Saturday, 2021-12-11

programmerjakejust saw:
lkclwe're google, we're github/microsoft, let us help you, the open source community, by becoming YOUR GOD as far as identity is concerned09:57
lkclyou can trust us, we are the ultimate GOD and AUTHORITY of proof of identity09:58
octavius"Don't be evil", right? It is interesting how companies develops very primitive, animalistic behaviour once their organisation becomes large enough. I guess the distance between responsibility and authority is increased11:18
lkcldebian has this one solved. it's a SEVENTEEN stage process however that is considered quotes too complicated quotes14:00
lkclof couuurse we can do better, said npm, archlinux, node.js, pypi, nix, and half a dozen others14:01
tplatenadded a skeleton for test_loadstore1_ifetch_multi()16:03
tplatenlkcl, had a look at your commits 104 min ago16:23
tplatenI found some dead links on, im going to fix that17:30
octaviusGiven that Debian has existed for as long as it did, they must've done a lot of things right17:56
tplatenI get an error in libreriscv$ git push18:02
tplatenFATAL: W any libreriscv tobias DENIED by fallthru18:02
octaviusAre you using this url
tplatenyes ssh://
tplatenwhich redirects to libre-soc, the other urls there work18:31
octaviusProbably a permission issue then19:16
octaviusI remember asking several times for permissions for different repos19:17
octaviusEach one is separate from the others (probably for the best)19:17
*** tplaten <tplaten!> has left #libre-soc19:24
lkclah he's dropped off irc.  probably the wiki git repo.19:35
lkcli had to remove everyone after a force-master-push19:35
octaviusSo lkcl, this BS_SAMPLE and BS_EXTEST/BS_INTEST are really cool20:10
octaviusIn BS_SAMPLE, the core and pad remain connected, thus setting the an input pad (such as UART rx), you'll see the signal propagate through to the core.20:12
octaviusIn this case (as UART tx connected to rx), the JTAG will see both rx/tx pads as asserted20:12
octaviusWhereas the EXTEST/INTEST (they use the same address), actually disconnect the core/pad mux. So if UART rx pad is asserted, tx pad will NOT be asserted unless it is set via TDI20:13
octaviusThe GTKWave show this pretty well (in the EXTEST case, you can see the core/pad connection break once the JTAG circuitry starts working20:14
lkcli know, it's pretty awesome, right?20:31
lkclbeing able to separate and independently test the entire ASIC's pads and its functions20:31
lkclso even if the pads are utterly broken, as long as the JTAG port is fine you can still test the functions20:32
lkcland even if the functions are utterly broken, you can still test the pads20:32
octaviusIs there a way to run the test within a test? For example, I'd like to run the GPIO test with the JTAG one, before I check the BS_SAMPLE data21:19
octaviusOr if not, make the JTAG test process wait until the GPIO test has been executed?21:19
lkcldo a separate test_xxxx() function which gets called after the first21:45
lkclhave a look at this to see how it's done21:48
lkclnote a setup function which returns a thing (the dut)21:48
lkclfollowed by that setup function being called in multiple sequential tests21:48
lkclhenriok, ping, trust all is well.  we should do a wikipedia page update for Libre-SOC now that e.g. the NGI POINTER Grant is underway, for the second ASIC22:02
*** A_Dragon is now known as Festive_Dragon22:19
octaviusThanks, lkcl, I'll continue the test tomorrow. Before Tuesday I'd like to make as much progress as possible. Tuesday onwards I'll be preparing for my driving test, so will need all the concentration I can get :)22:26
lkclniice - yeah i have often gotten visually distracted, dreaming of PCB layouts in full colour, one of the very rare times i could recall colour (mostly i visualise black and white)22:58
lkclya don't want lines of code scrolling across your visual cortex whilst trying to remember "Mirror Signal Mirror"...22:59

Generated by 2.17.1 by Marius Gedminas - find it at!