[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20200713140624.GA17904@localhost.localdomain>
Date: Mon, 13 Jul 2020 17:06:24 +0300
From: Alexey Dobriyan <adobriyan@...il.com>
To: Peter Zijlstra <peterz@...radead.org>
Cc: mingo@...hat.com, will@...nel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] rwsem: fix commas in initialisation
On Mon, Jul 13, 2020 at 01:51:41PM +0200, Peter Zijlstra wrote:
> On Sat, Jul 11, 2020 at 05:59:54PM +0300, Alexey Dobriyan wrote:
> > Leading comma prevents arbitrary reordering of initialisation clauses.
> > The whole point of C99 initialisation is to allow any such reordering.
>
> I'm conflicted on this argument, the only reason I'd be inclined to take
> this patch is that it allows fixing the initialization order to not be
> random.
Yes, this is how the issue was noticed.
> That is, I'd fold in the below.
>
> --- a/include/linux/rwsem.h
> +++ b/include/linux/rwsem.h
> @@ -89,9 +89,9 @@ static inline int rwsem_is_locked(struct
> #define __RWSEM_INITIALIZER(name) \
> { __RWSEM_INIT_COUNT(name), \
> .owner = ATOMIC_LONG_INIT(0), \
> - .wait_list = LIST_HEAD_INIT((name).wait_list), \
> - .wait_lock = __RAW_SPIN_LOCK_UNLOCKED(name.wait_lock),\
> __RWSEM_OPT_INIT(name) \
> + .wait_lock = __RAW_SPIN_LOCK_UNLOCKED(name.wait_lock),\
> + .wait_list = LIST_HEAD_INIT((name).wait_list), \
One less chunk to compile with g++, a billion to go :^)
Powered by blists - more mailing lists