Subject: Re: compat code restructuring status
To: None <christos@astron.com>
From: Izumi Tsutsui <tsutsui@ceres.dti.ne.jp>
List: tech-kern
Date: 09/17/2005 17:35:47
FYI, now I'm testing stuff on mips.

In article <dggcmo$6pq$1@sea.gmane.org>
christos@astron.com wrote:

> >> Create subtree for your arch under compat/arch
> >> 
> >> Steal compat/arch/i386/Makefile.inc
> >> 
> >> Convert "sys" subdirectory:
> >> 
> >> . In lib/libc/sys/Makefile.inc there's a list of files under
> >>   !defined(COMPATARCHDIR).
> >
> >How about arch/$ARCH/sys/__sigtramp1.S ?
> >This is also moved to compat/arch/$ARCH/sys/compat___sigtramp2.S .
> 
> I assume that the `2' in the second line is a typo.

Oops, yes my typo.
"Move arch/$ARCH/sys/__sigtramp1.S to
 compat/arch/$ARCH/sys/compat___sigtramp1.S"

> Well, without compat-1.6 you don't have sigreturn, so __sigtramp1.S
> does not work, that is why I moved it.
> 
> >> . Move each file $foo to compat/arch/$ARCH/sys/compat_$foo
> >> 
> >> . Adjust arch/$ARCH/sys/Makefile.inc accordingly
> >
> >There is no arch/$ARCH/sys/Makefile.inc.
> >Adjust arch/$ARCH/Makefile.inc for __sigtramp1.S ?
> 
> Right, this is what he meant.
> 
> >Some archs have __longjmp.c, which contains an entriy for __longjmp14().

Umm more typo, it should be __longjmp14.c, 

> >Should we also rename it longjmp.c ?
> 
> Yes, for consistency.

but thanks for confirmation anyway.

BTW, how about __sigaction14_sigtramp.c ?
---
Izumi Tsutsui