[PATCH] libcfs: Fix a sleep-in-atomic bug in cfs_wi_sched_destroy

From: Jia-Ju Bai
Date: Wed May 31 2017 - 04:05:08 EST


The driver may sleep under a spin lock, and the function call path is:
cfs_wi_sched_destroy (acquire the lock by spin_lock)
LASSERT
lbug_with_loc
libcfs_debug_dumplog
schedule and kthread_run --> may sleep

To fix it, all "LASSERT" is placed out of the spin_lock and spin_unlock.

Signed-off-by: Jia-Ju Bai <baijiaju1990@xxxxxxx>
---
drivers/staging/lustre/lnet/libcfs/workitem.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/staging/lustre/lnet/libcfs/workitem.c b/drivers/staging/lustre/lnet/libcfs/workitem.c
index dbc2a9b..e0424f6 100644
--- a/drivers/staging/lustre/lnet/libcfs/workitem.c
+++ b/drivers/staging/lustre/lnet/libcfs/workitem.c
@@ -302,11 +302,12 @@ static int cfs_wi_scheduler(void *arg)
return;
}

- LASSERT(!list_empty(&sched->ws_list));
sched->ws_stopping = 1;

spin_unlock(&cfs_wi_data.wi_glock);

+ LASSERT(!list_empty(&sched->ws_list));
+
i = 2;
wake_up_all(&sched->ws_waitq);

--
1.7.9.5