Re: 3.4 i386 code generation problem with leal

[Brad Lucier <lucier at math dot purdue dot edu>]

> 
> > With
> > 
> > gcc -v
> > Reading specs from /usr/local/gcc-3.4/lib/gcc-lib/i686-pc-linux-gnu/3.4/specs
> > Configured with: ../configure --prefix=/usr/local/gcc-3.4 --enable-languages=c --disable-checking
> > Thread model: posix
> > gcc version 3.4 20030403 (experimental)
> > 
> > and the latest gnu binutils, I get the following message when trying to compile
> > 
> > http://www.math.purdue.edu/~lucier/library.i.gz
> > 
> > with
> > 
> > gcc -DHAVE_CONFIG_H -I. -I. -I. -DC_STACK_GROWS_DOWNWARD=1 -DC_INSTALL_LIB_HOME=\"/usr/local/chicken-1.0/lib/chicken\" -DC_USE_C_DEFAULTS -DC_NO_PIC_NO_DLL -DC_BUILDING_LIBCHICKEN -O3 -fomit-frame-pointer -DHAVE_ALLOCA_H -fstrict-aliasing -c library.c -fPIC -DPIC -o .libs/libchicken_la-library.lo -save-temps
> > library.s: Assembler messages:
> > library.s:52511: Error: suffix or operands invalid for `lea'
> > 
> > The offending instruction is
> > 
> > 	leal	28(%ebp), %esi
> 
> Strange, the instruction is valid and is accepted by my gas.  Perhaps
> something is wrong with your assembler?

Yes, it is strange, but it happened.  And it's frustrating as hell.
Perhaps I have too low a limit on stack space or something on my machine?

I have a RedHat 8.0 installation, with gcc-3.4.  With the delivered binutils,
and with gcc-3.4 built with the delivered binutils, I got that message.
Thinking that I might need newer binutils, I built and installed
binutils-2.13.2, and rebuilt gcc-3.4 to use those binutils.  And then
I got the same error message (a few hours after the first error message
on my 350MHz PII).

Anyway, can someone try to compile the actual file I used and if it works
for you, I'll start trying system-level things rather than installing
newer versions of the software.

Brad