]> git.proxmox.com Git - mirror_ubuntu-kernels.git/commitdiff
kernel/debug/debug_core.c: more properly delay for secondary CPUs
authorDouglas Anderson <dianders@chromium.org>
Wed, 14 Dec 2016 23:05:49 +0000 (15:05 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Thu, 15 Dec 2016 00:04:08 +0000 (16:04 -0800)
We've got a delay loop waiting for secondary CPUs.  That loop uses
loops_per_jiffy.  However, loops_per_jiffy doesn't actually mean how
many tight loops make up a jiffy on all architectures.  It is quite
common to see things like this in the boot log:

  Calibrating delay loop (skipped), value calculated using timer
  frequency.. 48.00 BogoMIPS (lpj=24000)

In my case I was seeing lots of cases where other CPUs timed out
entering the debugger only to print their stack crawls shortly after the
kdb> prompt was written.

Elsewhere in kgdb we already use udelay(), so that should be safe enough
to use to implement our timeout.  We'll delay 1 ms for 1000 times, which
should give us a full second of delay (just like the old code wanted)
but allow us to notice that we're done every 1 ms.

[akpm@linux-foundation.org: simplifications, per Daniel]
Link: http://lkml.kernel.org/r/1477091361-2039-1-git-send-email-dianders@chromium.org
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Reviewed-by: Daniel Thompson <daniel.thompson@linaro.org>
Cc: Jason Wessel <jason.wessel@windriver.com>
Cc: Brian Norris <briannorris@chromium.org>
Cc: <stable@vger.kernel.org> [4.0+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
kernel/debug/debug_core.c

index 0874e2edd2756bcbe5542a496fb6b131b092b3ea..79517e5549f119aff663ff5572507c70671a2b41 100644 (file)
@@ -598,11 +598,11 @@ return_normal:
        /*
         * Wait for the other CPUs to be notified and be waiting for us:
         */
-       time_left = loops_per_jiffy * HZ;
+       time_left = MSEC_PER_SEC;
        while (kgdb_do_roundup && --time_left &&
               (atomic_read(&masters_in_kgdb) + atomic_read(&slaves_in_kgdb)) !=
                   online_cpus)
-               cpu_relax();
+               udelay(1000);
        if (!time_left)
                pr_crit("Timed out waiting for secondary CPUs.\n");