locking/lockdep: Fix overflow in presentation of average lock-time
authorChris Wilson <chris@chris-wilson.co.uk>
Sat, 25 Jul 2020 18:51:10 +0000 (19:51 +0100)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Thu, 3 Sep 2020 09:21:18 +0000 (11:21 +0200)
[ Upstream commit a7ef9b28aa8d72a1656fa6f0a01bbd1493886317 ]

Though the number of lock-acquisitions is tracked as unsigned long, this
is passed as the divisor to div_s64() which interprets it as a s32,
giving nonsense values with more than 2 billion acquisitons. E.g.

  acquisitions   holdtime-min   holdtime-max holdtime-total   holdtime-avg
  -------------------------------------------------------------------------
    2350439395           0.07         353.38   649647067.36          0.-32

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20200725185110.11588-1-chris@chris-wilson.co.uk
Signed-off-by: Sasha Levin <sashal@kernel.org>
kernel/locking/lockdep_proc.c

index 75d8080..09bad6c 100644 (file)
@@ -425,7 +425,7 @@ static void seq_lock_time(struct seq_file *m, struct lock_time *lt)
        seq_time(m, lt->min);
        seq_time(m, lt->max);
        seq_time(m, lt->total);
-       seq_time(m, lt->nr ? div_s64(lt->total, lt->nr) : 0);
+       seq_time(m, lt->nr ? div64_u64(lt->total, lt->nr) : 0);
 }
 
 static void seq_stats(struct seq_file *m, struct lock_stat_data *data)