Messages in this thread | | | Date | Fri, 10 May 2024 20:25:22 +0900 | Subject | Re: [PATCH (REPOST)] profiling: initialize prof_cpu_mask from profile_online_cpu() | From | Tetsuo Handa <> |
| |
On 2024/05/05 15:18, Tetsuo Handa wrote: > On 2024/04/27 15:51, Tetsuo Handa wrote: >> Can somebody test this patch? I don't know how not using >> cpu_possible_mask affects expected profile data collection >> (especially when written to /sys/kernel/profiling interface >> when some of CPUs are offline). > > I confirmed that writing to /sys/kernel/profiling while some of > /sys/devices/system/cpu/cpu$num/online are 0 will not affect profile data > collection, for profile_online_cpu() is called (and corresponding bit is set) > when /sys/devices/system/cpu/cpu$num/online becomes 1. Thus, I consider that > this patch itself is correct and safe.
Can someone take this patch and https://lkml.kernel.org/r/20230423012744.24320-1-chenzhongjin@huawei.com ? These bugs remain unfixed due to lack of someone who can take patches...
| |