aboutsummaryrefslogtreecommitdiffstats
path: root/drivers
diff options
context:
space:
mode:
authorMichael Ellerman <michael@ellerman.id.au>2005-08-31 21:29:00 -0400
committerJeff Garzik <jgarzik@pobox.com>2005-08-31 22:37:56 -0400
commit58c5900bdaffbf76afd7ad5e053410cb95eb3169 (patch)
tree1d6e6eba8392f496f9ac2f632fdda8aa48b0b732 /drivers
parentabfda4719c61550be4efaf277d4a904a7930d410 (diff)
[PATCH] iseries_veth: Try to avoid pathological reset behaviour
The iseries_veth driver contains a state machine which is used to manage how connections are setup and neogotiated between LPARs. If one side of a connection resets for some reason, the two LPARs can get stuck in a race to re-setup the connection. This can lead to the connection being declared dead by one or both ends. In practice the connection is declared dead by one or both ends approximately 8/10 times a connection is reset, although it is rare for connections to be reset. (an example here: http://michael.ellerman.id.au/files/misc/veth-trace.html) The core of the problem is that the end that resets the connection doesn't wait for the other end to become aware of the reset. So the resetting end starts setting the connection back up, and then receives a reset from the other end (which is the response to the initial reset). And so on. We're severely limited in what we can do to fix this. The protocol between LPARs is essentially fixed, as we have to interoperate with both OS/400 and old Linux drivers. Which also means we need a fix that only changes the code on one end. The only fix I've found given that, is to just blindly sleep for a bit when resetting the connection, in the hope that the other end will get itself sorted. Needless to say I'd love it if someone has a better idea. This does work, I've so far been unable to get it to break, whereas without the fix a reset of one end will lead to a dead connection ~8/10 times. Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
Diffstat (limited to 'drivers')
-rw-r--r--drivers/net/iseries_veth.c25
1 files changed, 23 insertions, 2 deletions
diff --git a/drivers/net/iseries_veth.c b/drivers/net/iseries_veth.c
index c19b32e0a5ad..db83b0d31327 100644
--- a/drivers/net/iseries_veth.c
+++ b/drivers/net/iseries_veth.c
@@ -324,8 +324,14 @@ static void veth_take_monitor_ack(struct veth_lpar_connection *cnx,
324 324
325 spin_lock_irqsave(&cnx->lock, flags); 325 spin_lock_irqsave(&cnx->lock, flags);
326 veth_debug("cnx %d: lost connection.\n", cnx->remote_lp); 326 veth_debug("cnx %d: lost connection.\n", cnx->remote_lp);
327 cnx->state |= VETH_STATE_RESET; 327
328 veth_kick_statemachine(cnx); 328 /* Avoid kicking the statemachine once we're shutdown.
329 * It's unnecessary and it could break veth_stop_connection(). */
330
331 if (! (cnx->state & VETH_STATE_SHUTDOWN)) {
332 cnx->state |= VETH_STATE_RESET;
333 veth_kick_statemachine(cnx);
334 }
329 spin_unlock_irqrestore(&cnx->lock, flags); 335 spin_unlock_irqrestore(&cnx->lock, flags);
330} 336}
331 337
@@ -483,6 +489,12 @@ static void veth_statemachine(void *p)
483 489
484 if (cnx->state & VETH_STATE_RESET) 490 if (cnx->state & VETH_STATE_RESET)
485 goto restart; 491 goto restart;
492
493 /* Hack, wait for the other end to reset itself. */
494 if (! (cnx->state & VETH_STATE_SHUTDOWN)) {
495 schedule_delayed_work(&cnx->statemachine_wq, 5 * HZ);
496 goto out;
497 }
486 } 498 }
487 499
488 if (cnx->state & VETH_STATE_SHUTDOWN) 500 if (cnx->state & VETH_STATE_SHUTDOWN)
@@ -667,6 +679,15 @@ static void veth_stop_connection(u8 rlp)
667 veth_kick_statemachine(cnx); 679 veth_kick_statemachine(cnx);
668 spin_unlock_irq(&cnx->lock); 680 spin_unlock_irq(&cnx->lock);
669 681
682 /* There's a slim chance the reset code has just queued the
683 * statemachine to run in five seconds. If so we need to cancel
684 * that and requeue the work to run now. */
685 if (cancel_delayed_work(&cnx->statemachine_wq)) {
686 spin_lock_irq(&cnx->lock);
687 veth_kick_statemachine(cnx);
688 spin_unlock_irq(&cnx->lock);
689 }
690
670 /* Wait for the state machine to run. */ 691 /* Wait for the state machine to run. */
671 flush_scheduled_work(); 692 flush_scheduled_work();
672 693