Re: [RFC PATCH 0/6] x86/mm: Flush remote and local TLBs concurrently

2019-05-27 Thread Peter Zijlstra
On Sat, May 25, 2019 at 01:21:57AM -0700, Nadav Amit wrote: > Currently, local and remote TLB flushes are not performed concurrently, > which introduces unnecessary overhead - each INVLPG can take 100s of > cycles. This patch-set allows TLB flushes to be run concurrently: first > request the

Re: [RFC PATCH 0/6] x86/mm: Flush remote and local TLBs concurrently

2019-05-27 Thread Peter Zijlstra
On Sat, May 25, 2019 at 01:21:57AM -0700, Nadav Amit wrote: > The proposed changes should also improve the performance of other > invocations of on_each_cpu(). Hopefully, no one has relied on the > behavior of on_each_cpu() that functions were first executed remotely > and only then locally. Oh

[RFC PATCH 0/6] x86/mm: Flush remote and local TLBs concurrently

2019-05-25 Thread Nadav Amit
Currently, local and remote TLB flushes are not performed concurrently, which introduces unnecessary overhead - each INVLPG can take 100s of cycles. This patch-set allows TLB flushes to be run concurrently: first request the remote CPUs to initiate the flush, then run it locally, and finally wait