]> git.baikalelectronics.ru Git - kernel.git/commit
usb: xhci: dbc: Fix lockdep warning
authorLu Baolu <baolu.lu@linux.intel.com>
Thu, 8 Mar 2018 15:17:15 +0000 (17:17 +0200)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Thu, 8 Mar 2018 17:06:53 +0000 (09:06 -0800)
commit9b639bc48dd45c47aa8ebf4e46a808e792cdc7e3
tree4bfd0bd7ff912a89b5f3367b599aa1e3130aaccf
parentd557c850a0f9d0486bff65fef024744788fca73e
usb: xhci: dbc: Fix lockdep warning

The xHCI DbC implementation might enter a deadlock situation because
there is no sufficient protection against the shared data between
process and softirq contexts. This can lead to the following lockdep
warnings. This patch changes to use spin_{,un}lock_irq{save,restore}
to avoid potential deadlock.

[ 528.248084] ================================
[ 528.252914] WARNING: inconsistent lock state
[ 528.257756] 4.15.0-rc1+ #1630 Not tainted
[ 528.262305] --------------------------------
[ 528.267145] inconsistent {SOFTIRQ-ON-W} -> {IN-SOFTIRQ-W} usage.
[ 528.273953] ksoftirqd/1/17 [HC0[0]:SC1[1]:HE0:SE0] takes:
[ 528.280075] (&(&port->port_lock)->rlock){+.?.}, at: [<ffffffff815396a8>] dbc_rx_push+0x38/0x1c0
[ 528.290043] {SOFTIRQ-ON-W} state was registered at:
[ 528.295570] _raw_spin_lock+0x2f/0x40
[ 528.299818] dbc_write_complete+0x27/0xa0
[ 528.304458] xhci_dbc_giveback+0xd1/0x200
[ 528.309098] xhci_dbc_flush_endpoint_requests+0x50/0x70
[ 528.315116] xhci_dbc_handle_events+0x696/0x7b0
[ 528.320349] process_one_work+0x1ee/0x6e0
[ 528.324988] worker_thread+0x4a/0x430
[ 528.329236] kthread+0x13e/0x170
[ 528.332992] ret_from_fork+0x24/0x30
[ 528.337141] irq event stamp: 2861
[ 528.340897] hardirqs last enabled at (2860): [<ffffffff810674ea>] tasklet_action+0x6a/0x250
[ 528.350460] hardirqs last disabled at (2861): [<ffffffff817dc1ef>] _raw_spin_lock_irq+0xf/0x40
[ 528.360219] softirqs last enabled at (2852): [<ffffffff817e0e8c>] __do_softirq+0x3dc/0x4f9
[ 528.369683] softirqs last disabled at (2857): [<ffffffff8106805b>] run_ksoftirqd+0x1b/0x60
[ 528.379048]
[ 528.379048] other info that might help us debug this:
[ 528.386443] Possible unsafe locking scenario:
[ 528.386443]
[ 528.393150] CPU0
[ 528.395917] ----
[ 528.398687] lock(&(&port->port_lock)->rlock);
[ 528.403821] <Interrupt>
[ 528.406786] lock(&(&port->port_lock)->rlock);
[ 528.412116]
[ 528.412116] *** DEADLOCK ***
[ 528.412116]
[ 528.418825] no locks held by ksoftirqd/1/17.
[ 528.423662]
[ 528.423662] stack backtrace:
[ 528.428598] CPU: 1 PID: 17 Comm: ksoftirqd/1 Not tainted 4.15.0-rc1+ #1630
[ 528.436387] Call Trace:
[ 528.439158] dump_stack+0x5e/0x8e
[ 528.442914] print_usage_bug+0x1fc/0x220
[ 528.447357] mark_lock+0x4db/0x5a0
[ 528.451210] __lock_acquire+0x726/0x1130
[ 528.455655] ? __lock_acquire+0x557/0x1130
[ 528.460296] lock_acquire+0xa2/0x200
[ 528.464347] ? dbc_rx_push+0x38/0x1c0
[ 528.468496] _raw_spin_lock_irq+0x35/0x40
[ 528.473038] ? dbc_rx_push+0x38/0x1c0
[ 528.477186] dbc_rx_push+0x38/0x1c0
[ 528.481139] tasklet_action+0x1d2/0x250
[ 528.485483] __do_softirq+0x1dc/0x4f9
[ 528.489630] run_ksoftirqd+0x1b/0x60
[ 528.493682] smpboot_thread_fn+0x179/0x270
[ 528.498324] kthread+0x13e/0x170
[ 528.501981] ? sort_range+0x20/0x20
[ 528.505933] ? kthread_delayed_work_timer_fn+0x80/0x80
[ 528.511755] ret_from_fork+0x24/0x30

Fixes: 2b51ddcdb96c ("usb: xhci: Add DbC support in xHCI driver")
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/usb/host/xhci-dbgcap.c
drivers/usb/host/xhci-dbgtty.c