Summary: | dev-util/strace-4.5.18 - segmentation fault in __canonicalize_funcptr_for_compare () | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Jeroen Roovers (RETIRED) <jer> |
Component: | Current packages | Assignee: | HPPA Porters <hppa> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | luckyluke |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
URL: | https://sourceforge.net/mailarchive/forum.php?thread_name=119aab440909091113jc5dc5f0m33f3be60bdd13eb5%40mail.gmail.com&forum_name=strace-devel | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | |||
Bug Blocks: | 241168 | ||
Attachments: |
Relevant dmesg output
strace's own output gdb output with strace built with -O0 ebuild patch using a git clone and the patch from the URL |
Description
Jeroen Roovers (RETIRED)
2008-12-11 18:04:48 UTC
Created attachment 174978 [details]
Relevant dmesg output
I'm not able to reproduce anymore with 2.6.28. Can you test too and confirm it works ? (In reply to comment #2) > I'm not able to reproduce anymore with 2.6.28. Can you test too and confirm it > works ? No, and I am still seeing it with 2.6.29_rc1, so could you please post your `emerge --info'? I am particularly interested in gcc/glibc versions. Created attachment 179190 [details]
strace's own output
Actually, strace -f segfault. Without -f it does work. (In reply to comment #4) > strace's own output strace /bin/ls will fail here at the same point (after futex() call). It segfaults with AND without the -f switch. linux-2.6.28.1 32bit UP (with linux-vserver patchset) on HPPA 2.0 (C3600) strace-4.5.18, gcc-4.2.4-r1, glibc-2.7-r2 Tried CFLAGS=CXXFLAGS="-pipe -O2 -march=2.0" and CFLAGS=CXXFLAGS="-pipe" with identical results (Segmentation fault). For comparison, the same strace-4.5.18 seems unaffected on hardened x86_64 (xen-sources-2.6.18-r12, gcc-3.4.6-r2, glibc-2.6.1) Created attachment 179357 [details]
gdb output with strace built with -O0
Kyle has done some where ... can you test these patches ? http://thread.gmane.org/gmane.linux.ports.parisc/256/focus=1661 (In reply to comment #8) > Kyle has done some where ... can you test these patches ? > > http://thread.gmane.org/gmane.linux.ports.parisc/256/focus=1661 The second patch[2] doesn't appear to be complete. It fails to apply in the first place, and after rewriting it and applying, I get: elmer ~ # strace ls -l upeek: ptrace(PTRACE_PEEKUSER,323608,104,0): No such process Compiling with only [1] gives me the same old segmentation faults. [1] http://userweb.kernel.org/~kyle/strace-add-hppa-syscalls.diff [2] http://userweb.kernel.org/~kyle/strace-fix-hppa-syscalls.diff If only the syscalls need updating, please tell me how to do it... the hppa guys have debugged this and posted a patch Thanks. The patch doesn't apply against 4.5.18 so we'd need some new version first. Created attachment 204114 [details, diff]
ebuild patch using a git clone and the patch from the URL
OK, a clone of the current git repo with the patch in the URL gives me a strace that runs well. I decided to call it 4.5.19_p20090914. Is that a good version to use in the tree?
strace-4.5.19 in the tree and it should work fine |