| *** tpb <[email protected]> has joined #litex | 00:00 | |
| *** toshywoshy <toshywoshy!~toshywosh@ptr-377wf33o3bnthuddmycb.18120a2.ip6.access.telenet.be> has quit IRC (Remote host closed the connection) | 00:06 | |
| *** toshywoshy <toshywoshy!~toshywosh@ptr-377wf33o3bnthuddmycb.18120a2.ip6.access.telenet.be> has joined #litex | 00:27 | |
| *** Degi_ <[email protected]> has joined #litex | 02:08 | |
| *** Degi <[email protected]> has quit IRC (Ping timeout: 256 seconds) | 02:09 | |
| *** Degi_ is now known as Degi | 02:09 | |
| *** TMM_ <[email protected]> has quit IRC (Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) | 02:53 | |
| *** TMM_ <[email protected]> has joined #litex | 02:53 | |
| *** peeps[zen] <peeps[zen]!~peepsalot@openscad/peepsalot> has joined #litex | 03:02 | |
| *** peepsalot <peepsalot!~peepsalot@openscad/peepsalot> has quit IRC (Ping timeout: 252 seconds) | 03:03 | |
| *** nrossi[m] <nrossi[m]!~nrossimat@2001:470:69fc:105::1:4527> has joined #litex | 07:01 | |
| *** FabM <[email protected]> has joined #litex | 08:07 | |
| *** cr1901_ <cr1901_!~cr1901@2601:8d:8600:911:3863:96e8:22ca:d681> has joined #litex | 08:12 | |
| *** cr1901 <cr1901!~cr1901@2601:8d:8600:911:dd88:331f:ddb3:df09> has quit IRC (Read error: Connection reset by peer) | 08:13 | |
| *** Wolf0 <Wolf0!~Wolf0@user/wolf0> has quit IRC (Ping timeout: 265 seconds) | 08:27 | |
| *** Wolf0 <Wolf0!~Wolf0@user/wolf0> has joined #litex | 08:27 | |
| tcal | tnt: that fixed it!!! | 08:28 |
|---|---|---|
| tnt | tcal: \o/ | 08:42 |
| *** jeffdi1 <[email protected]> has quit IRC (Quit: Leaving.) | 09:10 | |
| *** essele <[email protected]> has joined #litex | 09:25 | |
| leons | I've got to say, compared to how crazy running a bit-bang memory mapped I2C over wishbone over UART through a TCP connection tunneled through two SSH sessions with a latency of ~10ms sounds like, it works surprisingly well. Essentially doing the thing outlined here https://twitter.com/enjoy_digital/status/1382032355351801861. Thanks _florent_ for providing the infrastructure for that! | 09:44 |
| leons | _florent_: Would you mind placing https://gist.github.com/enjoy-digital/e40d296a852335acfcab89a686147272 under a free license? That way I can vendor it into my repository proper 🙂 | 09:45 |
| * geertu found "secret" code ;-) | 09:57 | |
| *** zjason` is now known as zjason | 12:58 | |
| zyp | leons, if you like, you can have this: https://paste.jvnv.net/view/oEaWv | 13:06 |
| tpb | Title: JVnV Pastebin – View paste – Untitled (at paste.jvnv.net) | 13:06 |
| zyp | consider it public domain, I wrote it | 13:06 |
| leons | zyp: Awesome, thanks! It works like a charm and is also significantly faster than the other one (presumably because of the microsecond-sleep delays which really aren't needed when you bitbang over wishbone-uart-tcp-ssh :D) | 13:32 |
| zyp | it's not needed for bitbanging over wishbone-usb either | 13:43 |
| zyp | i2c is usually doing 400 kHz without issue, where a half clock is 1.25us, good luck getting latency below that from a python script running on a regular OS :) | 13:46 |
| *** cr1901_ is now known as cr1901 | 13:47 | |
| leons | yeah i figure 🙂 | 13:47 |
| leons | oh my, I just found out I can tell Element to stop replacing my plain-text smileys with unicode emojis... not exactly a great idea to have that enabled by default :/ | 13:49 |
| *** jeffdi1 <[email protected]> has joined #litex | 15:02 | |
| *** FabM <FabM!~FabM@armadeus/team/FabM> has quit IRC (Quit: Leaving) | 15:51 | |
| *** jersey99 <[email protected]> has joined #litex | 17:26 | |
| *** TMM_ <[email protected]> has quit IRC (Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) | 19:11 | |
| *** TMM_ <[email protected]> has joined #litex | 19:11 | |
| bjonnh | I see that: https://github.com/litex-hub/litex-boards | 19:51 |
| bjonnh | doesn't have the Z7-20 | 19:52 |
| bjonnh | will it just require me to tweak some configs? | 19:52 |
| tpw_rules | i just cannot get lxterm to work properly. i'm trying linux-on-litex-vexriscv on my butterstick and i can use openocd to program the svf just fine. but lxterm just prints garbage. some letters on the keyboard echo a different letter back but it doesn't really make sense | 19:58 |
| tpw_rules | i get a blue pattern on the LEDs until lxterm starts then they shut off | 20:01 |
| bjonnh | (XC7Z020-1CLG400C) | 20:03 |
| bjonnh | I see that there are some XC7Z020 supported | 20:04 |
| bjonnh | I also found that: https://docs.google.com/spreadsheets/d/1q5CUI289TjGw9c66wAesOHlvH8RVo-OK6_yJ4btzZPA/edit#gid=150757850 | 20:08 |
| tpb | Title: litex (liteDRAM, liteVideo, liteEth) Supported boards & Interfaces - Google Tabellen (at docs.google.com) | 20:08 |
| bjonnh | does that mean that the DDR is on the ARM and not accessible directly from the FPGA? | 20:09 |
| tpw_rules | not entirely sure if ethernet isn't working either or i don't have the network configured correctly | 20:25 |
| tpw_rules | how did you get this to work? https://twitter.com/FlUxIuS/status/1461632857814749192 | 20:28 |
| tpw_rules | i did change the RAM because my board only has the 2gbits | 20:29 |
| tpw_rules | is the jtag uart just generally broken? i never got it to work right with my acorn either | 20:30 |
| tpw_rules | the design does instantiate a JTAGG block | 20:47 |
| tpw_rules | so it must be in there | 20:47 |
| _florent_ | tpw_rules: here is how I tested JTAG UART on the butterstick recently: | 20:49 |
| _florent_ | https://github.com/enjoy-digital/litex/pull/1087#issuecomment-952084304 | 20:49 |
| tpw_rules | ah hah, it needs --jtag-chain 49 . how am i supposed to know that? | 20:50 |
| tpw_rules | i guess that's how the hardware works out. it should certainly be documented somewhere | 20:51 |
| _florent_ | We need to document it yes, for now it was in this issue :) | 20:52 |
| _florent_ | for the test over Etherbone, the build command is the one at 12-13s of the video | 20:54 |
| tpw_rules | hmm, it seems the surgery to downgrade the ram did not work properly | 20:54 |
| _florent_ | tpw_rules: I just created https://github.com/enjoy-digital/litex/issues/1120 to look at it later | 20:59 |
| _florent_ | and https://github.com/litex-hub/litex-boards/issues/298 to look at the Butterstick DRAM | 21:03 |
| tpw_rules | _florent_: i replaced this line: https://github.com/litex-hub/litex-boards/blob/bf8b23c19f4299638f75274ee2d346ec99bf7572/litex_boards/targets/gsd_butterstick.py#L114 with MT41K64M16 but it did not work. do i need to change the second parameter too? | 21:18 |
| tpw_rules | the first 2 bytes work, the next 2 bytes don't, and the 2 bytes after that cause the bios to hang | 21:18 |
| tpw_rules | the bios does only say it has 131072KiB | 21:19 |
| tpw_rules | but i would expect to see 256MB | 21:20 |
| tpw_rules | can litex only use one chip? | 21:20 |
| _florent_ | Just changing to MT41K64M16 should be enough | 21:28 |
| tpw_rules | _florent_: https://pastebin.com/66cjLdpS | 21:42 |
| tpb | Title: Warn : gdb services need one or more targets defined __ _ __ - Pastebin.com (at pastebin.com) | 21:42 |
| _florent_ | tpw_rules: and it was OK with the MT41K256M16? | 21:47 |
| _florent_ | Can you try to increase the sys_clk_freq to 75MHz? | 21:48 |
| tpw_rules | idk, i didn't try it with that. my board only has the MT41K64M16 chip | 21:48 |
| tpw_rules | let me try with just the commands you listed | 21:49 |
| _florent_ | ok, I'm just trying to understand, but on my side I was testing with a 8gb version | 21:51 |
| tpw_rules | hm, the problem is with linux-on-litex-vexriscv specifically | 21:59 |
| tpw_rules | (which runs at 60MHz) | 22:03 |
| *** jersey99 <[email protected]> has quit IRC (Quit: Client closed) | 22:04 | |
| tpw_rules | ok, i tried rebuilding the linux-on-litex-vexriscv at 75MHz and it didn't help | 22:14 |
| _florent_ | tpw_rules: you could eventually try to build by adding this: https://github.com/litex-hub/linux-on-litex-vexriscv/blob/master/make.py#L377 (we saw issue with DM support on ECP5, which is used by SMP, using the l2 cache avoid using DM and could solve the issue if this the same thing we saw on OrangeCrab/ECPIX-5) | 22:14 |
| tpw_rules | does the l2 cache size make a difference? | 22:17 |
| tpw_rules | i'm trying it for experiments' sake but unless i am mistaken it is already defined here: https://github.com/litex-hub/litex-boards/blob/bf8b23c19f4299638f75274ee2d346ec99bf7572/litex_boards/targets/gsd_butterstick.py#L115 | 22:22 |
| tpw_rules | huh, sure enough it works with 2048 | 22:24 |
| tpw_rules | but not 8192 | 22:25 |
| tpw_rules | hm, now the --boot is failing with [LXTERM] Got unexpected response from device 'b'E'' | 22:31 |
| *** ewen <[email protected]> has joined #litex | 22:46 | |
| tpw_rules | i can't get tftp to work either but i think that's a network config issue | 22:48 |
| tpw_rules | i see ARP packets in wireshark but i guess the board is not receiving the replies? | 22:53 |
| tpw_rules | _florent_: how do you get bits onto this thing | 23:27 |
| tpw_rules | i resorted to writing all the images to flash then recompiling with flash support | 23:55 |
| tpw_rules | but it works!! https://pastebin.com/sr6k9LC4 | 23:55 |
| tpb | Title: boot 0x40f00000 - Pastebin.com (at pastebin.com) | 23:55 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!