pch_dma: Fix channel locking
authorAlexander Stein <alexander.stein@systec-electronic.com>
Wed, 22 Jun 2011 15:05:33 +0000 (17:05 +0200)
committerVinod Koul <vinod.koul@intel.com>
Wed, 13 Jul 2011 22:10:08 +0000 (03:10 +0530)
commit70f18915846f092e0e1c988f1726a532fa3ab3a1
tree2dad69276540ec682ae3f8013b01abf5658654da
parent4de1ba155bbe9b629b9fb03919c5d905b747e62f
pch_dma: Fix channel locking

Fix for the following INFO message

=================================
[ INFO: inconsistent lock state ]
2.6.39+ #89
---------------------------------
inconsistent {HARDIRQ-ON-W} -> {IN-HARDIRQ-W} usage.
rs232/822 [HC1[1]:SC0[0]:HE0:SE1] takes:
 (&(&pd_chan->lock)->rlock){?.....}, at: [<c123b9a1>] pdc_desc_get+0x16/0xab
{HARDIRQ-ON-W} state was registered at:
  [<c104fe28>] mark_irqflags+0xbd/0x11a
  [<c1050386>] __lock_acquire+0x501/0x6bb
  [<c1050945>] lock_acquire+0x63/0x7b
  [<c131c51d>] _raw_spin_lock_bh+0x43/0x51
  [<c123bee4>] pd_alloc_chan_resources+0x92/0x11e
  [<c123ad62>] dma_chan_get+0x9b/0x107
  [<c123b2d1>] __dma_request_channel+0x61/0xdc
  [<c11ba24b>] pch_request_dma+0x61/0x19e
  [<c11bb3b8>] pch_uart_startup+0x16a/0x1a2
  [<c11b8446>] uart_startup+0x87/0x147
  [<c11b9183>] uart_open+0x117/0x13e
  [<c11a5c7d>] tty_open+0x23c/0x34c
  [<c1097705>] chrdev_open+0x140/0x15f
  [<c10930a6>] __dentry_open.clone.14+0x14a/0x22b
  [<c1093dfb>] nameidata_to_filp+0x36/0x40
  [<c109f28b>] do_last+0x513/0x635
  [<c109f4af>] path_openat+0x9c/0x2aa
  [<c109f6e4>] do_filp_open+0x27/0x69
  [<c1093f02>] do_sys_open+0xfd/0x184
  [<c1093fad>] sys_open+0x24/0x2a
  [<c131d58c>] sysenter_do_call+0x12/0x32
irq event stamp: 2522
hardirqs last  enabled at (2521): [<c131ca3b>] _raw_spin_unlock_irqrestore+0x36/0x52
hardirqs last disabled at (2522): [<c131db27>] common_interrupt+0x27/0x34
softirqs last  enabled at (2354): [<c102fa11>] __do_softirq+0x10a/0x11a
softirqs last disabled at (2299): [<c10041a4>] do_softirq+0x57/0xa4

other info that might help us debug this:
2 locks held by rs232/822:
 #0:  (&tty->atomic_write_lock){+.+.+.}, at: [<c11a4b7a>] tty_write_lock+0x14/0x3c
 #1:  (&port_lock_key){-.....}, at: [<c11bad72>] pch_uart_interrupt+0x17/0x1e9

stack backtrace:
Pid: 822, comm: rs232 Not tainted 2.6.39+ #89
Call Trace:
 [<c1319f90>] ? printk+0x19/0x1b
 [<c104f893>] print_usage_bug+0x184/0x18f
 [<c104e5b1>] ? print_irq_inversion_bug+0x10e/0x10e
 [<c104f943>] mark_lock_irq+0xa5/0x1f6
 [<c104fc9c>] mark_lock+0x208/0x2d7
 [<c104fdc0>] mark_irqflags+0x55/0x11a
 [<c1050386>] __lock_acquire+0x501/0x6bb
 [<c10042ee>] ? dump_trace+0x92/0xb6
 [<c1050945>] lock_acquire+0x63/0x7b
 [<c123b9a1>] ? pdc_desc_get+0x16/0xab
 [<c131c2d0>] _raw_spin_lock+0x3e/0x4c
 [<c123b9a1>] ? pdc_desc_get+0x16/0xab
 [<c123b9a1>] pdc_desc_get+0x16/0xab
 [<c10504d8>] ? __lock_acquire+0x653/0x6bb
 [<c123bb2c>] pd_prep_slave_sg+0x7c/0x1cb
 [<c1006c3f>] ? nommu_map_sg+0x6e/0x81
 [<c11bace6>] dma_handle_tx+0x2cf/0x344
 [<c11bad72>] ? pch_uart_interrupt+0x17/0x1e9
 [<c11baebb>] pch_uart_interrupt+0x160/0x1e9
 [<c10642fb>] handle_irq_event_percpu+0x25/0x127
 [<c1064429>] handle_irq_event+0x2c/0x43
 [<c1065e0d>] ? handle_fasteoi_irq+0x84/0x84
 [<c1065eb9>] handle_edge_irq+0xac/0xce
 <IRQ>  [<c1003ecb>] ? do_IRQ+0x38/0x9d
 [<c131db2e>] ? common_interrupt+0x2e/0x34
 [<c105007b>] ? __lock_acquire+0x1f6/0x6bb
 [<c131ca3d>] ? _raw_spin_unlock_irqrestore+0x38/0x52
 [<c11b798b>] ? uart_start+0x2d/0x32
 [<c11b7998>] ? uart_flush_chars+0x8/0xa
 [<c11a7962>] ? n_tty_write+0x12c/0x1c6
 [<c1027a73>] ? try_to_wake_up+0x251/0x251
 [<c11a4d0b>] ? tty_write+0x169/0x1dc
 [<c11a7836>] ? n_tty_ioctl+0xb7/0xb7
 [<c1094841>] ? vfs_write+0x91/0x10d
 [<c11a4ba2>] ? tty_write_lock+0x3c/0x3c
 [<c1094a69>] ? sys_write+0x3e/0x63
 [<c131d58c>] ? sysenter_do_call+0x12/0x32

Signed-off-by: Alexander Stein <alexander.stein@systec-electronic.com>
Tested-by: Tomoya MORINAGA <tomoya-linux@dsn.okisemi.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
drivers/dma/pch_dma.c