[patch] Fix m68k syscalls

Lennart Sorensen lsorense at csclub.uwaterloo.ca
Thu Sep 10 15:14:17 UTC 2009


On Thu, Aug 13, 2009 at 07:00:55AM -0400, Mike Frysinger wrote:
> On Tuesday 04 August 2009 15:28:04 Maxim Kuvyrkov wrote:
> > Maxim Kuvyrkov wrote:
> > > Mike Frysinger wrote:
> > >> On Wednesday 22 July 2009 10:49:00 Maxim Kuvyrkov wrote:
> > >>> As described in thread
> > >>> http://lists.uclibc.org/pipermail/uclibc/2009-July/042670.html, uClibc
> > >>> build is currently broken on m68k.  The attached patch ports relevant
> > >>> pieces of syscall handling from recent GLIBC thus fixing the problem.
> > >>
> > >> please try the patch i posted as yours appears to contain a lot of
> > >> stuff that the unification patch was supposed to, well, unify:
> > >> http://lists.uclibc.org/pipermail/uclibc/2009-July/042714.html
> > >
> > > I didn't notice your patch before.  Do you plan to check it in soon?
> > >
> > >> also, do you really need your own syscall.[cS] ?  wont the common
> > >> syscall.c work for you now ?
> > >
> > > It should work fine.  In fact, my intention was to unify code from GLIBC
> > > and uClibc ;).
> >
> > I've added a couple of tweaks (removed syscall.c) to your patch and
> > verified that it builds on m68k.  I didn't do runtime testing at this
> > point, but I will sometime later.
> >
> > FWIW, I eyeballed the patch and the common code and didn't spot any
> > problems.  Do you necessarily need confirmation that runtime tests pass
> > or can you check in the patch as is?
> 
> i'd be nice, but really i want someone who actually understands m68k to say it 
> looks sane.  since it's been this long, i'll go ahead and merge it.

I just did a build yesterday for the coldfire (so m68k nommu), and I
didn't have to touch any code to make it build this time.  It runs too.
I still have the same unexplained crash problem I had before, but now
I am trying to upgrade the kernel to see if that helps any.

Anyhow, the issue I had in the past with daemon() using INTERNAL_SYSCALL
which wasn't on m68k is certainly fixed now.  I should try if the daemon()
call now works on m68k.  That would be neat.

-- 
Len Sorensen


More information about the uClibc mailing list