RE: [RFC] h8300 "info registers" fix

["Michael Snyder" <Michael dot Snyder at palmsource dot com>]

Grumble -- is the regcache kept in host order, or target order?
I can never remember.  Anyway, good call.  The read is done using
a gdb_byte[4], while the write is done using an unsigned long.
Surely they can't both be correct -- it should be one or the other.

Besides, a host unsigned long can't be right, 'cause we don't 
even know what size it is.  


-----Original Message-----
From: gdb-patches-owner@sourceware.org on behalf of Daniel Jacobowitz
Sent: Wed 9/13/2006 1:09 PM
To: gdb-patches@sourceware.org
Subject: Re: [RFC] h8300 "info registers" fix
 
On Wed, Sep 13, 2006 at 01:00:17PM -0700, Michael Snyder wrote:
> 
> At Wed, 13 Sep 2006 00:04:58 +0900,
> Yoshinori Sato wrote:
> > I modified it in technique to think that I was appropriate.
> 
> OK, much better.  Just one more question.
> 
> Your change to h8300_pseudo_register_read (and write): is it because
> someone is calling the function with a one-byte buffer?  I assume so, 
> but who is making that call?

I'm not familiar with this target, but are there host endianness
problems with this patch?  Casting an unsigned long * pointer to
a gdb_byte * pointer is very suspicious.

-- 
Daniel Jacobowitz
CodeSourcery