powerpc: Fix unpaired probe_hcall_entry and probe_hcall_exit
commitec0d3233562d591c892841fcfa73edd3700a6ca6
authorLi Zhong <zhong@linux.vnet.ibm.com>
Sun, 18 Dec 2011 16:03:04 +0000 (18 16:03 +0000)
committerGreg Kroah-Hartman <gregkh@suse.de>
Thu, 12 Jan 2012 19:35:03 +0000 (12 11:35 -0800)
tree600c51c819116e9b79f8608570b1b0329bee06bd
parentbfaebb8af017d443937022cf3d5735d726e4f711
powerpc: Fix unpaired probe_hcall_entry and probe_hcall_exit

commit e4f387d8db3ba3c2dae4d8bdfe7bb5f4fe1bcb0d upstream.

Unpaired calling of probe_hcall_entry and probe_hcall_exit might happen
as following, which could cause incorrect preempt count.

__trace_hcall_entry => trace_hcall_entry -> probe_hcall_entry =>
get_cpu_var => preempt_disable

__trace_hcall_exit => trace_hcall_exit -> probe_hcall_exit =>
put_cpu_var => preempt_enable

where:
A => B and A -> B means A calls B, but
=> means A will call B through function name, and B will definitely be
called.
-> means A will call B through function pointer, so B might not be
called if the function pointer is not set.

So error happens when only one of probe_hcall_entry and probe_hcall_exit
get called during a hcall.

This patch tries to move the preempt count operations from
probe_hcall_entry and probe_hcall_exit to its callers.

Reported-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Li Zhong <zhong@linux.vnet.ibm.com>
Tested-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
arch/powerpc/platforms/pseries/hvCall_inst.c
arch/powerpc/platforms/pseries/lpar.c