linux-kernel - Re: [PATCH 9/9] rcu: Switch kvfree_rcu() to new rcu

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: <mc5lknkeda3nifrter5hdkqus7set2gnay4qhhvfxdml5o6axy@cnwuyybrizbc>
Date: Mon, 19 Aug 2024 19:05:33 -0400
From: Kent Overstreet <kent.overstreet@...ux.dev>
To: "Paul E. McKenney" <paulmck@...nel.org>
Cc: rcu@...r.kernel.org, linux-kernel@...r.kernel.org, 
	linux-arch@...r.kernel.org
Subject: Re: [PATCH 9/9] rcu: Switch kvfree_rcu() to new rcu_pending

On Mon, Aug 19, 2024 at 03:18:02PM GMT, Paul E. McKenney wrote:
> On Mon, Aug 19, 2024 at 12:59:35PM -0400, Kent Overstreet wrote:
> > This nets us a slight performance increase, and converts to common
> > code.
> 
> How did you measure the performance, and what was the actual amount
> of the increase?
> 
> Either way, the performance increase needs to be verified by the people
> for whom the current code works well.  As noted a couple months ago,
> I am having a hard time imagining a tree beating a linked list of pages
> of pointers in terms of cache locality.  I added linux-arch on CC in
> case I am out of date on how computer systems work.

The point isn't performance, I was just noting what I saw, and the small
performance increase just came from carefull optimization, nothing
special. The old and new versions are close enough to not matter.

The point of the patchset is to consolidate kvfree_rcu() and
SLAB_TYPESAFE_BY_RCU into something more general, and dare I say it,
something a bit saner.