Messages in this thread | | | Date | Fri, 20 May 2005 15:16:22 -0700 | From | Ashok Raj <> | Subject | [patch 0/4] CPU hot-plug support for x86_64 |
| |
Hi Andi
Attached are patches that would apply over 2.6.12-rc4-mm2 to provide cpu hotplug support. It seems to hold up to stress (make -j's) on a 4way HT system)
FYI: Iam sending this via quilt mail, hopefully it makes its way ok, otherwise please pardon, and i will resend my usuall way again if this doesn't work.
- I tested with booting maxcpus=1, then you can kick each cpu online. * I also manually migrate interrupts to the new CPU, and ensure they are off when that same cpu is being offlined. * Seems to have trouble with CONFIG_SCHED_SMT, i need to look into this later. - Offline works as well
Andi: You had mentioned that you would not prefer to replace the broadcast IPI with the mask version for performance. Currently this seems to be the most optimal way without putting a sledge hammer on the cpu_up process.
If there are any known performance issues that we are worried about we should look for alternative mechanism.
TBD: 0. Verify with CONFIG_NUMA (not done yet) 1. ACPI integration for physical cpu hotplug support. 2. replace fixup_irqs() with a correct implementation. This works fine for now but has potential to miss interrupts. (in works).
Cheers, Ashok Raj - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |