Re: [Qemu-devel] [PATCH] Porting TCG to alpha platform

[Edgar E. Iglesias]

On Sat, Jan 30, 2010 at 12:04:29AM +0100, Stefan Weil wrote:
> Edgar E. Iglesias schrieb:
> > On Fri, Jan 29, 2010 at 09:04:58AM -0800, Richard Henderson wrote:
> >> On 01/28/2010 05:55 PM, identifier scorpio wrote:
> >>> Your code is more skilled and I learn a lot from it, but I haven't yet
> >>> recognized the key difference between your version and mine. In last
> >>> mail, you've said that "Mainly, tcg_target_long was used places that
> >>> could be holding HOST values, which means that the host values would
> >>> have been truncated to 32 bits.", and I can't fully understand it, may
> >>> you put it more clearly.
> >> It was a mis-statement made while tired. Of course tcg_target_long is
> >> defined to int64_t when TCG_TARGET_REG_BITS == 64, so there was no
> >> truncation to 32 bits.
> >>
> >> Have you tested of any of the other embedded targets? You may be able
> >> to find the error with one of those easier than finding it emulating
> >> full-scale Windows. See the images at the bottom of the QEMU download
> >> page.
> >
> > Hi,
> >
> > Although I've never hacked on a TCG backend I'd still like to give some
> > advice.
> >
> > >From experience when developing both the CRIS and the MicroBlaze ISA
> > translators I very much suggest getting something very stable to work
> > in linux-user before going further.
> >
> > I used GCC's C torture tests and OpenSSL's bignum tests to verify the
> > ports. In particular, OpenSSL's bignum tests seem to excersice lot's of
> > cornercases.
> >
> > If you pass those, you've come a long way. In general it's much easier
> > to debug a guest testcase (bntest, etc) that fails compared to huge
> > propretiary blob that fails (windows).
> >
> > Just a suggestion, hope it helps.
> >
> > Cheers,
> > Edgar
> 
> 
> Hi Edgar,
> 
> that was a good suggestion. bntest raises a segmentation fault
> (NULL pointer, p == 0, see below) with qemu-x86_64 on a x86_64 host.
> 
> Regards,
> Stefan
> 
> obase=16
> ibase=16
> test BN_add
> print "test BN_add\n"


Hi Stefan,

I didn't expect x86_64 to bug out so fast. But, like I said.
The bignum tests are not a game. That testsuite is a serious
thing to pass. Both CRIS and MicroBlaze pass them both in
linux-user and in system emul. If not let me know.

Cheers,
Edgar