The variable ioc->taskmgmt_quiesce_io is often protected by the lock
ioc->taskmgmt_lock when is accessed. Here is an example in
mpt_SoftResetHandler():
spin_lock_irqsave(&ioc->taskmgmt_lock, flags);
...
ioc->taskmgmt_quiesce_io = 0;
...
spin_unlock_irqrestore(&ioc->taskmgmt_lock, flags);
However, ioc->taskmgmt_quiesce_io is set to 1 without holding the lock
ioc->taskmgmt_lock in mpt_ioc_reset(), which can cause possible data
races:
case MPT_IOC_SETUP_RESET:
ioc->taskmgmt_quiesce_io = 1;
To fix this possible data race, WRITE_ONCE() is used to access the
variable ioc->taskmgmt_lock.
Reported-by: BassCheck <[email protected]>
Signed-off-by: Tuo Li <[email protected]>
---
v2:
* Replace lock operations with a WRITE_ONCE to avoid possible data
races.
Thank Bart Van Assche for helpful advice.
---
drivers/message/fusion/mptbase.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/message/fusion/mptbase.c b/drivers/message/fusion/mptbase.c
index 4bf669c55649..6c67eec9ab85 100644
--- a/drivers/message/fusion/mptbase.c
+++ b/drivers/message/fusion/mptbase.c
@@ -6563,7 +6563,7 @@ mpt_ioc_reset(MPT_ADAPTER *ioc, int reset_phase)
{
switch (reset_phase) {
case MPT_IOC_SETUP_RESET:
- ioc->taskmgmt_quiesce_io = 1;
+ WRITE_ONCE(ioc->taskmgmt_quiesce_io, 1);
dtmprintk(ioc, printk(MYIOC_s_DEBUG_FMT
"%s: MPT_IOC_SETUP_RESET\n", ioc->name, __func__));
break;
--
2.34.1