[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] RE: [Xen-devel] [PATCH][HVMLOADER][DISCUSS] 8 way default rombios
Thank you winston! we were performing the same experiments yesterday after our regression *finally* found the same problem... a bit late though, after the patch was applied to both trees. > The best way would be to do the same as the for our ACPI > tables -- dynamically generate the CPU entries (LAPIC entries > in case of ACPI > tables) when we know how many VCPUs the guest has been allocated. > > This would get rid of multiple statically-compiled rombios > images, each of which only differs in CPU info in the MPS > table. It could even be hacked up without modifying rombios > image -- e.g., hvmloader could scan for and then modify the > MPS table to insert secondary-processor entries. It wouldn't > be very hard to do. > > In any case, I'll revert the earlier patch. Using a MPS table > that advertises 8 processors when a guest actually has fewer > doesn't seem a very wise idea. Agreed. Reversion is probably best here for both unstable and 3.0-testing. I'll take a gander at adding the tables dynamically in hvmloader, and see if I can get a preliminary patch out in a couple of days - need to test quite a few configurations. Thanks Tom _______________________________________________ Xen-devel mailing list Xen-devel@xxxxxxxxxxxxxxxxxxx http://lists.xensource.com/xen-devel
|
![]() |
Lists.xenproject.org is hosted with RackSpace, monitoring our |