Holger Eitzenberger <holger@xxxxxxxxxxxxxxxxxx> writes: > Same result as with SuSE gcc 4.1.0. > > Now I'll go for bootstrapping 4.1.2, will be available soon. Debian gcc > 4.1.2 prerelease does fine BTW, however that might be patched heavily. holger:[gcc-4.1.2-build] $oct-gcc-4.1.2 -v Using built-in specs. Target: i686-pc-linux-gnu Configured with: ../gcc-4.1.2/configure --program-prefix=oct- --program-suffix=-4.1.2 --enable-languages=c,c++,treelang --disable-multilib --enable-threads --with-cpu=i686 --with-tune=i686 --with-dwarf2 Thread model: posix gcc version 4.1.2 Disassembly looks better now: c0110090 <wait_hpet_tick>: c0110090: 53 push %ebx c0110091: b8 08 01 00 00 mov $0x108,%eax c0110096: e8 d5 ff ff ff call c0110070 <hpet_readl> c011009b: 89 c3 mov %eax,%ebx c011009d: 8d 76 00 lea 0x0(%esi),%esi c01100a0: b8 08 01 00 00 mov $0x108,%eax c01100a5: e8 c6 ff ff ff call c0110070 <hpet_readl> c01100aa: 39 c3 cmp %eax,%ebx c01100ac: 74 f2 je c01100a0 <wait_hpet_tick+0x10> c01100ae: 5b pop %ebx c01100af: 90 nop c01100b0: c3 ret c01100b1: eb 0d jmp c01100c0 <hpet_timer_stop_set_go> c01100b3: 90 nop c01100b4: 90 nop ... The calls are no longer optimized away. Can someone please give the bugzilla entry for that? /holger
