Re: Updated i386 bootcode rewrite patch...

Alan Modra (alan@SPRI.Levels.UniSA.Edu.Au)
Fri, 15 Oct 1999 09:02:09 +0930 (CST)


On Thu, 14 Oct 1999, David Dyck wrote:

> > [referring to new boot code in 2.3.21]
> > Right, because 2.9.1.0.7 is the absolute minimum; anything below that is
> > *KNOWN TO MISCOMPILE* this code.
>
> So you are saying, before this patch 2.8.1.0.23 was the minimum, but
> now that THIS code is added that 2.9.1.0.7 is the minimum.

I recommend 2.9.5.x.x ftp.varesearch.com/pub/support/hjl/binutils/

> Was there anything wrong with 2.8.1.0.23 except for the boot code?

Try running the gas testsuite through the old assembler ;-)

The major changes in the x86 assembler over the last year or so are:

o Greatly improved syntax checking to help catch hand-assembly errors
o New .intel_syntax mode
o Fixed badly broken 16-bit mode assembly - .code16 and .code16gcc
o Recognise new instructions - AMD, PIII, MMX, XMM etc.
o Assembly bugs fixed in a number of infrequently used instructions:
o fwait form of floating point insn with another prefix.
o "mov %al,(,1)", and other unusual index/base combinations
o probably others, I forget.
o Don't emit prefixes when they will be ignored.
o Always use smallest form of index/base alternatives.

> > I wouldn't be surprised if other asms
> > around the kernel were compiled wrong (or at least sub-optimally) with
> > older binutils.
>
> Could you explain what you mean here a bit more please?

I assume Chris means my last two points above. As far as I know,
2.8.1.0.23 didn't completely mis-assemble anything used by the kernel.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/