linux-kernel - Re: [PATCH] riscv: support CPUs having only "zalrsc" but no "zaamo"

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 for Android: free password hash cracker in your pocket

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

Message-ID: <CAHk-=wi2Y3MuMb+HMROL1+Jx8khu8QQnSX-_dmqTWX-kw7i8dw@mail.gmail.com>
Date: Mon, 19 Jan 2026 09:40:36 -0800
From: Linus Torvalds <torvalds@...ux-foundation.org>
To: Gary Guo <gary@...yguo.net>
Cc: Vladimir Kondratiev <Vladimir.Kondratiev@...ileye.com>, Paul Walmsley <pjw@...nel.org>, 
	Palmer Dabbelt <palmer@...belt.com>, Will Deacon <will@...nel.org>, 
	Peter Zijlstra <peterz@...radead.org>, Boqun Feng <boqun.feng@...il.com>, 
	Mark Rutland <mark.rutland@....com>, Albert Ou <aou@...s.berkeley.edu>, 
	Alexandre Ghiti <alex@...ti.fr>, Yury Norov <yury.norov@...il.com>, 
	Rasmus Villemoes <linux@...musvillemoes.dk>, Chao-ying Fu <cfu@...s.com>, 
	Aleksandar Rikalo <arikalo@...il.com>, Aleksa Paunovic <aleksa.paunovic@...cgroup.com>, 
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>, 
	"linux-riscv@...ts.infradead.org" <linux-riscv@...ts.infradead.org>, "olof@...om.net" <olof@...om.net>
Subject: Re: [PATCH] riscv: support CPUs having only "zalrsc" but no "zaamo"

On Mon, 19 Jan 2026 at 09:17, Gary Guo <gary@...yguo.net> wrote:
>
> I'd be curious to know if any kind of performance evaluation has been done.
> atomic_inc is quite heavily used in the kernel.

LR/SC isn't generally a huge problem in itself - the problem is
typically that it requires a somewhat smart cache coherency to not get
into nasty "almost livelock" situations where heavy contention can
make things very unfair, and while the system on the whole makes
progress, some cores may end up starved.

That typically isn't seen unless you have lots of cores, though.

Of course, any CPU design that then left out AMO instructions in the
name of simplicity, probably *also* didn't put a lot of effort into
making the cache coherency have any kind of fairness guarantees.

So the platform is probably garbage for various other reasons.

             Linus