sched_clock: Disable seqlock lockdep usage in sched_clock()
authorJohn Stultz <john.stultz@linaro.org>
Thu, 2 Jan 2014 23:11:14 +0000 (15:11 -0800)
committerIngo Molnar <mingo@kernel.org>
Sun, 12 Jan 2014 09:14:00 +0000 (10:14 +0100)
Unfortunately the seqlock lockdep enablement can't be used
in sched_clock(), since the lockdep infrastructure eventually
calls into sched_clock(), which causes a deadlock.

Thus, this patch changes all generic sched_clock() usage
to use the raw_* methods.

Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
Reported-by: Krzysztof Hałasa <khalasa@piap.pl>
Signed-off-by: John Stultz <john.stultz@linaro.org>
Cc: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Cc: Willy Tarreau <w@1wt.eu>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1388704274-5278-2-git-send-email-john.stultz@linaro.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
kernel/time/sched_clock.c

index 68b79937598106918748832a396848f7fd22f870..0abb364642818e4ef842ab2f3631d15cb80400f2 100644 (file)
@@ -74,7 +74,7 @@ unsigned long long notrace sched_clock(void)
                return cd.epoch_ns;
 
        do {
-               seq = read_seqcount_begin(&cd.seq);
+               seq = raw_read_seqcount_begin(&cd.seq);
                epoch_cyc = cd.epoch_cyc;
                epoch_ns = cd.epoch_ns;
        } while (read_seqcount_retry(&cd.seq, seq));
@@ -99,10 +99,10 @@ static void notrace update_sched_clock(void)
                          cd.mult, cd.shift);
 
        raw_local_irq_save(flags);
-       write_seqcount_begin(&cd.seq);
+       raw_write_seqcount_begin(&cd.seq);
        cd.epoch_ns = ns;
        cd.epoch_cyc = cyc;
-       write_seqcount_end(&cd.seq);
+       raw_write_seqcount_end(&cd.seq);
        raw_local_irq_restore(flags);
 }