[PATCH 0 of 5] RFC: Handling of sparse, memoryless, and cpuless nodes

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



In response to a bug report where 'numactl --hardware' was segfaulting
with distro 2.0.3 version I've dug into the current state of
libnuma/numactl after not having looked at it in quite some time.  The
following series of patches resulted and hopefully clarifies things a bit
for the next person who comes along and accounts for the possibility of
sparsely numbered, memoryless and cpuless nodes in 'numactl --hardware'.
There are certainly some things in the patches that aren't quite optimal
and other things I saw that I didn't try to tackle (eg: there is a lot
of seeming inconsistency around the different "all" node pointers),
but the updated code works for me and I was trying to limit changes to
not start driving in the direction of a v3 library (or hwloc?).

The first 4 patches are relatively minor cleanups.

The bulk is in patch 5 which then deals with better handling sparse node
numbering and fixes parse_numbers() in particular which was experiencing
a buffer overflow in the face of this topology:

available: 2 nodes (0,3)
node 0 cpus: none
node 0 size: 0 MB
node 0 free: 0 MB
node 3 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
node 3 size: 15360 MB
node 3 free: 12524 MB
node distances:
node 0 3
 0: 10 40
 3: 40 10

I've tested the changes on "interesting" topology hardware to which I
have access, but would greatly appreciate if others might do similar.

-- 
Tim Pepper  <lnxninja@xxxxxxxxxxxxxxxxxx>
IBM Linux Technology Center
--
To unsubscribe from this list: send the line "unsubscribe linux-numa" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html


[Index of Archives]     [Linux Kernel]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux SCSI]     [Devices]

  Powered by Linux