Tomorrow's hotplug SIG is canceled, below is a status of OSDL's
activities.
Our next meeting will be February 14th, 2006 at 11:00am PST.
With respect to getting the existing hotplug cpu regression tests
running, we are still stuck on the gating factor: a machine with enough
cpus that can be rebooted in the event of a failed kernel.
Bryce did take a look at the ppc64 last week, and reactivated the queue
processing. He also did some upgrades on the machine to keep it in sync
with some of the recent changes to his test framework.
He created some scripts to help manage the automated testing. These are
in /root/bin on dev4-013:
bin/status - Shows what all the test machines are doing
bin/lock - Locks dev4-013 so it won't run tests
bin/unlock - Unlocks dev4-013 so it'll start running tests
Mary volunteers to give them a try.
Progress is being made to hire a temp to unload some of Bryce's sys
admin responsibilities. We are currently looking for candidates and
have an interview team identified.
Meantime, Mary has modified memtoy to run without numa or memory
migration. It compiles and loads. Now she needs to experiment with
what functionality remains (see help list below). The test examples Lee
provided were mostly ones involving NUMA migration (no surprise).
One thing we could work on while we wait for the test framework to
function: a step by step for how to implement the memory
regression test plan. We can do memory regression tests on
the newer equipment we received, since it does not require more than one
cpu, just memory. By next SIG meeting Mary plans to be ready with some
knowledge of how we can use memtoy.
FYI: Memtoy (without migration and without numa) commands:
quit - just what you think
EOF on stdin has the same effect
help - show this help
help <command> - display help for just <command>
pid - show process id of this session
pause - pause program until signal -- e.g., INT, USR1
show [<name>] - show info for segment[s]; default all anon <seg-name>
<seg-size>[k|m|g|p] [<seg-share>] -
define a MAP_ANONYMOUS segment of specified size
<seg-share> := private|shared - default = private file
<pathname> [<offset>[k|m|g|p] <length>[k|m|g|p]] [<seg-share>] -
define a mapped file segment of specified length starting at the
specified offset into the file. <offset> and <length> may be
omitted and specified on the map command.
<seg-share> := private|shared - default = private shm <seg-name>
<seg-size>[k|m|g|p] -
define a shared memory segment of specified size.
You may need to increase limits [/proc/sys/kernel/shmmax].
Use map/unmap to attach/detach
remove <seg-name> [<seg-name> ...] - remove the named segment[s] map
<seg-name> [<offset>[k|m|g|p] <length>[k|m|g|p]] [<seg-share>] -
mmap()/shmat() a previously defined, currently unmapped()
segment.
<offset> and <length> apply only to mapped files.
Use <length> of '*' or '0' to map to the end of the file.
unmap <seg-name> - unmap specified segment, but remember name/size/...
touch <seg-name> [<offset>[k|m|g|p] <length>[k|m|g|p]] [read|write] -
read [default] or write the named segment from <offset> through
<offset>+<length>. If <offset> and <length> omitted, touches
all
of mapped segment.
OLS 2006 participation:
A proposal for a paper comparing the balloon driver approach to memory
hotplug for handling memory management in virtualization has been
submitted for OLS 2006, the authors are Joel Schopp, Keir Fraser and
Martine Silbermann.
Martine
