bootstrap regression on sparc

[Joel Sherrill]

On 11/11/2011 10:53 PM, David Miller wrote:
> From: David Miller<davem@davemloft.net>
> Date: Fri, 11 Nov 2011 20:41:23 -0500 (EST)
>
>> I haven't looked more deeply at it, but the first recent suspicious change
>> are the basic block handling changes Alan made two days ago:
>>
>> 2011-11-09  Alan Modra<amodra@gmail.com>
>>
>> 	* function.c (bb_active_p): Delete.
>   ...
>
> I've verified that the bootstrap succeeds with this change reverted.
FWIW somewhere in the past few days, a few of the CPU
architectures that support RTEMS have broken.  Yesterday's
update did not build for sparc, i386, m68k, mips, sparc64,

That is in addition to existing regressions from 4.6
on avr, bfin, lm32, and powerpc.

Only arm, h8300, m32c (which now ICEs build rtems),
m32r, and sh build gcc and RTEMS at this point.

 From my perspective, the head doesn't look so good. :(

--joel