[Libre-soc-dev] NLnet Ongoing Grant

Luke Kenneth Casson Leighton lkcl at lkcl.net
Sun Sep 3 08:23:31 BST 2023


---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68

On Sun, Sep 3, 2023 at 6:58 AM Jacob Lifshay <programmerjake at gmail.com> wrote:
> before we do that, if PO9 will ever have any 32-bit insns

never. that decision has already been made.

> (icr if luke's latest reversion of PO9 encodings is only 64-bit -- been a while since I looked),
>I think we need to make space in bit 30-31 of the prefix for things like Rc,

absolutely not. again: decision has already been made

> Those edits should only take me a small amount of time, if luke lets me do them.

i cannot stop you from doing them, but they will be wasting
your time as they will not be used.  that decision was final
and was part of commercially-confidential discussions that
i cannot share at this time. or publicly.

>> * creating EXT200-263 under PO9
>> * moving all LD/ST-post-update to EXT200-263, with their exact
>>  respective *pre-update* EXT000-063 encodings
>>  (example: lhz is EXT040, therefore lhzp *is* EXT240, no
>>   arguing, no question: it just *is*)
>
>
> I don't disagree here, but this needs a better rationale than
> "because I said so",

well-spotted. sorry, i am dealing with an ongoing situation at home,
of which you are aware, and it is causing me to have limited time.

> may I suggest "because symmetry and reducing decoder complexity"

symmetry - RISC-paradigm-decoding - means one single bit
goes through, EXTnXX where "if n == 1 it's post-inc else pre-inc".

yes. glad you inferred this correctly.

> Everything sounds good, though I am a bit disappointed
> that I couldn't start working on implementing all the nice
> FP instructions for a change of scenery.

i know. we're in "delivery" mode now, not "research" mode, and
are approximately 2 years behind where "delivery" should be.

> Also, this means that #1026 can't have any FP ops, which means we
> will have a much shorter list to choose from.

good.

the focus there needs to be on the cryptographic primitives.
again for commercially-confidential reasons i cannot say why.


On Sun, Sep 3, 2023 at 7:05 AM Jacob Lifshay <programmerjake at gmail.com> wrote:

> I think we should leave #1025 as all the FP stuff, change its budget parent
> to be Future with EUR 0 assigned, and create a new bug to replace it
> in the ongoing grant.

yes definitely. it's been a lot of work and the context / discussion
needs to be preserved intact.

> As a reminder, please no reusing bugs for different things.

... because you can't remove all comments.  the only time this
is an exception is *if* there are no comments, no links to other
bugs, nothing. even then it's dicey, given that we have a hard
rule about not creating "Orphaned" bugs, and cross-referencing
requirements: tracking down all external links is all but impossible,
hence "no reusing".

l.



More information about the Libre-soc-dev mailing list