[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAAhV-H5h9zNCmsOumR=_Bryj3-Qarbq7FH=uYbT_3++-8pwfPw@mail.gmail.com>
Date: Wed, 22 Sep 2021 10:43:09 +0800
From: Huacai Chen <chenhuacai@...il.com>
To: Richard Henderson <rth@...ddle.net>
Cc: Arnd Bergmann <arnd@...db.de>,
"Eric W. Biederman" <ebiederm@...ssion.com>,
Huacai Chen <chenhuacai@...ngson.cn>,
Andy Lutomirski <luto@...nel.org>,
Thomas Gleixner <tglx@...utronix.de>,
Peter Zijlstra <peterz@...radead.org>,
Andrew Morton <akpm@...ux-foundation.org>,
David Airlie <airlied@...ux.ie>,
Jonathan Corbet <corbet@....net>,
Linus Torvalds <torvalds@...ux-foundation.org>,
linux-arch <linux-arch@...r.kernel.org>,
"open list:DOCUMENTATION" <linux-doc@...r.kernel.org>,
LKML <linux-kernel@...r.kernel.org>,
Xuefeng Li <lixuefeng@...ngson.cn>,
Yanteng Si <siyanteng@...ngson.cn>,
Jiaxun Yang <jiaxun.yang@...goat.com>
Subject: Re: [PATCH V3 14/22] LoongArch: Add signal handling support
Hi, Richard,
On Tue, Sep 21, 2021 at 5:14 AM Richard Henderson <rth@...ddle.net> wrote:
>
> On 9/19/21 7:36 PM, Huacai Chen wrote:
> > Hi, Arnd,
> >
> > On Sun, Sep 19, 2021 at 5:59 PM Arnd Bergmann <arnd@...db.de> wrote:
> >>
> >> On Sat, Sep 18, 2021 at 9:12 AM Huacai Chen <chenhuacai@...il.com> wrote:
> >>> On Sat, Sep 18, 2021 at 5:10 AM Eric W. Biederman <ebiederm@...ssion.com> wrote:
> >>>> For example does LoongArch have a version without built in floating
> >>>> point support?
> >>>
> >>> Some of these structures seems need rethinking, But we really have
> >>> LoongArch-based MCUs now (no FP, no SMP, and even no MMU).
> >>
> >> NOMMU Linux is kind-of on the way out as interest is fading, so I hope you
> >> don't plan on supporting this in the future.
> >>
> >> Do you expect to see future products with MMU but no FP or no SMP?
> > OK, we will not care no-MMU hardware in Linux, but no-FP and no-SMP
> > hardware will be supported.
>
> Please consider requiring the FP registers to be present even on no-FP hardware.
>
> With this plus the FP data movement instructions (FMOV, MOVGR2FR, MOVFR2GR, FLD, FST), it
> is possible to implement soft-float without requiring a separate soft-float ABI. This can
> vastly simplify compatibility and deployment.
OK, I'll send an updated version.
>
>
> r~
Powered by blists - more mailing lists