Subject: Re: ERROR: invalid spinlock number: 0
To: Michael Paquier <michael@paquier.xyz>,
 Thomas Munro <thomas.munro@gmail.com>
Cc: pgsql-hackers <pgsql-hackers@postgresql.org>
References: <7ef8708c-5b6b-edd3-2cf2-7783f1c7c175@oss.nttdata.com>
 <YCUpTL8bZHa+9K+0@paquier.xyz>
 <79698f73-7dd7-d9e3-5322-4032b4428653@oss.nttdata.com>
 <YCowabyl1dwan+JP@paquier.xyz>
 <CA+hUKG+tZHpADD_BsPp=c3EiYeenTTx5AxsZVGkv1mA0J3O21w@mail.gmail.com>
 <YCpQwdyXErcTRhVX@paquier.xyz>
From: Fujii Masao <masao.fujii@oss.nttdata.com>
Message-ID: <0c58fb4f-3492-c69e-c9fb-56dc252d22fd@oss.nttdata.com>
Date: Mon, 15 Feb 2021 20:49:08 +0900
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.14; rv:78.0)
 Gecko/20100101 Thunderbird/78.7.1
MIME-Version: 1.0
In-Reply-To: <YCpQwdyXErcTRhVX@paquier.xyz>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Precedence: bulk


On 2021/02/15 19:45, Michael Paquier wrote:
> On Mon, Feb 15, 2021 at 10:47:05PM +1300, Thomas Munro wrote:
>> Why not initialise it in WalRcvShmemInit()?
> 
> I was thinking about doing that as well, but we have no real need to
> initialize this stuff in most cases, say standalone deployments.  In
> particular for the fallback implementation of atomics, we would
> prepare a spinlock for nothing.

But on second thought, if we make WalRceiverMain() call pg_atomic_init_u64(),
the variable is initialized (i,e., SpinLockInit() is called in --disable-atomics)
every time walreceiver is started. That may be problematic? If so, the variable
needs to be initialized in WalRcvShmemInit(), instead.

BTW, the recent commit 46d6e5f567 has the similar issue. The variable
that commit added is initialized in InitProcess(), but maybe should be done
in InitProcGlobal() or elsewhere.

Regards,

-- 
Fujii Masao
Advanced Computing Technology Center
Research and Development Headquarters
NTT DATA CORPORATION