Re: [PATCH] percpu: add a schedule point in pcpu_balance_workfn()

2018-02-23 Thread Tejun Heo
On Fri, Feb 23, 2018 at 08:12:42AM -0800, Eric Dumazet wrote: > From: Eric Dumazet > > When a large BPF percpu map is destroyed, I have seen > pcpu_balance_workfn() holding cpu for hundreds of milliseconds. > > On KASAN config and 112 hyperthreads, average time to destroy a

[PATCH] percpu: add a schedule point in pcpu_balance_workfn()

2018-02-23 Thread Eric Dumazet
From: Eric Dumazet When a large BPF percpu map is destroyed, I have seen pcpu_balance_workfn() holding cpu for hundreds of milliseconds. On KASAN config and 112 hyperthreads, average time to destroy a chunk is ~4 ms. [ 2489.841376] destroy chunk 1 in 4148689 ns ... [