[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Diverse Targets to Try
From: |
Peter Maydell |
Subject: |
Re: Diverse Targets to Try |
Date: |
Fri, 25 Feb 2022 10:38:29 +0000 |
On Thu, 24 Feb 2022 at 21:57, Frederick Virchanza Gotham
<cauldwell.thomas@gmail.com> wrote:
>
> I want to test a program of mine on all kinds of computer. I've been
> able to build my program for 32-Bit ARM Big Endian and then run it in
> Qemu-user on a 64-Bit x86 Little Endian desktop computer.
>
> Do any of the targets of qemu-user have a signed number system
> other than Two's Complement? It would be cool to test out my program
> on a sign-magnitude or One's Complement machine.
Nope, we don't have any of those.
> If anyone can think of any other unique attributes of a particular
> target, please let me know.
One other interesting portability issue is assumptions about
alignment. In particular sparc requires exact alignment and will
raise SIGBUS if you for instance try to access a 32-bit value
that's not at a 4-byte-aligned address. However the support for
raising SIGBUS in those situations has only just landed in
QEMU's git tree so you'd need to build from the head of git
(or wait for 7.0 to release in a month or two).
thanks
-- PMM