[Libre-soc-isa] [Bug 686] create Power ISA test API

bugzilla-daemon at libre-soc.org bugzilla-daemon at libre-soc.org
Sat Sep 18 13:10:18 BST 2021


https://bugs.libre-soc.org/show_bug.cgi?id=686

--- Comment #54 from Luke Kenneth Casson Leighton <lkcl at lkcl.net> ---
(In reply to klehman9 from comment #53)
> https://git.libre-soc.org/?p=openpower-isa.git;a=commit;
> h=82545236fe16a9947d8871848cda26f38c106ab5
> 
> https://git.libre-soc.org/?p=soc.git;a=commit;
> h=9bfa69243b8d2d7f9bcef284e3387c81f935ac70
> 
> Added compare_mem as a separate function...felt it safer that way for now
> but could be tossed into compare I suppose.

yyeh, except it kinda feels better to do it explicitly,
so it's clear what is being done.  also...

> As for the compare, on the sim side memory is basically composed of a list. 
> So went with that method creating that out for the hdl side.  Because of
> this, if a location is zero it will be skipped.  Of course conceivably the
> sim side could store a zero to a location with a write, and this action
> would be disregarded in the compare operation.  Not ideal, but for now a
> good start.

... yeah nice.  also, at some point, we may wish to make the API compare
only those memory locations that were requested to be changed, by hooking
into the mem read/write and observing the actual addresses requested.
if done carefully it should avoid the problem that although instruction
X requested ST to Y but actually wrote to location Z, and the
comparison *completely missed it*.

missing changes is why i added code that just blithely compares
the entire memory (which is kept to only 64 bytes or so)

nicely done.

-- 
You are receiving this mail because:
You are on the CC list for the bug.


More information about the Libre-SOC-ISA mailing list