linux-rt-users.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Paul Gortmaker <paul.gortmaker@windriver.com>
To: <bigeasy@linutronix.de>
Cc: <linux-rt-users@vger.kernel.org>
Subject: [PATCH rt-v4.11] sched/clock: fix early boot splat on clock transition to unstable
Date: Mon, 2 Oct 2017 23:57:19 -0400	[thread overview]
Message-ID: <1507003039-16583-1-git-send-email-paul.gortmaker@windriver.com> (raw)

On an older machine with a Pentium(R) Dual-Core E5300 I see the
the following early (see time stamps) boot splat on clock transition
due to TSC unstable (indicated in the last line):

  [  2.487904] BUG: using smp_processor_id() in preemptible [00000000] code: swapper/0/1
  [  2.487909] caller is debug_smp_processor_id+0x17/0x20
  [  2.487911] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.11.12-rt14-00451
  [  2.487911] Hardware name: Dell Inc. OptiPlex 760                 /0M858N, BIOS A16 08/06/2013
  [  2.487912] Call Trace:
  [  2.487918]  dump_stack+0x4f/0x6a
  [  2.487919]  check_preemption_disabled+0xda/0xe0
  [  2.487921]  debug_smp_processor_id+0x17/0x20
  [  2.487924]  clear_sched_clock_stable+0x28/0x80
  [  2.487927]  mark_tsc_unstable+0x22/0x70
  [  2.487930]  acpi_processor_get_power_info+0x3e3/0x6a0
  [  2.487932]  acpi_processor_power_init+0x3a/0x1d0
  [  2.487933]  __acpi_processor_start+0x162/0x1b0
               ....
  [  2.487950]  acpi_processor_driver_init+0x20/0x96
  [  2.487951]  do_one_initcall+0x3f/0x170
  [  2.487954]  kernel_init_freeable+0x18e/0x216
  [  2.487955]  ? rest_init+0xd0/0xd0
  [  2.487956]  kernel_init+0x9/0x100
  [  2.487958]  ret_from_fork+0x22/0x30
  [  2.487960] sched_clock: Marking unstable (2488005383, -223143)<-(2590866395, -103084155)
  [  2.488004] tsc: Marking TSC unstable due to TSC halts in idle

(gdb) list *clear_sched_clock_stable+0x28
0xffffffff8108bbb8 is in clear_sched_clock_stable (kernel/sched/clock.c:114).

[...]

112     static inline struct sched_clock_data *this_scd(void)
113     {
114             return this_cpu_ptr(&sched_clock_data);
115     }

We now get this_scd with preemption disabled.  I also decided to pass
in the scd to __clear_sched_clock_stable in the hope it made it more
clear that the caller (currently only one) needs to get this_scd with
preemption disabled, even though that wasn't strictly required.

Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>

diff --git a/kernel/sched/clock.c b/kernel/sched/clock.c
index 11ad4bd995e2..32dcda23c616 100644
--- a/kernel/sched/clock.c
+++ b/kernel/sched/clock.c
@@ -155,9 +155,8 @@ static void __sched_clock_work(struct work_struct *work)
 
 static DECLARE_WORK(sched_clock_work, __sched_clock_work);
 
-static void __clear_sched_clock_stable(void)
+static void __clear_sched_clock_stable(struct sched_clock_data *scd)
 {
-	struct sched_clock_data *scd = this_scd();
 
 	/*
 	 * Attempt to make the stable->unstable transition continuous.
@@ -186,8 +185,14 @@ void clear_sched_clock_stable(void)
 
 	smp_mb(); /* matches sched_clock_init_late() */
 
-	if (sched_clock_running == 2)
-		__clear_sched_clock_stable();
+	if (sched_clock_running == 2) {
+		struct sched_clock_data *scd;
+
+		preempt_disable();
+		scd = this_scd();
+		preempt_enable();
+		__clear_sched_clock_stable(scd);
+	}
 }
 
 void sched_clock_init_late(void)
-- 
2.1.4


             reply	other threads:[~2017-10-03  3:57 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-10-03  3:57 Paul Gortmaker [this message]
2017-10-05 14:59 ` [PATCH rt-v4.11] sched/clock: fix early boot splat on clock transition to unstable Julia Cartwright
2017-10-06 10:09 ` Sebastian Andrzej Siewior
2017-10-06 12:31   ` Sebastian Andrzej Siewior
2017-10-11 13:30     ` Paul Gortmaker
2017-10-17 15:36       ` Sebastian Andrzej Siewior

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1507003039-16583-1-git-send-email-paul.gortmaker@windriver.com \
    --to=paul.gortmaker@windriver.com \
    --cc=bigeasy@linutronix.de \
    --cc=linux-rt-users@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).