sgi-xp: eliminate false detection of no heartbeat

After XPC has been up and running on multiple partitions for any length of
time, if XPC on one of the partitions is stopped and restarted (either by
a rmmod/insmod or a system restart), it is possible for the XPCs running
on the other partitions to falsely detect a lack of heartbeat from the XPC
that was just restarted.  This false detection will occur if the restarted
XPC comes up within the five-seconds preceding one of the other XPC's
heartbeat check (which occurs once every twenty seconds).

The detection of no heartbeat results in the detecting XPC deactivating
from the just restarted XPC.  The only remedy is to restart one of the
XPCs and hope that one doesn't hit this five-second window on any of the
other partitions.

Signed-off-by: Dean Nelson <dcn@sgi.com>
Signed-off-by: Robin Holt <holt@sgi.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

diff --git a/drivers/misc/sgi-xp/xpc_sn2.c b/drivers/misc/sgi-xp/xpc_sn2.c
index 73b7fb8de47a32e0e38b8abe048ede7333fbe832..82fb9958f22f7bc7b6bcfb03e31d7c66de7beb57 100644 (file)
--- a/drivers/misc/sgi-xp/xpc_sn2.c
+++ b/drivers/misc/sgi-xp/xpc_sn2.c
@@ -899,7 +899,7 @@ xpc_update_partition_info_sn2(struct xpc_partition *part, u8 remote_rp_version,
        dev_dbg(xpc_part, "  remote_vars_pa = 0x%016lx\n",
                part_sn2->remote_vars_pa);
 
-       part->last_heartbeat = remote_vars->heartbeat;
+       part->last_heartbeat = remote_vars->heartbeat - 1;
        dev_dbg(xpc_part, "  last_heartbeat = 0x%016lx\n",
                part->last_heartbeat);