emeb has joined ##openfpga
<rqou> prpplague: I'm sure it's stupidly obfuscated and not just me because I've successfully decrypted it :P
<prpplague> rqou: hehe
emeb has quit [Ping timeout: 260 seconds]
emeb has joined ##openfpga
lain has quit [Ping timeout: 246 seconds]
prpplague has quit [Quit: Leaving]
unixb0y has quit [Ping timeout: 240 seconds]
unixb0y has joined ##openfpga
prpplague has joined ##openfpga
X-Scale has quit [Ping timeout: 252 seconds]
<prpplague> ho ho hum
X-Scale has joined ##openfpga
GenTooMan has quit [Quit: Leaving]
<gruetzkopf> this is called a thermoluminescent ground fault indicator :D
laintoo has joined ##openfpga
digshadow has joined ##openfpga
laintoo has quit [Client Quit]
laintoo has joined ##openfpga
<pie_> gruetzkopf, jesus :D
<rqou> lol
<prpplague> pie_: doh
<gruetzkopf> also the reason why the old 4-pin-plus-metal-case threephase connectors are illegal now
<gruetzkopf> too easy to get electrocuted on live casing
digshadow has quit [Ping timeout: 260 seconds]
azonenberg_work has joined ##openfpga
<rqou> wait how does the legal 3ph connector work?
<gruetzkopf> the legal one is a CEEform variant. 3p+n+pe as pins, plastic case
scrts has quit [Ping timeout: 245 seconds]
scrts has joined ##openfpga
azonenberg_work has quit [Ping timeout: 264 seconds]
rohitksingh_work has joined ##openfpga
emeb has quit [Quit: Leaving.]
oeuf has joined ##openfpga
egg|egg has quit [Ping timeout: 245 seconds]
azonenberg_work has joined ##openfpga
Bike has quit [Quit: Lost terminal]
scrts has quit [Ping timeout: 244 seconds]
[X-Scale] has joined ##openfpga
X-Scale has quit [Ping timeout: 240 seconds]
[X-Scale] is now known as X-Scale
hackkitten has quit [Read error: Connection reset by peer]
hackkitten has joined ##openfpga
sgstair has quit [Read error: Connection reset by peer]
sgstair has joined ##openfpga
azonenberg_work has quit [Read error: Connection reset by peer]
azonenberg_work has joined ##openfpga
lain has joined ##openfpga
scrts has joined ##openfpga
azonenberg_work has quit [Read error: Connection reset by peer]
azonenberg_work has joined ##openfpga
rohitksingh_wor1 has joined ##openfpga
rohitksingh_work has quit [Ping timeout: 246 seconds]
azonenberg_work has quit [Read error: Connection reset by peer]
Patater_ is now known as Patater
<rqou> woot new nurdrage video
scrts has quit [Ping timeout: 245 seconds]
scrts has joined ##openfpga
azonenberg_work has joined ##openfpga
ayjay_t has quit [Read error: Connection reset by peer]
ayjay_t has joined ##openfpga
eightdot_ has quit [Quit: Reconnecting]
eightdot has joined ##openfpga
eightdot has quit [Quit: Changing server]
eightdot has joined ##openfpga
eightdot has quit [Client Quit]
eightdot has joined ##openfpga
s_frit has quit [Remote host closed the connection]
s_frit has joined ##openfpga
<keesj> can I drive like 20 (green) individual smd leds directly from iCE40-UP5K-SG48 or should I expect to need to add something inbetween? I know there is the 3 RGB leds output a 23mA but what about 20 leds?
<gruetzkopf> really depends on how bright you want them
<gruetzkopf> you can get some of the ultra-efficient LEDs to hurt your eyes at 1mA
<keesj> it is really for fun. it should be visible like little distance (e.g. like a watch)
<keesj> (it might end up being some sort of a badge) say https://www.tindie.com/products/MakersBox/smd-challenge/ but with iCE40-UP5K-SG48 and bit larger and.. using yet undefined power source
<keesj> I was reading up on the SPI flash boot process (and the possiblity of one time programming of the ice40) and looking at the tinyfpga schematics
<keesj> it is quite something (just just slapping the hardware on the board but also developing the bootloader)
<s_frit> with 0201 leds?
<keesj> No, I did manager the challenge myself but this time I want to play with using a stencil and hot air. More probably 0805 or 0603
<s_frit> I hadn't seen that challenge. I accidentally ordered some 0201 bypass caps recently, was "playing" with one last night.
<gruetzkopf> i'd say go with 0603 or possibly 0402
<s_frit> 0603 with a stencil is easy, i did it for the first time today with a hand cut stencil
<gruetzkopf> (but then i'm one of those people who'll hand-solder 0402 without magnification using a 10mm tip
<s_frit> erm, that's with an oven. no idea about hot air
<keesj> s_frit: inthe "final notes" there is also something about the Misery edition (01005)
<s_frit> keesj: but no 01005 LEDs apparently
<s_frit> keesj: what are your plans for the stencil? i'm still learning about stencil techniques
<keesj> s_frit: https://github.com/aspro648/KiCad/tree/master/projects/Attiny/AttinyChallengeMisery you are right (only 1 01005) but .. I felts proud/happy after the normal challenge
<keesj> I plan on ordering it at the same time as the pcb(I might try getting them from oshpark/ostencil) but being in .nl perhaps there is a better place
<s_frit> not sure, i'm in .au, got my last boards from jlcpcb in china
<keesj> I also have an old ir rework station that I can use to preheat the board and then hot air is really easy
<s_frit> nice. i have nothing like that
<s_frit> pcbshopper.com is probably worth trying
<keesj> I am also wondering about using one of those parts libraries (digikey/octopartks/seed studio) and get everything from a single source (possibly let them manufacture when / if it all works)
ayjay_t has quit [Read error: Connection reset by peer]
ayjay_t has joined ##openfpga
<s_frit> sounds good. i have yet to explore the parts libraries, although i've heard about them
<s_frit> keesj: with regards to your original question, you should be able to answer that by looking at the UP5k data sheet. work out what the (source/sink) current for the IO pins is (there should be a separate spec for individual pins, and the port as a whole). obviously the easiest thing is to then use a current limiting resistor in series with each LED. and maybe you want to sink the current to
<s_frit> the pins rather than drive the LEDs from the pins, not sure which is better.
ayjay_t has quit [Read error: Connection reset by peer]
ayjay_t has joined ##openfpga
<keesj> I think I will also want to prototype the setup before ordering (the dev board for iCE40-UP5K-SG48 is about 50 euro's so .. I would preffer to spend that on components)
<s_frit> for that project i think an upduino would be sufficient for prototyping
<s_frit> $16USD including shipping i think
<s_frit> i got one recently, arrived fairly quickly. they don't seem to charge extra for international shipping
<s_frit> but the dev board is almost certainly nicer
rohitksingh_wor1 has quit [Read error: Connection reset by peer]
<keesj> I will get one thanks
<s_frit> just be warned: there has been plenty of criticism of the pcb layout of the upduino (e.g. poor decoupling cap routing)
<keesj> how do you/people like the tinyfpga bx?
<s_frit> i don't own one yet, so i have no opinion really, except to say that from the outside it looks like tinyfpga is doing a great job
<s_frit> so far i have the Nandland Go Board and the upduino. my next board will either be an icebreaker or my own iCE40-UP5K-SG48 board, depending on which one is finished first
<keesj> I have to much already. my only real experience with FPGA was last year on implementing a can controller in vhdl (on a papilio-pro board xilinx sparatn..6?) I accidentally ordered two beaglebone BeagleWire boards
<keesj> I also did small things (implemeting triggers for SPI and similar by modifying code) but this .. open fpga thing .. really rocks
<s_frit> yeah, i'm still learning verilog, very much a beginner
<keesj> I felt I need to move to verilog because of the tools
rohitksingh has joined ##openfpga
emeb has joined ##openfpga
Bike has joined ##openfpga
genii has joined ##openfpga
wpwrak has quit [Ping timeout: 260 seconds]
wpwrak has joined ##openfpga
<whitequark> awygle: i want a 32 port glasgow :S
<whitequark> 32 pin even
<whitequark> 32 port would be cool but um.
<gruetzkopf> ENOPINS?
<whitequark> yeah
<whitequark> can't dump parallel flash
<whitequark> actually not even 32 pins would be quite enough
<whitequark> but at that point it would get... unwieldy
<whitequark> tbh this is another reason to go for hx8k
<keesj> doesn't it support a lower pin mode?
<whitequark> keesj: depnds on the flash
<whitequark> i have several that don't
<whitequark> one requires 40+ test points...
<keesj> https://elinux.org/images/7/74/SST39VF400.pdf (reading this one) it has like .. 18 address lines :/
<whitequark> mine has 22
<whitequark> and 16 data lines
<whitequark> and no 8 bit mode
<whitequark> look at this thing
<whitequark> sorry. 23
<keesj> for a hack .. you perhaps can read the data in like 4 times :) cool project
<gruetzkopf> the old completely parallel storage media get annoying in pin count very quickly
<TAL> a 64pin glasgow would be really nice to get a grasp whats happening in those old Vtech talking wiz computers :P
<pie_> so....mult-glasgow?
<pie_> *multi
<pie_> why make things easy when you can make them hard? time to clock-sync over multiple boards?
<pie_> :p
<awygle> we have that capability
<awygle> whitequark: lol. At some point I may be competing with my employer
<awygle> (not actually)
<awygle> we could do that for rev C, I guess. 95 pins on that package.
<balrog> whitequark: was there an attempt at an open fx3 sdk?
Miyu has joined ##openfpga
<balrog> a bunch of complaining about fx3 and cypress in #domesday86, lol
<q3k> whitequark: make an ECP5 glasgow ^^
rohitksingh has quit [Quit: Leaving.]
rohitksingh has joined ##openfpga
<awygle> we did discuss that
rohitksingh has quit [Quit: Leaving.]
rohitksingh has joined ##openfpga
emeb has quit [Ping timeout: 252 seconds]
emeb has joined ##openfpga
<azonenberg_work> whitequark: 32 pin glasgow
<azonenberg_work> you... just invented starshipraider again :p
<azonenberg_work> It
<azonenberg_work> It's plausible i could stretch it to >32 pins at reduced data rate, i think i could do 96 with some effort
<azonenberg_work> But at that point io connectors and level shifter pcb area would be real concerns
<azonenberg_work> And it wouldnt have all the bells and whistles of the current design
<Ultrasauce> an adapter board with an obnoxious number of shift registers might be an interesting addition
<azonenberg_work> yeah well thats the point of the design
<azonenberg_work> To be modular
<azonenberg_work> and let you hang whatever you want off it within limits of IO bandwidth
<pie_> too bad you cant hang more io banwidth off it
<awygle> download more RAM
<azonenberg_work> pie_: the limiting factor of the prototype board, at least, is the single io card connector
<azonenberg_work> which has 16 LVCMOS33 outputs and 8 LVDS inputs plus I2C for discovery and configuration
<azonenberg_work> Theoretically the outputs could become bidirectional, the LVDS termination is wired such that those pins are input only
<azonenberg_work> The proposed full starshipraider will have four of those connectors
<azonenberg_work> I might be able to fit more? it depends a lot on things like how big the io cards are (thus how dense i can pack them)
<azonenberg_work> But right now i can't do a ton without a functioning lab
<azonenberg_work> i have a prototype io buffer board sitting on a shelf right now waiting for me to solder it
<pie_> poor azonenberg_work people just wont leave him alone to do his damn house :P
<azonenberg_work> pie_: well actually we are making good progress, $wife's dad is in town helping out and has done a ton of work
<prpplague> tinyfpga: nice! i see you are up on sparkfun
<azonenberg_work> in the past two weeks we've almost finished two rooms (instead of our average of one room per 2 weeks) and he's only been here for one of those two weeks
<azonenberg_work> And we have rock up on a fair amount of the dining room too
<pie_> wewt
<azonenberg_work> Tonight i plan to extend the cable tray run out into the next room, hang the data tray
<azonenberg_work> And start hooking up power to more rooms
<azonenberg_work> Will probably have the electrical inspector come out soon to look at my work to date, so I can turn power on to more outlets
<azonenberg_work> Right now i am running everything off one circuit
<azonenberg_work> having 2-3 working even would be a huge help
azonenberg_work has quit [Ping timeout: 252 seconds]
azonenberg_work has joined ##openfpga
rohitksingh has quit [Quit: Leaving.]
<whitequark> awygle: i'm thinking revC would be 16ch and revD would be 32ch
<whitequark> due to cost etc
<whitequark> but we might go for 32ch earlier in principle
<whitequark> given how often i run out of pins
<awygle> sure, sounds reasonable. I was thinking maybe 16 buffered and 16 unbuffered, which only costs another header
<whitequark> i'm thinking that maybe fewer SKUs would be more important than a lower cost model
<whitequark> hmmmm
<whitequark> i think we need to buffer all
<whitequark> i have a lot of non-3V3 logic
<awygle> we weren't going to buffer 8 because we wanted lvds
<whitequark> like this flash is 5V
<whitequark> right, but that's different
<whitequark> i want 32 normal channels AND some LVDS channelss
<whitequark> :p
<awygle> so give you 8-bit lvds, and that's 16 unbuffered SE channels for 32 total
<awygle> Ah I see, it's greed then :-P
<whitequark> well yes :p
<whitequark> i really don't like the idea of normal unbuffered channels
<kc8apf> windows containers are such a massive hack
<awygle> fair enough
<awygle> a shift register as suggested above might not be terrible
<awygle> since I assume your old 5v flash is slow
<whitequark> q3k: after the revC or revD glasgow there will likely be a ecp5 based successor
<q3k> whitequark: neat
<whitequark> assuming there's toolchain support and rqou doesn't procrastinate too much :p
<rqou> i really need to get around to building Bus Armada
<rqou> lol
<whitequark> balrog: it will use fx3
<whitequark> plus ethernet
<awygle> An Infinity Of Glasgows
<rqou> yeah, i really need to fix being tired all the time
<q3k> a group of glasgows is called a ______
<whitequark> fx3 so that you can bootstrap the whole thing without an ftdi programmer or anything
<whitequark> like without reflashing it every time
<whitequark> and you also get a free usb3 interface
<rqou> awygle, whitequark: suggestions for feeling less tired and burned out?
<whitequark> rqou: amphetamines
<zkms> i'm so excited for the opensource ecp5 toolchain being a thing i cant wait
<q3k> rqou: work less
<whitequark> i'm serious btw
<rqou> heh, love the spectrum of replies here
<whitequark> you ideally want pure dextroamphetamine as it has fewer offtarget effects
<kc8apf> whitequark: can't disagree.
<whitequark> but racemate would work ok too
<whitequark> especially if you have high tolerance to like, tachycardia
<rqou> uh, so how would a person go about acquiring that? :P
<whitequark> well you could convince a doc that you have ADHD
<whitequark> or... i have no idea how the hell people acquire those elsewhere in the US
<zkms> you might even actually /have/ ADHD and it might not be diagnosed
<whitequark> i mean yes that's one way to convince them
<whitequark> have ADHD
<rqou> I've been evaluated as not having it
<kc8apf> US docs will tend to prescribe non-amphetamines for ADHD
<whitequark> kc8apf: mph works too
<whitequark> though ideally you want eph or ipph
<whitequark> you used to be able to buy those as rcs
<whitequark> but that loophole was closed a few months ago sadly
<whitequark> you could also fly to ru and help me synthesize ritalinic acid, which is not a scheduled compound and is one esterification away from eph
<whitequark> so like, mix it with ethanol and drain cleaner
<kc8apf> personally I'm on strattera which isn't a stimulant
<whitequark> strattera is a useless piece of shit
<whitequark> it's an nri
<whitequark> atomoxetine
<whitequark> "isn't a stimulant" isn't even true
<whitequark> it is a stimulant, just a really bad one
<whitequark> just enough to give idiots in the healthcare system a good feel about it
<rqou> ok whitequark I'm seriously considering visiting you in RU some time around CCC
<whitequark> welcome to "whitequark has opinions about healthcare"
<whitequark> rqou: i have some ideas for ritalinic acid too
<whitequark> i might want to develop those further?
<kc8apf> whitequark: I've been skeptical about that "not a stimulant" thing but I know next to nothing about pharmacology
<whitequark> kc8apf: there's no precise definition of a stimulant
<whitequark> if caffeine is a stimulant, atomoxetine is definitely a stimulant
<kc8apf> right
<whitequark> if you say that only dopaminergics are stimulants then no but like
<whitequark> snris are considered stimulating antidepressants
<pie_> they stimulate depression? :P
<whitequark> it's literally a classification whose only purpose is to convince people who think stimulants are bad to prescribe it
<whitequark> in case of atomoxetine that is
<kc8apf> I was prescribed atomoxetine due to concerns about addiction
<whitequark> idiotic
<whitequark> it's impossible to get addicted to therapeutic doses of dris if you have adhd
<kc8apf> given what I know about US docs, *shrug*
<whitequark> never happens
<whitequark> in fact as far as i know, you're generally less susceptible to addiction because of the pathophysiology of adhd, but this is something i'm less certain about and need more data
<whitequark> but addiction to adhd meds just isn't a thing
<kc8apf> well, addiction is part of my history so.....
<awygle> "high tolerance to tachycardia" turns out I have the opposite of this.
<whitequark> kc8apf: there are different mechanisms of addiction
<kc8apf> agreed
<whitequark> e.g. addiction to benzos and to amphetamines are essentially unrelated other than deltaFosB part of the mechanism
<whitequark> awygle: same really
<whitequark> i used to not be able to drink coffee at all, i'd feel like i'm dying
<whitequark> it got slightly better with age
<zkms> i get a bit of tolerance building up to adderall but i dont think i've ever had actual for realsies addiction? also the tolerance mostly evaporates after a day or two without taking the stuff
<whitequark> but i still have uh...
<awygle> I drink a lot of coffee but I topped out at like 10mg when trying to titrate up Adderall recently
<whitequark> like my normal dexamp dose while on clomipramine is 6-8 mg and if i take 10 mg i have a real bad time
<awygle> due to feeling like I'm dying
<whitequark> slightly more without clomipramine
<whitequark> zkms: yes, tolerance to dopaminergics is a thing
<whitequark> you can actually rapidly reverse it with L-DOPA
<zkms> what is L-DOPA and where do i get it
<whitequark> for adderall specifcally i mean
<whitequark> which is to say for all DRAs
<whitequark> it wouldn't work and isn't necessary for DRIs
<whitequark> L-DOPA is an antiparkinsonian medication
<whitequark> it's not scheduled or regulated though i'm not sure if it's prescription
<whitequark> it's a dopamine precursor
<rqou> whitequark: random question: how do you feel about (the stereotype of) college students abusing adderall to take finals?
<whitequark> amphetamine is a DRA that is a dopamine releasing agent
<whitequark> so it exhausts the intracellular reserves of dopamine, which are already low due to you having ADHD
<zkms> ic
<zkms> so i should get L-DOPA and take it?
<whitequark> L-DOPA restores them because it bypasses the bottleneck in dopamine synthesis pathway
<zkms> idk how to get it though >_<
<whitequark> yes
<whitequark> start with 20-30mg
<whitequark> not more
<whitequark> you can overdose on L-DOPA, it's not actually dangerous unless you take waaay too much but it's unpleasant
<zkms> ok
<whitequark> i'm sure you can find it easily, maybe in supplement shops?
<whitequark> it's a very simple and frequently used drug
<rqou> so uh, why do "real" doctors not prescribe this?
<whitequark> rqou: because they're idiots?
<whitequark> this is like biochem 101
<zkms> 20-30mg / day, ok
<whitequark> zkms: no not per day
<zkms> wait what
<whitequark> once when you determine that you get too much tolerance
<whitequark> so maybe once per 1-2 weeks?
<zkms> wait i'm confused
<zkms> what dose/interval should i take this stuff
<whitequark> so let's say you have your normal daily dose of adderall
<whitequark> at some point you discover that despite taking it as prescribed you are still tired / scatterbrained / etc
<whitequark> then you take one dose of L-DOPA
<whitequark> that reverses it for a number of days
<whitequark> rqou: i was throwing around ideas with printer about her tolerance to mph
<whitequark> and we discussed uh
<whitequark> 5-HTP
<whitequark> it's a serotonin precursor
* zkms nods
<whitequark> you can use it instead of an SSRI as a rapid-acting antidepressant, except it's metabolized very quickly and you get tolerance to it even quicker so it's not useful long-term
<whitequark> and she was like "can we do that for dopamine" and i instantly reply "sure there's L-DOPA"
<whitequark> we go to our usual pharmacy, buy it, test it on her
ym has quit [Quit: Leaving]
<whitequark> it works exactly like biochem 101 would make you think it works
<rqou> wow
<whitequark> bofh453 also thinks this is a brilliant idea, i think they've tested it too
<whitequark> but in any case it is "obvious"
<whitequark> so like i said
<whitequark> because they're stupidly risk-averse idiots :p
<rqou> am i the only one in this channel who isn't on any meds?
<whitequark> azonenberg?
<whitequark> isnt he healthy like a bull
<rqou> oh right azonenberg is also "normal" :P
<Bike> l-dopa was also what they used for that weird coma dealie
<Ultrasauce> he uses ropes to get high
<whitequark> ##openpharma
<awygle> I was about to say me, but I'm on stomach meds I guess. Nothing prescription currently tho
<whitequark> i mean that's literally what's ##whitequark is
<whitequark> awygle: huh, PPIs aren't prescription?
<awygle> whitequark: nope
<whitequark> US is weird
<awygle> i think they were? but now aren't
<awygle> hence "Prilosec OTC"
<whitequark> oh
<whitequark> dextroamphetamine otc when :p
<awygle> :p
<whitequark> i mean i literally think it is an excellent idea
<rqou> so i have a nice controversial question: what is the point of making people spend years of their life and go into massive debt just so that they can tell you to follow a procedure from a book that you can also read yourself?
<whitequark> rqou: there are several things here
<whitequark> first, systematic education versus being self-taught has fewer areas that you don't know you don't know
<whitequark> (and they are consistent between people)
<whitequark> e.g. unless you already know what malignant hyperthermia is, you're not likely to figure it out
<rqou> ok, i can definitely agree with that
<whitequark> second, in normal countries people don't have crushing student debt
<whitequark> you still waste at least a decade before you become a real doctor
<whitequark> third, many branches require extensive hands-on experience
<Ultrasauce> i definitely didn't have a handle on autodidactics until about halfway through my bachelors
<whitequark> fourth, most people can't follow the instructions that came with their USB drive that makes it work under Windows or whatever
<whitequark> and you expect them to handle medicine?
<whitequark> anyway if you're like, rqou
<whitequark> GPs are useless
<whitequark> specialists are only mostly useless
<whitequark> in the US
<whitequark> so feel free to ignore their existence in case you find that you can do these things just as well, because chances are, you can
<whitequark> well fifth, doctors provide the very important service of gatekeeping
<whitequark> can't have people just deciding they're trans
<whitequark> that would actually be humane
<whitequark> make war not informed consent
<rqou> right
<rqou> ##openhrt when? :p
<whitequark> uh
Maya-sama has joined ##openfpga
<cr1901_modern> There's a few projects re: that
<whitequark> you can literally just go look up WPATH guidelines
<whitequark> all HRT meds are cheap or moderately expensive COTS stuff
<whitequark> estradiol, GnRH agonists, SERMs even
<whitequark> tamoxifen plus a progestin is a useful choice for HRT if you don't wanna grow boobs
<rqou> what about the shortage you were talking about a while ago?
<whitequark> well, you could go progestin-only, that's actually even something a real doctor would prescribe
<whitequark> in seattle they do anyway
<whitequark> but i think progestin plus SERM might be a better idea in certain cases
<whitequark> shortage of what? injectable estradiol?
<rqou> yeah
<whitequark> most people can take oral estradiol, with the exception of the unlucky few that have unusually rapid first-pass metabolism
<whitequark> and there's a few other contraindications
<whitequark> injectable is strictly better pharmacologically but oral is way better than nothing
Miyu has quit [Ping timeout: 244 seconds]
<whitequark> the US also sucks in that it doesn't have GnRH agonists
<whitequark> like spironolactone is an amazingly shitty drug
<whitequark> it's only marginally better than nothing
<rqou> right i remember this rant
<whitequark> CPA also sucks but in different ways
<whitequark> GnRH agonists are mostly expensive because they're mostly cancer drugs
<whitequark> prostate cancer, even
<rqou> interestingly, i know someone (a girl) taking spiro to control acne and it seems to work fine for that
<whitequark> and obviously men would pay good money to avoid having an endoscope with a knife stuck up their dick
<whitequark> rqou: acne is more or less caused by testosterone
<whitequark> this is a simplification of course
<whitequark> antiandrogens, in general, work well against acne
<whitequark> printer used injectable T for a while to see if it helps (it did) and it was like second puberty
<whitequark> acne-wise
<rqou> helps what?
<gruetzkopf> nothing like reading ##openfpga scrollback in the train to $HACKERSPACE while eating fake-asian food
Maya-sama is now known as Miyu
<whitequark> psychological state and being weak af
<rqou> ah
<rqou> lol gruetzkopf
<whitequark> the muscle gain was absurd
<whitequark> she grew such biceps that she stopped fitting into old tshirts
<rqou> gruetzkopf: join the cat{girl|boy} fpga army :P
<whitequark> and that was two small doses. smaller than what came in the single-dose package
* Miyu checks the channel name O.o
<rqou> lol Miyu
<Bike> fpga with an endocrine system, for those very slow changing states
<rqou> looool
<zkms> field programmable endocrine system
<whitequark> lol that would be pretty rad
<whitequark> i mean either
<gruetzkopf> rqou: as long as inplantable uplink ports don't talk USB-PD i'm happy?
<rqou> doesn't hrt just field program your endocrine system? :P
<Bike> kind of a shitty programming interface
<Ultrasauce> why would you not want to be able to charge at 20v
<Bike> you could do like one of those wasps and inject into the pituitary or w/e but it would probably still suck
* whitequark reads "USB PD" and twitches
<rqou> only dragon lolis need that much charging :P
<whitequark> Bike: i think peroxycarbonate once suggested an uh
<whitequark> there's a thing similar to GnRH agonists but i think it has a chemical weapon attached to it
<whitequark> so it's permanent
<whitequark> like estramustine
<whitequark> it works on animals, don't think anyone tested it on humans yet
<whitequark> Bike: re: field programmable endocrine system
<gruetzkopf> Miyu: it's more like ##openeverythingwithterriblevendortools
<whitequark> Bike: there was a paper about reprogramming testes into ovaries by forcing FOXL2 to activate
<Bike> think i heard of that one
<whitequark> it's remarkable how much flexibility there's left
<whitequark> in what should be terminally differentiated
<Bike> i mean it's not super far fetched, fish and such do that stuff all the time
<whitequark> well sure
<whitequark> but you see parthenogenesis in fish or lizards all the time and i don't think anyone's done it in mammals
<whitequark> much less humans
<Bike> probably not. lemme check though
<whitequark> if we could do it we could just turn earth into gensokyo
<Bike> is quick sex changes what makes a place gensokyo
<whitequark> no i mean, everyone'd be a lesbian magical girl
<whitequark> parthenogenesis, not foxl2
<rqou> lol whitequark
<whitequark> zkms: that one is used
<Bike> oh
<whitequark> but i meant the one that kills cells with gnrh affinity
<Bike> fish, isopods, some tunicate, a starfish, a frog
<whitequark> lol tunicate
<Bike> fucking tunicates.
<Bike> actually fucking, in this case
<whitequark> i know right
<whitequark> wait
<awygle> i'm so glad i started hanging out in here lol
<awygle> i learn the most random shit all the time now
<whitequark> Bike: oh by the way the lesbian lizards
<whitequark> depending on where they are in the cycle one of them is the top and the other is the bottom
<whitequark> and they fuck
<whitequark> and that causes ovulation in the one that's bottom
<awygle> lmao that's amazing
<Bike> yeah, some stimulation thing
<whitequark> i know right?!
<Bike> there's some species where to spawn they need sperm but they just discard it
<Bike> like, ok.
<awygle> somebody must have the twitter handle @lesbianlizardbottom
<whitequark> they're also 3n and i spent a while figuring out how the hell that even works
<cr1901_modern> too many chars?
<whitequark> turns out it's basically incomplete meiosis
<awygle> idk how many chars are allowed
<awygle> it's fairly abbreviable tho. "lezlizbot"
<Miyu> the really fun stuff starts when you combine XX and XY cells in a single human being :)
<whitequark> you can sort of see it if you observe the first few divisions
<Bike> ploidy doesn't phase me since i found out wheat was 6n
<whitequark> Miyu: chimeras!
<Miyu> yup
<whitequark> Miyu: did you know that most women who gave birth to an XY fetus are chimeras
<Miyu> two sets of programming trying to figure out a way forward that doesn't involve everything dying off
<whitequark> something like 60%
<whitequark> i mean probably the ones that gave birth to an XX one too
<Miyu> as in true hermaphrodites?
<whitequark> we just can't detect it any easier
<whitequark> no
<Bike> i would be really surprised if 60% of women who gave birth to an xy fetus were hermaphrodites.
<whitequark> they have some fetal cells still stuck in them. Y is detectable in blood samples with PCR
<whitequark> Bike: this has actually happened more than once
<Miyu> ah
<Bike> what has happened
<Miyu> yeah, fetal tissue gets into the mother's system
<Bike> hermaphrodites giving birth?
<whitequark> XY women with mostly XX reproductive system gave birth
<whitequark> a lot
<Bike> oh, yeah.
<whitequark> and we mostly never found out because no one actually thought to try and detect it
<whitequark> because "obviously" that never happens
<gruetzkopf> iirc i saw a paper about that?
<Miyu> XX/XY hermaphrodites can be fertile sometimes
<whitequark> except it happens all the time
<Miyu> depends on what course the development took
<rqou> wait what
<Bike> my impression is that chimerism is underdetected
<Bike> because who cares, usually
<whitequark> yes, exactly
<gruetzkopf> with some family where that's been detected in like the third generation?
<whitequark> yes
<whitequark> in south america iirc
<Miyu> yeah, my own case wasn't detected either, because why would you test every single newborn? :)
<cr1901_modern> I'm lagging but what is "3n" vs "6n"?
<whitequark> Miyu: wait i feel like i've seen you already
<rqou> ah, so it's basically <insert diagram of wwii planes that survived here>?
<whitequark> rqou: similar
<rqou> cr1901_modern: ploidy (number of copies of chromosomes)
* Miyu is just a few characters on a display
<Miyu> :)
<whitequark> i mean, on the web elsewhere
<Bike> triploidy is kind of odd because it's odd.
<Miyu> whitequark: quite possible, I have somewhat of a presence :)
<whitequark> ah Maya right
<Miyu> yup
<Miyu> :)
<Bike> obviously male ants and stuff are one-ploid but that's because ants know what they are doing
<whitequark> i remember reading your answers on quora and having my mind kind of blown
<Miyu> :D
<rqou> wait what?
<whitequark> because i've never seen anyone like you while searching pubmed
<whitequark> Bike: lol ants know what they're doing
<Miyu> I have been trying to get doctors and researchers interested in my case, but they seem reluctant
<Miyu> kind of given up on it at this point
<rqou> Miyu: do you mind giving a tl;dr?
ym has joined ##openfpga
<whitequark> Miyu: yeah, i am vaguely aware of how horribly you were treated
<Miyu> rqou: I'm an XX/XY chimera, also known as a true hermaphrodite
<whitequark> i was pretty angry about that
* Miyu nods :(
<Miyu> currently going through a second puberty as the XX cells apparently are expressing themselves much more strongly now
<Miyu> for over two years now
<Miyu> hormones are funny
<Miyu> and it's funny how the body can reconfigure itself
<Miyu> :)
<Miyu> I was able to stop using HRT as a result
<whitequark> oh wow, that's remarkable
<Bike> how convenient
<Miyu> was getting the brown linea nigra stripe on my abdomen as if I was pregnant
<Miyu> that's how I found out that I was overdosing on estradiol
<Miyu> still trying to make sense of things
<Miyu> not getting a lot of help from doctors :(
<gruetzkopf> hm, the only bio/med research people i know are messing around with HLA stuff
<whitequark> VUC seems like an awfully gatekeepy and not very useful place to me
prpplague has quit [Quit: Leaving]
<rqou> so dumb question: why is there all this gatekeeping? like, what does it achieve? just to make people feel powerful?
<Miyu> no idea
<Miyu> I got thrown in with transgender people for the longest time =/
<whitequark> rqou: not exactly
<Miyu> got forced through the TS protocol
<Miyu> for real TS people it makes sense to have some gatekeeping, I guess
<whitequark> rqou: it's more because people already feel powerful (and are permitted to remain in that place socially) and therefore they have no impetus to get a clue
<whitequark> no it doesn't
* cr1901_modern knows another person who had similar issues
<whitequark> the risk/benefit ratio is way too low to bother with gatekeeping
<whitequark> in the current political climate anyway
<whitequark> especially true if you use gnrh agonists and not shitty not-really-antiandrogens like spironolactone
<whitequark> but i digress
<whitequark> Miyu: i kind of understand why you got grouped like that, given the lack of recorded cases of chrimerism similar to your case
<whitequark> but it's unforgivable that it went on for like a decade
<rqou> hey, anybody have a source for bulk USB-C to micro-B cables? (2.0 only, no superspeed)
<whitequark> taobao?
<rqou> i meant good ones :P
prpplague has joined ##openfpga
<whitequark> what's wrong with taobao
<whitequark> i bought acr supplise there and it was the purest copper i have ever had the pleasure of brazing
<whitequark> *way* purer than namebrand expensive copper pipe bought in RU
<whitequark> not to mention cheaper
<gruetzkopf> rqou: actual microb or microb IDd to host mode?
<rqou> actual micro B, to connect to a device
<gruetzkopf> spec compliant (though i don't know if whitelisted), but haven't seen a reliable seller
<rqou> there are some on amazon us but they're ridiculously expensive
<rqou> like $5 each
<prpplague> what is it you are looking for?
* prpplague is curious
<gruetzkopf> what's your OQ?
<rqou> "bulk" usb-c to usb-micro-b cables
<prpplague> i've had terrible luck with taobao
<prpplague> rqou: ahh
<rqou> and yes, i know this is spec compliant
<rqou> i try not to make devices that are particularly cursed
<prpplague> USB == Unlimited Supply of Bugs
* gruetzkopf checks the c spec for whitelist of legacy adapters
<Miyu> USB-C is a mistake
<Miyu> :)
* Miyu gets chased by a USB 1.0/2.0 stack :<
<prpplague> s/mistake/bastard
<gruetzkopf> oh, it's whitelisted under clause 3.5.7 of the usb-c connector 1.3 spec
<gruetzkopf> (so is a c->minib cable in 3.5.4)
<gruetzkopf> *3.5.5
<rqou> oh wtf
<rqou> $SERVICE is restricted to utf-8 BMP-only
<rqou> cursed
<gruetzkopf> is your usecase big enough to get someone to make them for you
<gruetzkopf> i mean, it's a trivial task for any company that builds usb cables
<rqou> nah, I'll just push the donglehell onto somebody else :p
<rqou> thanks applr
<rqou> *apple
<gruetzkopf> guess why my laptop bag contains a A->C cable plugged into a C->A adapter?
<gruetzkopf> turns out it's actually useful to know beforehand which way the link will come up :P
<gruetzkopf> (then again i also have cursed cables like microb(host)->microb(device) in that bag)
<cr1901_modern> wait how does that help you?
<cr1901_modern> to connect A->C->A
<gruetzkopf> C->A->C
<gruetzkopf> no usb-pd.
<gruetzkopf> no link role flip
<whitequark> wtf
<cr1901_modern> the hell is link role flip?
<whitequark> hahahahahhahhahaha
* whitequark nopes the fuck out
<cr1901_modern> I'm guessing I don't want to know. But I wouldn't figured that the device would remain, well the device. And the host would remain, well, the host.
<cr1901_modern> So the host would bring up the device
<kc8apf> and here I was just fighting with WinXP to see USB2 until I realized I installed the RTM rather than SP3
<gruetzkopf> cr1901_modern: yeah. now imagine plugging two identical cellphones into each other with a c-c cable
<gruetzkopf> both support both roles.
<gruetzkopf> theres three different ways this may be resolved.
<rqou> legacy HNP? :P :P :P
<gruetzkopf> "first to plug in" "random chance" "a-side flip xor b-side flip"
<cr1901_modern> What does "flip" mean? You manually answer a prompt on your screen telling which phone you want to be which role?
<cr1901_modern> It seems like "first to plug in"/"random chance" would have to apply first b/c you won't get to answer said prompt until _after_ the connection's made
<gruetzkopf> soo if i plug my nintendo switch into my phone
<gruetzkopf> if the switch is in standby when i connect, the phone will host USB and provide power
<gruetzkopf> if it is not, it's the other way around
<rqou> gruetzkopf: so what about legacy HNP?
<gruetzkopf> support is basically absent
<gruetzkopf> _never_ seen it implementet
<rqou> really
<rqou> i thought TI calculators did it?
<gruetzkopf> (after which i can manually flip it with android ui)
<rqou> also why does everybody advertise legacy HNP/SRP support?
<gruetzkopf> i'm not in ti calc land?
<cr1901_modern> gruetzkopf: So... what does "a-side flip xor b-side flip" actually mean specifically? "flip" to means "reversing the present role".
<cr1901_modern> "flip" to me*
<gruetzkopf> in this case it means "is the connector upside down or not"
<cr1901_modern> So one side always see the connector as upside down?
<awygle> i'm way late on this but can somebody please explain that "plane full of holes" picture to me? alt_kia retweets it all the time and i've never understood what it's trying to convey
<cr1901_modern> awygle: survivorship bias wiki article
<zkms> survivorship bias
<gruetzkopf> so if you have time for lovecraftian or greater horrors, read the USB-C and the USB-PD spec
<cr1901_modern> gruetzkopf: Can you give me a quick reference to the relevant sections? I'm still confused lmao
<zkms> namely that trying to find out where to up-armor aircraft is kinda non-intuitive because you're only looking at the aircraft that successfully returned from the sortie
<awygle> ohhhhhh okay
<awygle> that makes some sense at least
<cr1901_modern> Seems like which side "decides" to provide power is not only based on whether the connector is flipped but also which side is on standby (switch)
<cr1901_modern> USB is a clusterfuck. Right, the sky is also blue on a clear day
<gruetzkopf> it's not well defined at all which device will have which role before they are plugged together
<gruetzkopf> except for host-only and target-only devices
<gruetzkopf> also, power direction may be inverse of USB direction may be inverse of alt-mode direction.
<cr1901_modern> gruetzkopf: Just to be clear, if one side of a USB C connector is plugged in one way, the device on the other end will have the plug inverted?
<cr1901_modern> (or should*)
<cr1901_modern> and here I thought USB-C was symmetrical
<gruetzkopf> no
<gruetzkopf> you can turn both ends around independently
<gruetzkopf> but one orientation is where cable-cc connects to CC1 in the socket and the superspeed flip muxes are straight-through
<gruetzkopf> and the other is cc on cc2 and flip-muxes crossed
Miyu has quit [Ping timeout: 246 seconds]
<gruetzkopf> certain usb thingies will change their behaviour (like prefer source vs prefer sink) if you flip the plug connected to it
<zkms> awygle: namely (sorry for delay, laptop locked up) that you should focus on areas that are undamaged and not areas that are damaged, because the damage you see wasn't enough to prevent the aircraft from getting back, but the damage you don't see was enough to down the aircraft
<cr1901_modern> So "link role flip" refers to "if the plug is front-side up, do X. If the plug is upside down, do Y"
<cr1901_modern> (zero order approx)?
<awygle> zkms: i got it (found the wiki article) :) thanks!
<zkms> (more generally people use that meme to point out that someone's looking at the wrong population and shouldn't be so quick to draw conclusions on what works / what doesn't
<cr1901_modern> If I understand that correctly, that'll satisfy me for tonight
<gruetzkopf> for a certain , unenumerated subset of typec devices, yes
<cr1901_modern> unenumerated?
<gruetzkopf> there
<gruetzkopf> 's no list of devices which will behave this way
<gruetzkopf> it may also change with firmware versions.
<cr1901_modern> But is it standardized?
<cr1901_modern> Lemme guess- it's not, devices do it anyway
<gruetzkopf> it's not forbidden
<whitequark> so i'm doing my first ever bga reball
<gruetzkopf> on which subject?
Bike has quit [Ping timeout: 256 seconds]
<azonenberg_work> whitequark: yeah i am not on any long-term meds
<azonenberg_work> When i had my wisdom tooth pulled I was prescribed oxycodone/acetaminophen 5/325 plus 600mg ibuprofen
<azonenberg_work> I took a handful of 200 mg ibuprofens the day of the procedure
<azonenberg_work> and... i think one the day after?
<azonenberg_work> Just because i wanted to brush my teeth without too much discomfort when I was swishing around the area
<awygle> i need to find a drug disposal place for all my percocet >_>
<azonenberg_work> awygle: yeah the PD here has periodic take-backs
<azonenberg_work> Ironically an officer was recently fired and charged
<awygle> i'm sure there's one nearby, i just never think about it
<azonenberg_work> because they were pilfering the take-back box :p
* awygle is shocked, shocked i say.
<azonenberg_work> this wasnt even after they dumped it
<azonenberg_work> this was in the police station lobby in full view of like 4 cameras
<azonenberg_work> awygle: and my experience with cops this side of the water have mostly been quite positive
<azonenberg_work> Then again, the ones who volunteer to find missing persons are normally not the ones you need to be worried about
<azonenberg_work> There was a case not too long ago in which a patient from a nursing home wandered off
<azonenberg_work> When located, he was visibly confused and became a bit combative when they tried to pick him up
<azonenberg_work> They had a couple of officers trying their hardest to physically restrain him so he wouldn't wander deeper into the woods and freeze, without using so much force that they injured him further
<sorear> anyone think state v. valderrama is going to be important
<azonenberg_work> It's harder than it seems
<azonenberg_work> Incapacitating someone with extreme force isn't that hard, doing it with minimal force is tricky
<azonenberg_work> They did eventually get him back safe without causing significant harm though
<qu1j0t3> awygle | i need to find a drug disposal place for all my percocet >_> // What's bad is running out over a weekend, almost happened to my wife
<awygle> qu1j0t3: i got prescribed a hilarious amount for a wisdom tooth extraction, i'm not actually on it
<awygle> but yeah that kind of thing happens to my mother occasionally, which sucks
<qu1j0t3> awygle: Yes, this was for wisdom teeth
<qu1j0t3> awygle: i guess my point is too little is way worse than too much
<whitequark> azonenberg_work: lemme guess the guy was white as a sheet of paepr
<qu1j0t3> awygle: I would have kept it. Surprise needs for painkiller do arise :<
<awygle> qu1j0t3: all it really seemed to do was make me angry. i might keep a little bit though.
<qu1j0t3> wow.
<awygle> also i mis-remembered, it was for a sinoplasty, not for wisdom teeth (they just had me take ibuprofen for that)
<qu1j0t3> yeah i'm pretry sure we neede to dip back into Tramadol from wisdom teeth occasionally
<qu1j0t3> Percocet was probably all used up
<felix_> whitequark: now with an artix7 as comparator, i got this http://sigsegv.notmysegfault.de/intern/xps13-tb16.sr still doesn't look as i'd expect :/
<whitequark> felix_: what sample rate on LA?
<whitequark> this is real close to the reality
<whitequark> but you need a higher sample rate
<whitequark> 50M
<whitequark> actually no hang on
<whitequark> so on my captures the manchester encoding bit time is about 5us
<whitequark> on yours it's... very different
<whitequark> wtf
<felix_> i used 8mhz sampling rate; more than that makes pulseview crash very often
<whitequark> felix_: yeah i'm pretty sure 8M is not enough :/ :/
<whitequark> how about sigrok-cli?
<whitequark> does that crash too?
<whitequark> i think your bug was pulseview-specific
<felix_> have to leave the office now though; have to get up rather early to catch a flight :/
<felix_> will be back in office on tuesday
<whitequark> hahaha it worked
<whitequark> bga is ez
<awygle> woooo
<awygle> what did you reball?
<whitequark> awygle: some random gba cartridge
<whitequark> printer asked if it's reflashable so i found what proprietary markings on this intel flash mean and what are the testpoints
<awygle> sweet
<awygle> oh hey do you have an "automatically find me a JTAG" applet for glasgow?
<whitequark> nope
<whitequark> not sure how to implement it tbh
<whitequark> without potentially frying the board
<whitequark> i guess resistors in series with everything and then you can discern CMOS inputs vs CMOS outputs vs pullups by flipping the pin to output, then back to input and seeing if it settles and how fast
<awygle> you can get TDO (or at least outputs) with an ohmmeter
<cr1901_modern> How does that help you determine input? Charge on the line won't dissapate within the time you switch back to input and take a measurement?
<cr1901_modern> I did a horrible job asking that question, so nevermind for tonight lol
<whitequark> cr1901_modern: yes
<whitequark> it has a few pF of capacitance and that's well enough
<whitequark> that said, JTAG usually has pullups on inputs...
<cr1901_modern> Didn't think it'd be enough capacitance. I should measure it sometime on one of my own boards and see what happens
<whitequark> cr1901_modern: it's not that the bus capacitance is high
<whitequark> it's that typical cmos input capacitance is extremely low
<whitequark> so when you equalize charge between that and the input
<whitequark> it stays high
<whitequark> and impedance is high ofc.
<TD-Linux> speaking of whitequark's recent bga action... what's the easiest way to dump emmc flash? is there something aliexpress tier like the tl866?
<cr1901_modern> whitequark: Right, V = q/C, lower capacitance == less charge required to get a specific voltage.
<whitequark> TD-Linux: ...glasgow?
<whitequark> or i guess you could just use any old sd card reader
<TD-Linux> when can I order a pcb from awygle? :)
<cr1901_modern> I would've expected without bus capacitance taken into account, the little bit of charge required to get a hi voltage level would bleed out quickly (RC time constant)
<whitequark> well awygle made 5 and we only need 4
<whitequark> so the 5th can go to someone
<TD-Linux> I was imagining there would be zif emmc sockets by now
<whitequark> there are
<whitequark> hideously expensive
<TD-Linux> I can also just order myself some copies from pcbway, probably cheaper than you shipping them
<whitequark> sure, and everything it needs is on mouser
<whitequark> and solderable with a soldering iron in theory, though without hot air the qfn pad won't get attached properly
<TD-Linux> I can do hot air and qfn
<whitequark> then you can just build it yep
<whitequark> build it, run $ glasgow factory, you're done
<TD-Linux> awygle, does the latest rev have any errors?
<whitequark> not that we know! it was a bugfix rev essentially
<whitequark> it still uses the autosensing voltage translator with all its issues but for something like emmc it is not a problem at all
<whitequark> lemme know once you have it and i'll make an sd card applet real quick
<TD-Linux> oh okay. are you going to do a rev without autosense soon?
<whitequark> yes, but not sure when exactly
<whitequark> maybe within a month? it'll need some dfm and debug work done likely
<whitequark> so i'd say go ahead with revB for now, revA has been immensely useful to me
<TD-Linux> ok cool
<whitequark> also you're welcome to make like, five revBs and sell them to whoever wants
<whitequark> i am literally using CC0 so people can do this
<whitequark> awygle: btw regarding CC0 i was thinking today
<TD-Linux> my dream for all my oshw projects is that someone in china clones and mass produces them
<whitequark> awygle: that at some point we will likely encounter the problem of "board 100% identical to ours doesn't work because someone else fucked it up somehow"
<whitequark> since we don't trademark and use CC0, uh, I mean 0BSD, same difference really,
<whitequark> i think i'm going to add a little bit of ECC crypto and sign all "whitequark's" boards' serial number with my key, as a quick way to check where it came from that isn't susceptible to straight out cloning
<whitequark> and of course other vendors could use their signing key and send a PR to the frontend so that it recognizes those
<TD-Linux> a 100% clone could just use the same serial number over and over though
<whitequark> TD-Linux: yes but you don't want that because glasgows are cascadeable and having two with the same serial in a system breaks that
<whitequark> i *guess* i could in *theory* put a cheap secure element there but it seems like overkill
<whitequark> those chips are a pain in the ass to use anyway
<TD-Linux> whitequark, yeah, the purpose isn't drm of course
<TD-Linux> you could just blacklist certain serial numbers if it becomes a problem
<whitequark> basically yes
<TD-Linux> whitequark, maybe you could copy tinyfpga's spi flash metadata format?
<whitequark> oh? why?
<TD-Linux> future native usb? :)
<whitequark> nope
<whitequark> there will be no usb on the fpga
<whitequark> the use of fx2/fx3 is a major feature of glasgow
<whitequark> it means that you can build and place the applet gateware in less than 10 s
<whitequark> for most applets
<whitequark> the baseline fpga utilization is extremely low, the fx2 interface isn't complex at all
<TD-Linux> oh I just also realized that tinyfpga usb is only fs on ice40
<whitequark> yeah that too
<whitequark> but even if i could do hs on ecp5
<whitequark> or ss
<whitequark> i wouldn't
<TD-Linux> ecp5 can do ss
<whitequark> well for one ecp5 doesn't have partial reconfiguration (i think)
<TD-Linux> yeah not afaik
<whitequark> so if the thing that loads applets runs on the same fpga that runs applets
<rqou> whee, enterprise crapware of the day: WSDL
<whitequark> you need to reflash spi every time
<whitequark> don't want that
<whitequark> so, i'm thinking an fx3 would load a default bitstream with ethernet on startup and use the FPGA as an ethernet bridge essentially
<whitequark> then you could either load the bitstream to fx3 via usb or via ethernet
<TD-Linux> you could in theory implement software partial reconfigure. where the fpga has to fully reconfigure but the software only writes chunks of spi flash. but that's enormously complicated
<whitequark> then fx3 reconfigures the fpga and you use the applet
<TD-Linux> also most of the partial reconfiguration research seems oriented around evading xilinx toolchain licensing issues...
<whitequark> a magical udp packet (probably?) with minimal gateware to detect it triggers fx3 to go back online again so you can reflash it again
<whitequark> oh oh oh i know
<whitequark> i wonder if a phy can directly detect a wake-on-lan packet?
<whitequark> so you don't even need the fpga to be in a non-fucked-up state to reload
<whitequark> YEP WOL is a PHY feature
emeb has quit [Quit: Leaving.]
<whitequark> looks like you can even configure the magic frame so it can be UDP
<TD-Linux> now I'm trying to find the xilinx bitstream ORing paper