[Libre-soc-bugs] [Bug 50] nmigen pinmux

[bugzilla-daemon at libre-soc.org]

https://bugs.libre-soc.org/show_bug.cgi?id=50

--- Comment #12 from Luke Kenneth Casson Leighton <lkcl at lkcl.net> ---
(In reply to andrey from comment #11)
> (In reply to Luke Kenneth Casson Leighton from comment #10)
> > diagram explaining JTAG redirection:
> > https://libre-soc.org/shakti/m_class/JTAG/jtag-block.jpg
> > 
> > 3 phases needed here:
> > 
> > * phase 1 - simple dummy_pinset Resource list
> > 
> > a function is needed which turns the results of dummy_pinset()
> > into:
> > 
> > [UARTResource("uart", 0, tx=..., rx=..),
> >  I2CResource("i2c", 0, scl=..., sda=...),
> >  Resource("gpio", 0, Subsignal("i"...), Subsignal("o"...)
> >  Resource("gpio", 1, Subsignal("i"...), Subsignal("o"...)
> >  ...
> > ]
> 
> I created a file "testing_stage1.py" here:
> https://git.libre-soc.org/?p=pinmux.git;a=blob;f=src/spec/testing_stage1.py;
> h=39690b52aa97cc116d73eafd0e2b5e7c8de0beb9;
> hb=8cb658e7ba3174d976a900065b1e62fcd49de61c
> 
> The file extends the dummy function you mentioned in comment #9

ok, you'll like this: by using that larger function, you increase
the workload required for the follow-on iterative tasks :)

can i suggest now splitting things so that an additional (separate)
task is added which considers using that (much larger) function as
a separate incremental task?

besides, it may turn out that this is the wrong approach, and it would
be better to discover that with much smaller examples than with larger
ones, yeh?

the reason dummy pinset exists was very strategically picked IO:
it covers all 3 types of IOpads.  InType, OutType and InOutType,
in bare minimum.

the only thing it doesn't cover is "banked" IO where one direction
pin covers the direction of multiple pins (as a "bank"). adding sdmmc
would do the trick there.

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