| Summary: | >=sys-libs/glibc-2.7-r2 : SIGSEGV in gettimeofday | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | wbrana |
| Component: | New packages | Assignee: | Gentoo Toolchain Maintainers <toolchain> |
| Status: | RESOLVED NEEDINFO | ||
| Severity: | normal | CC: | csa |
| Priority: | High | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Attachments: |
strace top
dmesg kernel config strace top |
||
|
Description
wbrana
2008-06-10 19:02:10 UTC
How about running strace `which top` Did you upgrade anything else, perhaps sys-kernel/linux-headers? In what order? Did you try if rebuilding sys-process/procps and x11-apps/mesa-progs helps? You may also want to post dmesg for that system and perhaps your kernel config. Created attachment 156401 [details]
strace top
(In reply to comment #2) > Did you upgrade anything else, perhaps sys-kernel/linux-headers? In what order? > Did you try if rebuilding sys-process/procps and x11-apps/mesa-progs helps? > I upgraded in following order 1. linux-headers-2.6.25-r4 2. glibc-2.8 3. procps with debugging symbols 4. mesa-progs Crash also doesn't happen with reemerged glibc 2.6.1. Created attachment 156523 [details]
dmesg
Created attachment 156525 [details]
kernel config
"Reopen bug" comment Created attachment 160639 [details]
strace top
I have had the same thing happen. My box uses hardened, however. I believe this is something related to the kernel. I have two hardened servers on glibc-2.7-r1, one with hardened-sources-2.6.24-r2 and the other with hs-2.6.25-r1, neither experience this. When I upgraded to hs-2.6.25 (which pulls in 2.6.25.10) top segfaults on me now. `strace top` in attachment above. (In reply to comment #11) > hs-2.6.25-r1, neither experience this. When I upgraded to hs-2.6.25 (which > pulls in 2.6.25.10) top segfaults on me now. `strace top` in attachment above. This should read "When I upgraded to hs-2.6.25-r2..." Apologies. SIGSEGV disappeared when I compiled glibc 2.8 using gcc 3.4.6 instead of 4.2.4. test with glibc-2.8 and gcc-4.3.2 ... I had the same problem. Seems there is Kernel/GLibc incompatibility. Using gettimeofday with not NULL second parameter causes segmentation fault if >=Glibc-2.7 is used with kernel 2.6.25 (haven't check other versions). Kernel upgrade to 2.6.28 have solved problem for me. |