[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAFULd4bAZ3iJqfDkV5bJ+1PrjaNMZyqpKCB7D9Fq1q5kQBc7rw@mail.gmail.com>
Date: Fri, 20 Oct 2023 09:57:46 +0200
From: Uros Bizjak <ubizjak@...il.com>
To: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: peterz@...radead.org, Nadav Amit <namit@...are.com>,
"the arch/x86 maintainers" <x86@...nel.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Andy Lutomirski <luto@...nel.org>,
Brian Gerst <brgerst@...il.com>,
Denys Vlasenko <dvlasenk@...hat.com>,
"H . Peter Anvin" <hpa@...or.com>,
Thomas Gleixner <tglx@...utronix.de>,
Josh Poimboeuf <jpoimboe@...hat.com>,
Nick Desaulniers <ndesaulniers@...gle.com>
Subject: Re: [PATCH v2 -tip] x86/percpu: Use C for arch_raw_cpu_ptr()
On Thu, Oct 19, 2023 at 9:07 PM Linus Torvalds
<torvalds@...ux-foundation.org> wrote:
>
> On Thu, 19 Oct 2023 at 11:49, Linus Torvalds
> <torvalds@...ux-foundation.org> wrote:
> >
> > Honestly, I've actually never seen gcc rematerialize anything at all.
> >
> > I really only started worrying about remat issues in a theoretical
> > sense, and because I feel it would be relatively *easy* to do for
> > something where the source is a load.
>
> .. I started looking around, since I actually have gcc sources around.
>
> At least lra-remat.cc explicitly says
>
> o no any memory (as access to memory is non-profitable)
>
> so if we could just *rely* on that, it would actually allow us to use
> memory ops without the volatile.
>
> That would be the best of all worlds, of course.
I have made an experiment and changed:
#define __raw_cpu_read(qual, pcp) \
({ \
- *(qual __my_cpu_type(pcp) *)__my_cpu_ptr(&(pcp)); \
+ *(__my_cpu_type(pcp) *)__my_cpu_ptr(&(pcp)); \
})
#define __raw_cpu_write(qual, pcp, val) \
do { \
- *(qual __my_cpu_type(pcp) *)__my_cpu_ptr(&(pcp)) = (val); \
+ *(__my_cpu_type(pcp) *)__my_cpu_ptr(&(pcp)) = (val); \
} while (0)
Basically, I removed "volatile" from read/write accessors. With all
new percpu patches in place the difference in all percpu accesses is:
Reference: 15990 accesses
Patched: 15976 accesses.
So, the difference is 14 fewer accesses. Waaay too low of a gain for a
potential pain.
The code size savings are:
text data bss dec hex filename
25476129 4389468 808452 30674049 1d40c81 vmlinux-new.o
25476021 4389444 808452 30673917 1d40bfd vmlinux-ref.o
So, 108 bytes for the default build.
Uros.
Powered by blists - more mailing lists