[Libre-soc-bugs] [Bug 982] Support PowerPC ABI in ISACaller

[bugzilla-daemon at libre-soc.org]

https://bugs.libre-soc.org/show_bug.cgi?id=982

--- Comment #72 from Luke Kenneth Casson Leighton <lkcl at lkcl.net> ---
(In reply to Dmitry Selyutin from comment #66)

> I'm
> currently concentrated on bug #981, and I assumed that you'll have to handle
> #982. Please let me know if this is not the case. I don't think #981 will
> take too long to be finished.

ok so from a technical perspective, making any changes or improvements to
cavatools is a non-starter at the moment (far too much to do to get everything
else for cavatools-ppc64 going: decoder, compiler, everything).

cavatools-RISC-V works very well: it is high-performance and if there are
syscall mis-matches then that is solved by running *on native hardware*.
if any "redirection" tables are added it *actually defeats the objective*
of cavatools by slowing down its high performance (50x to 100x faster than
qemu).

for the exact same reason, *when* we come to doing cavatools-ppc64
(which is not right now, that was what the wednesday meeting was about)
then the exact same trick can be done: "if problems on syscall mis-match
due to host-guest being different, then run on an IBM POWER9".

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