Re: avr-gcc making really bad inlining decisions.
Bill Westfield <[email protected]>
| Newsgroups | gmane.comp.hardware.avr.gcc |
|---|---|
| Message-ID | <CAPmNOeBZnWxRG7fcDRwdPveSQ+e+7OoNpmx6geJ2SB3vxdXL-w@mail.gmail.com> |
Another interesting data point or two: 1) This goes back to at least 4.7 2) It's not avr-specific (as you might begin to suspect, given previous discussion.) I can get the same sort of bad behavior out of gcc for x86: BillW-MacOSX-2<10405> /sw/lib/gcc4.7/bin/g++-fsf-4.7 -c -Os -w foo.cpp -fno-inline-small-functions BillW-MacOSX-2<10406> size foo.o __TEXT __DATA __OBJC others dec hex 8725 209 0 0 8934 22e6 BillW-MacOSX-2<10407> /sw/lib/gcc4.7/bin/g++-fsf-4.7 -c -Os -w foo.cpp BillW-MacOSX-2<10408> size foo.o __TEXT __DATA __OBJC others dec hex 29936 209 0 0 30145 75c1 (now, to build up the courage to bug the generic gcc people...)