Stormwind_mobile has quit [Remote host closed the connection]
Stormwind_mobile has joined #m-labs
<cr1901_modern>
rjo: Did you do any work on porting smoltcp to a nucleo board? I seem to recall you did, opening a PR for eth support, but... I don't see anything from you on PRs/Issues for https://github.com/stm32-rs/stm32-eth/
<cr1901_modern>
rjo: Ahh I see, in my case, I have an f7 nucleo board. I haven't actually looked to see whether the Ethernet drirvers are sufficiently different from f4 yet.
<cr1901_modern>
(for example, the USB IP are the same on f4, f7, and h7- the Synopsys Designware OTG core with different options enabled.
Stormwind_mobile has quit [Ping timeout: 268 seconds]
attie has joined #m-labs
Stormwind_mobile has joined #m-labs
Stormwind_mobile has quit [Ping timeout: 240 seconds]
attie has quit [Ping timeout: 240 seconds]
attie has joined #m-labs
rohitksingh has joined #m-labs
<adamgreig>
H7 Eth is very different iirc
<adamgreig>
F7 is very similar to f4
<cr1901_modern>
adamgreig: I see, that's good. My f4-disco board died. I have an F7 nucleo I bought but didn't really use
<cr1901_modern>
Examples in Rust world seem firmly for f4 land
<cr1901_modern>
for the most part*
<adamgreig>
Hmm I think there's one f7 Eth for that uni teaching project that is quite independent of the wg
<cr1901_modern>
I don't _mind_ porting the f4 driver to f7. I just wanted some idea of how much work it'll be. Because tbh, I'm only using the nucleo for a proof-of-concept
<cr1901_modern>
for a design that doesn't even _have_ Ethernet once the PoC is done
<adamgreig>
My guess is not much work to port
<adamgreig>
Make sure your f7 has the new silicon revision
<adamgreig>
There's a really awful errata in original silicon that will ruin your week
<adamgreig>
Don't ask how many hours I lost to that on my f7 nucleo, lol
<cr1901_modern>
Yes, my nucleo has that nasty errata
<cr1901_modern>
this is looking pretty dire already lmao
<cr1901_modern>
adamgreig: I have the nasty errata, but I'm going to charitably assume a port won't require me to single step
<cr1901_modern>
so I'll just compare the two sections in the manual for f4 anf f7 and hope for the best
<adamgreig>
No the ethernet one
<adamgreig>
Not affected by single stepping
<cr1901_modern>
Oh, I'm not aware of that one
<adamgreig>
F767?
<cr1901_modern>
I have an F746
<cr1901_modern>
adamgreig: Also, see my q in #rust-embedded
<adamgreig>
Ah maybe it's okay. The original 767 is what was on the f7 nucleo when I got one and it had an errata where it just corrupts odd bits on received packets in rmii mode
<adamgreig>
Mii is ok, rmii has an unavoidable, random, and pvt dependent corruption
<cr1901_modern>
Have I mentioned I have no idea how Ethernet works :)?
<adamgreig>
Guess you'll find out soon :p
<adamgreig>
Shout if you have Qs, but I don't know the answer to your one on #r-e
lkcl has joined #m-labs
* cr1901_modern
nods
<Astro-_>
cr1901_modern: I'd be happy about a patch to stm32-eth; perhaps get PHY communication working first, then (in any order) RX and TX
<cr1901_modern>
Astro-_: Oh I wasn't aware you were in #m-labs. Uhh, I'll look into it today. But if it's going to be a multi-day port, I will have to set it aside for now.
<cr1901_modern>
But that's good b/c I can you directly: Why is there a memory.x file in the stm32-eth repo?
<cr1901_modern>
And doesn't that limit the number of examples you can do to exactly one board?
<Astro-_>
no worries, it took me multiple days for the f4 too ;)
<Astro-_>
the memory.x is used for the examples
<cr1901_modern>
what if I wanted to add support for the board I have. It's an f746 nucleo
<Astro-_>
yes, I have one board so far... do you recommendations for cheap devboards?
<cr1901_modern>
any nucleo board lol :P
<Astro-_>
s/one board/one board type/
<Astro-_>
but very few come with ethernet
<cr1901_modern>
it would be easiest for me to directly add a f746 example to your repo, but I'm not sure how to do that since you already have a memory.x that's not going to be compatible
<cr1901_modern>
General consensus seems to be use a separate crate for each board :/
<cr1901_modern>
i.e. cargo conventions for examples fall apart here
<Astro-_>
sounds good to me
<Astro-_>
yup, that's the problem here
<cr1901_modern>
Astro-_: I can do this as a separate PR later, but any possibility that someone from stm32-rs group can add a stm32-eth-examples repo and move your f4 nucleo examples to there?
<cr1901_modern>
Astro-_: Well it's 3 hours later and I haven't even done a git checkout of stm32-rs. I would say this was a smashing success for productivity today :D