[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20231020-cee6ba8c9b3dc3c2a984fbb5@orel>
Date: Fri, 20 Oct 2023 12:27:16 +0200
From: Andrew Jones <ajones@...tanamicro.com>
To: Conor Dooley <conor@...nel.org>
Cc: Clément Léger <cleger@...osinc.com>,
Evan Green <evan@...osinc.com>,
linux-riscv@...ts.infradead.org, devicetree@...r.kernel.org,
linux-kernel@...r.kernel.org, linux-doc@...r.kernel.org,
Palmer Dabbelt <palmer@...osinc.com>,
Paul Walmsley <paul.walmsley@...ive.com>,
Rob Herring <robh+dt@...nel.org>,
Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
Albert Ou <aou@...s.berkeley.edu>,
Jonathan Corbet <corbet@....net>,
Samuel Ortiz <sameo@...osinc.com>
Subject: Re: [PATCH v2 01/19] riscv: hwprobe: factorize hwprobe ISA extension
reporting
On Thu, Oct 19, 2023 at 11:22:26AM +0100, Conor Dooley wrote:
> On Thu, Oct 19, 2023 at 09:26:31AM +0200, Clément Léger wrote:
...
> > BTW, wouldn't
> > it make more sense to get rid out of the unsupported extensions directly
> > at ISA string parsing ? ie, if kernel is compiled without V support,
> > then do not set the bits corresponding to these in the riscv_isa_ext[]
> > array ? But the initial intent was probably to be able to report the
> > full string through cpuinfo.
>
> Yeah, hysterical raisins I guess, it's always been that way. I don't
> think anyone originally thought about such configurations and that is
> how the cpuinfo stuff behaves. I strongly dislike the
> riscv_isa_extension_available() interface, but one of Drew's patches
> does at least improve things a bit. Kinda waiting for some of the
> patches in flight to settle down before deciding if I want to refactor
> stuff to be less of a potential for shooting oneself in the foot.
And I recall promising to try and do something with it too, but that
promise got buried under other promises... It's still on the TODO, at
least!
drew
Powered by blists - more mailing lists