Summary: | dev-lang/php-5.1.1 cli killed in hardened system | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Steve Yin <steve> |
Component: | New packages | Assignee: | PHP Bugs <php-bugs> |
Status: | RESOLVED TEST-REQUEST | ||
Severity: | major | ||
Priority: | High | ||
Version: | 2005.1 | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- |
Description
Steve Yin
2006-01-06 00:10:23 UTC
Hardened has nothing (apparent) to do with the crash, grsecurity is just reporting system events, see: http://www.grsecurity.net/wiki/index.php/GrsecurityFAQ Feel free to re-assign if anything turns up showing a link between hardened and the cause of the event causing php to attempt to dump core. Today, I found this: on my normal system(not using hardened toolchain), the php cli will segfault on it's exit with both php-5.1.1 and php-5.1.2(tested use overlay). So, Seems This is not a hardened problem, just php cli problem, seg fault on normal system, killed by hardened system. (In reply to comment #2) > Today, I found this: on my normal system(not using hardened toolchain), the php > cli will segfault on it's exit with both php-5.1.1 and php-5.1.2(tested use > overlay). I definitely don't get any CLI segfaults with php-5.1.1 on either hardened or unhardened system. Recompile your toolchain and dev-lang/php without -DNDEBUG and similar unsupported cruft in your C[XX]FLAGS and reopen if the issue still persists then. This makes no sense to me, I've rebuilt my php with CFLAGS="-march=nocona -O2 -pipe -fomit-frame-pointer", but it still killed by hardened system. So, if this is a problem of -DNDEBUG, why the apache2-sapi works really really fine? The whole system works perfect, only php-cli get seg fault? This both happend to my amd64 and x86, 3 server, plus 2 non-hardened desktop all the same, All built with -DNDEBUG, only >=php-5.1.1 get this problem. |