Summary: | <sys-libs/glibc-2.11.1 has a race in free() with malloc checking enabled | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Togge <togge.gentoo> |
Component: | [OLD] Core system | Assignee: | Gentoo Toolchain Maintainers <toolchain> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | axiator, betelgeuse, cyp561, daggs, deduktionstheorem, esigra, holger, ppurka, tampakrap |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
URL: | http://sourceware.org/bugzilla/show_bug.cgi?id=10282 | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | |||
Bug Blocks: | 289342 |
Description
Togge
2009-11-17 18:39:04 UTC
*** Bug 293552 has been marked as a duplicate of this bug. *** this isnt a critical issue if it only occurs when malloc checking is enabled (In reply to comment #2) > this isnt a critical issue if it only occurs when malloc checking is enabled > I believe malloc checking is enabled by default though. i dont think so ... the code does: s = getenv("MALLOC_CHECK_"); if(s && s[0]) { mALLOPt(M_CHECK_ACTION, (int)(s[0] - '0')); if (check_action != 0) __malloc_check_init(); } if MALLOC_CHECK_ isnt set in the environment (which nowhere in Gentoo does that), then the racy code isnt executed Well some how it is being run, I am hit by this bug everytime I start qt creator or kdevelop. Others report that it affects them on every time they use multimedia players. Could it be that the env var is being set by individual packages and that this then pollutes the environment? Sorry about the spam did not intend to change back to critical. (In reply to comment #5) > Well some how it is being run, I am hit by this bug everytime I start qt > creator or kdevelop. Others report that it affects them on every time they use > multimedia players. Could it be that the env var is being set by individual > packages and that this then pollutes the environment? > If your using a development version of kde it should be getting set by startkde. Thats how I'm hitting it here. I added the patch to kde-testing overlay, in masked sys-libs/glibc-2.11-r1, since this bug affects mostly the KDE development versions (snapshots and live ebuilds). We expect a new glibc in the end of the month, i guess there is no need to move it to tree and cause upgrade again and again to so many users. there's no need to create your own ebuilds. just put the patches you want into the normal /etc/portage/patches/ dir. added the patch to the 2.10 and 2.11 queues *** Bug 297319 has been marked as a duplicate of this bug. *** *** Bug 298375 has been marked as a duplicate of this bug. *** (In reply to comment #9) > there's no need to create your own ebuilds. just put the patches you want into > the normal /etc/portage/patches/ dir. > > added the patch to the 2.10 and 2.11 queues > where can one find that patch? this problem causes my nerves to spike... *** Bug 298375 has been marked as a duplicate of this bug. *** ok, assuming that the patch exists, it doesn't solve the issue, I have the same problem with kdevelop from svn. |