This is the mail archive of the mailing list for the glibc project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug libc/21049] segfault in longjmp_chk() due to clobbered processor register

--- Comment #3 from dave.anglin at bell dot net ---
On 2017-01-13 3:54 PM, deller at gmx dot de wrote:
> One question before that:
> I did built the debian glibc package with my change.
> But it failed to build because there were 2-3 bugs while running the testsuite.
> One failure was tst-cancel2 (or something like that), the other one was
> some NIS testcase (although I can understand that, because I was building
> on a machine at SAP where NIS is active and as such it doesn't behaved like
> a standard debian buildd).
I have never done a build with NIS.  tst-cancel2 doesn't normally fail.
> What makes me wonder is, that glibc package is uploaded in debian, but
> in my build I saw the failure (I'm sure the failures are not due to that
> change).
> Did you uploaded the debian glibc packages manually?
I can't quite remember but I think that I added 
malloc/tst-malloc-thread-fail to the xfail list
as it shows as an XPASS in the last build.  We also had a random segv 
building intl/tst-gettext6
in one of the attempts.  I've not seen that before.

I've reached the conclusion that non of my kernel patches are helping 
with the random segmentation
faults.  Something else is causing them and possibly using the tmpfs for 
unpacking, etc, has mad things

I do build trunk outside the buildd.  Have a patch to fix pthread_cond_t 
issue and others.  Trunk built okay
a week or so ago.  However build last night failed due to a missing 
symbol link error (think symbol
was _init).


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

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]