Sunday, 2021-09-19

lkclcesar: :)10:24
lkclit was about time10:24
Las[m]lkcl: What version of litedram are you using?11:49
lkclLas[m], everything will be around the timestamp of when i was working on ls18012:14
lkclDate:   Fri Aug 7 23:14:09 2020 +020012:14
lkclcommit 198bcbab676e2b4065e5b6a7dc8a7733bae8315a12:14
lkclthe expectation - not explained *at all* - is that should be *altered* (specifically adapted, by you, and maintained as an *out-of-tree fork*)12:17
lkcland the sha1 checksums added12:17
Las[m]lkcl: Yeah I kind of assumed that it was something like that12:19
Las[m]I've gathered the hashes of the commits where the timestamp is <= the timestamp of the tagged commit12:19
Las[m]Practically, this means it's quite hard for an outsider to build Libre-SOC right now, because you'd have to find the right commits for every LiteX repository yourself..;.12:20
lkclLas[m], can you please submit a bugreport because that's really important12:41
lkclput the commits in it12:41
Las[m]lkcl: I think I need an account on Bugzilla?12:43
Las[m]lkcl: Have you tried running lately? I am getting odd errors that seem to be simple mistakes in the program itself, not an issue with my packaging, though I may be wrong of course.17:22
Las[m]Specifically, I get an error inside LiteX at;a=blob;f=litex/build/;h=8e93d19bd277ab55e332e81aa99a979b639c241c;hb=HEAD#l67, because it passes one argument too many.17:23
Las[m]I have a feeling that you're not actually using the tag, because this is fixed in 9950e7 on 2020-08-22...17:42
Las[m]Yep, using a slightly newer LiteX version fixes it...17:49
Las[m]I actually got it to build17:49
Las[m]lkcl: How do you want me to contribute patches to Libre-SOC?17:53
Las[m]Do you do PRs or do you use a patch-based workflow?17:54
lkclLas[m]: "fixed" with this18:27
lkcli just didn't commit it.18:27
lkcli just give people direct access to gitolite3 repos18:28
lkclsend me an ssh public key and i'll add you18:28
lkclLas[m]: i've added you to the wiki repo which you can test-commit to18:35
lkcl*be careful* when ssh'ing to git.libre-soc.org18:35
lkclthe fail2ban rules are quite draconian, to keep port-scanning at bay18:36
lkclwhich is of the order of hundreds of ssh accesses a day18:36
Las[m]Thanks, will do!18:37
Las[m]lkcl: Wasn't it using a non-standard port?19:12
Las[m]Ah nevermind, I just noticed it's on the wiki19:15
Las[m]I can pull at least!19:18
Las[m]Can also commit.19:19
cesarlkcl: The multiplication unit test (src/soc/fu/mul/test/ seems to fail on mulli, mulhd, mulhdu and mulhwu. Is this expected?23:07

Generated by 2.17.1 by Marius Gedminas - find it at!