[systemd-bugs] [Bug 86666] New: Use of gold linker breaks debugging
bugzilla-daemon at freedesktop.org
bugzilla-daemon at freedesktop.org
Mon Nov 24 10:31:26 PST 2014
https://bugs.freedesktop.org/show_bug.cgi?id=86666
Bug ID: 86666
Summary: Use of gold linker breaks debugging
Product: systemd
Version: unspecified
Hardware: x86-64 (AMD64)
OS: All
Status: NEW
Severity: normal
Priority: medium
Component: general
Assignee: systemd-bugs at lists.freedesktop.org
Reporter: peter at lekensteyn.nl
QA Contact: systemd-bugs at lists.freedesktop.org
Among the problems that the gold linker brings:
- The debugging section gets dropped, making the /usr/lib/debug/ files
useless.
- The .eh_frame section is bogus[1], resulting in crashes in
libunwind[2] and a truncated backtrace in gdb.
The issue "fixed" in that commit is gone as IFUNC is not used anymore
since version 210), so in theory we could drop the -Wl,-fuse-ld=gold
line...
...but in practice we run into problems because GNU ld sees that
symbols appear twice, once in the libsystemd-{internal,shared}.la,
and once in the compatibility library (libsystemd-id128.la).
I tried to remove libsystemd-internal.la from libsystemd_id128_LIBADD,
but then all symbols would disappear, leaving a pretty empty
libsystemd-id128.so file. (This presumably happens because there were
missing references.)
Now I would like to revert to the state before v208-1653-g39c4ead,
but meanwhile there have been a lot of other changes to the build system.
What would it take to avoid linking two archives with the same symbol?
That would also avoid problems such as
https://sourceware.org/bugzilla/show_bug.cgi?id=16504
since then there would only be one symbol.
[1]: https://sourceware.org/bugzilla/show_bug.cgi?id=17639
[2]:
http://lists.nongnu.org/archive/html/libunwind-devel/2014-11/msg00009.html
--
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/systemd-bugs/attachments/20141124/2a5e6c40/attachment.html>
More information about the systemd-bugs
mailing list