<whitequark> ok, off to the plane
<whitequark> or the airport rather
<rqou> HK->RU?
<awygle> Have a good flight
m_w has quit [Ping timeout: 240 seconds]
<whitequark> I'm still really confused about interrupt handling
<whitequark> it clearly never actually enters the handler
<whitequark> but, handlers with lower priority do not get executed anymore
<whitequark> the address is right...
DocScrutinizer05 has quit [Disconnected by services]
DocScrutinizer05 has joined ##openfpga
DocScrutinizer05 has quit [Disconnected by services]
DocScrutinizer05 has joined ##openfpga
<whitequark> okay, I disassembled the hex file and the handler is def in the right place
<whitequark> I even tried replacing the entire interrupt vector table with one giant nopsled (except for the USB interrupt), no dice
* qu1j0t3 sees a familiar escalation
GenTooMan has quit [Quit: Leaving]
sgstair has quit [Read error: Connection reset by peer]
bofh___ has joined ##openfpga
bofh_ has quit [Read error: Connection reset by peer]
sgstair has joined ##openfpga
pie__ has quit [Remote host closed the connection]
pie__ has joined ##openfpga
bofh___ has quit [Read error: Connection reset by peer]
bofh_ has joined ##openfpga
pie___ has joined ##openfpga
xdeller_ has joined ##openfpga
indy has quit [Ping timeout: 240 seconds]
pie__ has quit [Read error: Connection reset by peer]
xdeller has quit [Ping timeout: 240 seconds]
indy has joined ##openfpga
<awygle> icefuzz takes a long time to run
rohitksingh has joined ##openfpga
eduardo__ has joined ##openfpga
eduardo_ has quit [Ping timeout: 264 seconds]
rohitksingh has quit [Quit: Leaving.]
rohitksingh has joined ##openfpga
lexano has quit [Ping timeout: 240 seconds]
m_t has joined ##openfpga
<azonenberg> whitequark: have you confirmed interrupts are enabled?
<azonenberg> (the interrupt in question)
lexano has joined ##openfpga
rohitksingh has quit [Quit: Leaving.]
rohitksingh has joined ##openfpga
rohitksingh has quit [Remote host closed the connection]
rohitksingh has joined ##openfpga
digshadow has quit [Ping timeout: 256 seconds]
digshadow has joined ##openfpga
m_t has quit [Quit: Leaving]
eduardo__ has quit [Remote host closed the connection]
bitd has joined ##openfpga
rohitksingh has quit [Quit: Leaving.]
rohitksingh has joined ##openfpga
digshadow has quit [Quit: Leaving.]
digshadow has joined ##openfpga
rohitksingh has quit [Quit: Leaving.]
rohitksingh has joined ##openfpga
eddyb has quit []
eddyb has joined ##openfpga
Zorix has joined ##openfpga
rohitksingh has quit [Quit: Leaving.]
rohitksingh has joined ##openfpga
rohitksingh has quit [Quit: Leaving.]
soylentyellow_ has joined ##openfpga
soylentyellow has quit [Quit: Leaving]
soylentyellow_ has quit [Quit: Leaving]
soylentyellow has joined ##openfpga
Zorix has quit [Quit: Leaving]
X-Scale has joined ##openfpga
Zorix has joined ##openfpga
ym has joined ##openfpga
elms has quit []
elms has joined ##openfpga
<whitequark> azonenberg: yes
<whitequark> if I disable the interrupt, then nothing happens at all
<whitequark> and it's not an issue of e.g. ret vs reti
<whitequark> awygle: it was indeed a good flight
<whitequark> I haven't slept for like 36 hours before so I both felt really nice and slept very well
<whitequark> and it's a direct one too, bless Aeroflot
soylentyellow_ has joined ##openfpga
soylentyellow has quit [Read error: Connection reset by peer]
soylentyellow_ has quit [Remote host closed the connection]
soylentyellow has joined ##openfpga
soylentyellow has quit [Remote host closed the connection]
<awygle> Glad to hear it :-)
soylentyellow has joined ##openfpga
soylentyellow has quit [Remote host closed the connection]
soylentyellow has joined ##openfpga
ovf has quit []
ovf has joined ##openfpga
<daveshah> awygle: did icefuzz work in the end?
<awygle> daveshah: I did get it running thanks to the eth0 tip, I'm now figuring out how to add a new family
<daveshah> awygle: the first step is to get icepack unpacking and packing for the family
GenTooMan has joined ##openfpga
<sorear> i take it this is for adding LM to icestorm?
<awygle> Yup
<awygle> And thanks daveshah, I've got that open :-)
dingbat has quit []
dingbat has joined ##openfpga
<cr1901_modern> LM? I thought all devices were supported now
<sorear> no
<daveshah> cr1901_modern: LP, HX and UP devices are supported
<daveshah> LM, Ultra and UltraLite are not
<cr1901_modern> Maybe I should try Ultra/UltraLite. Last time I tried reading the icestorm docs, it didn't go well lol
<awygle> go for it!
<daveshah> cr1901_modern: I've started adding support for the Ultra to icepack already so there's a headstart there :) otoh the UltraLite might be easier as its IP is slightler closer to the UltraPlus and it may have fewer bugs to being newer in the generation (there are a few idiosyncrasies in the Ultra)
<awygle> it's crazy that UltraPlus support is (i think) less than a year old
<awygle> i was at the open source hardware summit last year and i met tannewt and at that time there was only very partial support
<awygle> and now it seems like it's in every new product and dev board
<daveshah> awygle: officially I would call UltraPlus support ~4 months old
<daveshah> although a lot of stuff was working from mid November last year it wasn't stable until end of the year or so
<Bike> a burrito is like a monad?
<awygle> pie___: i mean, i do that every day (usually to people who have no idea what i'm talking about)
<awygle> i frequently express a wish that restaurant menus were SQL databases
<awygle> so i could order by calories / dollar for example
<cr1901_modern> I wish copy and paste were real for when I have to write on an actual piece of paper
<pie___> oh man i know exactly what you mean...except just to figure out what i could possibly want to eat, im a slow chooser
<awygle> filter by "no tomatos"
<pie___> tablets as a menu card
<pie___> thats way too cheesy though
<awygle> or maybe "no more than two of these things i don't like, so i don't have to customize *too* much"
<pie___> also "what could possibly go wrong"
<Zorix> famous last words
pointfree has quit []
pointfree has joined ##openfpga
<qu1j0t3> awygle walks into cafe. "Hi! What's your wifi password? .... and do you have an API?"
<awygle> hit a REST endpoint to order more coffee
<rqou> HTTP 418
<sorear> ^
<qu1j0t3> hehe
<qu1j0t3> you're all nerds
* awygle glances around, sees the ##openfpga sign
<awygle> i mean, eys
<qu1j0t3> daveshah: Yeah... well. You deserve what you GET :)
<awygle> you only GET what you give
<awygle> ohhh no wonder everything is so slow. i built this VM with 1 CPU. whoops.
<rqou> lol
<rqou> also, is having a restaurant menu be a SQL database sufficient?
<rqou> don't you also want an SMT solver or something?
<rqou> to help you with the knapsack problem? :P
<awygle> if they have an accessible database or API, i can run my own SMT solver
<awygle> do i want PAE/NX? i have no idea what that does
<rqou> it makes exploiting bugs harder
<rqou> but you need to disable it to run chinese crapware that have been packed
<rqou> (no points for guessing how i found that out)
<rqou> also gives you >4GB ram on 32-bit for linux and non-consumer windows builds
<rqou> yes, these features are conflated, because intel
<awygle> lol
<rqou> 32-bit consumer windows builds will actually enable PAE to get NX but still won't use more than 4GB memory thanks to shitty drivers
<balrog> rqou: because they need to continue to support WinXP-era drivers?
<balrog> rqou: btw have you seen the driver for USB for cheapie plotter-style vinyl cutters?
<balrog> it's annoying enough that you're better off using a USB-serial adapter. lol
<rqou> lol
<rqou> i blame windows for having a super annoying to deal with driver ecosystem
<rqou> at least they have winusb now
<balrog> (fwiw that driver seems to be a barely-modified version of the MS hpgl plotter example code)
<rqou> lol
<rqou> i'm increasingly of the opinion that everything should be ethernet instead
<rqou> no funny driver issues with that
<balrog> who has a second ethernet port?
<rqou> um, me?
<balrog> (if you have a laptop, do you even have a single on? I don't)
<rqou> alternately you can use usb and then pretend to be CDC-ECM/RNDIS
<balrog> (so are you suggesting bundling a USB-ethernet adapter in everything?)
<balrog> well yeah there is that
<balrog> the beaglebone black does that
<rqou> yeah i know
<rqou> i've used ev3dev, the out-of-box experience was actually pretty good
<rqou> the only problem was the teammate that had some chinese crapware network manager software
<awygle> of course there's no part 3 of this series lol
<rqou> use the source luke :P
<awygle> daveshah: so i have icepack/iceunpack working (allegedly). i think i now need to convince icefuzz to fuzz the lm4k, right?
<rqou> go rtfm :P :P
<rqou> (yes, i know why this attitude isn't welcoming)
<awygle> rqou: i *wish* there was an m to rtf :p
<daveshah> awygle: yes, although I wouldn't add a new device database. It will almost certainly be the same as the 8k
<daveshah> you just want to run it to check
<daveshah> have a look at how the 5k works, obviously ignoring IPConnect and DSP
<daveshah> the expectation is you should not see a change to any of the DB files
<daveshah> first of all, have you manually checked a design with iceunpack and glbcheck?
<awygle> yes, but not a terribly complicated or interesting one
<awygle> (by which i mean, just logic, no BRAM or PLL or anything)
<daveshah> that's fine, chances are that means all is OK
<awygle> the DB files are all these .txt's?
<daveshah> yeah, and icebox/iceboxdb.py
<daveshah> the first step is to add the lm4k to https://github.com/cliffordwolf/icestorm/blob/master/icefuzz/icecube.sh
<awygle> i did the first bit already
<awygle> working on the second
<daveshah> also https://github.com/cliffordwolf/icestorm/blob/master/icefuzz/extract.py, you might want to create a RAM database for the lm4k just to check (if it is a strict subset of the 8k one, that's fine)
<daveshah> while you wait for icefuzz, you can also start adding the device to https://github.com/cliffordwolf/icestorm/blob/master/icebox/icebox.py
<daveshah> most of the information in icebox.py can be found from the iCECUBE floorplan viewer
<awygle> hmm, icefuzz won't run correctly with make -j4 :( bummer
<daveshah> weird
<daveshah> I've run it with up to -j5 or so
<daveshah> but then the OOM killer started kicking in
<daveshah> with 16GB of RAM, but some being used for other stuff
<awygle> it worked ok when i had 1k as the device class (i think) but now it's not working with 8k
<daveshah> what is it doing?
<awygle> "Feasibility check for IO Placement failed"
<daveshah> ah, I think that is a general icefuzz issue
<daveshah> nowt to do with jobs
<daveshah> just hitting it earlier with -j4
<awygle> i may actually be OOMing now that i look at it...
<daveshah> occasioanlly it randomly generates a design that icecube doesn't like
<awygle> i only gave this VM like 4g
<awygle> and it makes sense that 8k is bigger than 1k
<daveshah> yeah, more RAM might help
<awygle> err 8g apparently
<awygle> oh well, not a huge deal. time to cook lunch anyway.
<daveshah> enjoy your lunch! don't be too surprised if you run into icefuzz issues on the 8k, it's been a long time and several icecube versions now since anyone ran icefuzz for the 8k (I fixed a few issues while doing 5k stuff, but that might not solve all 8k issues)
* awygle imagines that somewhere whitequark is muttering about CI
<daveshah> awygle: btw I suggest creating a WIP pull request fairly soon with your work so far, so that Clifford and I can start reviewing it, which will make merging easier in the future
m_w has joined ##openfpga
<awygle> daveshah: will do
<daveshah> Thanks!
dingbat has quit [Quit: Connection closed for inactivity]
<daveshah> awygle: great, thanks
<daveshah> Looks good at a glance so far
<awygle> which makes sense given that so far i've literally just moved information from one screen to another :p
sgstair_ has joined ##openfpga
sgstair has quit [Disconnected by services]
sgstair_ is now known as sgstair
ZipCPU has quit [Ping timeout: 268 seconds]
ZipCPU_ has joined ##openfpga
ZipCPU has joined ##openfpga
ZipCPU_ has quit [Ping timeout: 260 seconds]
mumptai has joined ##openfpga
DocScrutinizer05 has quit [Disconnected by services]
DocScrutinizer05 has joined ##openfpga
<rqou> i still can't get over just how cool sliding mode control is
dingbat has joined ##openfpga
<awygle> hmm, that is cool
<awygle> as usual, the wiki page is like "here's a totally reasonable description" and then i get to the theory section and my head explodes
<rqou> shoulda taken EE222 :P
<awygle> i wanted to lol
mumptai has quit [Quit: Verlassend]
<awygle> daveshah: okay, so i ran icefuzz and succeeded (after some tweaks). it looks like all the "bitdata" txt files match the "cached" txt files - is that what i was looking for?
<daveshah> Yes that sounds good
<daveshah> If you run git status, have the cached files or iceboxdb.py changed?
<awygle> "cached_io.txt" has changed
<awygle> it looks like it's added a bunch of PLL config bits?
<daveshah> Interesting. They must be different in the lm4k then
<daveshah> Can you post a doff?
<daveshah> *diff
<daveshah> That is absolutely fine. It is just because the PLLs are in a different position so the names are new, but the bits are the same
<daveshah> There is no need to create a new db for that
<daveshah> So you can now go on to the next step
<daveshah> Which is to add the device throughout icebox.py
<awygle> arright, cool
<awygle> what are these dbs actually used for / how are they used? i'm feeling a little bit lost in the woods here
<daveshah> awygle: the DBs contain a list of mappings between bit position (the tuple) and function name from the glb file
<daveshah> These are then subsequently normalised and processed, to produce iceboxdb.py
<daveshah> It might help to compare the two files to get an idea how they are used
<daveshah> Then icebox.py processes that, does some magic to work out how routing connects together
<daveshah> This can then output a chipdb for arachne-pnr, which is a description of both configuration bits and the routing graph for the chip
soylentyellow has quit [Remote host closed the connection]
soylentyellow has joined ##openfpga
zino has quit [Excess Flood]
zino has joined ##openfpga
feuerrot has quit [Ping timeout: 276 seconds]
zino has quit [Ping timeout: 260 seconds]
digshadow has quit [*.net *.split]
lexano has quit [*.net *.split]
scrts has quit [*.net *.split]
kuldeep has quit [*.net *.split]
pakesson has quit [*.net *.split]
moho1 has quit [*.net *.split]
digshadow has joined ##openfpga
pakesson has joined ##openfpga
moho1 has joined ##openfpga
kuldeep has joined ##openfpga
scrts has joined ##openfpga
lexano has joined ##openfpga
zino has joined ##openfpga
kuldeep_ has joined ##openfpga
bitd has quit [Remote host closed the connection]
kuldeep has quit [Excess Flood]
feuerrot has joined ##openfpga
Zorix has quit [Remote host closed the connection]
Zorix has joined ##openfpga
knielsen has quit [Ping timeout: 264 seconds]
digshadow has quit [Ping timeout: 240 seconds]
<cr1901_modern> https://twitter.com/cr1901/status/993115928685576193 Excellent soldering technique
<pie___> that appears to be going it's rounds
<pie___> i dont even know if that part of the iron is even hot but i can already feel my fingers burning
<q3k> anyone at the risc-v workshop in barcelona starting tomorrow?
<rqou> too many things over in Europe
<rqou> thanks Tr*mp
<balrog> cr1901_modern: that's old
<balrog> lol
Zorix has quit [Remote host closed the connection]
<cr1901_modern> balrog: So? It was perfect for the hashtag
Zorix has joined ##openfpga
<cr1901_modern> And in a perfect world I'd have 3-digit RTs, but since I can't attain critical mass on well, anything, I'm hovering at single digit
<balrog> oh you meant the hashtag was popular :P
Zorix has quit [Remote host closed the connection]
Zorix has joined ##openfpga