Revision be14eb619108fa8b7120eb2c42d66d5f623ae10e authored by Brian King on 10 September 2010, 07:03:21 UTC, committed by Jens Axboe on 10 September 2010, 07:03:21 UTC
While testing CPU DLPAR, the following problem was discovered.
We were DLPAR removing the first CPU, which in this case was
logical CPUs 0-3. CPUs 0-2 were already marked offline and
we were in the process of offlining CPU 3. After marking
the CPU inactive and offline in cpu_disable, but before the
cpu was completely idle (cpu_die), we ended up in __make_request
on CPU 3. There we looked at the topology map to see which CPU
to complete the I/O on and found no CPUs in the cpu_sibling_map.
This resulted in the block layer setting the completion cpu
to be NR_CPUS, which then caused an oops when we tried to
complete the I/O.

Fix this by sanity checking the value we return from blk_cpu_to_group
to be a valid cpu value.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
1 parent edce682
History
File Mode Size
Kconfig -rw-r--r-- 3.0 KB
Makefile -rw-r--r-- 587 bytes
associola.c -rw-r--r-- 45.2 KB
auth.c -rw-r--r-- 23.9 KB
bind_addr.c -rw-r--r-- 13.8 KB
chunk.c -rw-r--r-- 9.6 KB
command.c -rw-r--r-- 2.3 KB
debug.c -rw-r--r-- 4.6 KB
endpointola.c -rw-r--r-- 13.0 KB
input.c -rw-r--r-- 31.0 KB
inqueue.c -rw-r--r-- 7.0 KB
ipv6.c -rw-r--r-- 27.9 KB
objcnt.c -rw-r--r-- 4.2 KB
output.c -rw-r--r-- 22.2 KB
outqueue.c -rw-r--r-- 53.2 KB
primitive.c -rw-r--r-- 7.6 KB
probe.c -rw-r--r-- 5.1 KB
proc.c -rw-r--r-- 13.5 KB
protocol.c -rw-r--r-- 37.0 KB
sm_make_chunk.c -rw-r--r-- 98.7 KB
sm_sideeffect.c -rw-r--r-- 47.4 KB
sm_statefuns.c -rw-r--r-- 190.2 KB
sm_statetable.c -rw-r--r-- 33.4 KB
socket.c -rw-r--r-- 186.8 KB
ssnmap.c -rw-r--r-- 3.5 KB
sysctl.c -rw-r--r-- 6.8 KB
transport.c -rw-r--r-- 19.3 KB
tsnmap.c -rw-r--r-- 9.7 KB
ulpevent.c -rw-r--r-- 29.6 KB
ulpqueue.c -rw-r--r-- 27.9 KB

back to top