Compile the following test (reduced from gcc.dg/torture/fp-int-convert-float128.c) on i686-pc-linux-gnu with -O -mmmx. It gives an ICE
z.c: In function 'main': z.c:9: internal compiler error: in cgraph_local_info, at cgraph.c:525 Please submit a full bug report, with preprocessed source if appropriate. See <URL:http://gcc.gnu.org/bugs.html> for instructions. I don't know if __float128 is meant to be supported on x86 (provided the user provides the external library with the relevant functions) or not (it's an optional part of the x86_64 ABI), but it should not ICE in any case; it should generate meaningful code or a meaningful error. (-mmmx is required for the __float128 type to be available on x86.) extern void abort (void); extern void exit (int); int main (void) { static volatile int i; static volatile __float128 a; a = i; if (0 != a) abort (); exit (0); } -- Summary: __float128 ICE on x86 Product: gcc Version: 4.2.0 Status: UNCONFIRMED Severity: normal Priority: P3 Component: target AssignedTo: unassigned at gcc dot gnu dot org ReportedBy: jsm28 at gcc dot gnu dot org GCC target triplet: i?86-*-* http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25042
