talsit has joined ##openfpga
bugzc has joined ##openfpga
digshadow has joined ##openfpga
eric_j has joined ##openfpga
<eric_j> hi azonenberg
amclain has quit [Quit: Leaving]
m_w has joined ##openfpga
<eric_j> so PSoC is on topic?
<azonenberg> eric_j: yes
<azonenberg> Any and all open source tools for any kind of FPGA/FPAA/CPLD or related
<azonenberg> whether standalone or part of a SoC like a zynq/psoc
<eric_j> great
<azonenberg> By way of introduction... clifford is the author of yosys, which pretty much all of the projects under this umbrella use for synthesis
<azonenberg> i have two projects, a back-burnered attempt to RE the xilinx coolrunner-2 (figured out ~80% of the bitstream and wrote part of a PAR, but at the time yosys didn't support CPLDs well so i stopped and never got back to it)
<azonenberg> and my active project is a full flow for silego greenpak4, primarily the slg46620v although i intend to support the whole family in the nearish future
<azonenberg> whitequark has submitted a lot of useful patches to it but isnt as active on that project as me
<eric_j> cool, hello clifford. yes i have heard of yosys although i have not tried it yet.
<azonenberg> pointfree and cyrozap are doing some good work on psoc
<azonenberg> still mostly bitstream RE, no toolchain yet
<azonenberg> digshadow was doing a bit of work on really old (xc2000 era) xilinx, not sure if he published any of it
<azonenberg> the goal there was understanding existing bitstreams, not writing a compiler
<azonenberg> wolfspraul did some spartan6 and 7 series work but ran out of time and thats kinda dead-ended for now
<azonenberg> then somebody, i can't recall who, was doing some stuff on a recent altera cyclone
<eric_j> sounds exciting
<azonenberg> So yeah, it looks like the open FPGA world is really starting to take off and this is the place to be
<azonenberg> Whiiich reminds me, my recent DAC patches for greenpak seem to have caused a memory leak...
* azonenberg investigates
<openfpga-github> [openfpga] azonenberg pushed 2 new commits to master: https://git.io/vPrza
<openfpga-github> openfpga/master 15dcbcd Andrew Zonenberg: Merge branch 'master' of github.com:azonenberg/openfpga
<openfpga-github> openfpga/master 1dafc37 Andrew Zonenberg: updated to latest logtools
<azonenberg> ==30381== 339 (152 direct, 187 indirect) bytes in 1 blocks are definitely lost in loss record 4 of 4
<azonenberg> ==30381== at 0x4C29180: operator new(unsigned long) (vg_replace_malloc.c:324)
<azonenberg> ==30381== by 0x4D2177: Greenpak4NetlistModule::GetNode(int) (Greenpak4NetlistModule.cpp:201)
<azonenberg> hmm
<openfpga-github> [openfpga] azonenberg force-pushed gh-pages from f8dbce7 to 7aec978: https://git.io/v6vmV
<openfpga-github> openfpga/gh-pages 7aec978 Travis CI User: Update documentation
<travis-ci> azonenberg/openfpga#89 (master - 15dcbcd : Andrew Zonenberg): The build is still failing.
<pointfree> o/ eric_j, welcome!
<eric_j> thanks pointfree
<pointfree> Your corrections to the two middle rows of UDBs in the diagram make sense https://www.reddit.com/r/PSoC/comments/576in6/psoc_5lp_udb_placement_cheatsheet/d8r6jeo?context=1 I was just guessing at the middle two rows based on the outer rows (which is all I saw). Also, I suppose the whole diagram can be flipped to match their coordinates. I can update the SVG.
<pointfree> I hadn't looked at the datapath or carry chains yet. But you're right I think it would complete that part of the picture.
digshadow has quit [Quit: Leaving.]
digshadow1 has joined ##openfpga
<pointfree> The UDB array routing fabric and DSI have very similar structure. The routing is said to work by short circuiting.
<pointfree> The 24 wires of the HV tile (FIG6) must link up with the 24 wires of the HC tile (FIG16). Inside an HC tile of FIG16 are 0..7S wires with 2 bits connecting twice to the same horizontal wire.
<azonenberg> Just a reminder, i do have silicon photos of a psoc4 https://siliconpr0n.org/archive/doku.php?id=azonenberg:cypress:cy8c4245axi
<pointfree> Each of the 8 vertical interface in one direction are connected to switches in two places as well, intersecting horizontal with vertical. The interface lines connect to the PI (PI Tiles) and by extension the PLD's IT (Input Terms).
<azonenberg> not a ton, and not super high res
<azonenberg> may or may not be useful
m_w has quit [Quit: leaving]
<azonenberg> https://siliconpr0n.org/archive/lib/exe/fetch.php?cache=&media=azonenberg:cypress:cy8c4245axi:psoc4_08.jpg is, i think, the 2x2 UDB array
<azonenberg> mag is high enough to see every gate, but not transistors
<azonenberg> and none of the interconnect
<pointfree> At the meetup yesterday digshadow said he could take some photos of the PSoC 5LP, if necessary.
<azonenberg> Ok
<azonenberg> I'm just saying, i have some already
<azonenberg> They may or may not be useful but keep in mind that they exist
<pointfree> I think the UDB array of the PSoC 4 is pretty similar.
<azonenberg> Yeah, exactly
<azonenberg> The one i linked is a psoc4 with 2x2 udbs
<DocScrutinizer05> azonenberg: you got greenpak5 samples meanwhile?
<azonenberg> No
<azonenberg> i'm sure i would have no trouble if i asked
<azonenberg> But i've been super busy
<DocScrutinizer05> I'm going to order then, maybe you want some, I can spare a few of those 100
<azonenberg> i've been traveling something like 50% of the last 3 months
<azonenberg> Three conferences, a funeral, a career fair, and then getting married
<azonenberg> i have one more 1-2 day trip i might have to do for work in the next few weeks but things are settling down
<DocScrutinizer05> that's plenty. Congrats to the marriage!
<azonenberg> i made my first commit to openfpga in like two months today
<azonenberg> Sooo this is interesting
<azonenberg> looks like Dac net number 9 is leaking
<azonenberg> wut
<whitequark> DocScrutinizer05: /join #solvespace :)
<whitequark> we're making some good progress lately
<azonenberg> whitequark: btw i think i am going to switch openfpga over to the new indentation framework from splash soon
<azonenberg> makes error messages a lot more readable
<whitequark> sure
<azonenberg> should be a one-afternoon refactoring
<azonenberg> and the threading issues are no big deal
<azonenberg> Hmmm this is interesting
<azonenberg> whitequark: so i'm trying to track down that memory leak travis/asan is complaining about
<azonenberg> also LOL that url
* azonenberg wonders if Cray ever made a machine called the M9
<azonenberg> and if somebody has 3 of them
<azonenberg> Anyway, look at the pointers for Dac
<azonenberg> net numbers 2...8 are allocated and freed as expected
<azonenberg> then it allocates net #9
<azonenberg> 0x5d11c20
<azonenberg> at the end, it tries to free net 9
<azonenberg> But the pointer is 0x6384a70
<azonenberg> interestingly enough memcheck does NOT complain about me freeing a bad pointer
<DocScrutinizer05> ((Cray ever made)) even a cluster of 3x
<azonenberg> Which suggests that 0x6384a70 is a valid pointer, but not the one i intended to free
<DocScrutinizer05> aah you already picked up on that part
<DocScrutinizer05> ((valid pinter)) can you 'grep* for it?
<DocScrutinizer05> 'grep'
<DocScrutinizer05> also, when memcheck doesn't complain at end of *this* routine, it should complain about double feee() at another place later on
<azonenberg> ha!
<azonenberg> here we go
<azonenberg> AddNet: module Dac net 9 = 0x6384a70
<azonenberg> So i'm overwriting the valid pointer with another one somehow
<DocScrutinizer05> search the malloc?
<azonenberg> lol its an off by one
<azonenberg> gimme a sec to confirm
<DocScrutinizer05> off-by-one, the most popular error in coding ever
<DocScrutinizer05> maybe the most common error in human thinking, after "it was this time once, so it shall be like that twice"
<DocScrutinizer05> it was like this*
<openfpga-github> [openfpga] azonenberg pushed 2 new commits to master: https://git.io/vPrVg
<openfpga-github> openfpga/master 42ffd99 Andrew Zonenberg: greenpak4: Fixed off-by-one causing memory leak (and worse)
<openfpga-github> openfpga/master 3b98b49 Andrew Zonenberg: Removed trailing spaces
<DocScrutinizer05> hehe
<azonenberg> i spent forever looking at dac code
<azonenberg> turns out, it was totally unrelated
<azonenberg> the dac just happened to allocate net numbers in a sequence that triggered the bug
digshadow1 has quit [Read error: Connection reset by peer]
<azonenberg> Welp, that should be fixed
* azonenberg waits for travis to see what i missed
<openfpga-github> [openfpga] azonenberg force-pushed gh-pages from 7aec978 to b9db33d: https://git.io/v6vmV
<openfpga-github> openfpga/gh-pages b9db33d Travis CI User: Update documentation
<travis-ci> azonenberg/openfpga#90 (master - 42ffd99 : Andrew Zonenberg): The build is still failing.
digshadow has joined ##openfpga
<azonenberg> welp, one bug fixed but another one i missed
* azonenberg fixes
<openfpga-github> [openfpga] azonenberg pushed 1 new commit to master: https://git.io/vPrwt
<openfpga-github> openfpga/master 60a8fc8 Andrew Zonenberg: greenpak4: Removed vestigial member variable that was not being used
<openfpga-github> [openfpga] azonenberg force-pushed gh-pages from b9db33d to e8b0a32: https://git.io/v6vmV
<openfpga-github> openfpga/gh-pages e8b0a32 Travis CI User: Update documentation
<travis-ci> azonenberg/openfpga#91 (master - 60a8fc8 : Andrew Zonenberg): The build was fixed.
<azonenberg> :D
<whitequark> nice
<azonenberg> I have some other stuff to do today and prob cant do any new features
<azonenberg> but maybe tomorrow after work i can get back to implementation
<DocScrutinizer05> azonenberg: do you want me to send you a dozen SLG4653* ? Or even a one *DIP20 (on top)? Would order the latter if you want one
<DocScrutinizer05> order #672 (last was 451) - seems they don't do _that_ much of direct low quantity webshop business ;-)
<azonenberg> DocScrutinizer05: i do not need a dip breakout
<DocScrutinizer05> :-)
<DocScrutinizer05> so a handful of STQFN-20 ?
<azonenberg> just chips are fine, although again i will not be able to work on them for quite a while
<azonenberg> i have a bunch of refactoring to do first
<azonenberg> Do they fit the greenpak4 stqfn-20 socket?
<azonenberg> tqfn-20*
<azonenberg> or do i need a new zif
<DocScrutinizer05> I hope so :-o
<DocScrutinizer05> the socket seems to be exactly the same, though Silego isn't very obvious and explicit about GP1 to GP5 devboard being all the same, incl sockets as long as they have the same name
<DocScrutinizer05> since the socket itself has no bold label "GP4", I think they simply wanted to make it 'simpler' for customers to pick the right socket, but the product is same box in their shelf in the end
<DocScrutinizer05> nevertheless I love silego, they have the right spirit with their design and offers to engineers
<azonenberg> Yeah i like them
<DocScrutinizer05> just their website is a tiny bit clumsy
<azonenberg> i get the feeling the bulk of their business is contract work through OEMs
<azonenberg> doing a custom design they sell as an "ASIC"
<azonenberg> selling loose greenpaks is a drop in the bucket by comparison
<DocScrutinizer05> yep
<DocScrutinizer05> their link "Reference designs" (rightmost in navbar) seems pretty much pointing to wrong place
<azonenberg> At the same time
<azonenberg> they probably arent going to say no to a new market :p
DocScrutinizer05 has quit [Disconnected by services]
DocScrutinizer05 has joined ##openfpga
<openfpga-github> [openfpga] azonenberg pushed 1 new commit to master: https://git.io/vPrKp
<openfpga-github> openfpga/master 59996e5 Andrew Zonenberg: Refactoring: now using LogIndenter instead of explicit " " indentation
<azonenberg> whitequark: welp, that was less than an afternoon :)
<openfpga-github> [openfpga] azonenberg force-pushed gh-pages from e8b0a32 to 7f340dc: https://git.io/v6vmV
<openfpga-github> openfpga/gh-pages 7f340dc Travis CI User: Update documentation
bugzc has left ##openfpga [##openfpga]
<travis-ci> azonenberg/openfpga#92 (master - 59996e5 : Andrew Zonenberg): The build passed.
<whitequark> nice
clifford has quit [Ping timeout: 268 seconds]
digshadow has quit [Read error: Connection reset by peer]
digshadow1 has joined ##openfpga
digshadow1 has quit [Client Quit]
notorious_bibor has quit [Ping timeout: 260 seconds]
digshadow has joined ##openfpga
<azonenberg> Welp, it begins
<azonenberg> Had my first power outage of the night
<azonenberg> several brief flickers of the lights followed by a ~2 sec total outage
<azonenberg> UPS covered it no sweat, but i suspect that's only a taste of what's coming
bibor has joined ##openfpga
digshadow1 has joined ##openfpga
digshadow has quit [Ping timeout: 260 seconds]
<rqou> speaking of power problems, my apartment just tripped our breaker again
<rqou> we should replace it with a no-blow fuse :P
<rqou> (as far as we have been able to determine, the entire apartment runs on 240Vx30A
<whitequark> 30A not enough?
<rqou> not when you're running 3 gaming battlestations, a space heater, and a microwave
<whitequark> you mean four space heaters
<whitequark> yeah that's a problem
<rqou> i think part of it is that the loads aren't balanced on the two phases
<rqou> but we never bothered to determine which outlets are on which phase
<whitequark> ... oh right, weird USA 240V, not proper 240V
<rqou> yeah, no ridiculously huge bs1363 plugs here :P
<whitequark> fuck bs1363
<whitequark> i don't believe in it anyway, i use iec c13/c14 personally
<rqou> aren't you in HK with bs1363 plugs?
<whitequark> i am in hk with bs1363 sockets
<rqou> so you have to at least deal with it somewhat
<whitequark> i take an iec c13 power strip and plug it into the bs thing
<rqou> that's unusual
<whitequark> you have no idea how much i hate bs1363
<whitequark> to a large degree because of counterfeit bs1363 also
<whitequark> with a sleeve over the ground pin and exactly nine other issues, every single one of which is enough to make a sane person never use that thing
<rqou> at least bs1363 is supposed to have a ground
<whitequark> schuko ftw
<rqou> my entire apartment here is running off of cheater plugs
<rqou> there is even a preinstalled (illegal) fake grounded outlet
<whitequark> amazing
<rqou> which is in the bathroom
<rqou> :P
<whitequark> oh dear
<whitequark> so it doesn't have a working gfci?
<whitequark> personally, I would find the landlord, chain him to the radiator and beat until he puts out the cash to fix that
<rqou> afaik gfci isn't mandatory on old installations
<whitequark> what
<whitequark> thats fucked up
<rqou> you're not obligated to retrofit things that used to meet code but no longer do
<rqou> there's no gfci on that outlet at all
<whitequark> well you should if you're renting them
<whitequark> actually i think americans are insane because they have gfci *outlets*
<rqou> the landlords aren't the most ethical group
<whitequark> in russia you have gfci protecting the entire set of outlets, and sometimes also lighting
<whitequark> in new installations, indeed. old simply don't have ground
<whitequark> but in the bathroom it is not a problem to find ground.
<rqou> the landlords for this apartment are somehow indirectly owned/controlled by a convicted human trafficker
<whitequark> what
<rqou> he served his time :P
<whitequark> Reddy was released from prison in 2008.[7] He moved into a mansion that had been custom-built for him while he was imprisoned.[5] He continues to manage his over 1,000 apartment buildings in Berkeley through his real estate company Raj Properties.[9][not in citation given]
<whitequark> the fuck
<rqou> whoops :P
<rqou> gotta admit that he's a brilliant businessman though
<rqou> "He used its profits to become, by 2000, the second largest property owner in the city, second only to the University of California, by his acquiring over 1,000 rundown apartment buildings."
<cr1901_modern> "rqou: afaik gfci isn't mandatory on old installations" I think that's mandatory in the last 25 years or something? I used to know exactly for a uni project
<rqou> i thought it's mandatory only if you touch it
<rqou> afaik knob and tube is still legal if you don't do anything to it
<cr1901_modern> Hmmm... could be misremembering
<cr1901_modern> Said project required the 20 Amp outlet that hopefully any bathroom or kitchen has
<rqou> but yes, afaik for a new building gfci (at least in the bathroom) has been mandatory for forever
<whitequark> >At present, new knob and tube installations are permitted in the US only in a few very specific situations listed in the National Electrical Code, such as certain industrial and agricultural environments.
<whitequark> what
<rqou> "eh, good enough"
<rqou> gfci in the kitchen wasn't always mandatory
<rqou> even my parents' house doesn't have that
<rqou> whitequark: I thought russia/eastern europe was supposed to be known for things that are even worse?
<cr1901_modern> it's required in the 20 Amp outlets I think
<whitequark> rqou: well, in practice in all new (maybe last 15, 20 years) installations you will see GFCI on every outlet at least
<whitequark> that's in the code for quite a while
<whitequark> usually you'd have an input breaker (50A), then GFCI for outlets, then per-outlet 15A breakers, then 15A for all lighting
<whitequark> in a typical small apartment
<rqou> afaik in the US afci+gfci is now mandatory almost everywhere for new installations as of 2014
<rqou> but that doesn't help old installations
<whitequark> ah finally
<whitequark> afci is electronic?
<whitequark> huh, never even heard of these
<rqou> some kind of electronic arc fault detector thing
parasite__ has joined ##openfpga
<whitequark> in RU the input and per-outlet breakers are just bimetallic i think
<cr1901_modern> 50A? That seems small to me... (Does RU use 240V?)
<whitequark> 230V
<whitequark> 50A is a lot.
<whitequark> 6kW translates to 25A of current
<whitequark> you don't have anything in the apartment that eats 12kW, especially given central heating.
<cr1901_modern> I recall my home can do 200A, though I'm not about to test that. I wonder if I'm misremembering
<whitequark> a cooking range will perhaps top out at 4kW simultaneously, and a space heater another 4kW (though that's quite a large one)
<rqou> are ranges/heaters usually gas or electric in RU?
<rqou> US has both available and afaik varies depending on area
<whitequark> used to be gas everywhere, but the piping was shit and it exploded a bunch, i think
<whitequark> that caused it to be redone in old houses, to put it outside
<whitequark> new installations universally use electricity (which is generated from the same gas, on very large power stations on the outskirts)
<whitequark> so yeah, new installations are electric. ohmic heating if you're looking at govt issue. or induction if you're looking at something fancy
<nats`> in france for a full electric apartment you have maximum 40A
<whitequark> if you're looking at village houses then it's all gas
<whitequark> or in general remote areas
<rqou> cr1901_modern: why not? what's a safety?
<nats`> there are no way in residential lot you can get more than 60A
<nats`> I mean 200A is fucking huge
<rqou> but france is also 240V, right?
<nats`> yep
<rqou> US new installations are afaik 200A * 120V
<whitequark> wow
<cr1901_modern> I have 120V
<nats`> oO
<rqou> US power is weird
<whitequark> indeed
<nats`> what a stupid choice
<rqou> it's 2x 120V 180 degrees out of phase
<nats`> loss in conductor is utterly huge
<nats`> I guess the first error was to stay in 120V
<whitequark> yeah if US didn't fall into fascism, i might have moved there and fixed a house to run on 240V
<whitequark> with schuko plugs :p
<nats`> TBH going to 380V would be ideal
<whitequark> nats`: what about arc flash hazard
<nats`> there are no real difference from 120 to 380
<rqou> why not a JPN 100V :P
<nats`> I mean safety distance is not taht different
<whitequark> i trust a good old gfci but not so much an afci
<whitequark> gfci is simple as a brick
<cr1901_modern> rqou: I assume the 2x120V is what permits 240V outlets as well in US?
<cr1901_modern> Power was never my strong suit
<DocScrutinizer05> my RCD / gfci(?) tripped a few days ago, from 5 drops of water entering a triple-extender in kitchen, and it took the whole flat down (it's central, 30mA) so all 20 branches a 16A went blackout, and my uninteruptable power supply crapped pants and took pout my PC after a ~5s, thanks to dead batteries (they are already 18 months 'old')
<nats`> there are nasty phenomen when reaching 1.5kV
<whitequark> nats`: hmm
<whitequark> pretty sure 480v already has arc flash phenomena
<nats`> whitequark, I don't mean it's equaly safe but arc flashing is unlikely if used like in 250V system
<rqou> cr1901_modern: yeah, the 2x120V are out of phase giving a 240V across them
<nats`> sure even 12V can arc it's a matter of distance
<nats`> but my point is, it's not that hard to get safe distance at 380
<DocScrutinizer05> what's a gfci, afci?
<cr1901_modern> Ground Fault Circuit Interrupter I think
<rqou> i think GFCI = RCD outside of US
<whitequark> DocScrutinizer05: and afci is like gfci but interrupts on arc fault
<DocScrutinizer05> err
<whitequark> so not the tiny arc you get when the caps charge
<whitequark> but an actual arcing
<whitequark> gfci compares current in on live vs current out in neutral
<rqou> so it detects the imbalance of power between the hot/neutral and trips if the imbalance is > a threshold
<DocScrutinizer05> res, RCD
<DocScrutinizer05> Fehlerstrom-schutzschalter, delta-I
<cr1901_modern> "so not the tiny arc you get when the caps charge" What do you mean?
<DocScrutinizer05> indeed
<whitequark> when you plug your laptop thing or w/e into the outlet you get a small arc in it
<whitequark> because of the inrush current
<cr1901_modern> Oh, TIL :)
<cr1901_modern> Makes sense, never gave it much thought
<DocScrutinizer05> no idea, never heard of afci
<rqou> :P
<whitequark> DocScrutinizer05: never heard about it either, earlier than 10 minutes ago
<whitequark> looked it up
<whitequark> DocScrutinizer05: wiki says it's some fancy electronics inside https://en.wikipedia.org/wiki/Arc-fault_circuit_interrupter
<DocScrutinizer05> we have normal overcurrent "blowfuse" / resettable, and we got RCD
<whitequark> yup same in RU
<whitequark> looks like AFCI is searching for the horrible EMI that an arc causes
<whitequark> and that goes back into the mains
<DocScrutinizer05> hmmm interesting
<rqou> anyways, fun fact
<rqou> apparently a few years ago HK started cracking down on "Universal Outlet" power strips
<DocScrutinizer05> whitequark: (EMI) well, somewhat yep. Transients
<rqou> because they're not safe/compliant
<DocScrutinizer05> delta-I/delta-t
<whitequark> rqou: yeah they're shit
<rqou> they used to be all over the place
<whitequark> but they're not cracking down enough
<whitequark> I reported the shitty BS1363 to the shops that sold it
<whitequark> no response
<whitequark> complained multiple times
<rqou> hey, you have to go to Ap Liu Street now rather than a big store :P
<whitequark> now I want to sic whatever regulatory agency is supposed to guard against that on them
<rqou> but that'll require all of the sellers on Ap Liu St to shut down, and we can't have that! :P
<whitequark> hm never been there
<whitequark> reclamation st is my go-to place
<whitequark> and taobao
<rqou> no Huaqiangbei in SZ?
<whitequark> never been to SZ yet
<whitequark> don't have a mainland visa
<rqou> ah, right
<whitequark> should get it
<rqou> forgot that that's hard for most people
* rqou is Chinese-American
<whitequark> ohh
<whitequark> convenient
<rqou> yeah, probably makes much more sense now :P
<whitequark> well, mainland visa isn't hard to get afaik
<whitequark> I just never bothered to
<rqou> apparently it can vary
<whitequark> huh
<rqou> when "something big" happens there may be difficulty
<rqou> e.g. the Olympics, the HK protests, etc.
<whitequark> ah right. but i am not in a hurry anyway.
<DocScrutinizer05> fsckin RCD taking out my office (incl PC) on a fault GND current in kitchen is a nogo anyway
<rqou> anyways, due to a loophole I actually have an HKSAR passport and Home Return Permit
<whitequark> DocScrutinizer05: i like being alive
<rqou> so I don't need a mainland visa
<whitequark> i like being alive more than i dislike my PC being turned off
<whitequark> especially looking at how i use a laptop
<whitequark> :p
<whitequark> but anyway if you hate it that much, just install two RCDs for kitchen and office
<whitequark> i'm sure you have enough skill for that.
<whitequark> rqou: hmm HKSAR has passports? TIL
<DocScrutinizer05> whitequark: well, I survived a maybe 500 electric shocks. And I'd think one RCD for bathroom and one for kitchen would suffice, no sense in have one central one taking down whole flat at once
<whitequark> sure, that's just cost cutting in RU
<rqou> yeah, it's distinct from both the mainland passport and a HKID
<rqou> a HKSAR passport implies you have Chinese nationality
<whitequark> rqou: interesting. how does it relate to "right of abode" ?
<whitequark> oh
<rqou> um... please consult a lawyer :P
<whitequark> i'm just curious
<whitequark> the immigration dept website talks about right of abode a lot
<whitequark> but never about hksar passports
<whitequark> that i remember at least
<rqou> IANAL but "right of abode" essentially means that you are allowed to live in HK
<rqou> you don't have to be a Chinese national
<DocScrutinizer05> I wonder if I could get a two 10mA RCDs for ktichen and bathroom. Can't remove the 30mA central one
<rqou> there is also a "right to land" status that most chinese-americans who missed my particular loophole can get
<rqou> *chinese-americans with a connection to HK
<whitequark> mm, complicated
<rqou> yes, quite
<rqou> there's also Taiwan
* rqou ducks to avoid flamewars :P
<rqou> but my family and I have no connection to Taiwan, so I don't know much about that
<rqou> but yeah, tl;dr is that HK has some nice and convoluted immigration policies
<rqou> you have a) right to live in HK b) a Chinese national according to HKSAR govt. and c) a Chinese national according to PRC govt.
<rqou> b/c don't have to match btw :P
<rqou> even though they're applying the same law
<rqou> whitequark: still saner than US immigration policy? I remember you were complaining about that a while back
digshadow1 has quit [Ping timeout: 250 seconds]
digshadow has joined ##openfpga
digshadow has quit [Client Quit]
digshadow has joined ##openfpga
carl0s has joined ##openfpga
<carl0s> hi pointfree, Creator 4.0 just being released, idk if the new option "Export CYSCH to XML file" would be useful in the future. And also would like to thank again for the work you and cyrozap and more ppl are doing with PSoC here :).
<pointfree> carl0s, thanks for the pointer. Updating PSoC Creator to 4.0 right now...
<carl0s> same, want to try that feature
<pointfree> carl0s, so you're using the PSoC in your bachelor's thesis?
<carl0s> yes, but it's nothing fancy
<carl0s> compared to OOCD that was also a thesis my project is nothing :), have to switch to Windows to install Creator
carl0s has quit [Quit: Leaving]
carl0s has joined ##openfpga
<pointfree> I like the new "Design Wide Resources" sections.
Marex_ is now known as Marex
<nats`> rahhhhhhhhhh fuck
<nats`> I just installed vivado 2016.2 and here the .3
clifford has joined ##openfpga
<rqou> can someone please make a vivado webpack docker/AWS image/whatever cloud thing?
<carl0s> pointfree, did you tried the export to XML? it's on Proyect->preferences i think. I still on the installation process
<pointfree> carl0s: I see "Export to IDE" which allows me to export to Eclipse, IAR Embedded Workbench, uVision, Generated CMSIS-Pack, and Makefile. I don't see anything specifically saying "Export to XML" ...Exporting to Makefile sounds appealing.
<carl0s> Export to Makefile is a new feature aswell, "so we added a project property (Project->Properties…) to ask the tool to save a text-based version of the schematic alongside the CYSCH file. The format is XML and you can use your favorite “diff” utility to see what is different between two designs, or two versions of the same design."
<carl0s> well, gotta go, will check Creator tomorrow, hope you find the option.
carl0s has quit [Quit: Leaving]
<pointfree> carl0s: Just found it. Looking for the cysem file right now. I'll talk to you later then.
Bike has quit [Quit: moonshine]
azonenberg_work has quit [Ping timeout: 248 seconds]
azonenberg has quit [Ping timeout: 260 seconds]
tecepe has quit [Ping timeout: 260 seconds]
azonenberg has joined ##openfpga
azonenberg has quit [Ping timeout: 258 seconds]
parasite__ has quit [Read error: Connection reset by peer]
<whitequark> rqou: do you know the chinese name for 35mm DIN rail by any chance
<rqou> unfortunately no
<whitequark> it seems extremely common yet i can't seem to pin down the keywords :/
<whitequark> this happens to me a lot.
<rqou> unfortunately, i don't know too many technical terms in chinese
<rqou> doubly unfortunately, my family speaks cantonese, not mandarin
<rqou> so yeah, I mostly have a basic conversational level in cantonese and just-barely-conversational-but-not-quite level in mandarin
<rqou> the worst of both worlds :P
<rqou> extra disappointingly, my parents are fluent in mandarin :P
<whitequark> heh
<rqou> apparently this situation is quite common amongst American/British/Canadian-born Chinese
<rqou> parents will speak some combination of mandarin/cantonese/shanghainese/taiwanese/english; children will speak their parents' language poorly and predominantly english
<nats`> that works for a lot of culture
<nats`> myself I'm mixed blood marocco / france
<nats`> I speak french and english
<nats`> really low arabic
<nats`> I guess english is the only one mandatory when you need to work worldwide
<rqou> unless you're in asia :P
<nats`> then you take the language from the place you were born
<rqou> china insists on using chinese
<rqou> japan insists on using japanese
<nats`> inside border yes it's the same in france
<nats`> but when you're outside english is the common base
<rqou> from what i've seen china/japan are much less willing to use english than e.g. france
<nats`> you mean to work with other country ?
<nats`> because I don't have trouble with chineese on that point
<rqou> maybe it's just for "official" things
<rqou> i do remember though that when my father was managing a US based company that did all their engineering in china, the official requirement was that all documentation was supposed to be english. This became "all official documentation is in english, all internal documentation is in chinese"
* whitequark is using google translate a lot in person
<whitequark> that actually works remarkably well for what it is
<rqou> google translate works pretty well for technical documents actually
<whitequark> we probably would have never communicated "three-phase power" to the real estate agent otherwise, and we had a cantonese speaker with us
<whitequark> (as in L2 or L3 speaker)
<rqou> I usually use a combination of https://mandarinspot.com/annotate and https://www.mdbg.net, but this only works if you already know some chinese
<whitequark> yeah it does
<whitequark> i'm guessing the ideograms are easier for machine translation somehow
<rqou> the first is a tool that adds pronunciations (pinyin) to characters
<rqou> the second is a dictionary
<whitequark> ah i see
<whitequark> i never bothered to learn pinyin; i actually enjoy learning new languages but only until i have to produce sounds with meat flaps
<rqou> because (as is again typical for chinese-americans) the words that I know based on the sound exceeds the words I can recognize the characters for
<whitequark> at which point it becomes a ngihtmare
<whitequark> and especially recognize them by ear. i still have trouble understanding english
<whitequark> spoken english, that is.
<rqou> cantonese is actually my L1 language along with english
<rqou> but because I don't use it enough I don't know enough vocabulary
<whitequark> i feel like i'm not actually gonna benefit from learning cantonese
<rqou> i don't even know how a foreigner is even supposed to learn cantonese
<whitequark> there are books and such
<rqou> the curriculum isn't nearly as well developed as for mandarin
<whitequark> so i know a guy who's a permanent resident and he can speak cantonese alright
<whitequark> enough for day-to-day interactions. but no technical terms.
<whitequark> and i'm basically interested *only* in technical terms
<whitequark> however that's possibly the worst case
<rqou> I keep telling myself to use mandarin more so I can practice, but then I always get too lazy to actually do so
<rqou> I'm at the awkward level where I know all the "basic stuff" that you would usually learn in an introductory language course, but don't know enough to comfortably use the language continuously
<whitequark> right
<whitequark> aren't there large mandarin speaking communities in north america anyway?
<rqou> so technically, all of my roommates could also use mandarin practice, but we're all not fluent enough to make it not seem like a lot of effort :P
<whitequark> irritating!
<rqou> there's a large mandarin speaking community here, but the SF bay area is much more known for its cantonese community
<rqou> interestingly, I stand out in HK if I speak cantonese
<rqou> the overseas communities didn't experience some sound shifts that HK did
<rqou> so I sound like the older generation should sound
<whitequark> interesting
<whitequark> hm, so would translating solvespace into chinese count as practice? :p
<rqou> my ability to read/write chinese is also even worse than my ability to speak/understand
<rqou> welcome to ideographic languages
<whitequark> and i suppose mCAD vocabulary isn't your strong side either
<whitequark> some mainland uni is using solvespace as a part of curriculum (i think) but i was unable to communicate with them
<rqou> I haven't looked at solvespace at all
<rqou> mechanical CAD is on my giant todo list
<whitequark> wait, really? i thought you expressed interest in it
<whitequark> mayeb i misremember
<rqou> i don't think so
<whitequark> hmm must have been someone else from here
<rqou> the limit of my mechanical cad knowledge is that a friend gave me a "crash course in how to make mechanical drawings"
<rqou> which included a basic "how to make a shape in autodesk inventor"
<whitequark> ah
<rqou> more tools are always useful though
<rqou> except for the part where every new tool has its own annoying learning curve
<whitequark> indeed
<rqou> and the part where every tool seems to have difficulty interacting with other tools for some dumb reason
<whitequark> well, solvespace can now actually import a file format!
<whitequark> an entire one!
<rqou> .obj? :P
<whitequark> natch, .dxf
<whitequark> and .dwg i guess, which is the same thing if we're talking about the subset it can recognize
<rqou> so what's the reasoning for yet another cad tool?
<rqou> can't we fix kicad first :P
<whitequark> kicad is eCAD
<rqou> actually, now that I think about it
<whitequark> solvespace is mCAD. the only other FOSS tool in this area is FreeCAD, and FreeCAD doesn't work
<whitequark> this is basically the reason
<rqou> i can't think of any good mCAD tool
<rqou> that is FOSS
<whitequark> try solvespace :p
<rqou> everybody in the mCAD world seems to be content with proprietary tools like inventor
<whitequark> or onshape
<rqou> actually, I can think of a FOSS "mCAD" tool:
<rqou> blender :P
<whitequark> nah, non-parametric mCAD tools are almost useles
<whitequark> you'll spend all your time redrawing shit that could be done in seconds in a parametric tool
<rqou> (my only experience with blender was to convert .obj files to .x files while playing around with XNA many years ago)
<whitequark> I'd rather use pencil and paper than a non-parametric tool, at least pencil and paper is fast
<whitequark> b;ender is fine for what it is
<whitequark> but its not an mcad tool by any sane measure
<rqou> i'm probably going to need solvespace after I finally stop being lazy and fix my k40 laser cutter
<whitequark> hm, how large is that thing anyway?
<whitequark> 300x200mm
<rqou> so when are kicad and solvespace going to be able to talk to each other?
<rqou> e.g. draw a capacitor in solvespace, design a board in kicad, export the board back to solvespace and put an enclosure around it
<cr1901_modern> Gerber to Solvespace... now THAT is a feature that would be cool. 'Cept pretty sure I could easily create the constraints for that so not worth the effort.
<rqou> doesn't kicad have the ability to export its 3d model?
<rqou> or is this yet another instance of "interactions are hard?"
<rqou> is this related to how kicad grew a dependency for opencascade a while back?
<cr1901_modern> I have no idea, tbh. The "best" I could think of doing during an import is to add appropriate constraints the "cling to" the 3d model, and the user can change them as needed (such as adding a hole for a large capacitor to lower the top of the enclosure).
<rqou> why can't we create a universal file format that can describe anything? :P (https://xkcd.com/927/)
azonenberg_work has joined ##openfpga
tecepe has joined ##openfpga
azonenberg has joined ##openfpga
<whitequark> rqou: no idea. i don't use kicad. i find it completely insufferable compared to eagle.
<balrog> whitequark: lol
<balrog> I've used both. kicad was unusable several years ago so I ended up using eagle then, but these days kicad is so much more usable.
<cr1901_modern> Insufferable PCB tool is redundant.
<cr1901_modern> I kinda like boldport's idea, but unless someone writes an Inkscape plugin to refresh the vectorized output automatically (i don't even care if it uses dbus on Windows!), that's a dealbreaker for me
ylm has joined ##openfpga
carl0s has joined ##openfpga
carl0s has quit [Quit: Leaving]
ylm has quit [Ping timeout: 260 seconds]
azonenberg has quit [Ping timeout: 258 seconds]
azonenberg_work has quit [Ping timeout: 252 seconds]
azonenberg_work has joined ##openfpga
X-Scale has quit [Read error: Connection reset by peer]
X-Scale has joined ##openfpga
tecepe has quit [Read error: Connection reset by peer]
<azonenberg_work> Welp
<azonenberg_work> Looks like I'm running on LTE for the foreseeable future
<azonenberg_work> my cable modem isnt getting a signal
<azonenberg_work> Lines probably were on the same stretch of line that got hit whe nthe power went out
digshadow has quit [Quit: Leaving.]
<balrog> azonenberg_work: ugh :(
<balrog> called up the cable company already?
azonenberg has joined ##openfpga
<azonenberg_work> rqou, whitequark: i do not want gfci for all outlets until the false positives are fixed
<azonenberg_work> i have had appliances with weird PFC that tripped gfci's
<azonenberg_work> so 100% gfci makes them useless to me
<azonenberg_work> also, upon closer investigation it was not the cable modem
<balrog> azonenberg_work: does that mean they're leaking voltage to the gnd pin?
<azonenberg_work> it was shorewall being a derp and not re-enabling NAT after a reboot
<azonenberg_work> And, yes
<azonenberg_work> my understanding is that certain PFC has a voltage divider from hot to PE
<azonenberg_work> and this will trip sensitive GFCIs
<azonenberg_work> I do not know for a fact that this is the case with a particular server PSU i have
<balrog> it will because devices should not be using PE like that
<balrog> they should be using neutral
<azonenberg_work> what i do know is, preliminary metering did not detect any obvious voltage on the enclosure
<balrog> that's not a false positive, that's a real positive
<balrog> the enclosure may be isolated from physical earth
<azonenberg_work> But i dont trust the thing enough to ever actually use it at this point
<azonenberg_work> and well, i consider the mission of GFCIs to keep people from getting electrocuted
<azonenberg_work> not to keep current from flowing on PE
<azonenberg_work> as a result, any leakage from hot to PE that isn't through a person is a false positive in my book
<balrog> how do you tell where the current flowing on PE is coming from?
<azonenberg_work> You dont
<balrog> (through a person vs. through the device)
<azonenberg_work> My point is more, until people stop making such devices
<azonenberg_work> i dont want 100% GFCI
<azonenberg> aaand back up and running on the desktop
<balrog> people still make such devices?
<azonenberg> This was an ATX2U server PSU i bought from newegg a few months ago
<balrog> ugh. what brand so I can avoid them?
<azonenberg> Dont remember, have to check
<azonenberg> i just know that the server in question is now running open-cased with a full sized ATX PSU
<azonenberg> sticking out of the top
<azonenberg> because the circuit it's on is GFCI'd
<azonenberg> and trips as soon as i attempt booting
<azonenberg> or is it as soon as i plug it in?
<azonenberg> i forget
<balrog> and is it a GFCI or an AFCI?
<azonenberg> GFCI
<azonenberg> To my knowledge there are no AFCIs anywhere in the building
<balrog> would this equipment still work if the ground pin was removed?
<azonenberg> I dont know and i dont want to find out
<azonenberg> last thing i want is ground loops through usb shields etc
<balrog> that's assuming the shielding is wired correctly already
<azonenberg> True
<balrog> use of physical earth for active stuff in circuit is bad enough :(
<nats`> lol don't look at recent PLC system
<balrog> nats`: what do they do...?
<nats`> (powerline communication)
Bike has joined ##openfpga
<nats`> they use the earth as one of the pair part
digshadow has joined ##openfpga
<nats`> :p
<nats`> they use it for mimo plc
<azonenberg> welp that took waay too long to debug
<azonenberg> somehow shorewall6 was configured to disable ip forwarding
<azonenberg> so after a reboot i lost ipv6
amclain has joined ##openfpga
azonenberg_work has quit [Ping timeout: 260 seconds]
Bike has quit [Read error: Connection reset by peer]
azonenberg_work has joined ##openfpga
Bike has joined ##openfpga
Bike has quit [Read error: Connection reset by peer]
Bike has joined ##openfpga
Bike has quit [Read error: Connection reset by peer]
Bike has joined ##openfpga
mithro_ has joined ##openfpga
rqou_ has joined ##openfpga
azonenberg has quit [*.net *.split]
clifford has quit [*.net *.split]
mithro has quit [*.net *.split]
forrestv has quit [*.net *.split]
nats` has quit [*.net *.split]
rqou has quit [*.net *.split]
felix_ has quit [*.net *.split]
rqou_ is now known as rqou
azonenberg has joined ##openfpga
forrestv has joined ##openfpga
mithro_ is now known as mithro
clifford has joined ##openfpga
nats` has joined ##openfpga
felix_ has joined ##openfpga
dingbat has quit [Ping timeout: 258 seconds]
dingbat has joined ##openfpga
_whitelogger has joined ##openfpga
doomlord has joined ##openfpga
clifford has quit [Ping timeout: 268 seconds]
digshadow has quit [Ping timeout: 260 seconds]
digshadow has joined ##openfpga
carl0s has joined ##openfpga
carl0s has quit [Ping timeout: 258 seconds]
carl0s has joined ##openfpga
<cr1901_modern> balrog: Oh, I didn't know that you could encrypt them
<balrog> yeah most modern large FPGAs have some encryption feature
<azonenberg> yes, xilinx parts have both efuse and battery backed sram
<azonenberg> you jtag an aes key in
<azonenberg> then can load encrypted bitstreams
<azonenberg> you have the option of locking the device to only run encrypted bitstreams
<rqou> including sometimes backdoors (*cough* *cough* Actel)
<azonenberg> or allowing it to run both cleartex and encrypted
<azonenberg> and lol
<balrog> rqou: can you turn off jtag on those? lol
<rqou> idk, never actually used them
<rqou> some researchers used DPA to extract a backdoor AES key from some Actel part a few years back
<cr1901_modern> efuse holds the bitstream or the key only?
<balrog> key
<balrog> bitstream is loaded from sram the usual way
<balrog> but is encrypted
<balrog> err, to sram
<cr1901_modern> I was about to say- I didn't think Xilinx supported OTP like Lattice does
<rqou> I assume xilinx also doesn't have DPA countermeasures?
<digshadow> rqou: I'm starting to do some serious DPA type stuff
<digshadow> maybe in a few months I can extract that key easily :)
<rqou> I have a "super duper secret" DPA project in the todo list
<digshadow> rqou: suite A?
<cr1901_modern> I don't even know what DPA means :D
<rqou> nah, Nintendo 3DS :P
<digshadow> thats cool too :)
<digshadow> azonenberg: quend wanted to extract the rom right
<balrog> cr1901_modern: deep power analysis
<rqou> i was originally planning to do it when the arbitrary code execution right after bootrom bug was first discovered
<digshadow> did anything happen with that
<balrog> digshadow: no it was doom
<digshadow> oh yeah
<digshadow> anyway, same q
<rqou> rumor says the 3DS bootrom was dumped via reset glitching, but i'm skeptical
<cr1901_modern> balrog: Ack :)
<digshadow> (not to be confused with doomlord)
<rqou> anyways, procrastination is the enemy of progress :P
<digshadow> rqou: is that how you missed mtvre? :)
<rqou> funemployment messes with your sense of time/schedules :P
<digshadow> heh
<digshadow> I was thinking of becoming funemployed
<digshadow> but looks like things have settled down a bit
<cr1901_modern> This makes me glad I'm self employed/not in danger of eviction
<cr1901_modern> or hunger*
<digshadow> cr1901_modern: sure but you have your own responsibilities
<digshadow> to make sure you get contracts etc
<rqou> there's always the "move back with parents" option :P
<cr1901_modern> digshadow: Yes, and admittedly I could be doing a better job re: that. Hope the blog helps a bit. I also could tap into my network if I'm desperate.
<rqou> herp derp i should update my blog :P
<rqou> writing is a lot of effort
<cr1901_modern> Yes, I figured that out the hard way. Even this wasn't exactly a wonderful time (but definitely worth the effort)
<digshadow> rqou: where is your blog
<rqou> https://robertou.com/blog.html, but only has 3 useful posts
<rqou> i have a bad habit of not documenting anything
<rqou> the blog was an attempt to fix that
<rqou> but nope
<digshadow> cool
<digshadow> rqou: one of the reasons I tried twitter
<digshadow> was because it is, all things considered, much easier to just post a few pictures with a brief description
<digshadow> than to actually take time to write something up
<digshadow> plus, the expectation of utility is much lower
<rqou> i don't have twitter; maybe i should
<digshadow> rqou: I have a write only twitter account
amclain_ has joined ##openfpga
<digshadow> although I'd really like to follow just one person
<digshadow> just to send some sort of message
<digshadow> but haven't figured out who that should be yet
<cr1901_modern> I paid someone to help set up nginx. Now, I do the maintenance myself, but it was TOTALLY worth paying someone to help the startup
* cr1901_modern is looking at rqou's blog
<digshadow> heh
<digshadow> rqou: fwiw it definitely has paid off
<digshadow> I've gotten a bunch new contacts as a result
<digshadow> and a few gigs
<rqou> cr1901_modern: my nginx setup is weird
<cr1901_modern> And then I set up my own mail server... not the *worst* decision I ever made.
<rqou> lol i'm still working on that step :P
<cr1901_modern> I learned more about TCP/IP and DNS than I wanted to know though
<rqou> i already knew a decent amount
<rqou> hence things like my reverse dns here
<rqou> :P
amclain has quit [Ping timeout: 260 seconds]
<cr1901_modern> New blog post! Setting up a mail server T_T
<azonenberg> digshadow: the CIA?
<azonenberg> :P
<digshadow> azonenberg: nah, I already know some folk at inqtel
<azonenberg> FSB?
<azonenberg> :P
<digshadow> ha
<azonenberg> some Daesh warlord?
<digshadow> you know there was a phase
<digshadow> where iranians started to reach out to me to help with nuclear projects
<azonenberg> lolwut
<digshadow> and I decided I needed to cut back a bit
<azonenberg> closest i got was some infosec company in the UAE poking me on linkedin
<digshadow> iirc adc changed his nick
<digshadow> after north korean hackers started getting to close to him
<digshadow> too
<azonenberg> oh fun
<azonenberg> Havent had quite that level of attention yet but i have turned away a few folks who seemed to be coming from sketchy places
<rqou> heh, i revealed my dual nationality on an irc channel discussing reverse engineering. I guess I have even more fun to look forward to
<cr1901_modern> I've always managed to stay fairly private, tbh
<azonenberg> rqou: dual?
<azonenberg> remind me :p
<rqou> US + China (HK)
<azonenberg> ah, fun
<azonenberg> yeah i could see that getting some attention
<cr1901_modern> azonenberg: What's wrong with KiCad's symbols/footprints? And which document should I read to fix them?
<azonenberg> biggest problem? a) they lack a lot of of the components i use
<azonenberg> b) they lack stability
<azonenberg> sch symbols that change under your feet and break existing designs are not tolerable
<azonenberg> c) i'd have to hand verify each one before i trusted it
<cr1901_modern> azonenberg: Your symbols respect a specific standard. What is that standard called?
<digshadow> IPC maybe?