The patch titled Subject: mm/mglru: optimize deactivation has been added to the -mm mm-unstable branch. Its filename is mm-mglru-optimize-deactivation.patch This patch will shortly appear at https://git.kernel.org/pub/scm/linux/kernel/git/akpm/25-new.git/tree/patches/mm-mglru-optimize-deactivation.patch This patch will later appear in the mm-unstable branch at git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm Before you just go and hit "reply", please: a) Consider who else should be cc'ed b) Prefer to cc a suitable mailing list as well c) Ideally: find the original patch on the mailing list and do a reply-to-all to that, adding suitable additional cc's *** Remember to use Documentation/process/submit-checklist.rst when testing your code *** The -mm tree is included into linux-next via the mm-everything branch at git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm and is updated there every 2-3 working days ------------------------------------------------------ From: Yu Zhao <yuzhao@xxxxxxxxxx> Subject: mm/mglru: optimize deactivation Date: Thu, 5 Dec 2024 17:31:22 -0700 Do not shuffle a folio in the deactivation paths if it is already in the oldest generation. This reduces the LRU lock contention. Before this patch, the contention is reproducible by FIO, e.g., fio -filename=/dev/nvme1n1p2 -direct=0 -thread -size=1024G \ -rwmixwrite=30 --norandommap --randrepeat=0 -ioengine=sync \ -bs=4k -numjobs=400 -runtime=25000 --time_based \ -group_reporting -name=mglru 98.96%--_raw_spin_lock_irqsave folio_lruvec_lock_irqsave | --98.78%--folio_batch_move_lru | --98.63%--deactivate_file_folio mapping_try_invalidate invalidate_mapping_pages invalidate_bdev blkdev_common_ioctl blkdev_ioctl After this patch, deactivate_file_folio() bails out early without taking the LRU lock. A side effect is that a folio can be left at the head of the oldest generation, rather than the tail. If reclaim happens at the same time, it cannot reclaim this folio immediately. Since there is no known correlation between truncation and reclaim, this side effect is considered insignificant. Link: https://lkml.kernel.org/r/20241206003126.1338283-3-yuzhao@xxxxxxxxxx Signed-off-by: Yu Zhao <yuzhao@xxxxxxxxxx> Reported-by: Bharata B Rao <bharata@xxxxxxx> Closes: https://lore.kernel.org/CAOUHufawNerxqLm7L9Yywp3HJFiYVrYO26ePUb1jH-qxNGWzyA@xxxxxxxxxxxxxx/ Tested-by: Kalesh Singh <kaleshsingh@xxxxxxxxxx> Cc: David Stevens <stevensd@xxxxxxxxxxxx> Cc: Kairui Song <kasong@xxxxxxxxxxx> Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> --- mm/swap.c | 44 +++++++++++++++++++++++++++++++++++++------- 1 file changed, 37 insertions(+), 7 deletions(-) --- a/mm/swap.c~mm-mglru-optimize-deactivation +++ a/mm/swap.c @@ -379,7 +379,8 @@ static void __lru_cache_activate_folio(s } #ifdef CONFIG_LRU_GEN -static void folio_inc_refs(struct folio *folio) + +static void lru_gen_inc_refs(struct folio *folio) { unsigned long new_flags, old_flags = READ_ONCE(folio->flags); @@ -406,10 +407,30 @@ static void folio_inc_refs(struct folio new_flags |= old_flags & ~LRU_REFS_MASK; } while (!try_cmpxchg(&folio->flags, &old_flags, new_flags)); } -#else -static void folio_inc_refs(struct folio *folio) + +static bool lru_gen_clear_refs(struct folio *folio) +{ + int gen = folio_lru_gen(folio); + int type = folio_is_file_lru(folio); + struct lru_gen_folio *lrugen = &folio_lruvec(folio)->lrugen; + + set_mask_bits(&folio->flags, LRU_REFS_MASK | LRU_REFS_FLAGS, 0); + + /* whether can do without shuffling under the LRU lock */ + return gen == lru_gen_from_seq(READ_ONCE(lrugen->min_seq[type])); +} + +#else /* !CONFIG_LRU_GEN */ + +static void lru_gen_inc_refs(struct folio *folio) +{ +} + +static bool lru_gen_clear_refs(struct folio *folio) { + return false; } + #endif /* CONFIG_LRU_GEN */ /** @@ -428,7 +449,7 @@ static void folio_inc_refs(struct folio void folio_mark_accessed(struct folio *folio) { if (lru_gen_enabled()) { - folio_inc_refs(folio); + lru_gen_inc_refs(folio); return; } @@ -524,7 +545,7 @@ void folio_add_lru_vma(struct folio *fol */ static void lru_deactivate_file(struct lruvec *lruvec, struct folio *folio) { - bool active = folio_test_active(folio); + bool active = folio_test_active(folio) || lru_gen_enabled(); long nr_pages = folio_nr_pages(folio); if (folio_test_unevictable(folio)) @@ -589,7 +610,10 @@ static void lru_lazyfree(struct lruvec * lruvec_del_folio(lruvec, folio); folio_clear_active(folio); - folio_clear_referenced(folio); + if (lru_gen_enabled()) + lru_gen_clear_refs(folio); + else + folio_clear_referenced(folio); /* * Lazyfree folios are clean anonymous folios. They have * the swapbacked flag cleared, to distinguish them from normal @@ -657,6 +681,9 @@ void deactivate_file_folio(struct folio if (folio_test_unevictable(folio)) return; + if (lru_gen_enabled() && lru_gen_clear_refs(folio)) + return; + folio_batch_add_and_move(folio, lru_deactivate_file, true); } @@ -670,7 +697,10 @@ void deactivate_file_folio(struct folio */ void folio_deactivate(struct folio *folio) { - if (folio_test_unevictable(folio) || !(folio_test_active(folio) || lru_gen_enabled())) + if (folio_test_unevictable(folio)) + return; + + if (lru_gen_enabled() ? lru_gen_clear_refs(folio) : !folio_test_active(folio)) return; folio_batch_add_and_move(folio, lru_deactivate, true); _ Patches currently in -mm which might be from yuzhao@xxxxxxxxxx are mm-mglru-clean-up-workingset.patch mm-mglru-optimize-deactivation.patch mm-mglru-rework-aging-feedback.patch mm-mglru-rework-type-selection.patch mm-mglru-rework-refault-detection.patch mm-mglru-rework-workingset-protection.patch