powerpc/mce: fix off by one errors in mce event handling
authorDaniel Axtens <dja@axtens.net>
Tue, 12 May 2015 03:23:59 +0000 (13:23 +1000)
committerMichael Ellerman <mpe@ellerman.id.au>
Tue, 12 May 2015 09:44:01 +0000 (19:44 +1000)
commitffb2d78eca08a1451137583d4e435aecfd6af809
tree223ad3754bf3073f65509dfc045cb70290c429a2
parent7b868e81be38d5ad4f4aa4be819a5fa543cc5ee8
powerpc/mce: fix off by one errors in mce event handling

Before 69111bac42f5 ("powerpc: Replace __get_cpu_var uses"), in
save_mce_event, index got the value of mce_nest_count, and
mce_nest_count was incremented *after* index was set.

However, that patch changed the behaviour so that mce_nest count was
incremented *before* setting index.

This causes an off-by-one error, as get_mce_event sets index as
mce_nest_count - 1 before reading mce_event.  Thus get_mce_event reads
bogus data, causing warnings like
"Machine Check Exception, Unknown event version 0 !"
and breaking MCEs handling.

Restore the old behaviour and unbreak MCE handling by subtracting one
from the newly incremented value.

The same broken change occured in machine_check_queue_event (which set
a queue read by machine_check_process_queued_event).  Fix that too,
unbreaking printing of MCE information.

Fixes: 69111bac42f5 ("powerpc: Replace __get_cpu_var uses")
CC: stable@vger.kernel.org
CC: Mahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
CC: Christoph Lameter <cl@linux.com>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
arch/powerpc/kernel/mce.c