md: don't use flush_signals in userspace processes
authorMikulas Patocka <mpatocka@redhat.com>
Wed, 7 Jun 2017 23:05:31 +0000 (19:05 -0400)
committerShaohua Li <shli@fb.com>
Tue, 13 Jun 2017 17:18:02 +0000 (10:18 -0700)
commitf9c79bc05a2a91f4fba8bfd653579e066714b1ec
treeb5a229ab75dc24ca0387588e4ea38a9da23b1ce1
parentcc27b0c78c79680d128dbac79de0d40556d041bb
md: don't use flush_signals in userspace processes

The function flush_signals clears all pending signals for the process. It
may be used by kernel threads when we need to prepare a kernel thread for
responding to signals. However using this function for an userspaces
processes is incorrect - clearing signals without the program expecting it
can cause misbehavior.

The raid1 and raid5 code uses flush_signals in its request routine because
it wants to prepare for an interruptible wait. This patch drops
flush_signals and uses sigprocmask instead to block all signals (including
SIGKILL) around the schedule() call. The signals are not lost, but the
schedule() call won't respond to them.

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Cc: stable@vger.kernel.org
Acked-by: NeilBrown <neilb@suse.com>
Signed-off-by: Shaohua Li <shli@fb.com>
drivers/md/raid1.c
drivers/md/raid5.c