| *** tpb has joined #timvideos | 00:00 | |
| shenki | shorne: hello! did you make any progress with the uart on hardware?~ | 00:29 |
|---|---|---|
| shorne | shenki: yes, I got linux to actually start to show output finally | 00:45 |
| shorne | after switching from the hdmi2usb timevideos/master branch back to mithro/litex-linux | 00:46 |
| shorne | now I can finally work on the serial driver | 00:46 |
| shorne | any progress on ethernet? | 00:47 |
| shenki | shorne: not in the past few days | 00:48 |
| shenki | shorne: i've got my board out and have a few hours to work on it | 00:48 |
| shenki | shorne: regarding the uart, I was using the keep_bootcon kernel command line option to keep the boot console up after the real console driver is loaded | 00:50 |
| shenki | this has the upside of producing output furhter into the boot process, but it always stops with garbled output | 00:50 |
| shenki | a guess is this is two drivers (earycons, and the normal driver) both talking to the same hardware | 00:50 |
| shenki | and the garbled mess is due to the output being interleaved from each driver | 00:51 |
| shenki | so perhaps keep_bootcon is not a great idea | 00:51 |
| shenki | just wanted to let you know what I'd tried :) | 00:51 |
| shorne | shenki: sorry I had to put the baby to sleep. I had disabled keep_bootcon because I just wanted to ensure both drivers were working | 01:10 |
| shorne | https://gist.github.com/stffrdhrn/fa17dfa34ff6e924b688e744e9f37678 | 04:41 |
| tpb | Title: gist:fa17dfa34ff6e924b688e744e9f37678 · GitHub (at gist.github.com) | 04:41 |
| shorne | I notice that with the garbage that comes after switching to real uart, it prints in 16 character bursts | 04:41 |
| *** samsagaz_ has joined #timvideos | 05:20 | |
| *** samsagaz has quit IRC | 05:20 | |
| *** rohitksingh has joined #timvideos | 06:05 | |
| shenki | shorne: ah. 16 is the size of the fifo | 07:17 |
| shorne | oh.. maybe its not getting the signal of fifo full | 07:32 |
| shorne | probably fifo is not an issue on qemu | 07:32 |
| *** rohitksingh has quit IRC | 09:10 | |
| *** rohitksingh has joined #timvideos | 09:11 | |
| *** rohitksingh has quit IRC | 09:11 | |
| *** micolous has quit IRC | 09:26 | |
| *** micolous has joined #timvideos | 09:27 | |
| shorne | I think if correctly implemented the uart_tx_empty(), but still getting the same issue | 09:30 |
| shorne | (it was hacked to always be empty before, which I thought might explain the fifo overrun symptom) | 09:31 |
| shorne | FYI, I dont think the issue is with interrupts | 10:08 |
| shorne | I get the same issue disabling the interrupts | 10:09 |
| shorne | ahh, cool I got some small progress | 10:11 |
| shorne | in, litex_uart_console_wait_tx if I put a msleep(1), it slows down the putchar and I get full output | 10:11 |
| shorne | but I see we are getting an oops | 10:11 |
| shorne | compiling with symbols to see more details | 10:14 |
| shorne | so it looks like the code to wait for tx is nto really working | 10:15 |
| shorne | ok... I fixed it, its and issue reading the TXFULL register, again endian thing | 10:21 |
| shorne | but input not working | 10:22 |
| shorne | also the fix is a hack | 10:22 |
| shorne | need something better | 10:22 |
| shorne | https://gist.github.com/a7c94c9fc76463da494349007a27dfc3 | 10:22 |
| tpb | Title: - · GitHub (at gist.github.com) | 10:22 |
| shorne | mithro: shenki: thats my latest changes | 10:23 |
| shorne | line 63 it the fix | 10:24 |
| shorne | but its a hack! | 10:26 |
| shorne | ok.. its better, I switch everything to ioread32be / iowrite32be and getting almost to the prompt, input is not working | 12:07 |
| shorne | last line: | 12:07 |
| shorne | Please press Enter to activate this console. ip: socket: Function not implemented | 12:07 |
| shorne | ok, got input now | 12:09 |
| shorne | mithro: got serial working on arty | 12:10 |
| shorne | good and qemu is working too | 12:12 |
| shorne | one minor issue on arty , pressing enter I get the prompt 2 times | 12:12 |
| shorne | mithro: shenki: https://github.com/mithro/linux-litex/pull/2 | 12:28 |
| tpb | Title: uart: litex: Get working on real hardware by stffrdhrn · Pull Request #2 · mithro/linux-litex · GitHub (at github.com) | 12:28 |
| shenki | shorne: great work! i just got back to my laptop, i'll give it a go here | 12:37 |
| shenki | # uname -a | 12:47 |
| shenki | Linux buildroot 4.14.0-rc3-00016-g237af2cf1b71 #265 Sun Nov 12 23:11:40 ACDT 2017 openrisc GNU/Linux | 12:47 |
| shenki | # | 12:47 |
| shenki | # cat /proc/cpuinfo | 12:47 |
| shenki | cpu: OpenRISC-10 | 12:47 |
| shenki | \o/ | 12:47 |
| shenki | shorne: 🍻 | 12:47 |
| CarlFK[m] | HDMI2USB - Open Hardware that has a very close source requirement :( | 14:56 |
| CarlFK[m] | Libre Planet talk proposal - sound accurate ? | 14:56 |
| mithro | shorne / shenki: you guys are my heros! | 16:32 |
| *** CarlFK has quit IRC | 17:06 | |
| _florent_ | shenki/shorne: great work! | 17:15 |
| shorne | CarlFK[m]: the closed part is the fpga and bitstream compiler? | 21:56 |
| shorne | shenki: thanks for checking | 21:57 |
| CarlFK[m] | shorne: yes | 21:58 |
| shorne | mithro: thanks :) FYI, not sure if you saw above, I dont have success with the timvides/master HDMI2USB branch. | 21:58 |
| shorne | mithro: I had to switch back to the mithro/litex-linux branch | 21:59 |
| mithro | shorne: Sorry if I caused confusion, we haven't merged the linux stuff into master yet | 22:00 |
| *** techman83 has quit IRC | 22:02 | |
| shorne | mithro: ok... I thought you did mention a bit merge, it looks like the main diff is the litex module | 22:03 |
| *** techman83 has joined #timvideos | 22:04 | |
| shorne | mithro: Also, for qemu build we have to set CFLAGS=Wno-error. I found we can just pass --disable-werror to qemu configure | 22:04 |
| *** ChanServ sets mode: +v techman83 | 22:04 | |
| mithro | shorne: I did merge the arty / series-7 stuff into mainline | 22:05 |
| shorne | mithro: ok, I am just looking at the diff | 22:07 |
| shorne | mithro: it looks like the main diff is enabling MMU and TIMER in mor1kx | 22:08 |
| mithro | shorne: Oh, I need to merge https://github.com/timvideos/HDMI2USB-litex-firmware/pull/374 sometime soon | 22:08 |
| tpb | Title: Make it possible to start an image in qemu as a user. by tsukasa-au · Pull Request #374 · timvideos/HDMI2USB-litex-firmware · GitHub (at github.com) | 22:08 |
| mithro | shorne: https://github.com/timvideos/HDMI2USB-litex-firmware/pull/372/files | 22:08 |
| tpb | Title: WIP - scripts: Working on script to build Linux. by mithro · Pull Request #372 · timvideos/HDMI2USB-litex-firmware · GitHub (at github.com) | 22:08 |
| shorne | mithro: yeah, I noticed qemu user mode networking could be used, but I was fine with tap :) it would be good to try though (I didnt see this PR, but it looks ok to me) | 22:11 |
| *** CarlFK has joined #timvideos | 22:16 | |
| *** ChanServ sets mode: +v CarlFK | 22:16 | |
| shorne | mithro: ah, pr/372 is looks like just what we need (am not using the build script, I just use my own) | 22:22 |
| shorne | mithro: I just left you some comments on the build script | 22:22 |
| mithro | I started putting together instructions to give to people at https://docs.google.com/document/d/1b2rBju9Fxjlcv6y3Ib1WcrvmzPA63Pe-pnBaziJXZNY/edit# | 22:27 |
| tpb | Title: Miniconf Set Up Instructions - Google Docs (at docs.google.com) | 22:27 |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!