From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 10375 invoked by alias); 2 Jan 2010 16:44:39 -0000 Received: (qmail 10349 invoked by uid 48); 2 Jan 2010 16:44:29 -0000 Date: Sat, 02 Jan 2010 16:44:00 -0000 Message-ID: <20100102164429.10348.qmail@sourceware.org> X-Bugzilla-Reason: CC References: Subject: [Bug target/42536] [4.4/4.5 regression] ICE in spill_failure, at reload1.c:2141 In-Reply-To: Reply-To: gcc-bugzilla@gcc.gnu.org To: gcc-bugs@gcc.gnu.org From: "jakub at gcc dot gnu dot org" Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Archive: List-Post: List-Help: Sender: gcc-bugs-owner@gcc.gnu.org X-SW-Source: 2010-01/txt/msg00163.txt.bz2 ------- Comment #3 from jakub at gcc dot gnu dot org 2010-01-02 16:44 ------- Reduced testcase: /* { dg-do compile { target { { i?86-*-* x86_64-*-* } && ilp32 } } } */ /* { dg-options "-O2 -fno-gcse" } */ struct C; struct B { struct C *b; }; struct C { void (*baz) (struct B *, void *, int); }; typedef union { int f; void *e; } D; struct E { struct B *e; }; struct A { struct E *a1; D *a2; D *a3; }; void foo (long *x, long y) { *(long long *) x = y; } extern long fn1 (D); extern void fn2 (void); void _bar (struct A *x) { register int a asm ("esi"); register D *volatile b asm ("edi"); register int c asm ("ebx"); void *d; asm volatile ("" : "=r" (a), "=r" (b), "=r" (c) : : "memory"); if ((d = b[-2].e) != 0 && b [-2].e < d) { foo (&(((long *) d) + 1) [b[0].f], fn1 (b[-1])); x->a2 = (D *) c; x->a3 = b; fn2 (); } x->a1->e->b->baz (x->a1->e, (void *) (long) a, 1); } Wonder why RA doesn't try to force the memory address into register, that would free up one register from the 4 otherwise needed (2 for the address, 2 for DImode value being stored into the memory). -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42536