GCC build of HEAD failed for native with your patch on 2004-07-01T09:25:29Z.

[Paolo Bonzini]

These patches should be ruled out:

2004-07-01  Joseph S. Myers  <jsm@polyomino.org.uk> (c-format)
2004-07-01  Nick Clifton  <nickc@redhat.com> (new target hook)
2004-07-01  Paolo Bonzini  <bonzini@gnu.org> (new builtins)
2004-07-01  Richard Henderson  <rth@redhat.com> (two cleanups)
2004-06-30  Roger Sayle  <roger@eyesopen.com> (LSHIFT_EXPR)

These have a bigger chance of causing the failure:

>>+2004-07-01  Richard Henderson  <rth@redhat.com>
>>+
>>+	* cgraph.h (cgraph_build_static_cdtor): Declare.
>>+	* cgraphunit.c (cgraph_build_static_cdtor): New.
>>+	* c-objc-common.c (build_cdtor): Use it.
>>+	* coverage.c (create_coverage): Likewise.
>>+	* libfuncs.h (LTI_gcov_init, gcov_init_libfunc): Remove.
>>+	* optabs.c (init_optabs): Don't set gcov_init_libfunc.
>>+
>>+2004-06-28  Geoffrey Keating  <geoffk@apple.com>
>>+	    Andreas Tobler  <a.tobler@schweiz.ch>
>>+
>>+	PR 15813
>>+	* dwarf2out.c (reg_save): Output DW_CFA_same_value when a
>>+	register is saved in itself.
>>+	(initial_return_save): If the return address is a register,
>>+	it's already there, don't bother to mention it in the CFI.
>>+	(struct queued_reg_save): Add field saved_reg.
>>+	(struct reg_saved_in_data): New.
>>+	(regs_saved_in_regs): New.
>>+	(num_regs_saved_in_regs): New.
>>+	(queue_reg_save): Add extra parameter to specify register saved
>>+	in register.  Remove duplicate entries from queue.  Add comment
>>+	for function.
>>+	(flush_queued_reg_saves): Handle registers saved in registers.
>>+	Update regs_saved_in_regs.  Add comment for function.
>>+	(clobbers_queued_reg_save): Add comment for function.  Allow
>>+	for regs_saved_in_regs.
>>+	(reg_saved_in): New.
>>+	(dwarf2out_frame_debug_expr): Handle saving registers in other
>>+	registers.
>>+	(dwarf2out_frame_debug): Reset regs_saved_in_regs.
>>+	* unwind-dw2.c (execute_cfa_program): Correct handling of
>>+	DW_CFA_same_value.  Add FIXME comment about incorrect implementation
>>+	of DW_CFA_restore_extended.
>>+	* config/rs6000/rs6000.c (rs6000_emit_prologue): Let
>>+	dwarf2out_frame_debug_expr see instructions that save registers
>>+	in other registers or save those other registers in memory.
>>+
>>+	* unwind-dw2.c (DWARF_FRAME_REGISTERS): Move to unwind-dw2.h.
>>+	(_Unwind_FrameState): Likewise.
>>+	* unwind-dw2.h: New.
>>+	* Makefile.in (LIB2ADDEHDEP): Add unwind-dw2.h.
>>+	* config/rs6000/darwin-fallback.c: New file.
>>+	* config/rs6000/darwin.h (MD_FALLBACK_FRAME_STATE_FOR): Define.
>>+	* config/rs6000/t-darwin (LIB2FUNCS_EXTRA): Add darwin-fallback.o.
>>+
>>+2004-07-01  Alan Modra  <amodra@bigpond.net.au>
>>+
>>+	* config/rs6000/rs6000.c (rs6000_mixed_function_arg): Rewrite.
>>+	(function_arg): Use rs6000_arg_size rather than CLASS_MAX_NREGS in
>>+	calculating gpr size for altivec.  Simplify and correct
>>+	rs6000_mixed_function_arg calls.  Call rs6000_mixed_function_arg
>>+	for ABI_V4 gpr case too.  Fix off-by-one error in long double
>>+	reg test.  Generate the correct PARALLEL to handle long double
>>+	for ABI_AIX 32-bit.  Use this for -m32 -mpowerpc64 fpr case too.
>>+	(function_arg_partial_nregs): Align before calculating regs left.
>>+	Don't return info on partial fprs when we need info on gprs.
>>+	Correct long double fpr off-by-one error.

There are loads of C++ failures on x86 as well, I don't know if they're 
related to the PowerPC ICE.  If so, Alan's patch would obviously not be 
responsible as well.

I hope I won't have to embarrass myself... :-)

Paolo