bad shifts generated in sim code

bad shifts generated in sim code

[Doug Evans]

I found these in sim/m32r/decodex.c

          unsigned int val = (((insn >> -8) & (3 << 0)));

          unsigned int val = (((insn >> -12) & (1 << 0)));

Someone want to fess up?

Re: bad shifts generated in sim code

[Ben Elliston]

>>>>> "Doug" == Doug Evans <dje@transmeta.com> writes:

  Doug> I found these in sim/m32r/decodex.c
  Doug>           unsigned int val = (((insn >> -8) & (3 << 0)));
  Doug>           unsigned int val = (((insn >> -12) & (1 << 0)));

  Doug> Someone want to fess up?

I'm not sure there is any fessing to do.  Has this file been
regenerated lately?  There was a bug in decoder generation that
generated negative shifts which I thought we'd fixed.

Ben

Re: bad shifts generated in sim code

[Frank Ch. Eigler]

[-- Attachment #1: Type: text/plain, Size: 477 bytes --]

Hi -

On Tue, Dec 10, 2002 at 12:40:08AM -0800, Doug Evans wrote:
> I found these in sim/m32r/decodex.c
>           unsigned int val = (((insn >> -8) & (3 << 0)));
>           unsigned int val = (((insn >> -12) & (1 << 0)));
> [...]

Do these bad shifts still occur with current cgen?

FWIW, working on sim/m32r may be a short-lived pleasure, if the
target-obsoleting eagerness of gdb guys ends up getting that target
removed.  You're welcome to work in sid land. :-)


- FChE

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]