[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <6D0B6A98-9C4B-4889-896E-3C412BFC5397@vmware.com>
Date: Wed, 18 Oct 2023 18:29:06 +0000
From: Nadav Amit <namit@...are.com>
To: Uros Bizjak <ubizjak@...il.com>
CC: Linus Torvalds <torvalds@...ux-foundation.org>,
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>,
Peter Zijlstra <peterz@...radead.org>,
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 Oct 18, 2023, at 9:08 PM, Uros Bizjak <ubizjak@...il.com> wrote:
>
> We can use alias to __seg_gs non-const memory, so the value can be
> accessed without asm. __barrier() will then force reload. Please note
> that any memory clobber, hidden inside asm will also force reload.
For the record, at the time I tried to find a creative solution to
have fine granularity barrier control. I looked into address-spaces,
pure and const function attributes, and use of the restrict keyword.
I did not find a better solution. The behavior of most of these
mechanisms was non-intuitive for me and inconsistent across compilers.
You may succeed where I have failed as you seem more familiar with
the compiler code, but be aware it is a rabbit hole...
Powered by blists - more mailing lists