linux-kernel - Re: [RFC PATCH v0] RISCV: Report vector unaligned accesses hwprobe

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] [day] [month] [year] [list]

Date: Wed, 5 Jun 2024 17:34:51 +0100
From: Conor Dooley <conor@...nel.org>
To: Evan Green <evan@...osinc.com>
Cc: Jesse Taube <jesse@...osinc.com>, linux-riscv@...ts.infradead.org,
	Paul Walmsley <paul.walmsley@...ive.com>,
	Palmer Dabbelt <palmer@...belt.com>,
	Albert Ou <aou@...s.berkeley.edu>,
	Conor Dooley <conor.dooley@...rochip.com>,
	Charlie Jenkins <charlie@...osinc.com>,
	Andrew Jones <ajones@...tanamicro.com>,
	Clément Léger <cleger@...osinc.com>,
	Xiao Wang <xiao.w.wang@...el.com>, Andy Chiu <andy.chiu@...ive.com>,
	Costa Shulyupin <costa.shul@...hat.com>,
	Björn Töpel <bjorn@...osinc.com>,
	Ben Dooks <ben.dooks@...ethink.co.uk>,
	"Gustavo A. R. Silva" <gustavoars@...nel.org>,
	Alexandre Ghiti <alexghiti@...osinc.com>,
	Erick Archer <erick.archer@....com>, linux-kernel@...r.kernel.org
Subject: Re: [RFC PATCH v0] RISCV: Report vector unaligned accesses hwprobe

> On Wed, Jun 05, 2024 at 08:54:21AM -0700, Evan Green wrote:
> > On Tue, Jun 4, 2024 at 9:25 AM Jesse Taube <jesse@...osinc.com> wrote:
> > What value should be returned when V is not enabled in the kernel, or
> > V is not supported in the hardware? Currently in the code it would be
> > UNKNOWN, right? Is that what we want, or is it worth differentiating
> > "no support for V" from "I don't know the speed of misaligned loads"?
> > Maybe UNKNOWN is the right value, as there are other values to tell
> > you V is not enabled.
> 
> I think UNKNOWN is fine for !V, assuming identical definitions as scalar.

I dunno, maybe we should set it to UNSUPPORTED in that case, but there's
probably some funny behaviour around the v prctl we might need to be
aware of. If the system has vector, we should probably figure out the
actual speed is, but not report it while vector is disabled from the
prctl?

Download attachment "signature.asc" of type "application/pgp-signature" (229 bytes)