linux-mips
[Top] [All Lists]

Re: new asm-mips/io.h

To: Atsushi Nemoto <nemoto@toshiba-tops.co.jp>
Subject: Re: new asm-mips/io.h
From: Ralf Baechle <ralf@oss.sgi.com>
Date: Wed, 28 Nov 2001 01:19:07 +1100
Cc: linux-mips@oss.sgi.com
In-reply-to: <20011127.191022.27957874.nemoto@toshiba-tops.co.jp>; from nemoto@toshiba-tops.co.jp on Tue, Nov 27, 2001 at 07:10:22PM +0900
References: <20011126200946.A8408@dea.linux-mips.net> <20011127.130406.104026562.nemoto@toshiba-tops.co.jp> <20011127180648.H29424@dea.linux-mips.net> <20011127.191022.27957874.nemoto@toshiba-tops.co.jp>
Sender: owner-linux-mips@oss.sgi.com
User-agent: Mutt/1.2.5i
On Tue, Nov 27, 2001 at 07:10:22PM +0900, Atsushi Nemoto wrote:

> ralf> Well, talk to it's developers before it's too late.  Or as it
> ralf> has already happened for some hardware I think we should simply
> ralf> go with your suggestion and make all those functions vectors.
> 
> For me, currently it happens only in big endian and I can live happy
> in a little endian world.  I will create new patch when I REALLY need
> it.  Thank you.

Right now the Linux philosophy is that we don't provide any swapping
facility for I/O port accesses.  Some hardware supports byteswapping for I/O
port and memory access, others must do that in software't, so big endian
systems can set CONFIG_SWAP_IO_SPACE to enable swapping for I/O ports and
memory.  In the past there has been quite some confusion about this and
how to use this properly in drivers ...

  Ralf

<Prev in Thread] Current Thread [Next in Thread>