Re: Help!

Re: Help!

[Frank Ch. Eigler]

Hi -

On Thu, Jan 04, 2001 at 02:15:47PM +1100, Ben Elliston wrote:
: [for sim/common]
: In a pbb engine, the hardware PC is not updated until a branch (ie. if no
: one cares what the PC is throughout sequential flow in a pbb, why update
: it?).  What if I do want to access it?  In the case of a full run  (ie. -t),
: is it made available anywhere?

Yes, in several places.  The scache entries usually carry a copy
(see *vpc; abuf->addr); one argument of the sem-switch entry function
is the starting pc.  Your cgen semantics is supposed to be able to
read the correct pc in the middle of a pbb sequence.  Perhaps you mean
making a c-call -- then you should pass the pc explicitly as an
argument instead of pulling it out of the sim cpu state ("hardware PC"?).

Does this make sense?  Am I answering your question?

- FChE
-- 
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.4 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD4DBQE6U+y5VZbdDOm/ZT0RAoshAJ4usK2BDLKcq1bnemrFsLvjMQqekgCYjTBO
TdWVPSDOr7FAv06xf5clFA==
=3Idk
-----END PGP SIGNATURE-----

Re: PC in mid-pbb

[Ben Elliston]

   Yes, in several places.  The scache entries usually carry a copy (see
   *vpc; abuf->addr); one argument of the sem-switch entry function is
   the starting pc.  Your cgen semantics is supposed to be able to read
   the correct pc in the middle of a pbb sequence.  Perhaps you mean
   making a c-call -- then you should pass the pc explicitly as an
   argument instead of pulling it out of the sim cpu state ("hardware
   PC"?).

I'm hacking the execution loop to emit debug information at the end of every
cycle.  I have `vpc', but not a pc to access for part of the debug
information.  I'll try using one of the places you've mentioned.  Thanks,

Ben

Re: PC in mid-pbb

[Doug Evans]

Ben Elliston writes:
 >    Yes, in several places.  The scache entries usually carry a copy (see
 >    *vpc; abuf->addr); one argument of the sem-switch entry function is
 >    the starting pc.  Your cgen semantics is supposed to be able to read
 >    the correct pc in the middle of a pbb sequence.  Perhaps you mean
 >    making a c-call -- then you should pass the pc explicitly as an
 >    argument instead of pulling it out of the sim cpu state ("hardware
 >    PC"?).
 > 
 > I'm hacking the execution loop to emit debug information at the end of every
 > cycle.  I have `vpc', but not a pc to access for part of the debug
 > information.  I'll try using one of the places you've mentioned.  Thanks,

Define "execution loop".
What kind of debug info?

The pbb (/ threaded interpreter) engine has no real loop that
you can hook into after every cycle (nor do you want one).