Re: [Qemu-devel] [PATCH for-3.1] configure: Disable the nanoMIPS disassembler on big endian hosts

[Peter Maydell]

On Wed, 28 Nov 2018 at 14:59, Stefan Weil <address@hidden> wrote:
>
> Am 27.11.2018 um 15:50 schrieb Philippe Mathieu-Daudé:
> > Stefan noticed the nanoMIPS disassembler uses incorrect format strings
> > on big endian hosts.
> >
> > On big endian hosts, the output will appear as:
> >
> > $ qemu-system-mipsel -M malta -cpu I7200 -semihosting \
> >   -bios nanoboot.bin -d in_asm
> > 0xbfc0000: Asm output not supported on this arch
> >
> > Reported-by: Stefan Weil <address@hidden>
> > Signed-off-by: Philippe Mathieu-Daudé <address@hidden>
> > ---
> >  configure | 6 +++++-
> >  1 file changed, 5 insertions(+), 1 deletion(-)
> >
> > diff --git a/configure b/configure
> > index 0a3c6a72c3..83a7c4746e 100755
> > --- a/configure
> > +++ b/configure
> > @@ -7382,7 +7382,11 @@ for i in $ARCH $TARGET_BASE_ARCH ; do
> >    mips*)
> >      disas_config "MIPS"
> >      if test -n "${cxx}"; then
> > -      disas_config "NANOMIPS"
> > +      # The nanomips disassembler currently does not work correctly
> > +      # on big endian host.
> > +      if test "$bigendian" = "no" ; then
> > +        disas_config "NANOMIPS"
> > +      fi
> >      fi
> >    ;;
> >    moxie*)
>
>
> Thank you. This looks like a patch which might be acceptable for 3.1.
>
> Reviewed-by: Stefan Weil <address@hidden>

It seems a bit like overkill to me. We could just mark it
as a known issue for 3.1. After all it seems unlikely that
many people will be running TCG MIPS on a big-endian host,
and if they are then "debug disassembly with some immediate
values incorrect" might be more useful than "none at all".

thanks
-- PMM