linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Ingo Molnar <mingo@kernel.org>
To: Matteo Rizzo <matteorizzo@google.com>
Cc: "Lameter, Christopher" <cl@os.amperecomputing.com>,
	Dave Hansen <dave.hansen@intel.com>,
	penberg@kernel.org, rientjes@google.com, iamjoonsoo.kim@lge.com,
	akpm@linux-foundation.org, vbabka@suse.cz,
	roman.gushchin@linux.dev, 42.hyeyoo@gmail.com,
	keescook@chromium.org, linux-kernel@vger.kernel.org,
	linux-doc@vger.kernel.org, linux-mm@kvack.org,
	linux-hardening@vger.kernel.org, tglx@linutronix.de,
	mingo@redhat.com, bp@alien8.de, dave.hansen@linux.intel.com,
	x86@kernel.org, hpa@zytor.com, corbet@lwn.net, luto@kernel.org,
	peterz@infradead.org, jannh@google.com, evn@google.com,
	poprdi@google.com, jordyzomer@google.com,
	Linus Torvalds <torvalds@linux-foundation.org>
Subject: Re: [RFC PATCH 00/14] Prevent cross-cache attacks in the SLUB allocator
Date: Wed, 20 Sep 2023 09:44:33 +0200	[thread overview]
Message-ID: <ZQqi4RqpEM7PRGkF@gmail.com> (raw)
In-Reply-To: <CAHKB1wKneke-dyvMY0JtW-xwW8m=GaUdafoAqdCE0B9csY7_bw@mail.gmail.com>


* Matteo Rizzo <matteorizzo@google.com> wrote:

> On Mon, 18 Sept 2023 at 19:39, Ingo Molnar <mingo@kernel.org> wrote:
> >
> > What's the split of the increase in overhead due to SLAB_VIRTUAL=y, between
> > user-space execution and kernel-space execution?
> >
> 
> Same benchmark as before (compiling a kernel on a system running the patched
> kernel):
> 
> Intel Skylake:
> 
>       LABEL    | COUNT |   MIN    |   MAX    |   MEAN   |  MEDIAN  | STDDEV
> ---------------+-------+----------+----------+----------+----------+--------
> wall clock     |       |          |          |          |          |
> SLAB_VIRTUAL=n | 150   | 49.700   | 51.320   | 50.449   | 50.430   | 0.29959
> SLAB_VIRTUAL=y | 150   | 50.020   | 51.660   | 50.880   | 50.880   | 0.30495
>                |       | +0.64%   | +0.66%   | +0.85%   | +0.89%   | +1.79%
> system time    |       |          |          |          |          |
> SLAB_VIRTUAL=n | 150   | 358.560  | 362.900  | 360.922  | 360.985  | 0.91761
> SLAB_VIRTUAL=y | 150   | 362.970  | 367.970  | 366.062  | 366.115  | 1.015
>                |       | +1.23%   | +1.40%   | +1.42%   | +1.42%   | +10.60%
> user time      |       |          |          |          |          |
> SLAB_VIRTUAL=n | 150   | 3110.000 | 3124.520 | 3118.143 | 3118.120 | 2.466
> SLAB_VIRTUAL=y | 150   | 3115.070 | 3127.070 | 3120.762 | 3120.925 | 2.654
>                |       | +0.16%   | +0.08%   | +0.08%   | +0.09%   | +7.63%

These Skylake figures are a bit counter-intuitive: how does an increase of 
only +0.08% user-time - which dominates 89.5% of execution, combined with a 
+1.42% increase in system time that consumes only 10.5% of CPU capacity, 
result in a +0.85% increase in wall-clock time?

There might be hidden factors at work in the DMA space, as Linus suggested?

Or perhaps wall-clock time is dominated by the single-threaded final link 
time of the kernel, which phase might be disproportionately hurt by these 
changes?

(Stddev seems low enough for this not to be a measurement artifact.)

The AMD Milan figures are more intuitive:

> AMD Milan:
> 
>       LABEL    | COUNT |   MIN    |   MAX    |   MEAN   |  MEDIAN  | STDDEV
> ---------------+-------+----------+----------+----------+----------+--------
> wall clock     |       |          |          |          |          |
> SLAB_VIRTUAL=n | 150   | 25.480   | 26.550   | 26.065   | 26.055   | 0.23495
> SLAB_VIRTUAL=y | 150   | 25.820   | 27.080   | 26.531   | 26.540   | 0.25974
>                |       | +1.33%   | +2.00%   | +1.79%   | +1.86%   | +10.55%
> system time    |       |          |          |          |          |
> SLAB_VIRTUAL=n | 150   | 478.530  | 540.420  | 520.803  | 521.485  | 9.166
> SLAB_VIRTUAL=y | 150   | 530.520  | 572.460  | 552.825  | 552.985  | 7.161
>                |       | +10.86%  | +5.93%   | +6.15%   | +6.04%   | -21.88%
> user time      |       |          |          |          |          |
> SLAB_VIRTUAL=n | 150   | 2373.540 | 2403.800 | 2386.343 | 2385.840 | 5.325
> SLAB_VIRTUAL=y | 150   | 2388.690 | 2426.290 | 2408.325 | 2408.895 | 6.667
>                |       | +0.64%   | +0.94%   | +0.92%   | +0.97%   | +25.20%
>
> 
> I'm not exactly sure why user time increases by almost 1% on Milan, it 
> could be TLB contention.

The other worrying aspect is the increase of +6.15% of system time ... 
which is roughly in line with what we'd expect from a +1.79% increase in 
wall-clock time.

Thanks,

	Ingo


  parent reply	other threads:[~2023-09-20  7:44 UTC|newest]

Thread overview: 48+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-09-15 10:59 Matteo Rizzo
2023-09-15 10:59 ` [RFC PATCH 01/14] mm/slub: don't try to dereference invalid freepointers Matteo Rizzo
2023-09-15 20:50   ` Kees Cook
2023-09-30 11:04   ` Hyeonggon Yoo
2023-09-15 10:59 ` [RFC PATCH 02/14] mm/slub: add is_slab_addr/is_slab_page helpers Matteo Rizzo
2023-09-15 20:55   ` Kees Cook
2023-09-15 10:59 ` [RFC PATCH 03/14] mm/slub: move kmem_cache_order_objects to slab.h Matteo Rizzo
2023-09-15 20:56   ` Kees Cook
2023-09-15 10:59 ` [RFC PATCH 04/14] mm: use virt_to_slab instead of folio_slab Matteo Rizzo
2023-09-15 20:59   ` Kees Cook
2023-09-15 10:59 ` [RFC PATCH 05/14] mm/slub: create folio_set/clear_slab helpers Matteo Rizzo
2023-09-15 21:02   ` Kees Cook
2023-09-15 10:59 ` [RFC PATCH 06/14] mm/slub: pass additional args to alloc_slab_page Matteo Rizzo
2023-09-15 21:03   ` Kees Cook
2023-09-15 10:59 ` [RFC PATCH 07/14] mm/slub: pass slab pointer to the freeptr decode helper Matteo Rizzo
2023-09-15 21:06   ` Kees Cook
2023-09-15 10:59 ` [RFC PATCH 08/14] security: introduce CONFIG_SLAB_VIRTUAL Matteo Rizzo
2023-09-15 21:07   ` Kees Cook
2023-09-15 10:59 ` [RFC PATCH 09/14] mm/slub: add the slab freelists to kmem_cache Matteo Rizzo
2023-09-15 21:08   ` Kees Cook
2023-09-15 10:59 ` [RFC PATCH 10/14] x86: Create virtual memory region for SLUB Matteo Rizzo
2023-09-15 21:13   ` Kees Cook
2023-09-15 21:49     ` Dave Hansen
2023-09-18  8:54       ` Matteo Rizzo
2023-09-15 10:59 ` [RFC PATCH 11/14] mm/slub: allocate slabs from virtual memory Matteo Rizzo
2023-09-15 21:22   ` Kees Cook
2023-09-15 21:57   ` Dave Hansen
2023-10-11  9:17     ` Matteo Rizzo
2023-09-15 10:59 ` [RFC PATCH 12/14] mm/slub: introduce the deallocated_pages sysfs attribute Matteo Rizzo
2023-09-15 21:23   ` Kees Cook
2023-09-15 10:59 ` [RFC PATCH 13/14] mm/slub: sanity-check freepointers Matteo Rizzo
2023-09-15 21:26   ` Kees Cook
2023-09-15 10:59 ` [RFC PATCH 14/14] security: add documentation for SLAB_VIRTUAL Matteo Rizzo
2023-09-15 21:34   ` Kees Cook
2023-09-20  9:04   ` Vlastimil Babka
2023-09-15 15:19 ` [RFC PATCH 00/14] Prevent cross-cache attacks in the SLUB allocator Dave Hansen
2023-09-15 16:30   ` Lameter, Christopher
2023-09-18 12:08     ` Matteo Rizzo
2023-09-18 17:39       ` Ingo Molnar
2023-09-18 18:05         ` Linus Torvalds
2023-09-19 15:48           ` Matteo Rizzo
2023-09-19 16:02             ` Dave Hansen
2023-09-19 17:56               ` Kees Cook
2023-09-19 18:49             ` Linus Torvalds
2023-09-19 13:42         ` Matteo Rizzo
2023-09-19 15:56           ` Dave Hansen
2023-09-20  7:44           ` Ingo Molnar [this message]
2023-09-20  8:49       ` 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=ZQqi4RqpEM7PRGkF@gmail.com \
    --to=mingo@kernel.org \
    --cc=42.hyeyoo@gmail.com \
    --cc=akpm@linux-foundation.org \
    --cc=bp@alien8.de \
    --cc=cl@os.amperecomputing.com \
    --cc=corbet@lwn.net \
    --cc=dave.hansen@intel.com \
    --cc=dave.hansen@linux.intel.com \
    --cc=evn@google.com \
    --cc=hpa@zytor.com \
    --cc=iamjoonsoo.kim@lge.com \
    --cc=jannh@google.com \
    --cc=jordyzomer@google.com \
    --cc=keescook@chromium.org \
    --cc=linux-doc@vger.kernel.org \
    --cc=linux-hardening@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=luto@kernel.org \
    --cc=matteorizzo@google.com \
    --cc=mingo@redhat.com \
    --cc=penberg@kernel.org \
    --cc=peterz@infradead.org \
    --cc=poprdi@google.com \
    --cc=rientjes@google.com \
    --cc=roman.gushchin@linux.dev \
    --cc=tglx@linutronix.de \
    --cc=torvalds@linux-foundation.org \
    --cc=vbabka@suse.cz \
    --cc=x86@kernel.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