Summary: | media-sound/mumble: mumble-overlay wrapper crashes when sys-libs/glibc is built w/out SySV hashing (DT_HASH) | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Daniel Scharrer <daniel> |
Component: | Current packages | Assignee: | Lars Wendler (Polynomial-C) (RETIRED) <polynomial-c> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | ionen, pravorskyi, sam |
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
URL: | https://github.com/mumble-voip/mumble/issues/1709 | ||
See Also: |
https://bugs.gentoo.org/show_bug.cgi?id=347761 https://bugs.gentoo.org/show_bug.cgi?id=864100 https://bugs.gentoo.org/show_bug.cgi?id=863863 |
||
Whiteboard: | |||
Package list: | Runtime testing required: | --- |
Description
Daniel Scharrer
2014-10-30 17:29:23 UTC
there's no need to walk the hash table directly. use dlsym("dlsym", RTLD_NEXT) in order to find the original symbol. i think that fix should go to the mumble guys. on the toolchain side, this change might not have been intentional, but if this is the first problem we've seen in the ~2.5 years that this has been active, then let's just go with it. having DT_HASH available when DT_GNU_HASH is usable is just a waste of disk/memory. > use dlsym("dlsym", RTLD_NEXT) in order to find the original symbol.
For that you need the original dlsym, which is exactly what they are trying to locate here.
Is there an upstream bug filed to understand DT_GNU_HASH sections? Bug confirmed, thank you very much for the detailed report! It's fixed in https://github.com/mumble-voip/mumble/pull/4534. Sorry for the huge delay. The issue was reported to us back in 2015 (https://github.com/mumble-voip/mumble/issues/1709) but it went unnoticed until this year. Thank you! Fix was merged upstream a year ago. The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=e5afbd004d49ecaa3a05b192a8bb0c21ea9d2f0e commit e5afbd004d49ecaa3a05b192a8bb0c21ea9d2f0e Author: Sam James <sam@gentoo.org> AuthorDate: 2022-04-19 22:08:23 +0000 Commit: Sam James <sam@gentoo.org> CommitDate: 2022-04-19 22:08:23 +0000 sys-libs/glibc: re-enable DT_HASH It ends up breaking applications in unexpected ways (and it's not obvious to the user what's happening), and after speaking to upstream, it's not really even on their agenda to remove DT_HASH right now. Bug: https://bugs.gentoo.org/347761 Bug: https://bugs.gentoo.org/527504 Signed-off-by: Sam James <sam@gentoo.org> sys-libs/glibc/glibc-9999.ebuild | 5 ----- 1 file changed, 5 deletions(-) The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=8afecc68b8d689dbfdbff3b16ca50be66deb3cce commit 8afecc68b8d689dbfdbff3b16ca50be66deb3cce Author: Sam James <sam@gentoo.org> AuthorDate: 2022-04-24 23:32:16 +0000 Commit: Sam James <sam@gentoo.org> CommitDate: 2022-04-24 23:32:43 +0000 sys-libs/glibc: re-enable DT_HASH It ends up breaking applications in unexpected ways (and it's not obvious to the user what's happening), and after speaking to upstream, it's not really even on their agenda to remove DT_HASH right now. Given this seems to break not just some relatively niche (sorry!) things, but some prominent Steam games now too, let's accelerate plans to roll this out and do 2.34 (and 2.35, but that was more on the cards anyway) too. Bug: https://bugs.gentoo.org/347761 Bug: https://bugs.gentoo.org/527504 Bug: https://github.com/anyc/steam-overlay/issues/309 See: e5afbd004d49ecaa3a05b192a8bb0c21ea9d2f0e Signed-off-by: Sam James <sam@gentoo.org> sys-libs/glibc/glibc-2.34-r12.ebuild | 1617 ++++++++++++++++++++++++++++++++++ sys-libs/glibc/glibc-2.35-r4.ebuild | 1596 +++++++++++++++++++++++++++++++++ 2 files changed, 3213 insertions(+) |