fault-inject: fix wrong should_fail() decision in task context
authorAkinobu Mita <akinobu.mita@gmail.com>
Thu, 10 Aug 2017 22:23:53 +0000 (15:23 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Thu, 10 Aug 2017 22:54:06 +0000 (15:54 -0700)
Commit 1203c8e6fb0a ("fault-inject: simplify access check for fail-nth")
unintentionally broke a conditional statement in should_fail().  Any
faults are not injected in the task context by the change when the
systematic fault injection is not used.

This change restores to the previous correct behaviour.

Link: http://lkml.kernel.org/r/1501633700-3488-1-git-send-email-akinobu.mita@gmail.com
Fixes: 1203c8e6fb0a ("fault-inject: simplify access check for fail-nth")
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Reported-by: Lu Fengqi <lufq.fnst@cn.fujitsu.com>
Tested-by: Lu Fengqi <lufq.fnst@cn.fujitsu.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
lib/fault-inject.c

index 7d315fd..cf7b129 100644 (file)
@@ -110,10 +110,12 @@ bool should_fail(struct fault_attr *attr, ssize_t size)
        if (in_task()) {
                unsigned int fail_nth = READ_ONCE(current->fail_nth);
 
-               if (fail_nth && !WRITE_ONCE(current->fail_nth, fail_nth - 1))
-                       goto fail;
+               if (fail_nth) {
+                       if (!WRITE_ONCE(current->fail_nth, fail_nth - 1))
+                               goto fail;
 
-               return false;
+                       return false;
+               }
        }
 
        /* No need to check any other properties if the probability is 0 */