This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
re: SH4 regression testing
- From: Daniel Kegel <dank at kegel dot com>
- To: gcc at gcc dot gnu dot org
- Date: Tue, 27 Jan 2004 13:37:21 -0800
- Subject: re: SH4 regression testing
Julian Brown wrote:
Since my interests include both GCC development and console hacking, I
thought I'd mention something I found on the web here. Dan Potter has
released some tools on his web page which he hopes will help with
writing regressions for the SH4 GCC port. His project KallistiOS, an
(impressive!) homebrew operating system for the Sega Dreamcast, only
compiles reliably with old GCC releases (3.0.x) because of apparent
bugs in newer releases.
The tools include an emulator which will run SH4 code on a host PC,
farming off requests to Dreamcast hardware when it is accessed
(DreamIce), as well as an emulator which runs entirely on the host PC
with console output (DryIce).
Although I have got some GCC regressions working on my own port under
emulation (and, erm, failing quite a lot), but I don't know if I could
do anything useful to a more mature port. Perhaps someone else is
interested in these tools?
http://cadcdev.sourceforge.net/
How does his sh4 emulator compare to the one built into gdb,
http://sources.redhat.com/cgi-bin/cvsweb.cgi/src/sim/sh/?cvsroot=src
?
I agree it'd be nice to have more people running gcc regression
tests against sh4.
(I'm doing it now on real sh4 hardware, but
I will have to give that hardware back soon, and I'm not sure I'm
motivated enough to go rustle up more hardware. 'Course, if some
landed on my doorstep, I'd probably use it, hint hint beg beg.)
One of these days I really should write a script to build all the simulators
integrated into gdb and run the gcc regression tests against them...
- Dan