Skip to contentSkip to conversation

x86: unify x86_32 and 64 NUMA init paths, take#3

PATCH Tejun Heo 2 participants 2 messages Dec 16, 2010 1 message not in the archive

lore
Share & tools
lore

Tejun Heo

author
#1 lore raw
In reply to a message not in the archive

(cc'ing Shaohui)
Hello, Thoams.

Quoted 18 lines · Tejun Heo

This is the third take of unify-x86_32-and-64-NUMA-init-paths patchset. Most changes from the last take[L] are to reflect comments from tglx.

* Rebased on top of tip/x86/apic-cleanup.

* "x86: apic: Cleanup and simplify setup_local_APIC()" is already merged into x86/apic-cleanup and dropped from this series.

* Instead of repurposing apic->cpu_to_logical_apicid(), replace it with a new 32bit specific callback apic->x86_32_early_logical_apicid().

* For consistency, in 0012, apic->apicid_to_node() is renamed to x86_32_numa_cpu_node() and put inside CONFIG_X86_32.

* NULL assignments for unimplemented apic ops dropped.

Quoted 7 lines

* 0013 now explicitly describes why amd srat_detect_node() accesses __apicid_to_node[] directly. Also, add a comment in the function noting the ugliness.

* Patch description updated in 0014.

* Other misc updates.

Do the above changes agreeable to you? Shaohui is wondering what to do with the conflicts between the NUMA hotplug simulator and this series, so...

Thank you.

Signature · 1 line
tejun
#2 lore raw
Quoted 26 lines · Tejun Heo

(cc'ing Shaohui)
Hello, Thoams.

Quoted 22 lines · Tejun Heo

This is the third take of unify-x86_32-and-64-NUMA-init-paths patchset. Most changes from the last take[L] are to reflect comments from tglx.

* Rebased on top of tip/x86/apic-cleanup.

* "x86: apic: Cleanup and simplify setup_local_APIC()" is already merged into x86/apic-cleanup and dropped from this series.

* Instead of repurposing apic->cpu_to_logical_apicid(), replace it with a new 32bit specific callback apic->x86_32_early_logical_apicid().

* For consistency, in 0012, apic->apicid_to_node() is renamed to x86_32_numa_cpu_node() and put inside CONFIG_X86_32.

* NULL assignments for unimplemented apic ops dropped.

* 0013 now explicitly describes why amd srat_detect_node() accesses __apicid_to_node[] directly. Also, add a comment in the function noting the ugliness.

Quoted 7 lines
Quoted 3 lines

* Patch description updated in 0014.

* Other misc updates.

Do the above changes agreeable to you? Shaohui is wondering what to do with the conflicts between the NUMA hotplug simulator and this series, so...

I was looking with hpa into it, review still in progress.