From: Florian Westphal <fw@strlen.de>
To: Willy Tarreau <w@1wt.eu>
Cc: George Spelvin <lkml@sdf.org>,
netdev@vger.kernel.org, aksecurity@gmail.com,
torvalds@linux-foundation.org, edumazet@google.com,
Jason@zx2c4.com, luto@kernel.org, keescook@chromium.org,
tglx@linutronix.de, peterz@infradead.org, tytso@mit.edu,
lkml.mplumb@gmail.com, stephen@networkplumber.org, fw@strlen.de
Subject: Re: [DRAFT PATCH] random32: make prandom_u32() output unpredictable
Date: Mon, 10 Aug 2020 14:03:02 +0200 [thread overview]
Message-ID: <20200810120302.GD19310@breakpoint.cc> (raw)
In-Reply-To: <20200810114700.GB8474@1wt.eu>
Willy Tarreau <w@1wt.eu> wrote:
> On Sun, Aug 09, 2020 at 06:30:17PM +0000, George Spelvin wrote:
> > Even something simple like buffering 8 TSC samples, and adding them
> > at 32-bit offsets across the state every 8th call, would make a huge
> > difference.
>
> Doing testing on real hardware showed that retrieving the TSC on every
> call had a non negligible cost, causing a loss of 2.5% on the accept()
> rate and 4% on packet rate when using iptables -m statistics. However
> I reused your idea of accumulating old TSCs to increase the uncertainty
> about their exact value, except that I retrieve it only on 1/8 calls
> and use the previous noise in this case. With this I observe the same
> performance as plain 5.8. Below are the connection rates accepted on
> a single core :
>
> 5.8 5.8+patch 5.8+patch+tsc
> 192900-197900 188800->192200 194500-197500 (conn/s)
>
> This was on a core i7-8700K. I looked at the asm code for the function
> and it remains reasonably light, in the same order of complexity as the
> original one, so I think we could go with that.
>
> My proposed change is below, in case you have any improvements to suggest.
As this relates to networking, you could also hook perturbation into rx/tx
softirq processing. E.g. once for each new napi poll round or only once
for each softnet invocation, depending on cost.
IIRC the proposed draft left a unused prandom_seed() stub around, you could
re-use that to place extra data to include in the hash in percpu data.
next prev parent reply other threads:[~2020-08-10 12:03 UTC|newest]
Thread overview: 68+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-08-08 15:26 Flaw in "random32: update the net random state on interrupt and activity" George Spelvin
2020-08-08 17:07 ` Andy Lutomirski
2020-08-08 18:08 ` Willy Tarreau
2020-08-08 18:13 ` Linus Torvalds
2020-08-08 19:03 ` George Spelvin
2020-08-08 19:49 ` Andy Lutomirski
2020-08-08 21:29 ` George Spelvin
2020-08-08 17:44 ` Willy Tarreau
2020-08-08 18:19 ` Linus Torvalds
2020-08-08 18:53 ` Willy Tarreau
2020-08-08 20:47 ` George Spelvin
2020-08-08 20:52 ` Linus Torvalds
2020-08-08 22:27 ` George Spelvin
2020-08-09 2:07 ` Linus Torvalds
2020-08-11 16:01 ` Eric Dumazet
2020-08-08 19:18 ` Florian Westphal
2020-08-08 20:59 ` George Spelvin
2020-08-08 21:18 ` Willy Tarreau
2020-08-08 20:08 ` George Spelvin
2020-08-08 20:47 ` Linus Torvalds
2020-08-09 6:57 ` [DRAFT PATCH] random32: make prandom_u32() output unpredictable George Spelvin
2020-08-09 9:38 ` Willy Tarreau
2020-08-09 17:06 ` George Spelvin
2020-08-09 17:33 ` Willy Tarreau
2020-08-09 18:30 ` George Spelvin
2020-08-09 19:16 ` Willy Tarreau
2020-08-10 11:47 ` Willy Tarreau
2020-08-10 12:01 ` David Laight
2020-08-10 14:48 ` Willy Tarreau
2020-08-10 12:03 ` Florian Westphal [this message]
2020-08-10 14:53 ` Willy Tarreau
2020-08-10 16:31 ` Linus Torvalds
2020-08-10 16:58 ` Willy Tarreau
2020-08-10 17:45 ` Linus Torvalds
2020-08-10 18:01 ` Willy Tarreau
2020-08-10 21:04 ` Willy Tarreau
2020-08-11 5:26 ` George Spelvin
2020-08-11 5:37 ` Willy Tarreau
2020-08-11 3:47 ` George Spelvin
2020-08-11 3:58 ` Willy Tarreau
[not found] ` <fdbc7d7d-cba2-ef94-9bde-b3ccae0cfaac@gmail.com>
2020-08-09 21:10 ` Marc Plumb
2020-08-09 21:48 ` Linus Torvalds
2020-08-09 13:50 ` Randy Dunlap
[not found] ` <CANEQ_++a4YcwQQ2XhuguTono9=RxbSRVsMw08zLWBWJ_wxG2AQ@mail.gmail.com>
2020-08-09 16:08 ` George Spelvin
[not found] <CA+icZUVnsmf1kXPYFYufStQ_MxnLuxL+EWfDS2wQy1VbAEMwkA@mail.gmail.com>
2020-08-09 21:10 ` Sedat Dilek
[not found] ` <20200809235412.GD25124@SDF.ORG>
[not found] ` <20200810034948.GB8262@1wt.eu>
[not found] ` <20200811053455.GH25124@SDF.ORG>
[not found] ` <20200811054328.GD9456@1wt.eu>
[not found] ` <20200811062814.GI25124@SDF.ORG>
[not found] ` <20200811074538.GA9523@1wt.eu>
2020-08-11 10:51 ` Sedat Dilek
2020-08-11 11:01 ` Sedat Dilek
2020-08-12 3:21 ` Willy Tarreau
2020-08-13 7:53 ` Sedat Dilek
2020-08-13 8:06 ` Willy Tarreau
2020-08-13 8:13 ` Sedat Dilek
2020-08-13 8:27 ` Sedat Dilek
2020-08-13 14:00 ` Eric Dumazet
2020-08-13 16:02 ` Sedat Dilek
2020-08-14 15:32 ` Sedat Dilek
2020-08-14 16:05 ` Willy Tarreau
2020-08-14 16:17 ` Sedat Dilek
2020-08-16 15:01 ` Willy Tarreau
2020-08-16 16:48 ` Sedat Dilek
2020-08-20 3:05 ` Sedat Dilek
2020-08-20 4:33 ` Willy Tarreau
2020-08-20 4:42 ` Sedat Dilek
2020-08-20 6:08 ` Willy Tarreau
2020-08-20 6:58 ` Willy Tarreau
2020-08-20 8:05 ` Willy Tarreau
2020-08-20 12:08 ` Sedat Dilek
[not found] ` <CANEQ_+L+22Hkdqf38Zr0bfq16fcL1Ax2X9fToXV_niHKXCB8aA@mail.gmail.com>
2020-08-27 1:09 ` Willy Tarreau
2020-08-27 7:08 ` Sedat Dilek
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=20200810120302.GD19310@breakpoint.cc \
--to=fw@strlen.de \
--cc=Jason@zx2c4.com \
--cc=aksecurity@gmail.com \
--cc=edumazet@google.com \
--cc=keescook@chromium.org \
--cc=lkml.mplumb@gmail.com \
--cc=lkml@sdf.org \
--cc=luto@kernel.org \
--cc=netdev@vger.kernel.org \
--cc=peterz@infradead.org \
--cc=stephen@networkplumber.org \
--cc=tglx@linutronix.de \
--cc=torvalds@linux-foundation.org \
--cc=tytso@mit.edu \
--cc=w@1wt.eu \
/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).