[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: PowerPC port of PUPA
From: |
Jeroen Dekkers |
Subject: |
Re: PowerPC port of PUPA |
Date: |
Sat, 24 Jan 2004 11:20:46 +0100 |
User-agent: |
Mutt/1.5.5.1+cvs20040105i |
On Sat, Jan 24, 2004 at 01:58:16AM +0100, Johan Rydberg wrote:
> Marco Gerards <address@hidden> wrote:
>
> : Hi,
> :
> : (I hope you do not receive this mail twice, gnus is doing weird stuff
> : here..)
>
> I did. But who cares? I love getting mail. :)
>
> : - Write iso9660fs.
>
> This could should be sharable between architectures, shouldn't it?
Every filesystem should be sharable between architectures. By the
way, I'm already looking at iso9660fs and El Torito for x86 and will
probably implement it for PUPA next month. (At the moment I'm busy
with school)
> Just a few comments on the patch:
>
> --- ../pupa/conf/powerpc-nwmac.rmk 1970-01-01 01:00:00.000000000 +0100
> +++ ./conf/powerpc-nwmac.rmk 2004-01-24 00:15:36.000000000 +0100
> @@ -0,0 +1,49 @@
> +
> +# -*- makefile -*-
> +
> +COMMON_ASFLAGS = -nostdinc -fno-builtin
> +COMMON_CFLAGS = -fno-builtin
>
> Is there a reason for not having builtins? Is it to eliminate the need
> for functions such as memcpy() and puts() (they do not live in the PUPA
> namespace) ?
Those functions are there, they are only prefixed with pupa_ (probably
to be sure they don't conflict with the standard functions). But
IMHO we should use gcc builtins and even the optimized OS-independent
functions from libc (i.e. mem*() and str*()) because those functions
are way faster than what we have now.
Jeroen Dekkers
Re: PowerPC port of PUPA, Yoshinori K. Okuji, 2004/01/24