From: Harry Yoo <harry.yoo@oracle.com>
To: Vlastimil Babka <vbabka@suse.cz>
Cc: Andrew Morton <akpm@linux-foundation.org>,
Christoph Lameter <cl@gentwo.org>,
David Rientjes <rientjes@google.com>,
Roman Gushchin <roman.gushchin@linux.dev>,
Uladzislau Rezki <urezki@gmail.com>,
"Liam R. Howlett" <Liam.Howlett@oracle.com>,
Suren Baghdasaryan <surenb@google.com>,
Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
Alexei Starovoitov <ast@kernel.org>,
linux-mm@kvack.org, linux-kernel@vger.kernel.org,
linux-rt-devel@lists.linux.dev, bpf@vger.kernel.org,
kasan-dev@googlegroups.com
Subject: Re: [PATCH RFC 08/19] slab: handle kmalloc sheaves bootstrap
Date: Mon, 27 Oct 2025 15:12:01 +0900 [thread overview]
Message-ID: <aP8NMX48FLn8FPZD@hyeyoo> (raw)
In-Reply-To: <20251023-sheaves-for-all-v1-8-6ffa2c9941c0@suse.cz>
On Thu, Oct 23, 2025 at 03:52:30PM +0200, Vlastimil Babka wrote:
> Enable sheaves for kmalloc caches. For other types than KMALLOC_NORMAL,
> we can simply allow them in calculate_sizes() as they are created later
> than KMALLOC_NORMAL caches and can allocate sheaves and barns from
> those.
>
> For KMALLOC_NORMAL caches we perform additional step after first
> creating them without sheaves. Then bootstrap_cache_sheaves() simply
> allocates and initializes barns and sheaves and finally sets
> s->sheaf_capacity to make them actually used.
>
> Afterwards the only caches left without sheaves (unless SLUB_TINY or
> debugging is enabled) are kmem_cache and kmem_cache_node. These are only
> used when creating or destroying other kmem_caches. Thus they are not
> performance critical and we can simply leave it that way.
>
> Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
> ---
> mm/slub.c | 88 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++---
> 1 file changed, 84 insertions(+), 4 deletions(-)
>
> diff --git a/mm/slub.c b/mm/slub.c
> index 5d0b2cf66520..a84027fbca78 100644
> --- a/mm/slub.c
> +++ b/mm/slub.c
>
> static void free_empty_sheaf(struct kmem_cache *s, struct slab_sheaf *sheaf)
> {
> kfree(sheaf);
> @@ -8064,8 +8071,11 @@ static int calculate_sizes(struct kmem_cache_args *args, struct kmem_cache *s)
> if (s->flags & SLAB_RECLAIM_ACCOUNT)
> s->allocflags |= __GFP_RECLAIMABLE;
>
> - /* kmalloc caches need extra care to support sheaves */
> - if (!is_kmalloc_cache(s))
> + /*
> + * For KMALLOC_NORMAL caches we enable sheaves later by
> + * bootstrap_kmalloc_sheaves() to avoid recursion
> + */
> + if (!is_kmalloc_normal(s))
> s->sheaf_capacity = calculate_sheaf_capacity(s, args);
I was going to say we should differentiate KMALLOC_NORMAL caches that
are created for kmalloc buckets.... but no, they don't have the SLAB_KMALLOC
flag.
> /*
> @@ -8549,6 +8559,74 @@ static struct kmem_cache * __init bootstrap(struct kmem_cache *static_cache)
> return s;
> }
>
> +/*
> + * Finish the sheaves initialization done normally by init_percpu_sheaves() and
> + * init_kmem_cache_nodes(). For normal kmalloc caches we have to bootstrap it
> + * since sheaves and barns are allocated by kmalloc.
> + */
> +static void __init bootstrap_cache_sheaves(struct kmem_cache *s)
> +{
> + struct kmem_cache_args empty_args = {};
> + unsigned int capacity;
> + bool failed = false;
> + int node, cpu;
> +
> + capacity = calculate_sheaf_capacity(s, &empty_args);
> +
> + /* capacity can be 0 due to debugging or SLUB_TINY */
> + if (!capacity)
> + return;
I think pcs->main should still be !NULL in this case?
--
Cheers,
Harry / Hyeonggon
> +
> + for_each_node_mask(node, slab_nodes) {
> + struct node_barn *barn;
> +
> + barn = kmalloc_node(sizeof(*barn), GFP_KERNEL, node);
> +
> + if (!barn) {
> + failed = true;
> + goto out;
> + }
> +
> + barn_init(barn);
> + get_node(s, node)->barn = barn;
> + }
> +
> + for_each_possible_cpu(cpu) {
> + struct slub_percpu_sheaves *pcs;
> +
> + pcs = per_cpu_ptr(s->cpu_sheaves, cpu);
> +
> + pcs->main = __alloc_empty_sheaf(s, GFP_KERNEL, capacity);
> +
> + if (!pcs->main) {
> + failed = true;
> + break;
> + }
> + }
> +
> +out:
> + /*
> + * It's still early in boot so treat this like same as a failure to
> + * create the kmalloc cache in the first place
> + */
> + if (failed)
> + panic("Out of memory when creating kmem_cache %s\n", s->name);
> +
> + s->sheaf_capacity = capacity;
> +}
next prev parent reply other threads:[~2025-10-27 6:12 UTC|newest]
Thread overview: 61+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-10-23 13:52 [PATCH RFC 00/19] slab: replace cpu (partial) slabs with sheaves Vlastimil Babka
2025-10-23 13:52 ` [PATCH RFC 01/19] slab: move kfence_alloc() out of internal bulk alloc Vlastimil Babka
2025-10-23 15:20 ` Marco Elver
2025-10-29 14:38 ` Vlastimil Babka
2025-10-29 15:30 ` Marco Elver
2025-10-23 13:52 ` [PATCH RFC 02/19] slab: handle pfmemalloc slabs properly with sheaves Vlastimil Babka
2025-10-24 14:21 ` Chris Mason
2025-10-29 15:00 ` Vlastimil Babka
2025-10-29 16:06 ` Chris Mason
2025-10-23 13:52 ` [PATCH RFC 03/19] slub: remove CONFIG_SLUB_TINY specific code paths Vlastimil Babka
2025-10-24 22:34 ` Alexei Starovoitov
2025-10-29 15:37 ` Vlastimil Babka
2025-10-23 13:52 ` [PATCH RFC 04/19] slab: prevent recursive kmalloc() in alloc_empty_sheaf() Vlastimil Babka
2025-10-23 13:52 ` [PATCH RFC 05/19] slab: add sheaves to most caches Vlastimil Babka
2025-10-27 0:24 ` Harry Yoo
2025-10-29 15:42 ` Vlastimil Babka
2025-10-23 13:52 ` [PATCH RFC 06/19] slab: introduce percpu sheaves bootstrap Vlastimil Babka
2025-10-24 15:29 ` Chris Mason
2025-10-29 15:51 ` Vlastimil Babka
2025-10-23 13:52 ` [PATCH RFC 07/19] slab: make percpu sheaves compatible with kmalloc_nolock()/kfree_nolock() Vlastimil Babka
2025-10-24 14:04 ` Chris Mason
2025-10-29 17:30 ` Vlastimil Babka
2025-10-24 19:43 ` Alexei Starovoitov
2025-10-29 17:46 ` Vlastimil Babka
2025-10-23 13:52 ` [PATCH RFC 08/19] slab: handle kmalloc sheaves bootstrap Vlastimil Babka
2025-10-27 6:12 ` Harry Yoo [this message]
2025-10-29 20:06 ` Vlastimil Babka
2025-10-29 20:06 ` Vlastimil Babka
2025-10-30 0:11 ` Harry Yoo
2025-10-23 13:52 ` [PATCH RFC 09/19] slab: add optimized sheaf refill from partial list Vlastimil Babka
2025-10-27 7:20 ` Harry Yoo
2025-10-27 9:11 ` Harry Yoo
2025-10-29 20:48 ` Vlastimil Babka
2025-10-30 0:07 ` Harry Yoo
2025-10-30 13:18 ` Vlastimil Babka
2025-10-23 13:52 ` [PATCH RFC 10/19] slab: remove cpu (partial) slabs usage from allocation paths Vlastimil Babka
2025-10-24 14:29 ` Chris Mason
2025-10-29 21:31 ` Vlastimil Babka
2025-10-30 4:32 ` Harry Yoo
2025-10-30 13:09 ` Vlastimil Babka
2025-10-30 15:27 ` Alexei Starovoitov
2025-10-30 15:35 ` Vlastimil Babka
2025-10-30 15:59 ` Alexei Starovoitov
2025-11-03 3:44 ` Harry Yoo
2025-10-23 13:52 ` [PATCH RFC 11/19] slab: remove SLUB_CPU_PARTIAL Vlastimil Babka
2025-10-24 20:43 ` Alexei Starovoitov
2025-10-29 22:31 ` Vlastimil Babka
2025-10-30 0:26 ` Alexei Starovoitov
2025-10-23 13:52 ` [PATCH RFC 12/19] slab: remove the do_slab_free() fastpath Vlastimil Babka
2025-10-24 22:32 ` Alexei Starovoitov
2025-10-29 22:44 ` Vlastimil Babka
2025-10-30 0:24 ` Alexei Starovoitov
2025-10-23 13:52 ` [PATCH RFC 13/19] slab: remove defer_deactivate_slab() Vlastimil Babka
2025-10-23 13:52 ` [PATCH RFC 14/19] slab: simplify kmalloc_nolock() Vlastimil Babka
2025-10-23 13:52 ` [PATCH RFC 15/19] slab: remove struct kmem_cache_cpu Vlastimil Babka
2025-10-23 13:52 ` [PATCH RFC 16/19] slab: remove unused PREEMPT_RT specific macros Vlastimil Babka
2025-10-23 13:52 ` [PATCH RFC 17/19] slab: refill sheaves from all nodes Vlastimil Babka
2025-10-23 13:52 ` [PATCH RFC 18/19] slab: update overview comments Vlastimil Babka
2025-10-23 13:52 ` [PATCH RFC 19/19] slab: remove frozen slab checks from __slab_free() Vlastimil Babka
2025-10-24 23:57 ` [PATCH RFC 00/19] slab: replace cpu (partial) slabs with sheaves Alexei Starovoitov
2025-11-04 22:11 ` Christoph Lameter (Ampere)
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=aP8NMX48FLn8FPZD@hyeyoo \
--to=harry.yoo@oracle.com \
--cc=Liam.Howlett@oracle.com \
--cc=akpm@linux-foundation.org \
--cc=ast@kernel.org \
--cc=bigeasy@linutronix.de \
--cc=bpf@vger.kernel.org \
--cc=cl@gentwo.org \
--cc=kasan-dev@googlegroups.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-rt-devel@lists.linux.dev \
--cc=rientjes@google.com \
--cc=roman.gushchin@linux.dev \
--cc=surenb@google.com \
--cc=urezki@gmail.com \
--cc=vbabka@suse.cz \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox