| *** tpb has joined #timvideos | 00:00 | |
| *** CarlFK has joined #timvideos | 00:19 | |
| *** ChanServ sets mode: +v CarlFK | 00:19 | |
| *** mauz555 has joined #timvideos | 01:18 | |
| *** meldron4 has joined #timvideos | 01:36 | |
| *** meldron4 has quit IRC | 01:41 | |
| CarlFK | https://github.com/timvideos/litex-buildenv/wiki/Xilinx-Vivado-screen-shots | 03:50 |
|---|---|---|
| tpb | Title: Xilinx Vivado screen shots · timvideos/litex-buildenv Wiki · GitHub (at github.com) | 03:50 |
| *** rohitksingh_work has joined #timvideos | 03:54 | |
| *** futarisIRCcloud has joined #timvideos | 04:09 | |
| *** Sc00bz has quit IRC | 05:30 | |
| *** CarlFK has quit IRC | 05:35 | |
| cr1901_modern | mithro: http://ix.io/1n5B nextpnr can't find the correct version of libboost_python | 05:49 |
| cr1901_modern | From the build log output, it doesn't even look like yosys or nextpnr ran at all | 05:51 |
| cr1901_modern | I know in a previous build you created a dummy gateware. Is it possible it got cached and the build step was skipped? | 05:52 |
| cr1901_modern | gateware build step* | 05:52 |
| *** sc00bz has joined #timvideos | 05:59 | |
| cr1901_modern | Okay, making a whole new clean conda env fixes this | 06:01 |
| cr1901_modern | Maybe conda decided not to upgrade my existing libboost_python | 06:03 |
| mithro | cr1901_modern: looks like it ran to me | 06:04 |
| mithro | https://travis-ci.org/timvideos/litex-buildenv/jobs/430770726 | 06:04 |
| cr1901_modern | mithro: Yes I stand corrected, it did run. But I was expecting all the nextpnr [Info] messages | 06:05 |
| cr1901_modern | looks like you redirected them all to a file | 06:05 |
| cr1901_modern | I would like to know why there's a libboost mismatch tho w/ my current repo | 06:06 |
| mithro | cr1901_modern: the Travis wrapper turns them into dots | 06:06 |
| cr1901_modern | s/repo/env/ | 06:06 |
| mithro | Otherwise with Xilinx tools you end up with the log output being too much and Travis killing your build | 06:07 |
| *** Guest90018 has joined #timvideos | 06:08 | |
| *** Guest90018 has quit IRC | 06:13 | |
| cr1901_modern | mithro: Yea for some inexplicable reason, conda has decided in my existing env that nextpnr is tied to "libboost: 1.65.1-habcd387_4" (wrong version). | 06:14 |
| cr1901_modern | But when I create a fresh conda env, that dependency becomes: "libboost: 1.67.0-h46d08c1_4" | 06:14 |
| cr1901_modern | Obviously, "just nuke the old environment" is a solution. But I'm wondering if you know of a more gentle solution. Like is there a way to refresh how conda decides which version of a package to use (since my existing metadata appears to be stale)? | 06:15 |
| mithro | conda install nextpnr | 06:16 |
| *** sc00bz has quit IRC | 06:16 | |
| cr1901_modern | That doesn't update the deps | 06:16 |
| cr1901_modern | in fact, even if I uninstall libboost, and install nextpnr, it'll still use libboost 1.65.1 | 06:17 |
| *** CarlFK has joined #timvideos | 06:30 | |
| *** ChanServ sets mode: +v CarlFK | 06:30 | |
| *** sc00bz has joined #timvideos | 06:56 | |
| *** Teeed9 has joined #timvideos | 07:16 | |
| *** Teeed9 has quit IRC | 07:19 | |
| cr1901_modern | mithro: Solution for now is to make sure libboost is installed _before_ nextpnr | 07:21 |
| cr1901_modern | otherwise nextpnr will pull in the wrong version | 07:21 |
| cr1901_modern | just in case this bites someone else | 07:21 |
| xobs | That sounds so helpful. | 07:22 |
| * cr1901_modern invokes poe's law | 07:22 | |
| cr1901_modern | being serious or facetious :P? | 07:22 |
| xobs | libboost is magical, and I don't understand it. I'm having a similar problem installing Coriolis on 16.04 -- it seems confused as to whether it's linking against libboost_python2 or libboost_python3, and I've never figured out how to tell it what's what. | 07:23 |
| xobs | So genuine helpful. | 07:23 |
| mithro | cr1901_modern: need to check why conda isn't seeing the deps correctly | 07:23 |
| cr1901_modern | mithro: I wouldn't know how to do that | 07:25 |
| cr1901_modern | mithro: It's also worth noting that this does _not_ happen in a fresh environment | 07:25 |
| cr1901_modern | mithro: Btw, we should prob add tinyprog to litex-buildenv :) | 07:26 |
| cr1901_modern | (it works w/ the makefile, I've tried it) | 07:26 |
| mithro | Hrm, I was sure I had added that... | 07:28 |
| mithro | But it seems to have gotten lost... | 07:29 |
| cr1901_modern | xobs: Fair :D, was just wondering. B/c rereading my solution, it sounded pretty vague | 07:34 |
| xobs | cr1901_modern: I wonder if it isn't any better on 18.04. I say that because despite all of the other problems I've experienced under 18.04, libboost just worked there unlike in 16.04. | 07:35 |
| cr1901_modern | I wouldn't know, I don't upgrade Ubuntu until LTS ends :P | 07:36 |
| cr1901_modern | Which reminds me... I still have 14.04 on my dying netbook | 07:36 |
| cr1901_modern | which I somehow managed to revive today | 07:36 |
| xobs | It's what's installed on Windows. You can pick from Debian, 18.04, 16.04, or some Suse flavour. | 07:37 |
| cr1901_modern | Oh, yea I use 7. No plans to upgrade until I'm forcefully coerced into doing it. | 07:37 |
| xobs | It's kind of a weird mindset. They're more-or-less a chroot. | 07:38 |
| cr1901_modern | Oh _now_ the deps have been updated when I install nextpnr by itself | 07:44 |
| cr1901_modern | well problem solved I guess | 07:45 |
| cr1901_modern | mithro: Let me know when tinyprog is avail please? | 07:50 |
| cr1901_modern | (from conda) | 07:50 |
| *** jonh29 has joined #timvideos | 07:53 | |
| *** jonh29 has quit IRC | 07:57 | |
| cr1901_modern | ewen: What's the purpose of running "make firmware-cmd" after ./scripts/download-prebuilt.sh? How does rebuilding the firmware affect prebuilt gateware and vice versa? | 08:04 |
| cr1901_modern | mithro: This line never runs, does it? https://github.com/timvideos/litex-buildenv/blob/47afea2b3a444ec6d1ad9334e364e6361fd8c6bc/Makefile#L192 | 08:05 |
| tpb | Title: litex-buildenv/Makefile at 47afea2b3a444ec6d1ad9334e364e6361fd8c6bc · timvideos/litex-buildenv · GitHub (at github.com) | 08:05 |
| cr1901_modern | See: https://github.com/timvideos/litex-buildenv/blob/47afea2b3a444ec6d1ad9334e364e6361fd8c6bc/Makefile#L119 | 08:05 |
| tpb | Title: litex-buildenv/Makefile at 47afea2b3a444ec6d1ad9334e364e6361fd8c6bc · timvideos/litex-buildenv · GitHub (at github.com) | 08:05 |
| cr1901_modern | mithro: Wait ignore | 08:05 |
| *** spruit1112 has joined #timvideos | 08:07 | |
| cr1901_modern | mithro: What is the purpose of args.override_firmware to make.py? | 08:07 |
| cr1901_modern | When I set FIRMWARE=micropython, the stub firmware gets built anyway | 08:07 |
| cr1901_modern | (stub being default firmware for devices like tinyfpga) | 08:08 |
| *** Neg127 has joined #timvideos | 08:19 | |
| *** Neg127 has quit IRC | 08:22 | |
| cr1901_modern | "/home/william/Projects/litex-buildenv/build/conda/bin/../lib/gcc/lm32-elf/5.4.0/../../../../lm32-elf/bin/ld:litex.ld:19: warning: memory region `main_ram' not declared" Yes, thank you. I would've liked to know that _before_ I crashed | 08:23 |
| cr1901_modern | obviously we'll need a separate linker script for running from flash | 08:24 |
| cr1901_modern | for now I'm just gonna use the current one | 08:24 |
| *** futarisIRCcloud has quit IRC | 08:48 | |
| *** sc00bz has quit IRC | 08:51 | |
| *** sc00bz has joined #timvideos | 09:10 | |
| *** sc00bz has joined #timvideos | 09:16 | |
| cr1901_modern | mithro: For micropython to work on tinyfpga, we're gonna have to change the startup code in libbase | 09:24 |
| cr1901_modern | specifically, micropython has a .data section you can't opt out of- i.e. initialized memory. This is stored on the flash. Normally, we run firmware from DRAM, and we don't have to worry about manually initializing this code. | 09:25 |
| cr1901_modern | The bootloader in the BIOS will take care of this when copying the whole program (including the .data section) to DRAM | 09:26 |
| cr1901_modern | With tinyfpga, we don't have this luxury | 09:26 |
| cr1901_modern | My suggestion is to provide an alternate startup code in libbase for each arch that copies the .data section from flash to RAM (for targets not executing from DRAM) | 09:28 |
| cr1901_modern | cc: _florent_, have any issues w/ this? I understand RISC-V forces the use of a .data section, but Idk if it conflicts with my idea. | 09:28 |
| *** CarlFK has quit IRC | 09:40 | |
| *** MB_Adriaan has joined #timvideos | 11:24 | |
| *** MB_Adriaan has quit IRC | 11:27 | |
| *** cRaNnK5 has joined #timvideos | 11:55 | |
| *** cRaNnK5 has quit IRC | 11:56 | |
| *** rohitksingh_wor1 has joined #timvideos | 12:40 | |
| *** rohitksingh_work has quit IRC | 12:41 | |
| *** rohitksingh_wor1 has quit IRC | 12:43 | |
| *** CarlFK has joined #timvideos | 12:55 | |
| *** ChanServ sets mode: +v CarlFK | 12:55 | |
| *** rohitksingh has joined #timvideos | 13:44 | |
| CarlFK | xobs: did you want something tested on 18.04? I got a box... | 13:55 |
| xobs | CarlFK: thanks, I'll keep that in mind. For now I'm using the WSL Ubuntu 18.04 install, which I'm hoping is "good enough". | 13:56 |
| CarlFK | k - I have like 4-10 boxes that are used for recording now and then, but most of the time they sit around doing nothing. the recording setup is all automated, so I don't mind destroying them between shows | 13:58 |
| xobs | That's very good to know. | 13:58 |
| xobs | Nice that you have the reimaging process well defined, too. | 13:59 |
| CarlFK | pxe boot the installer, preseed the os install, ansible all the rest. | 14:01 |
| *** thaytan has quit IRC | 14:05 | |
| *** thaytan has joined #timvideos | 14:06 | |
| *** ChanServ sets mode: +v thaytan | 14:06 | |
| CarlFK | "The Crickit is powered by seesaw, our I2C-to-whatever bridge firmware." https://www.adafruit.com/product/3093 | 14:11 |
| tpb | Title: Adafruit CRICKIT for Circuit Playground Express ID: 3093 - $29.95 : Adafruit Industries, Unique & fun DIY electronics and kits (at www.adafruit.com) | 14:11 |
| CarlFK | I2C-to-whatever bridge :p | 14:11 |
| *** rohitksingh has quit IRC | 14:50 | |
| *** rohitksingh has joined #timvideos | 14:54 | |
| *** rohitksingh has quit IRC | 15:08 | |
| *** mauz555 has joined #timvideos | 15:45 | |
| *** mauz555 has quit IRC | 16:37 | |
| *** waldo323 has quit IRC | 16:55 | |
| *** waldo323 has joined #timvideos | 16:57 | |
| *** waldo323 has quit IRC | 17:03 | |
| *** waldo323 has joined #timvideos | 17:10 | |
| *** tac-tics has joined #timvideos | 17:16 | |
| *** L0j1k4 has joined #timvideos | 17:21 | |
| *** L0j1k4 has quit IRC | 17:23 | |
| *** Kripton has quit IRC | 17:43 | |
| *** mauz555 has joined #timvideos | 17:56 | |
| *** Kripton has joined #timvideos | 17:56 | |
| *** john_doe_jr5 has joined #timvideos | 19:06 | |
| *** mauz555 has quit IRC | 19:11 | |
| *** rme16 has joined #timvideos | 19:12 | |
| *** dcousens5 has joined #timvideos | 20:33 | |
| *** dcousens5 has quit IRC | 20:37 | |
| *** Throttled0 has joined #timvideos | 21:43 | |
| *** Throttled0 has quit IRC | 21:51 | |
| *** nitdega4 has joined #timvideos | 22:15 | |
| *** g11 has joined #timvideos | 22:26 | |
| *** g11 has quit IRC | 22:31 | |
| *** wildermind has joined #timvideos | 22:42 | |
| *** mauz555 has joined #timvideos | 23:56 | |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!