There are ieee fp standards ambiguities, where several implementations are
acceptable. Alpha chose one way, Sparc chose another.
I wouldn't expect this as much on an 21264 because more fp is built into the
chip, but I'm betting you have a 21064 or 21164.
This will probably be pretty tricky to track down.
However, why don't you try to "strace" em86 on both of the kernels and see
if anything different shows up.
Actually, it is probably something more insidious, like the emulator relying on
a certain fp behavior.
Just a guess.
--Phil
Compaq: Business Critical Servers Division/Benchmark Performance Engineering
---------------- Alpha, The Fastest Processor on Earth --------------------
Phillip.Ezolt@compaq.com |C|O|M|P|A|Q| ezolt@perf.zko.dec.com
On Tue, 14 Sep 1999, Michal Jaegermann wrote:
>
> To add more information on my posting on the subject from yesterday.
> Alpha-specific differences between kernel 2.2.11, for which em86
> emulator works, and 2.2.12 where em86 is broken consist basically
> from two sets - "core_mcpcia/setup" and changes to "math-emu"
> support. Backing out the first set provides ends up with a kernel
> which is unbootable for some reasons so I tried to revert only
> "math-emu" mods. To my surprise with 2.2.12 kernel and 2.2.13pre7,
> where a contents of arch/alpha/math-emu/ is replaced with
> what was in 2.2.11, em86 - or at least Netscape on an emulator -
> works just fine.
>
> Anybody can shed a light on reasons? This is quite important issue
> to a number of people.
>
> Michal
>
>
> -
> 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/
>
>
-
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/