Mary, You should also post that to lhms which I think is a more appropriate target for this request than lhcs. Thanks - Martine -----Original Message----- From: hotplug_sig-bounces@xxxxxxxxxxxxxx [mailto:hotplug_sig-bounces@xxxxxxxxxxxxxx] On Behalf Of Mary Edie Meredith Sent: Wednesday, October 12, 2005 4:44 PM To: lhcs-devel@xxxxxxxxxxxxxxxxxxxxx Cc: Hotplug SIG Subject: [Hotplug_sig] Memory Hotplug regression testing In order to create some effective regression tests for memory hotplug, the Hotplug SIG has been doing some test planning and test framework construction. What we have is a test framework that should be easy to down-load and run for developers (see Bryce's announcement for the first version that already has cpu hotplug tests included). http://lists.osdl.org/pipermail/hotplug_sig/2005-September/000504.html Based on how the memory hotplug patches are moving into the kernel, we will create tests first for Fragmentation elimination while we decide how (on paper) to test for memory add, memory migration, memory remove, and tools. Feedback on the efficiency of this approach is most welcome. We have described what we know at this point here: http://developer.osdl.org/maryedie/HOTPLUG/planning/hotplug_memory_test_ plan_status.html Naturally, we'd rather not reinvent the wheel. If you have tests or test plans in any of these areas you can share, please do. We are interested in common regressions to look for. Please suggest anything that comes to mind. So that we can get something useful out quickly, we would like suggestions for simple, easy to implement, valuable tests that don't require instrumenting the kernel with patches or test drivers. Most of our ideas now are not simple (groan). And please understand, being the scribe, I'm not an expert in hotplug, so please forgive any ignorant mistakes in the descriptions so far. Any and all advice much, much appreciated. TIA -- Mary Edie Meredith Initiative Manager Open Source Development Labs maryedie@xxxxxxxx 503-906-1942
