Skip to contentSkip to conversation

staging: lustre: o2iblnd: Fix crash in kiblnd_handle_early_rxs()

PATCH In review Doug Oucharek 2 participants 2 messages May 10, 2018 – May 11, 2018 Staging

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 staging: lustre: o2iblnd: Fix crash in kiblnd_handle_early_rxs()

Under upstream staging commit 5a2ca43fa54f561c252c2, the list handling code in kiblnd_handle_early_rxs() got changed to list_for_each_safe(). That protects against the current thread from deleting the current entry it is looking at. It does not protect against another thread from deleting the next item in the list (which the tmp variable points to). The way this routine holds then releases a lock opens the door to other threads doing just that.

This patch reverts this commit on this routine.

Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-9886

Fixes 5a2ca43fa54f ("Staging: lustre: Iterate list using list_for_each_entry")

---
Change v2: Fixed list_entry parameter to be proper struct Change v1: Initial Patch

 drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c +3 -2
index 32fa8ca..47eb8b4 100644
@@ -1965,13 +1965,14 @@ static int kiblnd_resolve_addr(struct rdma_cm_id *cmid,
 {
 	unsigned long flags;
 	struct kib_rx *rx;
-	struct kib_rx *tmp;
 
 	LASSERT(!in_interrupt());
 	LASSERT(conn->ibc_state >= IBLND_CONN_ESTABLISHED);
 
 	write_lock_irqsave(&kiblnd_data.kib_global_lock, flags);
-	list_for_each_entry_safe(rx, tmp, &conn->ibc_early_rxs, rx_list) {
+	while (!list_empty(&conn->ibc_early_rxs)) {
+		rx = list_entry(conn->ibc_early_rxs.next,
+				struct kib_rx, rx_list);
 		list_del(&rx->rx_list);
 		write_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags);
 
Signature · 1 line
1.8.3.1
#1 lore raw
Quoted 11 lines · Doug Oucharek

Under upstream staging commit 5a2ca43fa54f561c252c2, the list handling code in kiblnd_handle_early_rxs() got changed to list_for_each_safe(). That protects against the current thread from deleting the current entry it is looking at. It does not protect against another thread from deleting the next item in the list (which the tmp variable points to). The way this routine holds then releases a lock opens the door to other threads doing just that.

This patch reverts this commit on this routine.

Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-9886

Please add a line like:

Fixes 5a2ca43fa54f ("Staging: lustre: Iterate list using list_for_each_entry")

when you know this information.

Also, I'm guessing you want this backported to the stable trees? Or does no one who actually runs lustre run a kernel.org stable kernel?

thanks,

greg k-h
_______________________________________________
devel mailing list
[email protected]
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel