usb:gadget:uvc Do not use worker thread to pump isoc usb requests") Signed-off-by: Avichal Rakesh <arakesh@google.com> Link: https://lore.kernel.org/r/20240104215009.2252452-2-arakesh@google.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
fe814b5b |
|
04-Jan-2024 |
Avichal Rakesh <arakesh@google.com> |
usb: gadget: uvc: Fix use are free during STREAMOFF There is a path that may lead to freed memory being referenced, causing kernel panics. The kernel panic has the following stack trace: Workqueue: uvcgadget uvcg_video_pump.c51fb85fece46625450f86adbf92c56c.cfi_jt pstate: 60c00085 (nZCv daIf +PAN +UAO -TCO BTYPE=--) pc : __list_del_entry_valid+0xc0/0xd4 lr : __list_del_entry_valid+0xc0/0xd4 Call trace: __list_del_entry_valid+0xc0/0xd4 uvc_video_free_request+0x60/0x98 uvcg_video_pump+0x1cc/0x204 process_one_work+0x21c/0x4b8 worker_thread+0x29c/0x574 kthread+0x158/0x1b0 ret_from_fork+0x10/0x30 The root cause is that uvcg_video_usb_req_queue frees the uvc_request if is_enabled is false and returns an error status. video_pump also frees the associated request if uvcg_video_usb_req_queue returns an error status, leading to double free and accessing garbage memory. To fix the issue, this patch removes freeing logic from uvcg_video_usb_req_queue, and lets the callers to the function handle queueing errors as they see fit. Fixes: 6acba0345b68 ("usb:gadget:uvc Do not use worker thread to pump isoc usb requests") Tested-by: Avichal Rakesh <arakesh@google.com> Signed-off-by: Avichal Rakesh <arakesh@google.com> Link: https://lore.kernel.org/r/20240104215009.2252452-1-arakesh@google.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
6acba034 |
|
19-Nov-2023 |
Jayant Chowdhary <jchowdhary@google.com> |
usb:gadget:uvc Do not use worker thread to pump isoc usb requests When we use an async work queue to perform the function of pumping usb requests to the usb controller, it is possible that amongst other factors, thread scheduling affects at what cadence we're able to pump requests. This could mean isoc usb requests miss their uframes - resulting in video stream flickers on the host device. To avoid this, we make the async_wq thread only produce isoc usb_requests with uvc buffers encoded into them. The process of queueing to the endpoint is done by the uvc_video_complete() handler. In case no usb_requests are ready with encoded information, we just queue a zero length request to the endpoint from the complete handler. For bulk endpoints the async_wq thread still queues usb requests to the endpoint. Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Signed-off-by: Jayant Chowdhary <jchowdhary@google.com> Suggested-by: Avichal Rakesh <arakesh@google.com> Suggested-by: Alan Stern <stern@rowland.harvard.edu> Link: https://lore.kernel.org/r/20231120062026.3759463-1-jchowdhary@google.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
da324ffc |
|
08-Nov-2023 |
Avichal Rakesh <arakesh@google.com> |
usb: gadget: uvc: Fix use-after-free for inflight usb_requests Currently, the uvc gadget driver allocates all uvc_requests as one array and deallocates them all when the video stream stops. This includes de-allocating all the usb_requests associated with those uvc_requests. This can lead to use-after-free issues if any of those de-allocated usb_requests were still owned by the usb controller. This is patch 2 of 2 in fixing the use-after-free issue. It adds a new flag to uvc_video to track when frames and requests should be flowing. When disabling the video stream, the flag is tripped and, instead of de-allocating all uvc_requests and usb_requests, the gadget driver only de-allocates those usb_requests that are currently owned by it (as present in req_free). Other usb_requests are left untouched until their completion handler is called which takes care of freeing the usb_request and its corresponding uvc_request. Now that uvc_video does not depends on uvc->state, this patch removes unnecessary upates to uvc->state that were made to accommodate uvc_video logic. This should ensure that uvc gadget driver never accidentally de-allocates a usb_request that it doesn't own. Link: https://lore.kernel.org/7cd81649-2795-45b6-8c10-b7df1055020d@google.com Reviewed-by: Daniel Scally <dan.scally@ideasonboard.com> Reviewed-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Suggested-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Tested-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Signed-off-by: Avichal Rakesh <arakesh@google.com> Link: https://lore.kernel.org/r/20231109004104.3467968-4-arakesh@google.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
2079b60b |
|
08-Nov-2023 |
Avichal Rakesh <arakesh@google.com> |
usb: gadget: uvc: move video disable logic to its own function This patch refactors the video disable logic in uvcg_video_enable into its own separate function 'uvcg_video_disable'. This function is now used anywhere uvcg_video_enable(video, 0) was used. Reviewed-by: Daniel Scally <dan.scally@ideasonboard.com> Suggested-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Signed-off-by: Avichal Rakesh <arakesh@google.com> Link: https://lore.kernel.org/r/20231109004104.3467968-3-arakesh@google.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
aeb686a9 |
|
08-Nov-2023 |
Avichal Rakesh <arakesh@google.com> |
usb: gadget: uvc: Allocate uvc_requests one at a time Currently, the uvc gadget driver allocates all uvc_requests as one array and deallocates them all when the video stream stops. This includes de-allocating all the usb_requests associated with those uvc_requests. This can lead to use-after-free issues if any of those de-allocated usb_requests were still owned by the usb controller. This patch is 1 of 2 patches addressing the use-after-free issue. Instead of bulk allocating all uvc_requests as an array, this patch allocates uvc_requests one at a time, which should allows for similar granularity when deallocating the uvc_requests. This patch has no functional changes other than allocating each uvc_request separately, and similarly freeing each of them separately. Link: https://lore.kernel.org/7cd81649-2795-45b6-8c10-b7df1055020d@google.com Reviewed-by: Daniel Scally <dan.scally@ideasonboard.com> Reviewed-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Suggested-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Tested-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Signed-off-by: Avichal Rakesh <arakesh@google.com> Link: https://lore.kernel.org/r/20231109004104.3467968-2-arakesh@google.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
991544dc |
|
08-Nov-2023 |
Avichal Rakesh <arakesh@google.com> |
usb: gadget: uvc: prevent use of disabled endpoint Currently the set_alt callback immediately disables the endpoint and queues the v4l2 streamoff event. However, as the streamoff event is processed asynchronously, it is possible that the video_pump thread attempts to queue requests to an already disabled endpoint. This change moves disabling usb endpoint to the end of streamoff event callback. As the endpoint's state can no longer be used, video_pump is now guarded by uvc->state as well. To be consistent with the actual streaming state, uvc->state is now toggled between CONNECTED and STREAMING from the v4l2 event callback only. Link: https://lore.kernel.org/20230615171558.GK741@pendragon.ideasonboard.com/ Link: https://lore.kernel.org/20230531085544.253363-1-dan.scally@ideasonboard.com/ Reviewed-by: Daniel Scally <dan.scally@ideasonboard.com> Reviewed-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Tested-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Signed-off-by: Avichal Rakesh <arakesh@google.com> Link: https://lore.kernel.org/r/20231109004104.3467968-1-arakesh@google.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
1053c4a4 |
|
05-Oct-2023 |
Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
Revert "usb: gadget: uvc: stop pump thread on video disable" This reverts commit 3a63f86c6a6cb0601f0563a81574745da2979e3b. Based on review comments, it was applied too soon and needs more work. Reported-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Link: https://lore.kernel.org/r/20231005081716.GA13853@pendragon.ideasonboard.com Cc: Michael Grzeschik <m.grzeschik@pengutronix.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
dddc00f2 |
|
05-Oct-2023 |
Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
Revert "usb: gadget: uvc: cleanup request when not in correct state" This reverts commit 52a39f2cf62bb5430ad1f54cd522dbfdab1d71ba. Based on review comments, it was applied too soon and needs more work. Reported-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Link: https://lore.kernel.org/r/20231005081716.GA13853@pendragon.ideasonboard.com Cc: Michael Grzeschik <m.grzeschik@pengutronix.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
e0fa80bb |
|
05-Oct-2023 |
Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
Revert "usb: gadget: uvc: rework pump worker to avoid while loop" This reverts commit bb00788bd62778ef80a97d67a0e3c569ac6be06f. Based on review comments, it was applied too soon and needs more work. Reported-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Link: https://lore.kernel.org/r/20231005081716.GA13853@pendragon.ideasonboard.com Cc: Michael Grzeschik <m.grzeschik@pengutronix.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
bb00788b |
|
11-Sep-2023 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: rework pump worker to avoid while loop The uvc_video_enable function is calling cancel_work_sync which will be blocking as long as new requests will be queued with the while loop. To ensure an earlier stop in the pumping loop in this particular case we rework the worker to requeue itself on every requests. Since the worker is already running prioritized, the scheduling overhad did not have real impact on the performance. Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20230911140530.2995138-4-m.grzeschik@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
52a39f2c |
|
11-Sep-2023 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: cleanup request when not in correct state The uvc_video_enable function of the uvc-gadget driver is dequeing and immediately deallocs all requests on its disable codepath. This is not save since the dequeue function is async and does not ensure that the requests are left unlinked in the controller driver. By adding the ep_free_request into the completion path of the requests we ensure that the request will be properly deallocated. Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20230911140530.2995138-3-m.grzeschik@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
3a63f86c |
|
11-Sep-2023 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: stop pump thread on video disable Since the uvc-video gadget driver is using the v4l2 interface, the streamon and streamoff can be triggered at any times. To ensure that the pump worker will be closed as soon the userspace is calling streamoff we synchronize the state of the gadget ensuring the pump worker to bail out. Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20230911140530.2995138-2-m.grzeschik@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
5ae8a354 |
|
02-Jun-2023 |
Avichal Rakesh <arakesh@google.com> |
usb: gadget: uvc: clean up comments and styling in video_pump This patch elaborates on some of the edge cases handled by video_pump around setting no_interrupt flag, and brings the code style in line with rest of the file. Link: https://lore.kernel.org/20230602151916.GH26944@pendragon.ideasonboard.com/ Signed-off-by: Avichal Rakesh <arakesh@google.com> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Message-ID: <20230602220455.313801-1-arakesh@google.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
c3ff12a9 |
|
08-May-2023 |
Avichal Rakesh <arakesh@google.com> |
usb: gadget: uvc: queue empty isoc requests if no video buffer is available ISOC transfers expect a certain cadence of requests being queued. Not keeping up with the expected rate of requests results in missed ISOC transfers (EXDEV). The application layer may or may not produce video frames to match this expectation, so uvc gadget driver must handle cases where the application is not queuing up buffers fast enough to fulfill ISOC requirements. Currently, uvc gadget driver waits for new video buffer to become available before queuing up usb requests. With this patch the gadget driver queues up 0 length usb requests whenever there are no video buffers available. The USB controller's complete callback is used as the limiter for how quickly the 0 length packets will be queued. Video buffers are still queued as soon as they become available. Link: https://lore.kernel.org/CAMHf4WKbi6KBPQztj9FA4kPvESc1fVKrC8G73-cs6tTeQby9=w@mail.gmail.com/ Signed-off-by: Avichal Rakesh <arakesh@google.com> Link: https://lore.kernel.org/r/20230508231103.1621375-1-arakesh@google.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
88c8e05e |
|
26-Oct-2022 |
Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
Revert "usb: gadget: uvc: limit isoc_sg to super speed gadgets" This reverts commit 19905240aef0181d1e6944070eb85fce75f75bcd. It was a new feature, and it doesn't even work properly yet, so revert it from this branch as it is not needed for 6.1-final. Cc: Michael Grzeschik <m.grzeschik@pengutronix.de> Cc: stable <stable@kernel.org> Fixes: 19905240aef0 ("usb: gadget: uvc: limit isoc_sg to super speed gadgets") Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
19905240 |
|
17-Oct-2022 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: limit isoc_sg to super speed gadgets The overhead of preparing sg data is high for transfers with limited payload. When transferring isoc over high-speed usb the maximum payload is rather small which is a good argument no to use sg. This patch is changing the uvc_video_encode_isoc_sg encode function only to be used for super speed gadgets. Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Cc: stable <stable@kernel.org> Link: https://lore.kernel.org/r/20221017221141.3134818-1-m.grzeschik@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
b57b08e6 |
|
18-Oct-2022 |
Jeff Vanhoof <qjv001@motorola.com> |
usb: gadget: uvc: fix sg handling during video encode In uvc_video_encode_isoc_sg, the uvc_request's sg list is incorrectly being populated leading to corrupt video being received by the remote end. When building the sg list the usage of buf->sg's 'dma_length' field is not correct and instead its 'length' field should be used. Fixes: e81e7f9a0eb9 ("usb: gadget: uvc: add scatter gather support") Cc: <stable@vger.kernel.org> Signed-off-by: Jeff Vanhoof <qjv001@motorola.com> Signed-off-by: Dan Vacura <w36195@motorola.com> Link: https://lore.kernel.org/r/20221018215044.765044-5-w36195@motorola.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
0a0a2760 |
|
18-Oct-2022 |
Dan Vacura <w36195@motorola.com> |
usb: gadget: uvc: fix sg handling in error case If there is a transmission error the buffer will be returned too early, causing a memory fault as subsequent requests for that buffer are still queued up to be sent. Refactor the error handling to wait for the final request to come in before reporting back the buffer to userspace for all transfer types (bulk/isoc/isoc_sg). This ensures userspace knows if the frame was successfully sent. Fixes: e81e7f9a0eb9 ("usb: gadget: uvc: add scatter gather support") Cc: <stable@vger.kernel.org> Signed-off-by: Dan Vacura <w36195@motorola.com> Link: https://lore.kernel.org/r/20221018215044.765044-4-w36195@motorola.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
8e8e923a |
|
18-Oct-2022 |
Dan Vacura <w36195@motorola.com> |
usb: gadget: uvc: fix dropped frame after missed isoc With the re-use of the previous completion status in 0d1c407b1a749 ("usb: dwc3: gadget: Return proper request status") it could be possible that the next frame would also get dropped if the current frame has a missed isoc error. Ensure that an interrupt is requested for the start of a new frame. Fixes: fc78941d8169 ("usb: gadget: uvc: decrease the interrupt load to a quarter") Cc: <stable@vger.kernel.org> Signed-off-by: Dan Vacura <w36195@motorola.com> Link: https://lore.kernel.org/r/20221018215044.765044-2-w36195@motorola.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
9b91a652 |
|
07-Sep-2022 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: increase worker prio to WQ_HIGHPRI This patch is changing the simple workqueue in the gadget driver to be allocated as async_wq with a higher priority. The pump worker, that is filling the usb requests, will have a higher priority and will not be scheduled away so often while the video stream is handled. This will lead to fewer streaming underruns. Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20220907215818.2670097-1-m.grzeschik@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
a725d0f6 |
|
29-May-2022 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: call uvc uvcg_warn on completed status instead of uvcg_info Likewise to the uvcvideo hostside driver, this patch is changing the usb_request message of an non zero completion handler call from dev_info to dev_warn. Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20220529223848.105914-4-m.grzeschik@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
c5d337a3 |
|
08-Jun-2022 |
Laurent Pinchart <laurent.pinchart@ideasonboard.com> |
usb: gadget: uvc: Fix comment blocks style The UVC gadget driver historically uses the /* Comment * style */ for multi-line block comments, which is frowned upon. Patches for the driver are required to use the more standard /* * Comment * style */ style. This result in inconsistencies. Fix it by converting all remaining instances of the old style. Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Link: https://lore.kernel.org/r/20220608174918.14656-1-laurent.pinchart@ideasonboard.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
96163f83 |
|
17-Jun-2022 |
Dan Vacura <w36195@motorola.com> |
usb: gadget: uvc: fix list double add in uvcg_video_pump A panic can occur if the endpoint becomes disabled and the uvcg_video_pump adds the request back to the req_free list after it has already been queued to the endpoint. The endpoint complete will add the request back to the req_free list. Invalidate the local request handle once it's been queued. <6>[ 246.796704][T13726] configfs-gadget gadget: uvc: uvc_function_set_alt(1, 0) <3>[ 246.797078][ T26] list_add double add: new=ffffff878bee5c40, prev=ffffff878bee5c40, next=ffffff878b0f0a90. <6>[ 246.797213][ T26] ------------[ cut here ]------------ <2>[ 246.797224][ T26] kernel BUG at lib/list_debug.c:31! <6>[ 246.807073][ T26] Call trace: <6>[ 246.807180][ T26] uvcg_video_pump+0x364/0x38c <6>[ 246.807366][ T26] process_one_work+0x2a4/0x544 <6>[ 246.807394][ T26] worker_thread+0x350/0x784 <6>[ 246.807442][ T26] kthread+0x2ac/0x320 Fixes: f9897ec0f6d3 ("usb: gadget: uvc: only pump video data if necessary") Cc: stable@vger.kernel.org Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Dan Vacura <w36195@motorola.com> Link: https://lore.kernel.org/r/20220617163154.16621-1-w36195@motorola.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
aef11279 |
|
02-Apr-2022 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: improve sg exit condition The exit condition to quit iterating over the sg_list, while encoding the sg entries, has to consider the case that the dma_len of the entry could be zero. This patch takes this condition to account. Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20220402232744.3622565-4-m.grzeschik@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
9b969f93 |
|
02-Apr-2022 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: giveback vb2 buffer on req complete On uvc_video_encode_isoc_sg the mapped vb2 buffer is returned to early. Only after the last usb_request worked with the buffer it is allowed to give it back to vb2. This patch fixes that. Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20220402232744.3622565-3-m.grzeschik@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
61aa709c |
|
02-Apr-2022 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: rework uvcg_queue_next_buffer to uvcg_complete_buffer The function uvcg_queue_next_buffer is used different than its name suggests. The return value nextbuf is never used by any caller. This patch reworks the function to its actual purpose, by removing the unused code and renaming it. The function name uvcg_complete_buffer makes it more clear that it is actually marking the current video buffer as complete. Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20220402232744.3622565-2-m.grzeschik@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
fd03af27 |
|
22-Oct-2021 |
Michael Olbrich <m.olbrich@pengutronix.de> |
usb: gadget: uvc: implement dwPresentationTime and scrSourceClock This patch adds the fields UVC_STREAM_PTS and UVC_STREAM_SCR to the uvc header, in case this data is available. It also enables the copy of the timestamp to the vb2_v4l2_buffer by setting V4L2_BUF_FLAG_TIMESTAMP_COPY in the queue.timestamp_flags. Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de> Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20211022093223.26493-2-m.grzeschik@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
f262ce66 |
|
22-Oct-2021 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: use on returned header len in video_encode_isoc_sg The function uvc_video_encode_header function returns the number of bytes used for the header. We change the video_encode_isoc_sg function to use the returned header_len rather than UVCG_REQUEST_HEADER_LEN and make the encode function more flexible. Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20211022093223.26493-1-m.grzeschik@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
f9897ec0 |
|
17-Oct-2021 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: only pump video data if necessary If the streaming endpoint is not enabled, the worker has nothing to do. In the case buffers are still queued, this patch ensures that it will bail out without handling any data. Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Reviewed-by: Paul Elder <paul.elder@ideasonboard.com> Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20211017215017.18392-6-m.grzeschik@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
5fc49d8b |
|
17-Oct-2021 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: only schedule stream in streaming state This patch ensures that the video pump thread will only be scheduled if the uvc is really in streaming state. This way the worker will not have to run on an empty queue. Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Reviewed-by: Paul Elder <paul.elder@ideasonboard.com> Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20211017215017.18392-5-m.grzeschik@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
38db3716 |
|
17-Oct-2021 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: test if ep->desc is valid on ep_queue The reason that the ep_queue has failed could be a disabled endpoint. In that case it is not guaranteed that the ep->desc is still valid. This patch adds a check for NULL. Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Reviewed-by: Paul Elder <paul.elder@ideasonboard.com> Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20211017215017.18392-4-m.grzeschik@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
859c675d |
|
18-Oct-2021 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: consistently use define for headerlen The uvc request headerlen of 2 was defined as UVCG_REQUEST_HEADER_LEN in commit e81e7f9a0eb9 ("usb: gadget: uvc: add scatter gather support"). We missed to use it consistently. This patch fixes that. Reviewed-by: Paul Elder <paul.elder@ideasonboard.com> Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20211018072059.11465-1-m.grzeschik@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
|
#
fc78941d |
|
28-Jun-2021 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: decrease the interrupt load to a quarter With usb3 we handle many more requests. Decrease the interrupt load by only enabling the interrupt every quarter of the allocated requests. Reviewed-by: Paul Elder <paul.elder@ideasonboard.com> Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> -- v1 -> v2: - edited patch description - removed extra parantheses - added a comment for the logic - using unsigned int instead of int - reinitializing req_int_count in uvcg_video_enable v2 -> v3: - Link: https://lore.kernel.org/r/20210628155311.16762-6-m.grzeschik@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
e81e7f9a |
|
28-Jun-2021 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: add scatter gather support This patch adds support for scatter gather transfers. If the underlying gadgets sg_supported == true, then the videeobuf2-dma-sg is used and the encode routine maps all scatter entries to separate scatterlists for the usb gadget. When streaming 1080p with request size of 1024 times 3 bytes top shows a difference of about 6.4% CPU load applying this patch: PID USER PR NI VIRT RES %CPU %MEM TIME+ S COMMAND 64 root 0 -20 0.0m 0.0m 7.7 0.0 0:01.25 I [kworker/u5:0-uvcvideo] 83 root 0 -20 0.0m 0.0m 4.5 0.0 0:03.71 I [kworker/u5:3-uvcvideo] 307 root -51 0 0.0m 0.0m 3.8 0.0 0:01.05 S [irq/51-dwc3] vs. 64 root 0 -20 0.0m 0.0m 5.8 0.0 0:01.79 I [kworker/u5:0-uvcvideo] 306 root -51 0 0.0m 0.0m 3.2 0.0 0:01.97 S [irq/51-dwc3] 82 root 0 -20 0.0m 0.0m 0.6 0.0 0:01.86 I [kworker/u5:1-uvcvideo] Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20210628155311.16762-5-m.grzeschik@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
9973772d |
|
28-Jun-2021 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: make uvc_num_requests depend on gadget speed While sending bigger images is possible with USB_SPEED_SUPER it is better to use more isochronous requests in flight. This patch makes the number uvc_num_requests dynamic by changing it depending on the gadget speed. Reviewed-by: Paul Elder <paul.elder@ideasonboard.com> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Link: https://lore.kernel.org/r/20210628155311.16762-3-m.grzeschik@pengutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
43cd0023 |
|
21-Apr-2020 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc_video: add worker to handle the frame pumping This patch changes the function uvc_video_pump to be a separate scheduled worker. This way the completion handler of each usb request and every direct caller of the pump has only to schedule the worker instead of doing the request handling by itself. Moving the request handling to one thread solves the locking problems between the three queueing cases in the completion handler, v4l2_qbuf and video_enable. Many drivers handle the completion handlers directly in their interrupt handlers. This patch also reduces the workload on each interrupt. Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Signed-off-by: Felipe Balbi <balbi@kernel.org>
|
#
dc0f755b |
|
10-Aug-2018 |
Laurent Pinchart <laurent.pinchart@ideasonboard.com> |
usb: gadget: uvc: Replace plain printk() with dev_*() Adding device context to the kernel log messages make them more useful. Add new uvcg_* macros based on dev_*() that print both the gadget device name and the function name. While at it, remove a commented out printk statement and an unused printk-based macro. Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Reviewed-by: Paul Elder <paul.elder@ideasonboard.com> Tested-by: Paul Elder <paul.elder@ideasonboard.com> Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
|
#
8dbf9c7a |
|
10-Aug-2018 |
Laurent Pinchart <laurent.pinchart@ideasonboard.com> |
usb: gadget: uvc: Only halt video streaming endpoint in bulk mode When USB requests for video data fail to be submitted, the driver signals a problem to the host by halting the video streaming endpoint. This is only valid in bulk mode, as isochronous transfers have no handshake phase and can't thus report a stall. The usb_ep_set_halt() call returns an error when using isochronous endpoints, which we happily ignore, but some UDCs complain in the kernel log. Fix this by only trying to halt the endpoint in bulk mode. Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Reviewed-by: Paul Elder <paul.elder@ideasonboard.com> Tested-by: Paul Elder <paul.elder@ideasonboard.com> Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
|
#
9d1ff5dc |
|
10-Aug-2018 |
Laurent Pinchart <laurent.pinchart@ideasonboard.com> |
usb: gadget: uvc: Factor out video USB request queueing USB requests for video data are queued from two different locations in the driver, with the same code block occurring twice. Factor it out to a function. Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Reviewed-by: Paul Elder <paul.elder@ideasonboard.com> Tested-by: Paul Elder <paul.elder@ideasonboard.com> Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
|
#
24ceb9c6 |
|
06-Nov-2017 |
Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
USB: gadget: function: Remove redundant license text Now that the SPDX tag is in all USB files, that identifies the license in a specific and legally-defined manner. So the extra GPL text wording can be removed as it is no longer needed at all. This is done on a quest to remove the 700+ different ways that files in the kernel describe the GPL license text. And there's unneeded stuff like the address (sometimes incorrect) for the FSF which is never needed. No copyright headers or other non-license-description text was removed. Cc: Oliver Neukum <oneukum@suse.com> Cc: Johan Hovold <johan@kernel.org> Cc: Jerry Zhang <zhangjerry@google.com> Cc: John Keeping <john@metanate.com> Cc: Krzysztof Opasiak <k.opasiak@samsung.com> Cc: Abdulhadi Mohamed <abdulahhadi2@gmail.com> Cc: Matthew Wilcox <willy@linux.intel.com> Cc: Janusz Dziedzic <januszx.dziedzic@linux.intel.com> Acked-by: Felipe Balbi <felipe.balbi@linux.intel.com> Acked-by: Michal Nazarewicz <mina86@mina86.com> Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Acked-by: Vincent Pelletier <plr.vincent@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
5fd54ace |
|
03-Nov-2017 |
Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
USB: add SPDX identifiers to all remaining files in drivers/usb/ It's good to have SPDX identifiers in all files to make it easier to audit the kernel tree for correct licenses. Update the drivers/usb/ and include/linux/usb* files with the correct SPDX license identifier based on the license text in the file itself. The SPDX identifier is a legally binding shorthand, which can be used instead of the full boiler plate text. This work is based on a script and data from Thomas Gleixner, Philippe Ombredanne, and Kate Stewart. Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Kate Stewart <kstewart@linuxfoundation.org> Cc: Philippe Ombredanne <pombredanne@nexb.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Acked-by: Felipe Balbi <felipe.balbi@linux.intel.com> Acked-by: Johan Hovold <johan@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
#
eaa496ff |
|
27-Sep-2016 |
Felipe Balbi <felipe.balbi@linux.intel.com> |
usb: gadget: composite: always set ep->mult to a sensible value ep->mult is supposed to be set to Isochronous and Interrupt Endapoint's multiplier value. This value is computed from different places depending on the link speed. If we're dealing with HighSpeed, then it's part of bits [12:11] of wMaxPacketSize. This case wasn't taken into consideration before. While at that, also make sure the ep->mult defaults to one so drivers can use it unconditionally and assume they'll never multiply ep->maxpacket to zero. Cc: <stable@vger.kernel.org> Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
|
#
d8e96c4b |
|
17-Feb-2015 |
Hans Verkuil <hans.verkuil@cisco.com> |
[media] uvc gadget: switch to v4l2 core locking Switch this driver over to the V4L2 core locking mechanism in preparation for switching to unlocked_ioctl. Suggested by Laurent Pinchart. This patch introduces a new mutex at the struct uvc_video level and drops the old mutex at the queue level. The new lock is now used for all ioctl locking and in the release file operation (the driver always has to take care of locking in file operations, the core only serializes ioctls). Note that the mmap and get_unmapped_area file operations no longer take a lock. Commit f035eb4e976ef5a059e30bc91cfd310ff030a7d3 fixed a AB-BA deadlock by moving all the locking down into vb2, so the mmap and get_unmapped_area file operations should no longer do any locking before calling into vb2. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
|
#
70685711 |
|
05-Feb-2015 |
Lad, Prabhakar <prabhakar.csengg@gmail.com> |
usb: gadget: function: uvc: fix sparse warnings this patch fixes following sparse warnings: uvc_video.c:283:5: warning: symbol 'uvcg_video_pump' was not declared. Should it be static? uvc_video.c:342:5: warning: symbol 'uvcg_video_enable' was not declared. Should it be static? uvc_video.c:381:5: warning: symbol 'uvcg_video_init' was not declared. Should it be static? Signed-off-by: Lad, Prabhakar <prabhakar.csengg@gmail.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
|
#
d7577b38 |
|
29-Sep-2014 |
Felipe Balbi <balbi@ti.com> |
usb: gadget: function: uvc: conditionally dequeue We shouldn't try to dequeue a NULL pointer. Signed-off-by: Felipe Balbi <balbi@ti.com>
|
#
6dd5b021 |
|
16-Sep-2014 |
Laurent Pinchart <laurent.pinchart@ideasonboard.com> |
usb: gadget: uvc: Simplify uvcg_video_pump by using local variable Use the local queue variable instead of computing it every time. Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
|
#
3a83c16e |
|
08-Sep-2014 |
Andrzej Pietrasiewicz <andrzej.p@samsung.com> |
usb: gadget: uvc: separately compile some components of f_uvc Compile uvc_queue, uvc_v4l2, uvc_video separately so that later they can be all combined in a separately compiled f_uvc. Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com> Tested-by: Michael Grzeschik <m.grzeschik@pengutronix.de> [Make uvc_v4l2_ioctl_ops non-static] [Rename __UVC__V4L2__H__ and __UVC__VIDEO__H__] [Update MAINTAINERS] Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Acked-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
|
#
7ea95b11 |
|
08-Sep-2014 |
Andrzej Pietrasiewicz <andrzej.p@samsung.com> |
usb: gadget: uvc: rename functions to avoid conflicts with host uvc Prepare for separate compilation of uvc function's components. Some symbols will have to be exported, so rename to avoid conflicts with functions of the same name in host uvc. Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com> Tested-by: Michael Grzeschik <m.grzeschik@pengutronix.de> [Rename uvc_video_pump and uvc_queue_head as well] [Rename forgotten uvc_queue_cancel instance in a comment] Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Acked-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
|
#
6bc17375 |
|
21-Aug-2014 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: Change KERN_INFO to KERN_DEBUG on request shutdown The disconnect of the USB Device is a common pattern for an UVC Camera. In many cases this will give us an meaningless information for all buffers that couldn't be enqueued. That patch changes this to KERN_DEBUG. Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
|
#
bd52b813 |
|
08-Aug-2014 |
Michael Grzeschik <m.grzeschik@pengutronix.de> |
usb: gadget: uvc: fix possible lockup in uvc gadget If the pending buffers in the queue could not be pushed to the udc endpoint we have to cancel the uvc_queue. Otherwise the gadget will get stuck on this error. This patch calls uvc_queue_cancel if usb_ep_queue failed. Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Signed-off-by: Felipe Balbi <balbi@ti.com>
|
#
00a2430f |
|
15-Jul-2014 |
Andrzej Pietrasiewicz <andrzej.p@samsung.com> |
usb: gadget: Gadget directory cleanup - group usb functions The drivers/usb/gadget directory contains many files. Files which are related can be distributed into separate directories. This patch moves the USB functions implementations into a separate directory. Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com> Signed-off-by: Felipe Balbi <balbi@ti.com>
|