Received-SPF: pass (google.com: domain of linux-kernel+bounces-194096-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Message-ID: <7e2963a3-d471-4593-9170-7f59aa1ce038@huaweicloud.com>
Date: Wed, 29 May 2024 14:37:30 +0200
Precedence: bulk
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: [PATCH] tools/memory-model: Document herd7 (internal)
 representation
To: Boqun Feng <boqun.feng@gmail.com>
Cc: Andrea Parri <parri.andrea@gmail.com>,
 Hernan Ponce de Leon <hernan.poncedeleon@huaweicloud.com>,
 stern@rowland.harvard.edu, will@kernel.org, peterz@infradead.org,
 npiggin@gmail.com, dhowells@redhat.com, j.alglave@ucl.ac.uk,
 luc.maranget@inria.fr, paulmck@kernel.org, akiyks@gmail.com,
 dlustig@nvidia.com, joel@joelfernandes.org, linux-kernel@vger.kernel.org,
 linux-arch@vger.kernel.org
References: <20240524151356.236071-1-parri.andrea@gmail.com>
 <1a3c892c-903e-8fd3-24a6-2454c2a55302@huaweicloud.com>
 <ZlSKYA/Y/daiXzfy@andrea>
 <41bc01fa-ce02-4005-a3c2-abfabe1c6927@huaweicloud.com>
 <ZlYbXZSLPmjTKtaE@boqun-archlinux>
From: Jonas Oberhauser <jonas.oberhauser@huaweicloud.com>
In-Reply-To: <ZlYbXZSLPmjTKtaE@boqun-archlinux>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit


Am 5/28/2024 um 7:58 PM schrieb Boqun Feng:
> On Mon, May 27, 2024 at 03:40:13PM +0200, Jonas Oberhauser wrote:
>>
>>
>> Am 5/27/2024 um 3:28 PM schrieb Andrea Parri:
>>>>> +    |                smp_store_mb | W[once] ->po F[mb]                        |
>>>>
>>>> I expect this one to be hard-coded in herd7 source code, but I cannot find
>>>> it. Can you give me a pointer?
>>>
>>> smp_store_mb() is currently mapped to { __store{once}(X,V); __fence{mb}; } in
>>> the .def file, so it's semantically equivalent to "WRITE_ONCE(); smp_mb();".
>>
>> By the way, I experimented a little with these kind of mappings to see if we
>> can just explicitly encode the mapping there. E.g., I had an idea to use
>>      { __fence{mb-successful-rmw}; __cmpxchg{once}...;
>> __fence{mb-successful-rmw}; }
>>
>> for defining (almost) the current mapping of cmpxchg explicitly.
>>
>> But none of the changes I made were accepted by herd7.
>>
>> Do you know how the syntax works?
>>
> 
> This may not be trivial. Note that cmpxchg() is an expression (it has a
> value), so in .def, we want to define it as an expression. However, the
> C-like multiple-statement expression is not supported by herd parser, in
> other words we want:
> 
> 	{
> 		__fence{mb-successful-rmw};
> 		int tmp = __cmpxchg{once}(...);
> 		__fence{mb-successful-rmw};
> 		tmp;
> 	}

Oh, you're right. Then probably the rule I was violating is that 
value-returning macros can not be defined with {} at all.

Given herd's other syntactic limitations, perhaps the best way would be 
to introduce these macros as

	x = cmpxchg(...) {
		__fence{mb-successful-rmw};
  		x = __cmpxchg{once}(...);
  		__fence{mb-successful-rmw};
	}

since I think x = M(...) is the only way we are allowed to use these 
macros anyways.