Monday, 2019-04-08

*** tpb has joined #tomu00:00
*** xkapastel has quit IRC01:09
*** rohitksingh has quit IRC02:16
*** rohitksingh has joined #tomu02:17
xobsOkay, good morning all.  Time to fix this.02:51
xobsThough I've very quickly come to the decision that I've broken this Raspberry Pi somehow.  I think I fried a GPIO block or two.02:54
xobsHmm... no, I just hit a bug in fomu-flash.03:43
*** rohitksingh_work has joined #tomu04:40
xobsOkay, /that's/ weird.  Foboot is failing because the subsequent packet is getting stuck on the end of the previous packet.05:02
futarisIRCcloudxobs: getting stuck on the end of the previous packet?05:39
xobsI just managed to figure it out.05:39
xobsI was pre-queueing an "ACK" so that when the final "IN" token came in, I'd acknowledge it right away.05:39
xobsWell, because you can't NAK a SETUP packet, it would automatically queue the "ACK" to the "IN", then load the subsequent "SETUP" packet into the buffer.05:40
tntYeah, that whole thing "you can't NAK" setup was definitely a bit annoying to me too.06:25
xobsFortunately, it looks like a software problem, which means I can hook up the debugger.06:38
*** jas4711 has joined #tomu06:44
*** rohitksingh_work has quit IRC07:44
*** cdmatter has quit IRC07:44
*** rohitksingh_work has joined #tomu08:16
*** rohitksingh_work has quit IRC08:24
*** awe00 has joined #tomu09:19
*** AmosSam has left #tomu09:21
*** AmosSam has joined #tomu09:22
xobsOkay, I just pushed a version with a new USB state machine.09:38
xobsThis one seems more reliable, in that it does the right thing most of the time.09:39
xobsI did get two bits of corruption on SPI when uploading data, and that's a little concerning.  My guess there is that my SPI state machine isn't behaving right.09:39
xobsRather, when I did a "fomu-flash -a 262144 -v blink.bin", it said that one bit had been transposed in the resulting file.09:41
AmosSamWhat is "-a 262144" flag?09:44
xobsThat specifies the address to read from.09:45
xobsSince the bitstream goes at address 262144, I'm verifying the contents of "blink.bin" beginning at that address.09:45
AmosSamAhhh, so when I verified foboot it worked because foboot is at beggining?09:45
xobsMost likely!  The "verify" command is used to check that a given binary exists on the SPI flash at a given offset.  To check that it was written properly.09:46
xobsUseful for verifying that the initial Foboot is written, and also for checking that Foboot is doing the right thing with the images that it writes.09:47
AmosSamOne thing that crossed my mind last night, can rpi be connected to test points and at same time usb pins ? Or that won't work?09:48
xobsThat's how I'm testing things now.09:52
xobsIt's probably doing slightly bad things with two 5V sources fighting.09:53
AmosSamWell, if rpi is powered by laptop, and fomu usb is connected to same laptop, it shouldn't matter since grounds are on same potential?09:55
AmosSamBut, on test pads it should be 3.3V? Would 5V from usb leak to that point?09:55
xobsThat's not my setup right now, though.09:55
xobsYou could probably skip the 5V line.  You might even be able to skip ground, but that's not recommended in case it's isolated somehow.09:56
AmosSamI'll try in next few days, when I move my printer to new address...09:58
xobsAlright, thanks for the help. It would have been really bad to ship with that iteration of foboot.10:00
*** awe00 has quit IRC10:04
AmosSamNo problem! Did you pushed changes? I'll test this evening and report back10:06
xobsI pushed my changes.10:13
*** futarisIRCcloud has quit IRC10:19
*** awe00 has joined #tomu10:47
*** awe00 has quit IRC11:56
*** awe00 has joined #tomu12:09
*** awe00 has quit IRC12:19
*** AmosSam has left #tomu12:20
*** AmosSam has joined #tomu12:27
*** awe00 has joined #tomu12:37
*** awe00 has quit IRC12:46
*** AmosSam has left #tomu12:59
*** awe00 has joined #tomu12:59
*** awe00 has quit IRC13:09
*** rohitksingh has quit IRC13:24
*** awe00 has joined #tomu13:32
*** awe00 has quit IRC13:38
*** AmosSam has joined #tomu13:42
*** rohitksingh has joined #tomu13:48
*** awe00 has joined #tomu13:51
*** rohitksingh has quit IRC13:57
*** rohitksingh_ has joined #tomu13:57
*** awe00 has quit IRC14:04
*** awe00 has joined #tomu14:27
*** AmosSam has left #tomu14:39
*** xkapastel has joined #tomu14:43
*** AmosSam has joined #tomu14:47
*** AmosSam has left #tomu15:00
*** AmosSam has joined #tomu15:16
*** awe00 has quit IRC15:16
*** emeb has joined #tomu15:37
mithroxobs: Could we rename the fomu-repo repository to something like fomu-rpi-packages or fomu-rpi-dpkg ?15:46
xobsmithro: sure15:47
mithroxobs: fomu-raspbian-packages?15:48
xobsmithro: do you have the ability to test the latest Foboot?15:51
mithroxobs: On the hacker fomu, or otherwise?15:56
*** awe00 has joined #tomu15:56
xobsmithro: any will do15:57
mithroxobs: I can probably do that15:59
xobsmithro: Would you like to try building it yourself, or would you like a bitstream?16:09
mithroxobs: Both?16:09
xobsI wonder if will let me do this...16:13
* xobs posted a file: top-multiboot-hacker.bin (102KB) < >16:13
xobsmithro: ^ does that show up?16:13
mithroxobs: Yes16:14
*** awe00 has quit IRC16:16
xobsFoboot [currently] only loads bitstreams.  With something like the picorv32 spi block I could have it do RISC-V code, too.16:17
AmosSamxobs: does foboot builds for you with --with-pll switch?16:20
xobsAmosSam: looks like it doesn't meet timing, no :(16:24
xobsWhat version of nextpnr are you using?16:25
xobsIt works for me if I add "--placer heap". The "heap" placer was added to nextpnr 15 days ago.16:27
*** rohitksingh_ has quit IRC16:28
*** rohitksingh has joined #tomu16:29
AmosSamI installed nextpnr yesterday, and version from git (aur package)16:29
AmosSamOk, builded new foboot, and it works on all three usb ports on my laptop16:30
xobsGive "--placer heap" a try then!16:30
xobs\o/ Hooray!16:30
AmosSamBut, in one port after dfu-util -e it deatached device but didn't run bitstream... after reflashing in same port it worked, and didn't succeded to repeat that error again...16:31
xobsThat sounds similar to the problems I experienced where one bit got transposed.16:32
AmosSamI tested with blink example and rgbblink from mifune16:32
xobsOkay.  It sounds like the bare minimum is there, and it works a good chunk of the time.  If it doesn't work, you can try it again.16:34
xobsI'd like to try more USB ports, and I'll test every device I can get my hands on, but I'm reasonably happy with the USB stack as-is.16:34
AmosSam> Give "--placer heap" a try then!16:34
AmosSamWith --with-pll argument for Because it doesn't meet timings for 48MHz i think16:34
xobsIt's getting late here.  I think I'll head to bed.16:35
xobsAmosSam: yes, if you add "--with-pll --placer heap" it meets timing, at least on my system.16:35
xobsInfo: Max frequency for clock 'clk48_1': 49.95 MHz (PASS at 48.00 MHz)16:35
AmosSamI'll play with it this evening, and if I have any problems I'll write them here :-D16:36
AmosSamAnd --with-dsp ?16:36
xobsI haven't tried that, but I think that right now there's nothing to infer.16:36
AmosSamAhaa, because I didn't removed dsp switch... and, which one is better? (Last question, I promise!)16:38
xobsI haven't seen the pll make a difference.16:40
xobsThe dsp will become more important once it can infer the multiplier I have vexriscv. For now they don't really matter.16:41
*** xkapastel has quit IRC17:02
mithroxobs: I won't get a chance to test it until this afternoon at the earliest17:23
mifunexobs: with me it has a 75% success rate, the other cases i get: dfu-util: Error during download get_status17:50
mifunedfu-util: Error during download get_status17:50
mifuneit doesnt do that with the bitstream from xobs18:11
*** tannewt has quit IRC18:28
*** esden has quit IRC18:28
*** tannewt has joined #tomu18:28
*** esden has joined #tomu18:29
mifuneok, bit too soon, same issues with that bitstream18:41
*** hl has quit IRC19:55
*** hl has joined #tomu20:09
*** xkapastel has joined #tomu21:29
mifuneit doesn't work consistantly, but compared to the pogo pin programming method it is so fun to play with.21:42
mifuneand an updated rgb blink:
tpbTitle: GitHub - pimdegroot/Fomu-rgbblink: RGB blink code for the Fomu hacker board (at
tntmifune: you don't happen to have a bunch of fomy you're not using btw ? :p21:58
*** TheJJ has quit IRC22:26
*** TheJJ has joined #tomu22:30
mithrotnt: I have Fomu Hacker boards...22:30
tntmithro: but you're half a world away22:33
mithrotnt: Were in the world are you?22:33
mithrotnt: Easy for me to ship anywhere in the world however22:33
tntwhat I'm more worried about is customs/import ... lately they've pretty much got every package I got from outside.22:37
*** jayvee has quit IRC22:38
*** jayvee has joined #tomu22:40
emebtape it inside a birthday card? :P23:09
tntYeah I guess an unmarked enveloppe would be a good bet23:12

Generated by 2.13.1 by Marius Gedminas - find it at!