On Fri, Jan 19, 2024 at 3:22 AM Chengming Zhou <zhouchengming@xxxxxxxxxxxxx> wrote: > > Each swapfile has one rb-tree to search the mapping of swp_entry_t to > zswap_entry, that use a spinlock to protect, which can cause heavy lock > contention if multiple tasks zswap_store/load concurrently. > > Optimize the scalability problem by splitting the zswap rb-tree into > multiple rb-trees, each corresponds to SWAP_ADDRESS_SPACE_PAGES (64M), > just like we did in the swap cache address_space splitting. > > Although this method can't solve the spinlock contention completely, it > can mitigate much of that contention. Below is the results of kernel build > in tmpfs with zswap shrinker enabled: > > linux-next zswap-lock-optimize > real 1m9.181s 1m3.820s > user 17m44.036s 17m40.100s > sys 7m37.297s 4m54.622s > > So there are clearly improvements. If/when you respin this, can you mention that testing was done with a single swapfile? I assume the improvements will be less with multiple swapfiles as lock contention should be better. > > Acked-by: Johannes Weiner <hannes@xxxxxxxxxxx> > Acked-by: Nhat Pham <nphamcs@xxxxxxxxx> I think the diff in zswap_swapoff() should be much simpler with the tree(s) cleanup removed. Otherwise LGTM. Acked-by: Yosry Ahmed <yosryahmed@xxxxxxxxxx>
