linux-mips
[Top] [All Lists]

Re: CVS Update@linux-mips.org: linux

To: Thiemo Seufer <ths@networkno.de>
Subject: Re: CVS Update@linux-mips.org: linux
From: "Maciej W. Rozycki" <macro@linux-mips.org>
Date: Mon, 4 Apr 2005 12:26:38 +0100 (BST)
Cc: linux-mips@linux-mips.org
In-reply-to: <20050402101135.GB1641@hattusa.textio>
Original-recipient: rfc822;linux-mips@linux-mips.org
References: <20050401175340Z8226142-1340+5040@linux-mips.org> <20050402101135.GB1641@hattusa.textio>
Sender: linux-mips-bounce@linux-mips.org
On Sat, 2 Apr 2005, Thiemo Seufer wrote:

> > Log message:
> >     Remove useless casts.  Fix formatting.
> 
> This patch leads for 64bit kernels to:
> 
>   CC      arch/mips/mm/pg-sb1.o
> arch/mips/mm/pg-sb1.c: In function `sb1_dma_init':
> arch/mips/mm/pg-sb1.c:220: warning: cast from pointer to integer of different 
> size
> arch/mips/mm/pg-sb1.c:225: warning: passing arg 2 of `__raw_writeq' discards 
> qualifiers from pointer target type
> arch/mips/mm/pg-sb1.c:226: warning: passing arg 2 of `__raw_writeq' discards 
> qualifiers from pointer target type
> arch/mips/mm/pg-sb1.c:227: warning: passing arg 2 of `__raw_writeq' discards 
> qualifiers from pointer target type

 Thanks for pointing this out.  That "const" shouldn't be on "base_reg" 
there, of course.  I'm committing a fix right now.  My apologies for 
inadequate testing.

> arch/mips/mm/pg-sb1.c: In function `clear_page':
> arch/mips/mm/pg-sb1.c:233: warning: cast from pointer to integer of different 
> size
> arch/mips/mm/pg-sb1.c:237: warning: cast from pointer to integer of different 
> size
> arch/mips/mm/pg-sb1.c: In function `copy_page':
> arch/mips/mm/pg-sb1.c:257: warning: cast from pointer to integer of different 
> size
> arch/mips/mm/pg-sb1.c:258: warning: cast from pointer to integer of different 
> size
> arch/mips/mm/pg-sb1.c:262: warning: cast from pointer to integer of different 
> size
> arch/mips/mm/pg-sb1.c:263: warning: cast from pointer to integer of different 
> size

 These are unrelated.  Essentially "CPHYSADDR(foo)" expands to 
"(int)(foo)" (that is, after having removed some unrelated bits) and it's 
not going to work in a portable way if "foo" is a pointer...  Thanks for 
your report though -- this code needs a rewrite for a proper 64-bit 
support and I'll try to have a look at it.

  Maciej

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