https://sourceware.org/bugzilla/show_bug.cgi?id=33852

--- Comment #7 from Alan Modra <amodra at gmail dot com> ---
(In reply to Jens Remus from comment #6)
> Did you identify this via the relocation and looking at the code?
Yes.  I've seen this sort of thing before so it didn't take long. 
MALLOC_PERTURB_ and valgrind are good tools to use.

$ cat aff.s
 .quad x@ntpoff
$ gas/as-new -m64 aff.s -o t.o
The above is enough to recreate the problem, no debug sections involved.  I'm
going to commit my fix as I'm 100% sure it is correct.  I'll also note this
problem ought to have been seen some time in the last few years with asan
testsuite runs (ld-s390/reloctlsle-1.s), but it seems asan doesn't pick up this
particular occurrence of uninitialised memory.

The bug has been there for 23 years since commit 69fc87f180.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

Reply via email to