[media] m2m: fix bad unlock balance
This patch removes unnecessary mutex queue unlock/lock sequence causing bad
unlock balance in v4l2_m2m_poll when the last buffer on the destination
queue has been dequeued and adds spin lock protection for the done list
list_empty calls.
[  144.990873] =====================================
[  144.995584] [ BUG: bad unlock balance detected! ]
[  145.000301] 
4.1.0-00137-ga105070 #98 Tainted: G        W
[  145.006140] -------------------------------------
[  145.010851] demux:sink/487 is trying to release lock (&dev->dev_mutex) at:
[  145.017785] [<
808cc578>] mutex_unlock+0x18/0x1c
[  145.022322] but there are no more locks to release!
[  145.027205]
[  145.027205] other info that might help us debug this:
[  145.033741] no locks held by demux:sink/487.
[  145.038015]
[  145.038015] stack backtrace:
[  145.042385] CPU: 2 PID: 487 Comm: demux:sink Tainted: G        W       
4.1.0-00137-ga105070 #98
[  145.051089] Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
[  145.057622] Backtrace:
[  145.060102] [<
80014a4c>] (dump_backtrace) from [<
80014cc4>] (show_stack+0x20/0x24)
[  145.067679]  r6:
80cedf78 r5:
00000000 r4:
00000000 r3:
00000000
[  145.073421] [<
80014ca4>] (show_stack) from [<
808c61e0>] (dump_stack+0x8c/0xa4)
[  145.080661] [<
808c6154>] (dump_stack) from [<
80072b64>] (print_unlock_imbalance_bug+0xb8/0xe8)
[  145.089277]  r6:
808cc578 r5:
ac6cd050 r4:
ac38e400 r3:
00000001
[  145.095020] [<
80072aac>] (print_unlock_imbalance_bug) from [<
80077db4>] (lock_release+0x1a4/0x250)
[  145.103983]  r6:
808cc578 r5:
ac6cd050 r4:
ac38e400 r3:
00000000
[  145.109728] [<
80077c10>] (lock_release) from [<
808cc470>] (__mutex_unlock_slowpath+0xc4/0x1b4)
[  145.118344]  r9:
acb27a41 r8:
00000000 r7:
81553814 r6:
808cc578 r5:
60030013 r4:
ac6cd01c
[  145.126190] [<
808cc3ac>] (__mutex_unlock_slowpath) from [<
808cc578>] (mutex_unlock+0x18/0x1c)
[  145.134720]  r7:
00000000 r6:
aced7cd4 r5:
00000041 r4:
acb87800
[  145.140468] [<
808cc560>] (mutex_unlock) from [<
805a98b8>] (v4l2_m2m_fop_poll+0x5c/0x64)
[  145.148494] [<
805a985c>] (v4l2_m2m_fop_poll) from [<
805955a0>] (v4l2_poll+0x6c/0xa0)
[  145.156243]  r6:
aced7bec r5:
00000000 r4:
ac6cc380 r3:
805a985c
[  145.161991] [<
80595534>] (v4l2_poll) from [<
80156edc>] (do_sys_poll+0x230/0x4c0)
[  145.169391]  r5:
00000000 r4:
aced7be4
[  145.173013] [<
80156cac>] (do_sys_poll) from [<
801574a8>] (SyS_ppoll+0x1d4/0x1fc)
[  145.180414]  r10:
00000000 r9:
aced6000 r8:
00000000 r7:
00000000 r6:
75c04538 r5:
00000002
[  145.188338]  r4:
00000000
[  145.190906] [<
801572d4>] (SyS_ppoll) from [<
800108c0>] (ret_fast_syscall+0x0/0x54)
[  145.198481]  r8:
80010aa4 r7:
00000150 r6:
75c04538 r5:
00000002 r4:
00000008
Signed-off-by: Zahari Doychev <zahari.doychev@linux.com>
Acked-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>