From: Vlastimil Babka <vbabka@suse.cz>
To: Hyeonggon Yoo <42.hyeyoo@gmail.com>
Cc: Christoph Lameter <cl@linux.com>,
Pekka Enberg <penberg@kernel.org>,
David Rientjes <rientjes@google.com>,
Joonsoo Kim <iamjoonsoo.kim@lge.com>,
Andrew Morton <akpm@linux-foundation.org>,
Roman Gushchin <roman.gushchin@linux.dev>,
Joe Perches <joe@perches.com>,
Vasily Averin <vasily.averin@linux.dev>,
Matthew WilCox <willy@infradead.org>,
linux-kernel@vger.kernel.org, linux-mm@kvack.org
Subject: Re: [PATCH v3 08/15] mm/slab_common: kmalloc_node: pass large requests to page allocator
Date: Tue, 2 Aug 2022 11:32:41 +0200 [thread overview]
Message-ID: <321b8b3e-9d06-b01c-d871-1f7ca35ce91e@suse.cz> (raw)
In-Reply-To: <Yujnihj5YVPP2LjA@hyeyoo>
On 8/2/22 10:59, Hyeonggon Yoo wrote:
> On Mon, Aug 01, 2022 at 04:44:22PM +0200, Vlastimil Babka wrote:
>>
>
> Yeah, uninlining __kmalloc_large_node saves hundreds of bytes.
> And the diff below looks good to me.
>
> By The Way, do you have opinions on inlining slab_alloc_node()?
> (Looks like similar topic?)
>
> AFAIK slab_alloc_node() is inlined in:
> kmem_cache_alloc()
> kmem_cache_alloc_node()
> kmem_cache_alloc_lru()
> kmem_cache_alloc_trace()
> kmem_cache_alloc_node_trace()
> __kmem_cache_alloc_node()
>
> This is what I get after simply dropping __always_inline in slab_alloc_node:
>
> add/remove: 1/1 grow/shrink: 3/6 up/down: 1911/-5275 (-3364)
> Function old new delta
> slab_alloc_node - 1356 +1356
> sysfs_slab_alias 134 327 +193
> slab_memory_callback 528 717 +189
> __kmem_cache_create 1325 1498 +173
> __slab_alloc.constprop 135 - -135
> kmem_cache_alloc_trace 909 196 -713
> kmem_cache_alloc 937 191 -746
> kmem_cache_alloc_node_trace 1020 200 -820
> __kmem_cache_alloc_node 862 19 -843
> kmem_cache_alloc_node 1046 189 -857
> kmem_cache_alloc_lru 1348 187 -1161
> Total: Before=32011183, After=32007819, chg -0.01%
>
> So 3.28kB is cost of eliminating function call overhead in the
> fastpath.
>
> This is tradeoff between function call overhead and
> instruction cache usage...
We can investigate this aftewards, with proper measurements etc. I think
it's more sensitive than kmalloc_large_node.
next prev parent reply other threads:[~2022-08-02 9:32 UTC|newest]
Thread overview: 49+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-07-12 13:39 [PATCH v3 00/15] common kmalloc v3 Hyeonggon Yoo
2022-07-12 13:39 ` [PATCH v3 1/15] mm/slab: move NUMA-related code to __do_cache_alloc() Hyeonggon Yoo
2022-07-12 14:29 ` Christoph Lameter
2022-07-13 9:39 ` Hyeonggon Yoo
2022-07-12 13:39 ` [PATCH v3 2/15] mm/slab: cleanup slab_alloc() and slab_alloc_node() Hyeonggon Yoo
2022-07-12 13:39 ` [PATCH v3 03/15] mm/slab_common: remove CONFIG_NUMA ifdefs for common kmalloc functions Hyeonggon Yoo
2022-07-12 13:39 ` [PATCH v3 04/15] mm/slab_common: cleanup kmalloc_track_caller() Hyeonggon Yoo
2022-07-12 13:39 ` [PATCH v3 05/15] mm/sl[au]b: factor out __do_kmalloc_node() Hyeonggon Yoo
2022-07-28 14:45 ` Vlastimil Babka
2022-07-12 13:39 ` [PATCH v3 06/15] mm/slab_common: fold kmalloc_order_trace() into kmalloc_large() Hyeonggon Yoo
2022-07-28 15:23 ` Vlastimil Babka
2022-08-01 13:26 ` Hyeonggon Yoo
2022-08-01 13:36 ` Vlastimil Babka
2022-08-02 2:54 ` Hyeonggon Yoo
2022-07-12 13:39 ` [PATCH v3 07/15] mm/slub: move kmalloc_large_node() to slab_common.c Hyeonggon Yoo
2022-07-12 13:39 ` [PATCH v3 08/15] mm/slab_common: kmalloc_node: pass large requests to page allocator Hyeonggon Yoo
2022-07-28 16:09 ` Vlastimil Babka
2022-08-01 14:37 ` Hyeonggon Yoo
2022-08-01 14:44 ` Vlastimil Babka
2022-08-02 8:59 ` Hyeonggon Yoo
2022-08-02 9:32 ` Vlastimil Babka [this message]
2022-07-12 13:39 ` [PATCH v3 09/15] mm/slab_common: cleanup kmalloc_large() Hyeonggon Yoo
2022-07-28 16:13 ` Vlastimil Babka
2022-07-12 13:39 ` [PATCH v3 10/15] mm/slab: kmalloc: pass requests larger than order-1 page to page allocator Hyeonggon Yoo
2022-07-28 16:25 ` Vlastimil Babka
2022-07-12 13:39 ` [PATCH v3 11/15] mm/sl[au]b: introduce common alloc/free functions without tracepoint Hyeonggon Yoo
2022-07-29 9:49 ` Vlastimil Babka
2022-07-12 13:39 ` [PATCH v3 12/15] mm/sl[au]b: generalize kmalloc subsystem Hyeonggon Yoo
2022-07-29 10:25 ` Vlastimil Babka
2022-07-12 13:39 ` [PATCH v3 13/15] mm/slab_common: unify NUMA and UMA version of tracepoints Hyeonggon Yoo
2022-07-29 10:52 ` Vlastimil Babka
2022-07-12 13:39 ` [PATCH 14/16] mm/slab_common: drop kmem_alloc & avoid dereferencing fields when not using Hyeonggon Yoo
2022-07-29 11:23 ` Vlastimil Babka
2022-08-02 9:22 ` Hyeonggon Yoo
2022-07-12 13:39 ` [PATCH 15/16] mm/slab_common: move definition of __ksize() to mm/slab.h Hyeonggon Yoo
2022-07-29 11:47 ` Vlastimil Babka
2022-08-02 9:25 ` Hyeonggon Yoo
2022-07-12 13:39 ` [PATCH 16/16] mm/sl[au]b: check if large object is valid in __ksize() Hyeonggon Yoo
2022-07-12 15:13 ` Christoph Lameter
2022-07-13 9:25 ` Hyeonggon Yoo
2022-07-13 10:07 ` Christoph Lameter
2022-07-13 10:33 ` Marco Elver
2022-07-14 9:15 ` Christoph Lameter
2022-07-14 10:30 ` Marco Elver
2022-07-20 10:05 ` Hyeonggon Yoo
2022-07-29 11:50 ` Vlastimil Babka
2022-07-29 15:08 ` [PATCH v3 00/15] common kmalloc v3 Vlastimil Babka
2022-08-14 10:06 ` Hyeonggon Yoo
2022-08-15 12:59 ` Vlastimil Babka
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=321b8b3e-9d06-b01c-d871-1f7ca35ce91e@suse.cz \
--to=vbabka@suse.cz \
--cc=42.hyeyoo@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=cl@linux.com \
--cc=iamjoonsoo.kim@lge.com \
--cc=joe@perches.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=penberg@kernel.org \
--cc=rientjes@google.com \
--cc=roman.gushchin@linux.dev \
--cc=vasily.averin@linux.dev \
--cc=willy@infradead.org \
/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;
as well as URLs for NNTP newsgroup(s).