Re: [RFC][PATCH] struct kernel_stat -> struct cpu_stat[NR_CPUS]

Zack writes: > These per cpu statistics are reported via a new /proc/cpustat, a quick > tool for processing that output, vmstat-style, can be found near Could you consider /proc/cpu/0/stats or similar? It is much nicer than polluting the top-level /proc directory, and I believe there are a

Re: [RFC][PATCH] struct kernel_stat - struct cpu_stat[NR_CPUS]

Zack writes: These per cpu statistics are reported via a new /proc/cpustat, a quick tool for processing that output, vmstat-style, can be found near Could you consider /proc/cpu/0/stats or similar? It is much nicer than polluting the top-level /proc directory, and I believe there are a bunch

[RFC][PATCH] struct kernel_stat -> struct cpu_stat[NR_CPUS]

Currently struct kernel_stat has a few pre cpu arrays. This creates cacheline exchange noise as the cpus update their entries in each array. This patch creates an array of per cpu structs. The structure is padded to the length of a cacheline. The meat of the patch against 2.4.6-pre8 is

[RFC][PATCH] struct kernel_stat - struct cpu_stat[NR_CPUS]

Currently struct kernel_stat has a few pre cpu arrays. This creates cacheline exchange noise as the cpus update their entries in each array. This patch creates an array of per cpu structs. The structure is padded to the length of a cacheline. The meat of the patch against 2.4.6-pre8 is