Description of problem: ia32el doesn't use kernel parameters related to core file when ia32el creates a elf32 core file. Version-Release number of selected component (if applicable): ia32el-1.1-13.ia64.rpm How reproducible: always Steps to Reproduce: 1. set kernel parameters # echo /tmp/mycore-%e-%p-%t > /proc/sys/kernel/core_pattern # echo 0 > /proc/sys/kernel/core_uses_pid # # cat /proc/sys/kernel/core_pattern /tmp/mycore-%e-%p-%t # cat /proc/sys/kernel/core_uses_pid 0 2. compile test program <elf64> compile on ia64 machine # cat test.c main() { int *p, *q; p=0; *q=*p; } # cc -o test_ia64 test.c <elf32> compile on i32 machine # cc -o test_ia32 test.c 3. run test program # pwd /var/tmp/work # ./test_ia64 Segmentation fault (core dumped) # ./test_ia32 Segmentation fault (core dumped) # ls /tmp mycore-test_ia64-17731-1104145866 # ls /var/tmp/work core.17732 Actual results: # ls /var/tmp/work core.17732 Expected results: # ls /tmp mycore-test_ia32-17732-xxxxxxxxxxx Additional info:
*** Bug 143767 has been marked as a duplicate of this bug. ***
The severity and priority of this entry are both set to normal. Is this correct ? Thanks, Yoav.
Maybe low is more appropriate, not sure.
we already have a fix for this problem internaly, but it obviously won't make it into RHEL4 at this point. i suggest we resolve this entry with 'closed deffered'. any objections to that ? thanks, yoav.
It doesn't matter to me. But could you tell me when the fixed version is released? Is it included in RHEL4 RC or GM? Thanks,
it missed the train for RHEL4. it will probably be distributed in one of the updates. thanks, yoav.
I guess it is a RHEL4 U1 candidate. For that we should keep the bug open though and as soon as bugzilla has RHEL4 as Product/Version pair this bug should be moved there.
Add RHEL4-L3support team in Japan
Has the fixed version been registered as RHEL4 U1 candidate? Thanks,
ia32el-1.2-2 has the changes in.
Changing this to MODIFIED so that we can test. The package is built into 4E-lacd-U1.
Unfortunately what is in ia32el-1.2-3 doesn't match kernel's core_pattern handling. ia32el handles core_pattern as a string and if /proc/sys/kernel/core_uses_pid is not 0, appends .%p expanded to it. But kernel expands %h in /proc/sys/kernel/core_pattern into hostname, %p into pid, %u into uid, etc., and if /proc/sys/kernel/core_uses_pid is not 0 or the process has multiple threads, appends pid if core_pattern did not contain %p. I think for U1 we need to keep this as is, but would be good to fix it for U2.
Moving this to the U2 tracker list.
I believe this issue is fixed in U2. Fujitsu - can you please verify and close ? Jay - I could not find the corresponding tracker. Do you know what it is ? Thanks, Yoav.
It is definitely not fixed in U2, nor in 5.3.102.39.23. There were some changes in U1, but not sufficient, see comment #13. I haven't looked into newer sources yet to see if it got fixed there.
Your'e right - the fix was in 5.3.104.39.24. Sorry for the confusion. Do you know what's the issue-tracker # for this ? Thanks, Yoav.
IT#66088
ia32el-1.3-2 built into 4E-lacd-qu-candidate.
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on the solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHBA-2006-0131.html