Thursday, 2017-12-14

*** tpb has joined #timvideos00:00
felix_one moment; i'll make a pull request for that00:01
shenkiso when I do this step, it gives me the "failed to identify flash" message:00:04
shenkijtagspi_init 0 {third_party/flash_proxies/bscan_spi_xc7a35t.bit}00:04
shenkiError: Unknown flash device (ID 0x00ffffff)00:05
felix_you have the spi proxy bitstreams on your machine? iirc they aren't shipped within openocd00:05
shenkiyeah, i do00:10
shenkimithro ships them as part of the litex environment we use00:10
shenkiit comes from
tpbTitle: GitHub - jordens/bscan_spi_bitstreams (at
felix_made a pull request for that; would be good if someone could verify that this doesn't break anything (it shouldn't, but i don't have an arty here to test)
tpbTitle: openocd/digilent-arty: use existing digilent-hs1.cfg as programmer by felixheld · Pull Request #23 · timvideos/conda-hdmi2usb-packages · GitHub (at
felix_upstreaming that would be best, but the openocd people seem to be quite slow with looking at patches; i submitted openocd support for the ac701 maybe three weeks ago and still didn't get any response or review00:19
mithroSee also ->
tpbTitle: Gerrit Code Review (at
mithrofelix_: I've found the best way is to join #openocd and poke PaulFester to get stuff reviewed00:23
shenkimithro: how do we debug the failure to identify flash?00:24
mithroshenki: Not sure...00:24
mithroshenki: Sorry, have stuff I need to be doing...00:24
felix_the patch on the openocd gerrit misses the spi flash programming include00:24
felix_mithro: ok, thx; i put that on my todo list00:25
felix_currently packing the stuff i need for the hackathon in vienna00:26
mithroThis year is disappearing soooooo fast....00:32
mithrofelix_: Do you have an Arty A7 board?00:37
*** CarlFK has quit IRC00:37
*** shorne_ has joined #timvideos00:40
felix_i currently have a zybo, an ac701, an icoboard, a pipistrello, some custom fpga board and maybe some old fpga boards i don't remember exactly (maybe spartan 3 and virtex 4)00:41
felix_ah and one trenz artix7 micromodule with base board00:42
*** shorne has quit IRC00:43
shenkimithro: my cpu_reset hacks did not work00:48
shenkimithro: when you have a moment please tell me what I did wrong00:48
tpbTitle: arty: add csr for cpu_reset · shenki/[email protected] · GitHub (at
mithroshenki: What happens?00:49
mithroshenki: Nothing or the system locks up?00:50
mithroshenki: Can you test by using bios mr command?00:50
shenkii can try that. i was using devmem from userspace which does the same thing00:51
shenkioh. that made the console print lots of "ef bf bd ef"00:57
shenkiand then it hung00:57
*** CarlFK has joined #timvideos00:57
*** ChanServ sets mode: +v CarlFK00:57
mithroshenki: Well that is kind of what I expected....01:01
*** wcmckee has quit IRC01:01
shenkithat's not what i'm after01:02
shenkii want it to reboot the cpu :)01:02
mithroshenki: You need to do something a little bit more advanced because doing the reset causes the clock to go away which means you never come out of reset...01:02
mithro  arch/openrisc/kernel/head.o: In function `evbar_setup':01:23
mithro>> (.head.text+0xa8): undefined reference to `PHYSICAL_START'01:23
shenkigit describe?01:23
felix_something seems to be wrong with the openocd buildbot from conda-hdmi2usb-packages; it fails with "Source download failed. Please investigate."
shenkimithro: what does 'git describe' say?01:26
mithroshenki: It's in a random email from kbuild-test-robot01:26
shenkioh. ignore it01:26
shenkimy github tree gets build by the intel 0day infrastructure :)01:26
shenkifree build-test coverage of litex development \o/01:27
mithrofelix_: Open the last fold on that page01:27
mithroApplying patch: '/home/travis/build/timvideos/conda-hdmi2usb-packages/openocd/digilent-arty.patch'01:27
mithroPatch level ambiguous, selecting least deep01:27
mithropatching file tcl/board/digilent_arty.cfg01:27
mithro/usr/bin/patch: **** unexpected end of file in patch01:27
felix_oops. well i'll fix it tomorrow01:29
shenkigenerate the patch with `git format-patch`01:30
shenkiif you haven't committed it, then generate it with `diff -up`01:30
shenkibut i recommend committing it :)01:30
felix_ok, will do. have to get some sleep now though to be awake when i have to catch my train to the airport01:31
shenkino worries. enjoy your hackfest!01:32
felix_thx :)01:34
*** [d__d] has quit IRC02:34
*** jea has quit IRC02:34
*** [d__d] has joined #timvideos02:38
*** jea has joined #timvideos02:38
*** CarlFK has quit IRC02:57
*** CarlFK has joined #timvideos03:00
*** ChanServ sets mode: +v CarlFK03:00
*** CarlFK has quit IRC03:31
*** rohitksingh_work has joined #timvideos03:49
*** rohitksingh_wor1 has joined #timvideos04:07
*** rohitksingh_work has quit IRC04:09
*** sb0 has quit IRC04:57
shorne_shenki: I will start looking at the patches for evbar06:41
shenkishorne_: cool06:42
shorne_as per mithro's mail updates for: networking, evbar patches, next steps...06:42
shorne_Is there anything you want me to take over?06:42
shenkilet me check the email06:43
shorne_I have been busy the last few days/weeks with gdb upstreaming (or1k port is upstream now)  and the uclibc fixes06:43
shenkishorne_: congrats on the gdb upstreaming work!06:43
shenkishorne_: there are a few things in head.S we need to fix for litex support06:48
shenkishorne_: one is using evbar for saving registers06:48
shenkishorne_: another changing the way the uart is used in head.S06:49
shenkishorne_: the options there would be to write a driver for the litex uart, and use that when we're on a litex platform06:49
shenkishorne_: the other is to remove the use of it from the assembly, and rely on a C driver06:49
shenkido you have thoughts?06:49
shorne_shenki: thanks on gdb its been a long work in progress07:13
shorne_on uart, I am fine removing in head.S, usually thats just there for some really early time debugging07:14
shorne_its just edited when needed07:14
shorne_Is it needed/used in litex?07:14
shenkishorne_: i don't think it's required. need to double check08:12
*** sb0 has joined #timvideos09:59
*** tsglove has joined #timvideos12:05
*** rohitksingh_wor1 has quit IRC12:35
*** rohitksingh has joined #timvideos13:20
*** CarlFK has joined #timvideos13:26
*** ChanServ sets mode: +v CarlFK13:26
*** sb0 has quit IRC13:29
*** sb0 has joined #timvideos14:53
*** FelixVi has joined #timvideos15:29
*** CarlFK has quit IRC15:45
*** CarlFK has joined #timvideos15:51
*** ChanServ sets mode: +v CarlFK15:51
*** springermac has quit IRC16:01
*** FelixVi has quit IRC16:01
cr1901_modernmithro: the needs work detecting the python version. On my machine "python" is for python 2, not python 316:23
cr1901_modernI'm not sure how to proceed considering I almost never use enter-env, and I'm only doing so b/c I suspect I broke travis by not assuming was run16:23
cr1901_modernmithro: Disregard, I didn't follow directions16:28
cr1901_modernActually, maybe not. Do you support either running from conda or your distro-provided python?16:30
*** samsagaz has joined #timvideos16:36
*** rohitksingh has quit IRC16:43
*** rohitksingh has joined #timvideos16:44
mithrocr1901_modern: no17:21
mithrocr1901_modern: we only support running inside of the conda environment...17:21
cr1901_modernmithro: I see. Well in any case, that wasn't the issue17:22
cr1901_modernmithro: Check #j-core for what the real issue was, since I asked resident Unix expert dalias a question about what I was seeing in travis17:22
cr1901_modernOkay _finally_ the travis builds (except for arty so far) are passing17:27
cr1901_modern(and it's a missing #ifdef, so that's simple enough)17:28
mithroHow hard is it to avoid the meson dependency until we use meson for everything?17:29
mithroOr an alternative question, how hard would it to be to convert our makefile + scripts/ into meson?17:30
cr1901_modernthat would also be a bit of work (but it would make things a lot cleaner when its done)17:31
cr1901_modernI don't see you getting around scripts/ tho17:32
cr1901_modernperhaps talk to jpakkane and see what he thinks?17:33
*** FelixVi has joined #timvideos17:52
*** tsglove has quit IRC18:11
*** tsglove has joined #timvideos18:14
*** rohitksingh has quit IRC18:15
cr1901_modernOkay, that should be the last commit:
tpbTitle: Add writable SPI flash and XMODEM xfer support by cr1901 · Pull Request #370 · timvideos/HDMI2USB-litex-firmware · GitHub (at
cr1901_modernNow, would anybody be willing to try out the spiflash XMODEM upload and report any issues to me? Compiling should work fine now that the travis build passes18:47
cr1901_moderncc: mithro^^18:47
CarlFK[m]cr1901_modern (IRC): wanna try it on my box?  (Im into something else right now)18:53
CarlFK[m]like monkeypatching print so that it errrors because I can't find what is printing !!!18:54
cr1901_modern^ Was that my fault ._.18:55
cr1901_modernCarlFK[m] And sure, but I don't have an acct on your box18:56
CarlFK[m]what is your github username?18:56
*** samsagaz has quit IRC19:15
*** MaZderMind has joined #timvideos20:25
cr1901_modernmithro: How do I handle this is switching to serial and back to jtag doesn't remedy the issue?21:28
tpbTitle: HDMI2USB-litex-firmware/ at master · timvideos/HDMI2USB-litex-firmware · GitHub (at
shorne_shenki: I posted some comments on the patch set on your github23:22
shorne_on the linux/litex-4.15 branch23:22
shorne_I dont think most of the evbar patches are needed in head.S anylonger23:23
_florent_shenki: n_rst.eq(~rst) --> self.comb += n_rst.eq(~rst)23:38
mithro_florent_: Is just that enough?23:52

Generated by 2.13.1 by Marius Gedminas - find it at!