[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAHk-=whD_bqOrBx96DViC_ZrNjg8isyQbTZbx1StxFFOQW+u_A@mail.gmail.com>
Date: Mon, 7 Apr 2025 11:45:07 -0700
From: Linus Torvalds <torvalds@...ux-foundation.org>
To: Guenter Roeck <linux@...ck-us.net>
Cc: Miri Korenblit <miriam.rachel.korenblit@...el.com>,
Johannes Berg <johannes.berg@...el.com>, Oak Zeng <oak.zeng@...el.com>,
Matthew Brost <matthew.brost@...el.com>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>, "the arch/x86 maintainers" <x86@...nel.org>
Subject: Re: Linux 6.15-rc1
On Mon, 7 Apr 2025 at 10:15, Guenter Roeck <linux@...ck-us.net> wrote:
>
> > Hmm. Why didn't that show up on x86-32? Or maybe it did and you're
> > just quoting the xtensa case.
>
> I didn't see the problem on x86. I didn't check others.
> Let's see.... openrisc and parisc are affected as well. I don't build
> allmodconfig for all architectures, so that may miss some.
I think what is going on is that gcc is smart enough to turn the
64-bit divide by a constant into a series of shifts and adds and a
32x32->64 multiply instruction.
Presumably on some other 32-bit architectures, that optimization
doesn't trigger - perhaps due to them not having the 32x32->64
multiply?
Linus
Powered by blists - more mailing lists