[PATCH net] ibmvnic: Skip fatal error reset after passive init

[PATCH net] ibmvnic: Skip fatal error reset after passive init

[Juliet Kim]

During MTU change, the following events may happen.
Client-driven CRQ initialization fails due to partner’s CRQ closed,
causing client to enqueue a reset task for FATAL_ERROR. Then passive
(server-driven) CRQ initialization succeeds, causing client to
release CRQ and enqueue a reset task for failover. If the passive
CRQ initialization occurs before the FATAL reset task is processed,
the FATAL error reset task would try to access a CRQ message queue
that was freed, causing an oops. The problem may be most likely to
occur during DLPAR add vNIC with a non-default MTU, because the DLPAR
process will automatically issue a change MTU request.

Fix this by not processing fatal error reset if CRQ is passively
initialized after client-driven CRQ initialization fails.

Signed-off-by: Juliet Kim <julietk@linux.vnet.ibm.com>
---
 drivers/net/ethernet/ibm/ibmvnic.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/ibm/ibmvnic.c b/drivers/net/ethernet/ibm/ibmvnic.c
index 4bd33245bad6..3de549c6c693 100644
--- a/drivers/net/ethernet/ibm/ibmvnic.c
+++ b/drivers/net/ethernet/ibm/ibmvnic.c
@@ -2189,7 +2189,8 @@ static void __ibmvnic_reset(struct work_struct *work)
 				rc = do_hard_reset(adapter, rwi, reset_state);
 				rtnl_unlock();
 			}
-		} else {
+		} else if (!(rwi->reset_reason == VNIC_RESET_FATAL &&
+				adapter->from_passive_init)) {
 			rc = do_reset(adapter, rwi, reset_state);
 		}
 		kfree(rwi);
-- 
2.16.4

Re: [PATCH net] ibmvnic: Skip fatal error reset after passive init

[David Miller]

From: Juliet Kim <julietk@linux.vnet.ibm.com>
Date: Thu, 30 Apr 2020 13:22:11 -0500

> During MTU change, the following events may happen.
> Client-driven CRQ initialization fails due to partner’s CRQ closed,
> causing client to enqueue a reset task for FATAL_ERROR. Then passive
> (server-driven) CRQ initialization succeeds, causing client to
> release CRQ and enqueue a reset task for failover. If the passive
> CRQ initialization occurs before the FATAL reset task is processed,
> the FATAL error reset task would try to access a CRQ message queue
> that was freed, causing an oops. The problem may be most likely to
> occur during DLPAR add vNIC with a non-default MTU, because the DLPAR
> process will automatically issue a change MTU request.
> 
> Fix this by not processing fatal error reset if CRQ is passively
> initialized after client-driven CRQ initialization fails.
> 
> Signed-off-by: Juliet Kim <julietk@linux.vnet.ibm.com>

Applied, thanks.