Messages in this thread |  | | Date | Sat, 21 Jun 2014 09:30:30 -0400 | Subject | More 3.16 vdso fallout | From | Josh Boyer <> |
| |
We've had a report[1] of d-bus getting a SEGV only on i686 machines with the 3.16 kernel. It's a bit light on specifics, but the relevant detail seems to be:
"I'm pretty sure kernel is to blame here.
The crash occurs when signal handler for RT signal 33 (used internaly by NPTL/pthreads) returns to address zero (not sure why).
This only happens with 3.16 kernel, works fine with 3.15."
and
"Seems to be related to v3.16's arch/x86 VDSO changes. Works fine when booting with vdso=0."
There isn't an actual backtrace for the SEGV in the bug, but we can ask for one. The kernels in question should be very recent, but lack the 5 patch series Andy recently sent as those aren't in Linus' tree yet. Not sure if those would impact anything or not.
josh
[1] https://bugzilla.redhat.com/show_bug.cgi?id=1110968
|  |