[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <YuAv+lV324G7pmlk@yury-laptop>
Date: Tue, 26 Jul 2022 11:18:34 -0700
From: Yury Norov <yury.norov@...il.com>
To: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: Dennis Zhou <dennis@...nel.org>,
Guenter Roeck <linux@...ck-us.net>,
Russell King - ARM Linux <linux@...linux.org.uk>,
Catalin Marinas <catalin.marinas@....com>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Geert Uytterhoeven <geert@...ux-m68k.org>,
linux-m68k@...ts.linux-m68k.org
Subject: Re: Linux 5.19-rc8
+ Geert Uytterhoeven <geert@...ux-m68k.org>
+ linux-m68k@...ts.linux-m68k.org
On Tue, Jul 26, 2022 at 10:51:01AM -0700, Linus Torvalds wrote:
> On Tue, Jul 26, 2022 at 10:39 AM Dennis Zhou <dennis@...nel.org> wrote:
> >
> > Are we okay with adding the contract find_*_bit() operations must handle
> > asking for past size properly? FWIW, we'd have to modify most of the
> > iterators in find.h.
>
> So I think we're ok with it, if only it makes the macros simpler.
>
> I also think we should probably look at the m68k case, because while
> that one seems to not have the bug that the arm case had, if we remove
> the arm case the m68k code is now the only non-generic case remaining.
>
> And it just makes me go "maybe we should get rid of the whole
> 'override the generic code' thing entirely?"
>
> I don't think that inlining the first word (like the m68k code does)
> is worth it, but it *is* possible that the architecture-specific
> functions generate better code for some common cases,
We have find_bit_benchmark to check how it works in practice. Would
be great if someone with access to the hardware can share numbers.
> so I think this
> is a "needs looking at the generated code" and not just a blind
> removal.
>
> Linus
Powered by blists - more mailing lists