Description of Problem: Running "strace -f host www.redhat.com" never finishes, and I have to kill -9 the strace process. Version-Release number of selected component (if applicable): [bill@pikachu bill]$ rpm -qf $( which strace ) strace-4.4-3 [bill@pikachu bill]$ rpm -qf $( which host ) bind-utils-9.2.0-8 [bill@pikachu bill]$ rpm -q glibc glibc-2.2.5-30 [bill@pikachu bill]$ uname -a Linux pikachu.bchsz.org.uk 2.4.18-0.12 #1 Wed Mar 27 16:18:42 EST 2002 i686 unknown
Created attachment 51924 [details] strace output
Created attachment 51925 [details] snippet from output of "ps auxww"
killall -9 host yields the following after the strace output instead of "Killed" -- Unable to wait on inferior: No child processes Unable to wait on inferior: No child processes
This looks very similar to bug #62176
... bug #62176 which I can't read ... this is really frustrating, yanno?
*** Bug 62176 has been marked as a duplicate of this bug. ***
If I debug a process using "strace -f" and the process spawns a child and waits for it, strace would freeze. ps would show both the child and the parent in the "T" state and if will stay that way untill kill -9. I am seeing it with strace-4.4-3 glibc-2.2.5-30 kernel-2.4.18-0.13 It didn't exist in 7.2
Are you really sure this did not exist in 7.2? I just tried strace -f -o log ./ex3 (where ex3 is libc/linuxthreads/Examples/ex3.c, quite simple pthread test). strace needs to be kill -9ed with glibc-2.2.5-32+strace-4.3-2+kernel-2.4.9-21smp, glibc-2.2.4-24+strace-4.3-2+kernel-2.4.17-0.13smp and glibc-2.2.5-32+strace-4.4-3+kernel-2.4.17-0.13smp combinations.
Yes, pretty sure. My test is strace -f sh -c 'TEST=`/bin/echo xyz`; exec /bin/echo xyz' It works OK on glibc-2.2.4-19.3 kernel-smp-2.4.9-31 strace-4.3-2 and glibc-2.2.4-19.3 kernel-smp-2.4.9-21 strace-4.3-2 but gets stuck (and requires kill -9) on glibc-2.2.5-30 kernel-2.4.18-0.13 strace-4.4-3 I will go upgrade one of 7.2 machines to the latest glibc and report back.
(disregard the kernel-smp-2.4.9-31 line in the above - that machine still runs -21). Works OK: glibc-2.2.4-19.3 kernel-2.4.9-31 strace-4.2.20010119-3 (RH7.1+updates) After upgrading to glibc-2.2.4-24 both the above and the kernel-smp-2.4.9-21 strace-4.3-2 one (but not rebooting it), it still works.
This means there are two separate problems. strace -f sh -c 'TEST=`/bin/echo xyz`; exec /bin/echo xyz' which is caused by the patch which was added in strace-4.3-3, and problems with threaded programs, which existed in 7.2 and probably earlier too.
*** Bug 63370 has been marked as a duplicate of this bug. ***
I reported this problem in bug 63370 as a problem with skipjack.... Sorry -- I limited my search only to the beta release and not rawhide. Oops. In any case, since this does impact the version of strace in the beta release, does it make sense to change the Product of this bug?
Should be fixed in strace-4.4-4.
This problem appears to be fixed in RedHat 7.3, but the problem I reported in 63370, which was marked as a duplicate of this bug, has not been entirely fixed. I am reopening that bug, as it is different from this one, changing it to a Valhalla problem, and updating it with new information. I don't see why this specific bug, which no longer appears in 7.3 with strace-4.4-4, can't be closed, but I am obviously not the one who should make that judgment call being neither a RedHat person nor the original poster....
This problem appears to be fixed in limbo. Maybe someone should close it.
Fix verified, closing. (resolution set to RAWHIDE due to product version of defect, likely fixed much earlier.)