PATCH: [3.3 branch] IA64 bootstrap failure

[Andrew Pinski]

> I am testing this patch now.
>
>
> H.J.
> <gcc-3.3-const.patch>

This patch will slow down the compiler iff you bootstrap with a non gcc 
3.0.1.
If this patch goes in I will file a compile time regression bug and 
that will
get a very high Priority as it is a regression from a dot release.

Also repeating what Zack said the stage 1 compiler will never have the 
asm as
the stage 2 or 3 compilers, why do you think this is causing your 
problem?

As other points out this looks like a scheduling problem or a reload 
problem or
the machine dependent reorganisation pass (see 
<http://gcc.gnu.org/PR11320>).

Thanks,
Andrew Pinski