[PATCH 0/3] watchdog/softlockup: Make softlockup reports more reliable and useful

From: Petr Mladek
Date: Mon Aug 19 2019 - 06:47:39 EST


( Resending this as a proper patch with updated commit messages.
The original was
https://lkml.kernel.org/r/20190605140954.28471-1-pmladek@xxxxxxxx )

We were analyzing logs with several softlockup reports in flush_tlb_kernel_range().
They were confusing. Especially it was not clear whether it was deadlock,
livelock, or separate softlockups.

It went out that even a simple busy loop:

while (true)
cpu_relax();

is able to produce several softlockups reports:

[ 168.277520] watchdog: BUG: soft lockup - CPU#1 stuck for 22s! [cat:4865]
[ 196.277604] watchdog: BUG: soft lockup - CPU#1 stuck for 22s! [cat:4865]
[ 236.277522] watchdog: BUG: soft lockup - CPU#1 stuck for 23s! [cat:4865]
^^^^

This patchset fixes the problem in two steps:

+ 1st patch prevents restart of the watchdog from unrelated locations.
Each softlockup is reported only once:

[ 320.248948] watchdog: BUG: soft lockup - CPU#2 stuck for 26s! [cat:4916]


+ 2nd patch helps to distinguish several possible situations by
regular reports. The report looks like:

[ 480.372418] watchdog: BUG: soft lockup - CPU#2 stuck for 26s! [cat:4943]
[ 508.372359] watchdog: BUG: soft lockup - CPU#2 stuck for 52s! [cat:4943]
[ 548.372359] watchdog: BUG: soft lockup - CPU#2 stuck for 89s! [cat:4943]
[ 576.372351] watchdog: BUG: soft lockup - CPU#2 stuck for 115s! [cat:4943]
^^^^^

3rd patch provides a sample code to trigger a softlockup.


Petr Mladek (3):
watchdog/softlockup: Preserve original timestamp when touching
watchdog externally
watchdog/softlockup: Report the same softlockup regularly
Test softlockup

fs/proc/consoles.c | 5 ++++
fs/proc/version.c | 7 +++++
kernel/watchdog.c | 85 +++++++++++++++++++++++++++++++-----------------------
3 files changed, 61 insertions(+), 36 deletions(-)

--
2.16.4