This bug is very similar to http://gcc.gnu.org/PR21616 but is not fixed with
the patch associated with that bug. Using the latest 3_4 branch code, I'm
seeing the following ICE (compiles fine using GCC 3.3.3 and 4.1):
[EMAIL PROTECTED]> ./install-3_4/bin/gcc -v -Os -m64 -c -msoft-float
-mminimal-toc raid6int8.i
Reading specs from
/home/bergner/install-3_4/lib/gcc/powerpc64-linux/3.4.6/specs
Configured with: /home/bergner/gcc-20060308-3_4/configure
--target=powerpc64-linux --host=powerpc64-linux --build=powerpc64-linux
--with-cpu=default32 --enable-threads=posix --enable-shared
--enable-languages=c --disable-checking --prefix=/home/bergner/install-3_4
Thread model: posix
gcc version 3.4.6
/home/bergner/install-3_4/libexec/gcc/powerpc64-linux/3.4.6/cc1 -fpreprocessed
raid6int8.i -quiet -dumpbase raid6int8.i -m64 -msoft-float -mminimal-toc
-auxbase raid6int8 -Os -version -o /tmp/cc83eiKe.s
GNU C version 3.4.6 (powerpc64-linux)
compiled by GNU C version 3.3.3 (SuSE Linux).
GGC heuristics: --param ggc-min-expand=100 --param ggc-min-heapsize=131072
raid6int8.i: In function `raid6_int8_gen_syndrome':
raid6int8.i:134: error: unable to find a register to spill in class
`FLOAT_REGS'
raid6int8.i:134: error: this is the insn:
(insn:HI 622 624 643 5 (set (mem:DI (plus:DI (reg/v/f:DI 122 [ p ])
(reg/v:DI 66 ctr [orig:124 d ] [124])) [9 S8 A64])
(reg/v:DI 129 [ wp0 ])) 320 {*movdi_internal64} (nil)
(expr_list:REG_DEAD (reg/v:DI 129 [ wp0 ])
(nil)))
raid6int8.i:134: confused by earlier errors, bailing out
--
Summary: [3.4 only] ICE: unable to find a register to spill in
class `FLOAT_REGS'
Product: gcc
Version: 3.4.6
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: target
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: bergner at vnet dot ibm dot com
GCC build triplet: powerpc64-unknown-linux-gnu
GCC host triplet: powerpc64-unknown-linux-gnu
GCC target triplet: powerpc64-unknown-linux-gnu
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26610
