whitequark changed the topic of #glasgow to: glasgow debug tool · code https://github.com/GlasgowEmbedded/Glasgow · logs https://freenode.irclog.whitequark.org/glasgow
<_whitenotifier-3> [whitequark/libfx2] whitequark pushed 1 commit to master [+0/-0/±2] https://git.io/fjyBp
<_whitenotifier-3> [whitequark/libfx2] whitequark 0ca7f8d - Simplify. NFC.
<_whitenotifier-3> [whitequark/libfx2] whitequark pushed 1 commit to master [+0/-0/±2] https://git.io/fjyRe
<_whitenotifier-3> [whitequark/libfx2] whitequark 5fec719 - Simplify. NFC.
_whitelogger has joined #glasgow
_whitelogger has joined #glasgow
_whitelogger has joined #glasgow
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark pushed 3 commits to master [+0/-0/±3] https://git.io/fjyu4
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark 1d93929 - applet.sensor.scd30: simplify. NFC.
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark 7777839 - applet.interface.jtag_probe: fix scan_xr_length on overlong XR.
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark d19696e - applet.interface.jtag_pinout: support more than 8 pins.
ali_as has joined #glasgow
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark pushed 1 commit to master [+0/-0/±1] https://git.io/fjyze
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark 001c19c - applet.program.nrf24l: unmark as preview.
<whitequark> tnt: question, about program-ice40-sram
<whitequark> does it ever make sense to not have a reset when using that applet?
<whitequark> isn't that the whole point
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark pushed 1 commit to master [+0/-0/±1] https://git.io/fjyzm
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark 4b74048 - applet.program.ice40_sram: simplify. NFC.
<whitequark> actually, it's required in that one
<whitequark> what about -flash?
<tnt> whitequark: with the sram one, maybe it makes less sense.
<tnt> For the flash one, I left it optional because I have at least one board where the "programming connector" doesn't have reset on it, I just press a button to hold it in reset while I use the applet ...
<tnt> but realistically I could just use the spi-flash applet for that board ...
<whitequark> i sort of see it but also it sort of doesn't make a lot of sense
<whitequark> i have no real opinion on it i guess
<marcan> whitequark: btw, selftest on my revC*0* complains now
<whitequark> marcan: complains how
<marcan> pins shorted or something
<marcan> let me try again
<whitequark> oh yeah that
<whitequark> it's extraordinarily sensitive to noise
<whitequark> like moisture
<marcan> ah
<whitequark> and flux residue
<whitequark> can you post a verbose log?
<whitequark> i've tried to reduce false positives but they still occur often, it needs more work
<whitequark> marcan: btw i'm reviewing your pushback changes right now
<whitequark> looking very good so far
<marcan> btw I pushed a quick hack that I needed to origin/marcan-wip, not sure if you saw that
<marcan> I needed to be able to use it on a host without the ice40 toolchain
<whitequark> er
<marcan> since apparently the way it was previously you could load a bitstream but not, like, actually run an applet?
<marcan> or maybe I just failed?
<whitequark> i'm confused
<marcan> whitequark: about soft_flush?
<whitequark> the soft parameter is ignored, twice
<marcan> uhh sorry that's just me being dumb
<whitequark> the semantics looks good the implementation um
<marcan> I started with a parameter and then made it two different methods
<marcan> and forgot the parameter
<whitequark> ack. will you fix or shall I cherry-pick and fix?
<marcan> either works
<marcan> I: g.applet.internal.selftest: self-test: PASS
<marcan> lol it works now
<whitequark> marcan: yeah the heuristic for using selftest is "does it *ever* PASS?"
<marcan> lol
<whitequark> like i said, it needs improvemet
<whitequark> hm
<whitequark> re: soft flush
<whitequark> I think I would prefer something like .flush(block=False) ?
<whitequark> the "soft" isn't super informative
<marcan> sure, whatever works
<marcan> that sounds good
<whitequark> or even wait=
<marcan> btw I did give that glasgow talk. only had the ws2812 thing at hand for a demo, but I live-coded a shitty logic analyzer (just dumps samples into the fifo) and it worked
<marcan> (and then just touched pins to test it, lol)
<whitequark> nice :D
<whitequark> how'd the response go
<marcan> as expected I guess? interest, some decent questions
<marcan> I think there were 20 or so people, IIRC
<whitequark> ah
thaytan has quit [Ping timeout: 272 seconds]
thaytan has joined #glasgow
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark pushed 3 commits to master [+0/-0/±3] https://git.io/fjyzC
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] marcan 80099fb - access.direct.demultiplexer: introduce non-blocking flushes
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] marcan 60052f4 - applet.video.ws2812_output: use non-blocking flush
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] marcan 8efdb8a - applet.video.ws2812_output: reduce reset time to 300µs as documented
<whitequark> marcan: ok, cherry-picked some obviously correct things
<whitequark> reviewing the rest now
<marcan> whitequark: thanks[
<marcan> this stupid hack from yesterday, not sure if there's another way of doing this I just don't understand https://github.com/GlasgowEmbedded/Glasgow/commit/fa0e8339d57bfec0eec989c1a2130f122fca3a73
<marcan> I just wanted to be able to use a prebuilt bitstream for an applet
<whitequark> marcan: I've actually hit this just yesterday as well
<whitequark> trying to use a specific bitstream with benchmark applet
<whitequark> to triage my revC1 issue
<whitequark> so yes, I think --bitstream and APPLET args should have an "at least one of" relationship, not "exactly one"
<whitequark> wtf does `async with cond:` do
<whitequark> oh
<whitequark> oh it has a lock inside
<whitequark> wait
<whitequark> marcan: isn't the asyncio condition... weird
<whitequark> is it literally just a (event,lock) tuple
<whitequark> why is it so boneless
<whitequark> nvm
<whitequark> i forgot how condition variables worked
<marcan> yeah I figured the implementation should be pretty simple
<whitequark> marcan: should there be a message on IN pushback too?
<whitequark> not just OUT pushback
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark pushed 3 commits to master [+0/-0/±3] https://git.io/fjyzN
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] marcan b855a71 - access.direct.demultiplexer: introduce software side queue length limits
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] marcan a321cb3 - support.endpoint: introduce queue_size argument
<whitequark> marcan: yeah the rest looks good, published with minor revisions. er. merged
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] marcan dfa6383 - applet.video.ws2812_output: implement pushback
<marcan> whitequark: thanks
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark pushed 2 commits to master [+0/-0/±2] https://git.io/fjygU
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark 6315ec1 - cli: remove `run --bitstream`, add `run-prebuilt [BITSTREAM]`.
<whitequark> marcan: added your hack too, in a slightly nicer way
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark 73f79e6 - cli: help wording. NFC.
<whitequark> marcan: btw once we migrate to nmigen i believe the "bitstream cache" becomes a reasonably good idea
<whitequark> the nmigen version is included in the netlist text already, as well as any pin attributes
<whitequark> and i'm going to hash the entire zip file with the netlist and constraints and build script
<whitequark> (nmigen provides a way to get a canonical archive of all build inputs)
<whitequark> then it's as easy as adding yosys and nextpnr version strings to the mix
<_whitenotifier-3> [Glasgow] whitequark created branch wip-applet.video.smia - https://git.io/fhhGp
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark deleted branch wip-smia
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark pushed 1 commit to wip-applet.video.smia [+4/-0/±1] https://git.io/fjv9P
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark 282d894 - applet.smia: new (partially functional) applet.
<_whitenotifier-3> [Glasgow] whitequark deleted branch wip-smia - https://git.io/fhhGp
<_whitenotifier-3> [Glasgow] whitequark created branch wip-applet.interface.lpc - https://git.io/fhhGp
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark pushed 1 commit to wip-applet.interface.lpc [+1/-0/±1] https://git.io/fjv96
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark a4e8419 - wip
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark deleted branch wip-lpc
<_whitenotifier-3> [Glasgow] whitequark deleted branch wip-lpc - https://git.io/fhhGp
<_whitenotifier-3> [Glasgow] Error. The Travis CI build could not complete due to an error - https://travis-ci.org/GlasgowEmbedded/Glasgow/builds/564367494?utm_source=github_status&utm_medium=notification
<marcan> whitequark: yeah that makes sense
<whitequark> marcan: think i pulled everything from your branches
<_whitenotifier-3> [Glasgow] whitequark commented on issue #143: Spy-Bi-Wire support - https://git.io/fjyg8
<tnt> whitequark: The MSP-EXP430FR2433 is the board I mostly use, it's 10$. Doesn't make much sense for me to ship one as it costs more in shipping than the board value, but I have no issue sponsoring a few of them for testing. Not sure what electronics distributors you usually use ?
<whitequark> tnt: local ones
<whitequark> mhhh let's see which preposterously expensive devboards they have.
<whitequark> oh yeah there it is, EXP430
<whitequark> except it's $22
<tnt> which one ? (because there is a bunch of exp430xxxx models)
<whitequark> MSP-EXP430G2
<whitequark> it has MSP430G2553IN20 and MSP430G2452IN20 in the box
<whitequark> does that work
<tnt> Yeah, I have a bunch of those as well.
<tnt> They use the same debug protocol.
carl0s has joined #glasgow
carl0s has quit [Remote host closed the connection]
verg has joined #glasgow
qazwsx57 has joined #glasgow
qazwsx57 is now known as qazwsxal
<marcan> whitequark: yeah I have a bunch of MSP crap too but there's no point in shipping it
<marcan> I got some of it for free even
_whitelogger has joined #glasgow
luis65 has joined #glasgow
ebb has joined #glasgow
luis65 has quit [Remote host closed the connection]
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark pushed 3 commits to master [+0/-0/±3] https://git.io/fjyrG
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark cea1ad8 - cli: simplify after 6315ec10. NFC.
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark 4fd2224 - cli: display a prominent warning in log for preview applets.
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark cf220b6 - target.analyzer: add basic infrastructure for generic events.
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark pushed 1 commit to master [+1/-0/±1] https://git.io/fjyoJ
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark f1abbf5 - applet.interface.ps2_host: new applet.
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark pushed 1 commit to master [+1/-0/±1] https://git.io/fjyoU
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark c81d189 - applet.interface.ps2_host: new applet.
<_whitenotifier-3> [Glasgow] whitequark deleted branch wip-applet.interface.ps2_host - https://git.io/fhhGp
<_whitenotifier-3> [GlasgowEmbedded/Glasgow] whitequark deleted branch wip-applet.interface.ps2_host
_whitelogger has joined #glasgow