On Fri, Mar 14, 2025 at 02:33:28PM +0100, Vlastimil Babka wrote: > On 3/14/25 07:15, Shakeel Butt wrote: > > This is a cleanup series (first 7 patches) is trying to simplify the > > memcg stock code, particularly it tries to remove unnecessary > > dependencies. The last 3 patches are a prototype to make per-cpu memcg > > stock work without disabling irqs. > > > > My plan is to send out the first 7 cleanup patches separately for the > > next release window and iterate more on the last 3 patches plus add > > functionality for multiple memcgs. > > > > This series is based on next-20250313 plus two following patches: > > > > Link: https://lore.kernel.org/all/20250312222552.3284173-1-shakeel.butt@xxxxxxxxx/ > > Link: https://lore.kernel.org/all/20250313054812.2185900-1-shakeel.butt@xxxxxxxxx/ > > > > to simply the memcg stock code > > Hi, > > I've been looking at this area too, and noticed a different opportunity for > cleanup+perf improvement yesterday. I rebased it on top of patch 7 as it > would make sense to do it before changing the locking - it reduces the > number of places where the local_trylock is taken. If it's ok to you, please > incorporate to your series. Thanks a lot Vlastimil, I will take it and add review tag after reviewing it. Andrew, I will post only the cleanup series soon. It will be based on next-20250314. If you decide to take it for upcoming open window (6.15-rc) then some coordination with bpf tree would be needed. However for the next window (6.16-rc), I don't expect conflicts.
