The patch titled
Subject: mm, slub: move disabling irqs closer to get_partial() in ___slab_alloc()
has been added to the -mm tree. Its filename is
mm-slub-move-disabling-irqs-closer-to-get_partial-in-___slab_alloc.patch
This patch should soon appear at
https://ozlabs.org/~akpm/mmots/broken-out/mm-slub-move-disabling-irqs-closer-to-get_partial-in-___slab_alloc.patch
and later at
https://ozlabs.org/~akpm/mmotm/broken-out/mm-slub-move-disabling-irqs-closer-to-get_partial-in-___slab_alloc.patch
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 and is updated
there every 3-4 working days
------------------------------------------------------
From: Vlastimil Babka <vbabka@xxxxxxx>
Subject: mm, slub: move disabling irqs closer to get_partial() in ___slab_alloc()
Continue reducing the irq disabled scope. Check for per-cpu partial slabs
with first with irqs enabled and then recheck with irqs disabled before
grabbing the slab page. Mostly preparatory for the following patches.
Link: https://lkml.kernel.org/r/20210805152000.12817-15-vbabka@xxxxxxx
Signed-off-by: Vlastimil Babka <vbabka@xxxxxxx>
Cc: Christoph Lameter <cl@xxxxxxxxx>
Cc: David Rientjes <rientjes@xxxxxxxxxx>
Cc: Jann Horn <jannh@xxxxxxxxxx>
Cc: Jesper Dangaard Brouer <brouer@xxxxxxxxxx>
Cc: Joonsoo Kim <iamjoonsoo.kim@xxxxxxx>
Cc: Mel Gorman <mgorman@xxxxxxxxxxxxxxxxxxx>
Cc: Mike Galbraith <efault@xxxxxx>
Cc: Pekka Enberg <penberg@xxxxxxxxxx>
Cc: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx>
Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
mm/slub.c | 34 +++++++++++++++++++++++++---------
1 file changed, 25 insertions(+), 9 deletions(-)
--- a/mm/slub.c~mm-slub-move-disabling-irqs-closer-to-get_partial-in-___slab_alloc
+++ a/mm/slub.c
@@ -2679,11 +2679,6 @@ reread_page:
if (unlikely(node != NUMA_NO_NODE &&
!node_isset(node, slab_nodes)))
node = NUMA_NO_NODE;
- local_irq_save(flags);
- if (unlikely(c->page)) {
- local_irq_restore(flags);
- goto reread_page;
- }
goto new_slab;
}
redo:
@@ -2724,6 +2719,7 @@ redo:
if (!freelist) {
c->page = NULL;
+ local_irq_restore(flags);
stat(s, DEACTIVATE_BYPASS);
goto new_slab;
}
@@ -2753,12 +2749,19 @@ deactivate_slab:
goto reread_page;
}
deactivate_slab(s, page, c->freelist, c);
+ local_irq_restore(flags);
new_slab:
- lockdep_assert_irqs_disabled();
-
if (slub_percpu_partial(c)) {
+ local_irq_save(flags);
+ if (unlikely(c->page)) {
+ local_irq_restore(flags);
+ goto reread_page;
+ }
+ if (unlikely(!slub_percpu_partial(c)))
+ goto new_objects; /* stolen by an IRQ handler */
+
page = c->page = slub_percpu_partial(c);
slub_set_percpu_partial(c, page);
local_irq_restore(flags);
@@ -2766,6 +2769,16 @@ new_slab:
goto redo;
}
+ local_irq_save(flags);
+ if (unlikely(c->page)) {
+ local_irq_restore(flags);
+ goto reread_page;
+ }
+
+new_objects:
+
+ lockdep_assert_irqs_disabled();
+
freelist = get_partial(s, gfpflags, node, &page);
if (freelist) {
c->page = page;
@@ -2798,15 +2811,18 @@ new_slab:
check_new_page:
if (kmem_cache_debug(s)) {
- if (!alloc_debug_processing(s, page, freelist, addr))
+ if (!alloc_debug_processing(s, page, freelist, addr)) {
/* Slab failed checks. Next slab needed */
+ c->page = NULL;
+ local_irq_restore(flags);
goto new_slab;
- else
+ } else {
/*
* For debug case, we don't load freelist so that all
* allocations go through alloc_debug_processing()
*/
goto return_single;
+ }
}
if (unlikely(!pfmemalloc_match(page, gfpflags)))
_
Patches currently in -mm which might be from vbabka@xxxxxxx are
mm-slub-fix-slub_debug-disablement-for-list-of-slabs.patch
mm-slub-dont-call-flush_all-from-slab_debug_trace_open.patch
mm-slub-allocate-private-object-map-for-debugfs-listings.patch
mm-slub-allocate-private-object-map-for-validate_slab_cache.patch
mm-slub-dont-disable-irq-for-debug_check_no_locks_freed.patch
mm-slub-remove-redundant-unfreeze_partials-from-put_cpu_partial.patch
mm-slub-unify-cmpxchg_double_slab-and-__cmpxchg_double_slab.patch
mm-slub-extract-get_partial-from-new_slab_objects.patch
mm-slub-dissolve-new_slab_objects-into-___slab_alloc.patch
mm-slub-return-slab-page-from-get_partial-and-set-c-page-afterwards.patch
mm-slub-restructure-new-page-checks-in-___slab_alloc.patch
mm-slub-simplify-kmem_cache_cpu-and-tid-setup.patch
mm-slub-move-disabling-enabling-irqs-to-___slab_alloc.patch
mm-slub-do-initial-checks-in-___slab_alloc-with-irqs-enabled.patch
mm-slub-move-disabling-irqs-closer-to-get_partial-in-___slab_alloc.patch
mm-slub-restore-irqs-around-calling-new_slab.patch
mm-slub-validate-slab-from-partial-list-or-page-allocator-before-making-it-cpu-slab.patch
mm-slub-check-new-pages-with-restored-irqs.patch
mm-slub-stop-disabling-irqs-around-get_partial.patch
mm-slub-move-reset-of-c-page-and-freelist-out-of-deactivate_slab.patch
mm-slub-make-locking-in-deactivate_slab-irq-safe.patch
mm-slub-call-deactivate_slab-without-disabling-irqs.patch
mm-slub-move-irq-control-into-unfreeze_partials.patch
mm-slub-discard-slabs-in-unfreeze_partials-without-irqs-disabled.patch
mm-slub-detach-whole-partial-list-at-once-in-unfreeze_partials.patch
mm-slub-separate-detaching-of-partial-list-in-unfreeze_partials-from-unfreezing.patch
mm-slub-only-disable-irq-with-spin_lock-in-__unfreeze_partials.patch
mm-slub-dont-disable-irqs-in-slub_cpu_dead.patch
mm-slab-make-flush_slab-possible-to-call-with-irqs-enabled.patch
mm-slub-optionally-save-restore-irqs-in-slab_lock.patch
mm-slub-make-slab_lock-disable-irqs-with-preempt_rt.patch
mm-slub-protect-put_cpu_partial-with-disabled-irqs-instead-of-cmpxchg.patch
mm-slub-use-migrate_disable-on-preempt_rt.patch
mm-slub-convert-kmem_cpu_slab-protection-to-local_lock.patch
