Summary: | glibc-2.3.5 on PAX system fails to build sanely - execution attempt in: <anonymous mapping> terminating task | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Michael Crawford (ali3nx) <mcrawford> |
Component: | [OLD] Core system | Assignee: | Gentoo Toolchain Maintainers <toolchain> |
Status: | RESOLVED DUPLICATE | ||
Severity: | critical | CC: | hardened |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- |
Description
Michael Crawford (ali3nx)
2005-04-27 01:54:26 UTC
grep of the kernel logfile only returned about 2000 linuxthreads processes killed by pax while compiling glibc-2.3.5. Nice!? o lastly before i turn in i just noticed this seems to have occured on the previous glibc build aswell with ld.so appearing to fail with some mysterious elf address error. ld does however still work correctly. uberboxen ~ # ld ld: no input files Mar 23 20:27:16 uberboxen PAX: terminating task: /var/tmp/portage/glibc-2.3.4.20050125-r1/work/build-default-i686-pc-linux-gnu-linuxthreads/elf/ld.so(ld.so):894, uid/euid: 0/0, PC: 5e20ae94, SP: 5e20acd4 Mar 23 20:27:19 uberboxen PAX: terminating task: /var/tmp/portage/glibc-2.3.4.20050125-r1/work/build-default-i686-pc-linux-gnu-linuxthreads/elf/ld.so(ld.so):29687, uid/euid: 0/0, PC: 5c403e24, SP: 5c403c64 Mar 23 20:27:20 uberboxen PAX: terminating task: /var/tmp/portage/glibc-2.3.4.20050125-r1/work/build-default-i686-pc-linux-gnu-linuxthreads/elf/ld.so(ld.so):6166, uid/euid: 0/0, PC: 5cdba3e4, SP: 5cdba224 Apr 27 00:19:55 uberboxen PAX: terminating task: /var/tmp/portage/glibc-2.3.5/work/build-default-i686-pc-linux-gnu-linuxthreads/elf/ld.so(ld.so):20127, uid/euid: 0/0, PC: 5cb3626c, SP: 5cb360e8 Apr 27 00:19:57 uberboxen PAX: terminating task: /var/tmp/portage/glibc-2.3.5/work/build-default-i686-pc-linux-gnu-linuxthreads/elf/ld.so(ld.so):18078, uid/euid: 0/0, PC: 5d1c662c, SP: 5d1c64a8 Apr 27 00:19:59 uberboxen PAX: terminating task: /var/tmp/portage/glibc-2.3.5/work/build-default-i686-pc-linux-gnu-linuxthreads/elf/ld.so(ld.so):12129, uid/euid: 0/0, PC: 5c718d2c, SP: 5c718ba8 glibc uses it's newly created ld.so to exec another program (rpcgen -Y) while making locales. If your system lacks the resources to build a ton of locales at one time this operation will fail with random null pointer execution as you see. Solution: use hardened profile or enable userlocales. # grep ^en_US /etc/locales.build en_US/ISO-8859-1 en_US.UTF-8/UTF-8 solar@gentoo.org: "If your system lacks the resources to build a ton of locales at one time..." ???? - I have one giga ram in my box, and no server programs run except samba. - The build process builds all of the locales simultaneously? i don't think so. - This bug shows up since the 2.3.4.20050125-r1 version. this is a fact. so the problem is in the changes of the previous versions and that version. Not our systems! "Solution: use hardened profile or enable userlocales." ??????? Your kidding? You think that i don't use the hardened profile? Please.. We use hardened profile and hardened kernel. And i tell you, if i use the userlocales use-flag, the problem is still exists. Probe it if you don't believe me. toth just a suggestion. if you expect support from developers perhaps mind your manners =] |