Red Hat Bugzilla – Bug 456443
ptrace: reparent-zombie testcase FAILs
Last modified: 2009-05-19 11:43:07 EDT
Description of problem:
The testcase program
should PASS as on F-9 (utrace) kernels and (reportedly) upstream 2.6.26+.
(/proc/PID/status should not be left Zombie for parents with SIG_IGN SIGCHLD.)
Version-Release number of selected component (if applicable):
Steps to Reproduce:
wget -O reparent-zombie.c
gcc -o reparent-zombie reparent-zombie.c -Wall -ggdb2 -D_GNU_SOURCE -pthread;
./reparent-zombie; echo $?
24649 left in State: Z (zombie)
It was reported by Roland at http://sourceware.org/systemtap/wiki/utrace/tests
as "fixed upstream after 2.6.26".
reparent-zombie-clone testcase also FAILs but only with the same result as this
Bug's testcase reparent-zombie with no kernel crash. IMO kernel crash was the
reason for reparent-zombie-clone so it should get fixed along by this Bug.
The bug is not reproducible not only in upstream kernel (2.6.27-rc1), but in F-9
(188.8.131.52-97) and RHEL-5 (2.6.18-92.1.6) also. I see it only in RHEL-4. Are you
sure about 2.6.26+?
(In reply to comment #1)
> The bug is not reproducible not only in upstream kernel (2.6.27-rc1), but in
> F-9 (184.108.40.206-97) and RHEL-5 (2.6.18-92.1.6) also. I see it only in RHEL-4.
> Are you sure about 2.6.26+?
(My comment #0)
> It was reported by Roland at http://sourceware.org/systemtap/wiki/utrace/tests
> as "fixed upstream after 2.6.26".
I do not see any contradictions.
I agree with you that:
RHEL-4 is broken - This Bug.
RHEL-5 is OK (expecting due to utrace - instead of upstream ptrace)
F-9 is OK (expecting due to utrace - instead of upstream ptrace)
upstream 2.6.27-rc1 is OK (as Roland said it is fixed upstream after 2.6.26)
My test on 220.127.116.11-55.vanilla.fc9.x86_64 shows it is broken (as expected)
Updating PM score.
move back to ASSI, the patch posted need follow up after Roland's feedback.