Re: [Cooker] gcc 2.96

[Thierry Vignaud]

Francis Galiegue <[EMAIL PROTECTED]> writes:
> > 
> > The kernel can compile with these patches  :
> > 
> 
> [...]
> 
> OK, but the kernel should not have to be patched in order to compile. What's
> more, the result is unpredictable. And having to redefine such basic functions
> makes me seriously doubt about the compiler consistency!

all bugs reported on gcc because of the kernel buid failled ended in
kernel bugs. That's the point.
As for the non inlined memcpy in chmouel patch, it should not exists.
It's only here because gcc-2.96 optimize copy operator when the data size
is over 4 bytes on ix86 by inserting a mmecpy call, _after_ the cpp traverse,
so this new memcpy call isn't inlined as others calls were.
fix bad drivers, not provide such file.
this has been done for 2.4.0-testX kernel, not for the 2.2.X.

-- 
www.linux-mandrake.com
somewhere between the playstation and the craystation
        Thierry