[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <27c2c661-43f2-1297-cc64-9ed3a7c10e26@prevas.dk>
Date: Tue, 31 Mar 2020 21:05:45 +0200
From: Rasmus Villemoes <rasmus.villemoes@...vas.dk>
To: Joe Perches <joe@...ches.com>,
Vegard Nossum <vegard.nossum@...cle.com>,
Andrew Morton <akpm@...ux-foundation.org>
Cc: linux-kernel@...r.kernel.org,
Daniel Santos <daniel.santos@...ox.com>,
Rasmus Villemoes <linux@...musvillemoes.dk>,
Masahiro Yamada <yamada.masahiro@...ionext.com>,
Ian Abbott <abbotti@....co.uk>
Subject: Re: [PATCH] compiler.h: fix error in BUILD_BUG_ON() reporting
On 31/03/2020 21.00, Joe Perches wrote:
> On Tue, 2020-03-31 at 20:56 +0200, Rasmus Villemoes wrote:
>> On 31/03/2020 20.20, Joe Perches wrote:
>>> On Tue, 2020-03-31 at 13:26 +0200, Vegard Nossum wrote:
>>>> #define compiletime_assert(condition, msg) \
>>>> - _compiletime_assert(condition, msg, __compiletime_assert_, __LINE__)
>>>> + _compiletime_assert(condition, msg, __compiletime_assert_, __COUNTER__)
>>>
>>> This might be better using something like __LINE__ ## _ ## __COUNTER__
>>>
>>> as line # is somewhat useful to identify the specific assert in a file.
>>>
>>
>> Eh, if the assert fires, doesn't the compiler's diagnostics already
>> contain all kinds of location information?
>
> I presume if that were enough,
> neither __LINE__ nor __COUNTER__
> would be useful.
Not only useful, necessary: They are used to create a unique identifier.
Which turns out to not be unique when one uses __LINE__, causing the
problem Vegard saw and fixes.
Rasmus
Powered by blists - more mailing lists