Notices by ✧✦✶✷Catherine✷✶✦✧ (whitequark@mastodon.social), page 10

Lattice iCE40 doesn't have IODELAY but we have IODELAY at home

(yes, this actually does delay the clock somewhat reliably, before/after for length=8)

✔ make one part of a RGMII MAC work by adding LUT4s to a design that do nothing
✔ make the other part of the RGMII MAC work by rearranging some of the wires to connect the exact same pins but with different wires adjacent to each other

today is sorcery day

we tried to make the laptop on the other side of this contraption route through the glasgownet and the linux kernel panicked

i don't know what's worse, that i understand signal integrity well enough to make this work, or that i continue to make assemblies like this despite understanding signal integrity enough to make them work

and i had to give TXCLK (yellow) its own ground (orange) or i'd get Bonus Nibbles every time the data bus would transition 0->f

look at how incredibly crude it is! it's barely even _doing_ anything. it's using a bunch of LUTs to delay the clocks. it's doing things with clock domains that aren't discussed in polite audience. it completely disregards the boundaries of Ethernet frames and just kind of shoves them into the other pipe. the PHY is connected to the MAC using wiring so shoddy half the time the bug was "oh, TXCTL fell out of the socket again". there's only one ground pin connected (the one paired with TXCLK)!

going to 100 Mbps makes packets arrive "too quickly" less often and makes it not merely work, but be _usable_

you can _watch video_. also you get the right blinkenlights :D

incredible. i built an Ethernet MAC that doesn't even packetize things properly and TCP/IP still functions well using it

i love technology

this is @mwk logging in through the glasgownet, running in 10 Mbps mode with a quirk

the quirk is that if you send two packets "too quickly" they get glued into a single packet, either direction

somehow, Element (two copies, concurrently), Akkoma, Discord, and other webpages still manage to load in this environment

i wired this up and wrote half this code so high on progesterone i can only somewhat remember what i was doing and struggled with the list comprehension syntax at one point, and now Wanda is like "I have absolutely no reason to connect to WiFi, this works just fine"

incredible

and yet, it works

https://github.com/GlasgowEmbedded/glasgow/blob/78abdeeb0b8b524bf6e7b8da0564fc6001f270c5/software/glasgow/applet/interface/ethernet_rgmii/__init__.py

@rcombs micro-B is rated for 10x the cycles and also the wear parts are on the cable rather than the device! I question the judgement of anyone preferring mini-B

*looks at Lattice devboard dated 2019* why did they go _back_ from micro-B to mini-B?!

what on earth would possess somebody to put a mini-B connector on a new devboard from 2019? did they run out of micro-B's and found a crate of mini-B's? did they get a kickback? so few options

@foone udevadm info -q all
https://lab.whitequark.org/notes/2016-11-20/writing-udev-rules-for-development-boards/

girl who interprets "the Intel architecture" to refer to iAPX 432

i love it when hardware RST# input is unconditionally multiplexed with an unrelated data pin

falling asleep cuddling a girlthing in the same way you would fall asleep cuddling a rifle (and for the same reason too: it is keeping you safe)

i know multiple entities who have fallen asleep cuddling their rifle. personally i prefer cuddling weapon girls; not only it provides a comparable degree of safety, but also yuri is largely deregulated

Never touch the catgirl.
Please don't. Cause of failure
This may be the cause

USB socks https://www.amazon.co.uk/Rechargeable-6000mAh-Washable-Electric-Settings/dp/B0CK2H66CN

a 10x programmer is a programmer who has ten evil exes