Re: [parisc-linux] new gcc-default for hppa

[Randolph Chung <tausq@debian.org>]

> You could try using "objdump -d" on the .o's, then diff'ing.  However
> the diff might be large.  Probably, comparing .S files would be
> best.  Don't use "-g".  However, if you find a difference that looks
> suspicious, "-g" will help you locate the source line.

well.... Grant and I did some more debugging on this.... here's what
we've tried so far:

- pty.c compiled completely with 3.2.2 => crash
- pty.c compiled completely with 3.0.4 => ok
- all the functions except pty_init compiled with 3.2.2, and with
pty_init compiled with 3.0 => ok
- pty.c compiled with 3.2.2 with only one of -fno-schedule-insns
  or -fno-schedule-insns2 => crash
- pty.c compiled with 3.2.2 with both -fno-schedule-insns and
  -fno-schedule-insns2 => ok

For reference, this is the original compile line:
gcc -D__KERNEL__ -I/usr/src/linux/include -Wall -Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer -D__linux__ -pipe -fno-strength-reduce -mno-space-regs -mfast-indirect-calls -mdisable-fpregs -ffunction-sections -march=2.0 -mschedule=8000   -nostdinc -I /usr/lib/gcc-lib/hppa-linux/3.2.2/include -DKBUILD_BASENAME=pty  -DEXPORT_SYMTAB -c pty.c

here are the corresponding assembly files:
without -fno-schedule-insns -fno-schedule-insns2
    http://www.parisc-linux.org/~tausq/pty.s-opt
with -fno-schedule-insns -fno-schedule-insns2
    http://www.parisc-linux.org/~tausq/pty.s-noopt
diff -ywb output
    http://www.parisc-linux.org/~tausq/pty.s.diff

still need to look at the asm output a bit more carefully....

randolph
-- 
Randolph Chung
Debian GNU/Linux Developer, hppa/ia64 ports
http://www.tausq.org/