Mark Hatle | 29 Aug 01:51

GCC 3.2

(Since I havn't seen anything else pop up from this list, I thought I'd get the 
ball rolling)

I'm just curious if anyone else here has worked on migrating forward to GCC 3.2 
(on x86) and what problems they have run into.  Especially in the area of 
building the kernel.

So far userland seems to be stable, but the kernel isn't building right..  our 
guess is that it is our fault and not something in the upstream kernel.  (Kernel 
2.4.18 is what we're trying to build right now)

--Mark
MontaVista Software, Inc.

Warly | 29 Aug 10:23

Re: GCC 3.2

Mark Hatle <fray@...> writes:

> (Since I havn't seen anything else pop up from this list, I thought
> I'd get the ball rolling)
>
> I'm just curious if anyone else here has worked on migrating forward
> to GCC 3.2 (on x86) and what problems they have run into.  Especially
> in the area of building the kernel.
>
> So far userland seems to be stable, but the kernel isn't building
> right..  our guess is that it is our fault and not something in the
> upstream kernel.  (Kernel 2.4.18 is what we're trying to build right
> now)

It builds fine here.

--

-- 
Warly

Mark Hatle | 4 Sep 17:24

Re: GCC 3.2

Warly wrote:
> Mark Hatle <fray@...> writes:
 >
>>I'm just curious if anyone else here has worked on migrating forward
>>to GCC 3.2 (on x86) and what problems they have run into.  Especially
>>in the area of building the kernel.
>>
>>So far userland seems to be stable, but the kernel isn't building
>>right..  our guess is that it is our fault and not something in the
>>upstream kernel.  (Kernel 2.4.18 is what we're trying to build right
>>now)
> 
> 
> It builds fine here.
> 

just as an update.. we found the problem.  (We were actually building BOTH 
2.4.17 and 2.4.18) anyway.. apparently the usage of asmlinkage and FASTCALL in 
the do_signal call in the kernel was corrupting registers/stack (or something 
like that) 2.4.18pre1 apparently fixes that problem.

Anyway, I thought you folks might find that interesting.

--Mark


Gmane