On 01/28/2018 11:23 PM, Sakari Ailus wrote: > Hi Devin, > > On Sun, Jan 28, 2018 at 09:12:44AM -0500, Devin Heitmueller wrote: >> Hello all, >> >> I recently updated to the latest kernel, and I am seeing the following >> dumped to dmesg with both au0828 and em28xx based devices whenever I >> exit tvtime (my kernel is compiled with CONFIG_VIDEO_ADV_DEBUG=y by >> default): > > Thanks for reporting this. Would you be able to provide the full dmesg, > with VB2 debug parameter set to 2? > > I can't immediately see how you'd get this, well, without triggering a > kernel warning or two. The code is pretty complex though. This might happen if you queue up buffers and then close the file descriptor without ever calling STREAMON. Calling QBUF will call __buf_prepare which will call the 'prepare' memop for the buffer. But the corresponding 'finish' memop is only done in vb2_buffer_done. But if the buffer is never passed to the driver with __enqueue_in_driver, then vb2_buffer_done is also never called and the 'prepare' and 'finish' memops are now unbalanced. It's easy to test: modprobe vivid no_error_inj=1 v4l2-compliance -s And it will log this: [782114.469159] videobuf2_core: counters for queue 00000000d2727d92, buffer 0: UNBALANCED! [782114.469164] videobuf2_core: buf_init: 1 buf_cleanup: 1 buf_prepare: 2 buf_finish: 2 [782114.469167] videobuf2_core: buf_queue: 0 buf_done: 0 [782114.469169] videobuf2_core: alloc: 1 put: 1 prepare: 2 finish: 0 mmap: 0 [782114.469171] videobuf2_core: get_userptr: 0 put_userptr: 0 [782114.469173] videobuf2_core: attach_dmabuf: 0 detach_dmabuf: 0 map_dmabuf: 0 unmap_dmabuf: 0 [782114.469175] videobuf2_core: get_dmabuf: 0 num_users: 0 vaddr: 0 cookie: 0 [782114.469178] videobuf2_core: counters for queue 00000000d2727d92, buffer 1: UNBALANCED! [782114.469180] videobuf2_core: buf_init: 1 buf_cleanup: 1 buf_prepare: 2 buf_finish: 2 [782114.469182] videobuf2_core: buf_queue: 0 buf_done: 0 [782114.469184] videobuf2_core: alloc: 1 put: 1 prepare: 2 finish: 0 mmap: 0 [782114.469186] videobuf2_core: get_userptr: 0 put_userptr: 0 [782114.469188] videobuf2_core: attach_dmabuf: 0 detach_dmabuf: 0 map_dmabuf: 0 unmap_dmabuf: 0 [782114.469190] videobuf2_core: get_dmabuf: 0 num_users: 0 vaddr: 0 cookie: 0 prepare: 2 finish: 0 Regards, Hans > > Cc Hans. > >> >> [ 129.219666] vb2: counters for queue ffff88026463ac48: >> [ 129.219667] vb2: setup: 1 start_streaming: 2 stop_streaming: 2 >> [ 129.219667] vb2: wait_prepare: 0 wait_finish: 0 >> [ 129.219668] vb2: counters for queue ffff88026463ac48, buffer 0: UNBALANCED! >> [ 129.219669] vb2: buf_init: 1 buf_cleanup: 1 buf_prepare: 14 >> buf_finish: 14 >> [ 129.219669] vb2: buf_queue: 13 buf_done: 13 >> [ 129.219673] vb2: alloc: 1 put: 1 prepare: 14 finish: 13 mmap: 1 >> [ 129.219674] vb2: get_userptr: 0 put_userptr: 0 >> [ 129.219674] vb2: attach_dmabuf: 0 detach_dmabuf: 0 map_dmabuf: >> 0 unmap_dmabuf: 0 >> [ 129.219675] vb2: get_dmabuf: 0 num_users: 0 vaddr: 13 cookie: 0 >> [ 129.219676] vb2: counters for queue ffff88026463ac48, buffer 1: UNBALANCED! >> [ 129.219676] vb2: buf_init: 1 buf_cleanup: 1 buf_prepare: 13 >> buf_finish: 13 >> [ 129.219677] vb2: buf_queue: 12 buf_done: 12 >> [ 129.219678] vb2: alloc: 1 put: 1 prepare: 13 finish: 12 mmap: 1 >> [ 129.219678] vb2: get_userptr: 0 put_userptr: 0 >> [ 129.219679] vb2: attach_dmabuf: 0 detach_dmabuf: 0 map_dmabuf: >> 0 unmap_dmabuf: 0 >> [ 129.219679] vb2: get_dmabuf: 0 num_users: 0 vaddr: 12 cookie: 0 >> [ 129.219680] vb2: counters for queue ffff88026463ac48, buffer 2: UNBALANCED! >> [ 129.219680] vb2: buf_init: 1 buf_cleanup: 1 buf_prepare: 13 >> buf_finish: 13 >> [ 129.219681] vb2: buf_queue: 12 buf_done: 12 >> [ 129.219682] vb2: alloc: 1 put: 1 prepare: 13 finish: 12 mmap: 1 >> [ 129.219682] vb2: get_userptr: 0 put_userptr: 0 >> [ 129.219683] vb2: attach_dmabuf: 0 detach_dmabuf: 0 map_dmabuf: >> 0 unmap_dmabuf: 0 >> [ 129.219683] vb2: get_dmabuf: 0 num_users: 0 vaddr: 12 cookie: 0 >> [ 129.219684] vb2: counters for queue ffff88026463ac48, buffer 3: UNBALANCED! >> [ 129.219684] vb2: buf_init: 1 buf_cleanup: 1 buf_prepare: 13 >> buf_finish: 13 >> [ 129.219685] vb2: buf_queue: 12 buf_done: 12 >> [ 129.219686] vb2: alloc: 1 put: 1 prepare: 13 finish: 12 mmap: 1 >> [ 129.219686] vb2: get_userptr: 0 put_userptr: 0 >> [ 129.219687] vb2: attach_dmabuf: 0 detach_dmabuf: 0 map_dmabuf: >> 0 unmap_dmabuf: 0 >> [ 129.219687] vb2: get_dmabuf: 0 num_users: 0 vaddr: 12 cookie: 0 >> >> The above suggests that the prepare/finish calls are unbalanced. I >> added a bit of debug and identified that it only occurs with the video >> node; it's not happening with the VBI node (which when using tvtime >> makes use of the read() emulation done by videobuf2). >> >> I went through a git bisect, and came up with the following patch >> which introduced the issue: >> >> commit a136f59c0a1f1b09eb203951975e3fc5e8d3e722 >> Author: Sakari Ailus <sakari.ailus@xxxxxxxxxxxxxxx> >> Date: Wed May 31 11:17:26 2017 -0300 >> >> [media] vb2: Move buffer cache synchronisation to prepare from queue >> >> The buffer cache should be synchronised in buffer preparation, not when >> the buffer is queued to the device. Fix this. >> >> Mmap buffers do not need cache synchronisation since they are always >> coherent. >> >> Signed-off-by: Sakari Ailus <sakari.ailus@xxxxxxxxxxxxxxx> >> Acked-by: Hans Verkuil <hans.verkuil@xxxxxxxxx> >> Signed-off-by: Mauro Carvalho Chehab <mchehab@xxxxxxxxxxxxxxxx> >> >> It's not clear to me whether this is really a regression in videobuf2, >> or if it's a change in behavior that exposes some improper handling in >> em28xx/au0828 when the device node is closed. >> >> To reproduce: >> >> 1. compile kernel with CONFIG_VIDEO_ADV_DEBUG=y >> 2. Plug in au0828 or em28xx device (reproduced with HVR-950 and HVR-950q) >> 3. start tvtime >> 4. exit tvtime >> >> I don't claim to be videobuf2 expert, so any advice that could be >> offered with regards to a fix (either in videobuf2 or in au0828) would >> certainly be welcome. >> >> Thanks in advance, >> >> Devin >> >> -- >> Devin J. Heitmueller - Kernel Labs >> http://www.kernellabs.com >
