00:20 <gruetzkopf> E: glasgow.applet.benchmark: mode source failed!

00:20 <gruetzkopf> is that applet currently supposed to work

02:41 <whitequark> gruetzkopf: mode source and sink should work

02:41 <whitequark> mode loopback should not

06:57 <_whitenotifier-c> [whitequark/Glasgow] whitequark pushed 1 commit to master [+1/-0/±1] https://git.io/fhQLn

06:57 <_whitenotifier-c> [whitequark/Glasgow] whitequark 6c47f86 - applet.audio_output: new applet.

07:00 <_whitenotifier-c> [Glasgow] Success. The Travis CI build passed - https://travis-ci.org/whitequark/Glasgow/builds/491487257?utm_source=github_status&utm_medium=notification

11:01 <gruetzkopf> loop hangs, sink works fine

11:11 <whitequark> loop is unfortunately expected to hang, there's a timing issue i keep trying to fix

11:11 <whitequark> source should work on revB and revC

11:11 <whitequark> run with -vv and post the log please

11:15 <gruetzkopf> https://paste.debian.net/plainh/e682bc1e

11:18 <whitequark> gruetzkopf: that's really weird. can you do a diff between expected/actual?

11:22 <gruetzkopf> i have exactly this one device to test with, not sure?

11:23 <whitequark> no, it's in the applet

11:23 <whitequark> give me a moment

11:25 <whitequark> https://paste.debian.net/1066913/

11:29 <gruetzkopf> they're definitly different, find them at https://gruetzkopf.org/glasgow

11:31 <whitequark> expired certificate :S

11:35 <whitequark> gruetzkopf: try again on commit ec6e3ca0b25e8da6f3d58a80190c7b126a5fd568

11:35 <whitequark> I think you're hitting #89 in a particularly unlucky PVT corner

11:40 <whitequark> or, it could be something else

11:42 <gruetzkopf> argh, stupid certbot

11:42 <gruetzkopf> will do

11:42 <whitequark> https://paste.debian.net/1066918/

11:42 <whitequark> try selftest with this

11:44 <gruetzkopf> I: glasgow.applet.selftest: self-test: PASS

11:44 <whitequark> hm

11:45 <whitequark> oh, right, that wouldn't help a lot

11:45 <whitequark> ok, yeah, so try that earlier commit

11:45 <whitequark> it's also broken but in a different way

11:49 <gruetzkopf> yes, source and sink work on that commit

11:50 <whitequark> ok, so it is #89

11:50 <whitequark> do you feel like trying to fix it? I can explain what's needed

11:53 <gruetzkopf> was that the "fun timing between fx2 and ice40" issue with SB_GB_IO?

11:54 <whitequark> yes, but the solution doesn't involve timing at all

11:54 <whitequark> it's actually a few tweaks on FX2 side

11:55 <gruetzkopf> that i can try to do

11:56 <whitequark> ok, so

11:57 <whitequark> hm

11:57 <whitequark> hang on

12:11 <whitequark> gruetzkopf: ok, can you take a look at the semantics of flags in FX2?

12:11 <whitequark> specifically, let's say FF (full flag) with INFM1 bit set

12:12 <whitequark> and how it compares with PF and FIFO level set to FIFO depth - 1

12:12 <whitequark> I can't figure out for the life of me if these two are equivalent

12:14 <gruetzkopf> will do

13:25 <gruetzkopf> they read equivalently for DECIS=0, i should grab a LA and compare them (see if they rise on the same edge)

13:27 <whitequark> please do

20:34 <Yatekii> \o/ gruetzkopf is here :P

20:35 <gruetzkopf> hi!

20:35 <gruetzkopf> i'll grab a scope in about 2h

21:14 <gruetzkopf> can i (easily) make one of the IN endpoints not be used so i can steal its FLAG_FF pin in favor of the others FLAG_PF?

21:37 <gruetzkopf> okay, i can take out the ep4->ep8 loopback test from selftest for this

21:48 <Yatekii> gruetzkopf: you know Doeme aight?

21:49 <gruetzkopf> yep

21:49 <Yatekii> hah! i knew it

21:49 <Yatekii> didn't know u were part of this awesome project!

21:49 <Yatekii> discovered it a few days back :)

21:50 <gruetzkopf> i more-or-less stumbled upon it by it being discussed A LOT in another channel i was in

21:50 <gruetzkopf> (before this channel existed)

21:50 <gruetzkopf> and now i'm apparently hunting after bad documentation for parts used

21:52 <Yatekii> haha, #mosfetkiller? ^^

21:53 <gruetzkopf> yep.

21:54 <Yatekii> yeah, I basically did this as a bachelor thesis (FPGA IP, webserver on the ARM core, Webui for timedomain & frequency domain analysis) I used a red pitaya tho and ofc it had way less features, so I am really interested in this project :)

21:54 <Yatekii> will buy one for sure!