[Libre-soc-bugs] [Bug 238] POWER Compressed Formal Standard writeup

[bugzilla-daemon at libre-soc.org]

https://bugs.libre-soc.org/show_bug.cgi?id=238

--- Comment #163 from Jacob Lifshay <programmerjake at gmail.com> ---
(In reply to Jacob Lifshay from comment #162)
> (In reply to Alexandre Oliva from comment #160)
> > this speaks against trapping, too, unless you want to trap because it's an
> > unintended fallthrough.  yuck
> 
> It would trap due to an illegal combination, running the illegal instruction
> handler.

(In reply to Jacob Lifshay from comment #158)
> I think the issue was detecting the case:
> 1. 16-bit compressed next=standard-then-compressed
> 2. 10-bit next=standard
> 3. what mode are we now?
> 
> I'm proposing the mode at 3 should either be compressed or 2 should trap
> (trapping preferred).

Note that trapping would occur at instruction 2, where the illegal combination
was detected. it does not occur at 3, which would just be confusing.

An alternate plan is to just trap on any 10-bit instruction when the current
mode is standard-then-compressed (at location 2), since the previous
instruction (at location 1) could have instead switched to just standard mode
instead of standard-then-compressed mode.

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