This patch series sits on top of: https://patchwork.linuxtv.org/project/linux-media/list/?series=13513 This series makes it possible for drivers to drop the wait_prepare/finish callbacks if the queue lock pointer is set. Analyzing the code shows that in a few places the lock is set after the call to vb2_queue_init, so the first 7 patches move that lock assignment up. Then two additional WARN_ONs are added to vb2_queue_init to ensure that either q->lock is set, or wait_prepare/finish are set. The last patch will just unlock/lock the queue lock if wait_prepare/finish isn't set. Once this is in, we can start dropping the wait_prepare/finish ops in almost all drivers. Regards, Hans Hans Verkuil (9): media: staging: atomisp: set lock before calling vb2_queue_init() media: pwc: set lock before calling vb2_queue_init() media: msi2500: set lock before calling vb2_queue_init() media: hackrf: set lock before calling vb2_queue_init() media: airspy: set lock before calling vb2_queue_init() media: rcar_drif.c: set lock before calling vb2_queue_init() media: video-i2c: set lock before calling vb2_queue_init() media: vb2: vb2_core_queue_init(): sanity check lock and wait_prepare/finish media: vb2: use lock if wait_prepare/finish are NULL .../media/common/videobuf2/videobuf2-core.c | 21 ++++++++++++++++--- drivers/media/i2c/video-i2c.c | 2 +- drivers/media/platform/renesas/rcar_drif.c | 2 +- drivers/media/usb/airspy/airspy.c | 2 +- drivers/media/usb/hackrf/hackrf.c | 4 ++-- drivers/media/usb/msi2500/msi2500.c | 2 +- drivers/media/usb/pwc/pwc-if.c | 2 +- .../media/atomisp/pci/atomisp_subdev.c | 2 +- 8 files changed, 26 insertions(+), 11 deletions(-) -- 2.34.1
