Received-SPF: pass (google.com: domain of linux-kernel+bounces-135463-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Date: Mon, 08 Apr 2024 06:41:12 -0700
From: "H. Peter Anvin" <hpa@zytor.com>
To: Peter Zijlstra <peterz@infradead.org>,
        Linus Torvalds <torvalds@linux-foundation.org>
CC: Ingo Molnar <mingo@kernel.org>, Thomas Gleixner <tglx@linutronix.de>,
        the arch/x86 maintainers <x86@kernel.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Subject: Re: More annoying code generation by clang
User-Agent: K-9 Mail for Android
In-Reply-To: <20240408084934.GC21904@noisy.programming.kicks-ass.net>
References: <CAHk-=whHWjKK1TOMT1XvxFj8e-_uctJnXPxM=SyWHmW63B_EDw@mail.gmail.com> <20240408084934.GC21904@noisy.programming.kicks-ass.net>
Message-ID: <A120A8EC-AB1E-44FB-AEAB-09DFE282CCF8@zytor.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain;
 charset=utf-8
Content-Transfer-Encoding: quoted-printable

On April 8, 2024 1:49:34 AM PDT, Peter Zijlstra <peterz@infradead=2Eorg> wr=
ote:
>On Thu, Apr 04, 2024 at 03:53:49PM -0700, Linus Torvalds wrote:
>
>> diff --git a/arch/x86/include/asm/barrier=2Eh b/arch/x86/include/asm/ba=
rrier=2Eh
>> index 66e57c010392=2E=2E6159d2cbbfde 100644
>> --- a/arch/x86/include/asm/barrier=2Eh
>> +++ b/arch/x86/include/asm/barrier=2Eh
>> @@ -33,20 +33,15 @@
>>   * Returns:
>>   *     0 - (index < size)
>>   */
>> +#define array_index_mask_nospec(idx,sz) ({	\
>> +	typeof((idx)+(sz)) __idx =3D (idx);	\
>> +	typeof(__idx) __sz =3D (sz);		\
>> +	typeof(__idx) __mask;			\
>> +	asm volatile ("cmp %1,%2; sbb %0,%0"	\
>> +			:"=3Dr" (__mask)		\
>> +			:"ir"(__sz),"r" (__idx)	\
>> +			:"cc");			\
>> +	__mask; })
>
>Should this not carry a comment about the "ir" constraint wanting to be
>"g" except for clang being daft?
>
>(I really wish clang would go fix this, it keeps coming up time and
>again)=2E
>
>> =20
>>  /* Prevent speculative execution past this barrier=2E */
>>  #define barrier_nospec() asm volatile("lfence":::"memory")
>

If the only reason for "ir" as opposed to "g" (=3D "irm") is clang then it=
 really needs to be called out=2E Or better yet, don't do anything and let =
the clang people actually fix their code generation=2E