NetBSD-Bugs archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

kern/55948: threadpool_job_cancelthrash test randomly fails



>Number:         55948
>Category:       kern
>Synopsis:       threadpool_job_cancelthrash test randomly fails
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sat Jan 23 10:30:00 +0000 2021
>Originator:     Andreas Gustafsson
>Release:        NetBSD-current, source date >= 2021.01.16.23.51.51
>Organization:
>Environment:
System: NetBSD
Architecture: x86_64
Machine: amd64
>Description:

The threadpool_job_cancelthrash test case of the rump/rumpkern/t_threadpool
test program now randomly fails on my testbed running on real amd64 hardware.

Log output from one such failure is at:

  https://www.gson.org/netbsd/bugs/build/amd64-baremetal/2021/2021.01.17.21.56.20/test.html#rump_rumpkern_t_threadpool_threadpool_job_cancelthrash

It does not fail on the qemu/nvmm based TNF testbed.

I ran an automated bisection, running the test repeatedly to be sure
to catch the failure even though it happens only randomly, and this
identified the following as the commit where the problem started:

  2021.01.16.23.51.50 chs src/sys/arch/arm/arm/psci.c 1.5
  2021.01.16.23.51.50 chs src/sys/conf/files 1.1278
  2021.01.16.23.51.51 chs src/sys/lib/libkern/arch/hppa/bcopy.S 1.16
  2021.01.16.23.51.51 chs src/sys/lib/libkern/libkern.h 1.141
  2021.01.16.23.51.51 chs src/sys/sys/cdefs.h 1.156
  2021.01.16.23.51.51 chs src/sys/sys/queue.h 1.76

>How-To-Repeat:

cd /usr/tests/rump
for i in $(seq 100)
do
    if
        atf-run rumpkern >log
        ! egrep '^tc-end:.* threadpool_job_cancelthrash, passed' log
    then
        echo FAIL
        break
    fi
done

>Fix:



Home | Main Index | Thread Index | Old Index