sched/debug: Fix potential call to __ffs(0) in sched_show_task()

"struct task_struct"->state is "volatile long" and __ffs() warns that
"Undefined if no bit exists, so code should check against 0 first."

Therefore, at expression

  state = p->state ? __ffs(p->state) + 1 : 0;

in sched_show_task(), CPU might see "p->state" before "?" as "non-zero"
but "p->state" after "?" as "zero", which could result in
"state >= sizeof(stat_nam)" being true and bogus '?' is printed.

This patch changes "state" from "unsigned int" to "unsigned long" and
save "p->state" before calling __ffs(), in order to avoid potential call
to __ffs(0).

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/201412052131.GCE35924.FVHFOtLOJOMQFS@I-love.SAKURA.ne.jp
Signed-off-by: Ingo Molnar <mingo@kernel.org>

1 files changed, 3 insertions, 2 deletions

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 56c9b79772bd..816c17203c16 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -4508,9 +4508,10 @@ void sched_show_task(struct task_struct *p)
 {
         unsigned long free = 0;
         int ppid;
-        unsigned state;
+        unsigned long state = p->state;
 
-        state = p->state ? __ffs(p->state) + 1 : 0;
+        if (state)
+                state = __ffs(state) + 1;
         printk(KERN_INFO "%-15.15s %c", p->comm,
                 state < sizeof(stat_nam) - 1 ? stat_nam[state] : '?');
 #if BITS_PER_LONG == 32