linux-mips
[Top] [All Lists]

Re: your mail

To: "'linux-mips@linux-mips.org'" <linux-mips@linux-mips.org>
Subject: Re: your mail
From: Greg Lindahl <lindahl@keyresearch.com>
Date: Sat, 26 Oct 2002 13:32:09 -0700
In-reply-to: <37A3C2F21006D611995100B0D0F9B73CBFE312@tnint11.telogy.design.ti.com>; from nmckee@telogy.com on Sat, Oct 26, 2002 at 03:48:27PM -0400
Original-recipient: rfc822;linux-mips@linux-mips.org
References: <37A3C2F21006D611995100B0D0F9B73CBFE312@tnint11.telogy.design.ti.com>
Sender: linux-mips-bounce@linux-mips.org
User-agent: Mutt/1.2.5.1i
On Sat, Oct 26, 2002 at 03:48:27PM -0400, Zajerko-McKee, Nick wrote:

> I'm porting some code from x86 to mips(32) and noticed that in
> include/asm-mips/siginfo.h differs from include/asm-i386/siginfo.h in the
> order of elements of the sigchld structure.  Was this an oversight or a
> design decision?  I would think that it would be desirable to be almost the
> same as the x86 for userland ease of portability...

User programs normally get recompiled, so anything using the proper
includes IS portable.

The issue only appears if you are using binary translation of x86
programs on mips. For example, this is one:

http://www.transitives.com/products.htm

For this, you need to write a system call translation layer which
rearranges things appropriately. An existing example is the o32 layer
in mips64, and soon the n32 layer in mips64.

-- greg


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