linux-kernel - Re: [PATCH 3/5] regmap: optimize sync() and drop() regcache callbacks

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20200210131157.GE7685@sirena.org.uk>
Date:   Mon, 10 Feb 2020 13:11:57 +0000
From:   Mark Brown <broonie@...nel.org>
To:     Davidlohr Bueso <dave@...olabs.net>
Cc:     akpm@...ux-foundation.org, linux-kernel@...r.kernel.org,
        mcgrof@...nel.org, alex.williamson@...hat.com,
        Davidlohr Bueso <dbueso@...e.de>
Subject: Re: [PATCH 3/5] regmap: optimize sync() and drop() regcache callbacks

On Fri, Feb 07, 2020 at 10:03:03AM -0800, Davidlohr Bueso wrote:
> At a higher memory footprint (two additional pointers per node) we
> can get branchless O(1) tree iterators which can optimize in-order
> tree walks/traversals. For example, on my laptop looking at the rbtree
> debugfs entries:

It's not clear that this is a good optimization - we're increasing the
memory footprint for a bit of performance but our performance is all
relative to the I2C or SPI I/O we're most likely all in the noise here
whereas for larger maps the size cost looks like it could be quite bad
(but equally your case looks to have a lot of 2 register blocks which is
almost the worst case for overhead so...).

That said the code is fine so from that point of view:

Acked-by: Mark Brown <broonie@...nel.org>

Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)