Re: MIPS defines __kernel_uid

To:	Mika Kukkonen <mika@osdl.org>
Subject:	Re: MIPS defines __kernel_uid_t as int?
From:	Ralf Baechle <ralf@linux-mips.org>
Date:	Thu, 8 Jul 2004 02:59:45 +0200
Cc:	Andrew Morton <akpm@osdl.org>, linux-mips@linux-mips.org
In-reply-to:	<1089223996.20452.31.camel@miku.mobile.lnx.nokia.com>
Original-recipient:	rfc822;linux-mips@linux-mips.org
References:	<1089223996.20452.31.camel@miku.mobile.lnx.nokia.com>
Sender:	linux-mips-bounce@linux-mips.org
User-agent:	Mutt/1.4.1i

On Wed, Jul 07, 2004 at 11:13:16AM -0700, Mika Kukkonen wrote:

> I was doing
>       $ grep __kernel_uid_t include/*/posix_types.h
> 
> and noticed that MIPS is the only architecture that
> defines that to be signed (int) and not unsigned?
> 
> Same with __kernel_uid32_t. Is this intentional
> deviation or just an oversight?

Intentional but with a really weak reason.  Linux/MIPS uses the same
type definitions as SysV rsp. the MIPS ABI in it's EFT (Extended
Fundamental Types).  Not a great idea in retroperspective but that's a
choice made 10 years ago.

I don't think we'd break anything by changing this.  Objections to
changing it?

  Ralf

<Prev in Thread]	Current Thread	[Next in Thread>
Re: MIPS defines __kernel_uid_t as int?, Ralf Baechle <=

Previous by Date:	Re: [PATCH] MIPS getdomainname() off by 1;, Ralf Baechle
Next by Date:	Re: something like sparc32/setarch for mips?, Ralf Baechle
Previous by Thread:	Re: [PATCH] MIPS getdomainname() off by 1;, Ralf Baechle
Next by Thread:	Optimisation, Sridhar Adagada
Indexes:	[Date] [Thread] [Top] [All Lists]

Re: MIPS defines __kernel_uid_t as int?