Red Hat Bugzilla – Bug 241795
Bad numactl command hangs the system
Last modified: 2007-11-30 17:07:44 EST
Description of problem:
Executing a command like this: numactl --cpubind <N> --membind <N> <command> on
an interleaved system produces the attached register dump and the system hangs:
you can ping it but that's about it.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. On an interleaved IA64 system (I used an rx8640, but I have no
reason to believe it's any different on an rx8620 or a Superdome),
numactl --cpubind 0 --membind 0 ls
See attachment for console output.
Command should return an error.
Lee Schermerhorn tried a 2.6.22-rc2-mm1 kernel on the same machine - it behaves
properly: the command returns an error as it should and nothing bad happens.
Created attachment 155719 [details]
Was it just the kernel that was different between what you tried and what worked
for Lee or was it a different version of the numactl cmd line util also?
Doug, it was just the kernel. Everything else stayed the same.
I have verified that this is already fixed in the latest RHEL5.1 kernels.
Appears to be a dup of 229084.
*** This bug has been marked as a duplicate of 229084 ***