[uClibc]Modification of uClibc for SH architecture

AVENARD,JEAN-YVES (HP-Australia,ex2) jean-yves_avenard at hp.com
Fri May 25 02:11:02 UTC 2001


Hello Erik.

Ooops, I forgot to mentioned it.

There's no need to zero the BSS segment in uClibc in normal case, however,
in our usage with uClinux we have to. Unfortunately, I'm not able to give
you much more details about it
As it doesn't do any "bad" things (just redundant), I've integrated it in
the patch (so it's easier for me to update uClibc in the future).

Jean-Yves

-----Original Message-----
From: Erik Andersen [mailto:andersen at lineo.com]
Sent: Friday, May 25, 2001 11:33 AM
To: AVENARD,JEAN-YVES (HP-Australia,ex2)
Cc: uclibc at uclibc.org
Subject: Re: [uClibc]Modification of uClibc for SH architecture


On Thu May 24, 2001 at 12:36:13PM +1000, AVENARD,JEAN-YVES
(HP-Australia,ex2) wrote:
> Hello Eric.
> 
> Please find here the patch file for the SH architecture. It's been done
> against the latest update from CVS as of May, 24th 2001 at 12:33PM
(EST+10)
> 
> In the current configuration, the SH build is broken and will not compile.
> After this patch, everything should work fine again.
> 
> What's new:
> *vfork is now done in ML in a similar way as the i386 port (no direct call
> to sys_vfork)
> *Will work on SH2 (removed any typical SH3 only instructions). The SH2 is
a
> MMU-less version of the SH3
> *lot of clean-up

Cool, thanks.  Applied.

Was the .bss not being zeroed?
If the kernel's fs/binfmt_foo.c is not clearing the .bss for you,
I consider that a kernel bug.  crt0 shouldn't have to clear it.

 -Erik

--
Erik B. Andersen   email:  andersen at lineo.com
--This message was written using 73% post-consumer electrons--


_______________________________________________
uClibc mailing list
uClibc at uclibc.org
http://uclibc.org/mailman/listinfo/uclibc





More information about the uClibc mailing list