06:00.35 | *** join/#elinux vstehle (~vstehle@88.121.149.49) |
08:22.18 | *** join/#elinux diego_r (~diego@217-133-17-98.static.clienti.tiscali.it) |
09:17.16 | *** join/#elinux Bunio_FH (~bunio@81-18-201-214.static.chello.pl) |
11:22.39 | *** join/#elinux nsaenz (~nico@190.pool85-52-20.dynamic.orange.es) |
11:25.05 | *** join/#elinux kilovolt[m] (kilovoltma@gateway/shell/matrix.org/x-zacktrldjkuocsrq) |
11:28.54 | *** join/#elinux nsaenz (~nico@190.pool85-52-20.dynamic.orange.es) |
16:08.15 | *** join/#elinux tlwoerner (~Trevor@unaffiliated/tlwoerner) |
16:12.57 | tlwoerner | alan_o: this channel is quiet because embedded linux is so easy now! heck, thanks to yocto, a monkey with a malfunctioning keyboard could still get it right! on the first try even!! ;-) |
16:29.54 | alan_o | Yocto is a great Linux distribution! |
16:36.39 | tlwoerner | finally! someone who understands ;-) |
16:37.31 | L29Ah | a pity they haven't merged back with Gentoo |
16:57.57 | Forty-Bot | does anyone have advice for identifying appropriate device drivers for a (less-than-adequately-documented) SoC? |
16:58.39 | Forty-Bot | e.g. it has several uarts, and I need to figure out if I can use/modify an existing driver based on the register layout |
17:46.59 | tlwoerner | L29Ah: maybe it's gentoo that needs to upgrade? |
17:50.50 | L29Ah | tlwoerner: they both need |
17:51.27 | L29Ah | Gentoo brings lots of packages and extended depspecs, Yocto brings better support for cross-compilation |
21:20.00 | *** join/#elinux m4t (~matt@shenzhen.guangdong.systems) |
21:55.26 | alan_o | L29Ah: meta-debian will give us the best of both worlds |
21:56.01 | alan_o | Forty-Bot: by "identifying" you mean "which existing driver probably drives this hardware?" |
21:56.09 | Forty-Bot | yes |
21:56.32 | alan_o | Forty-Bot: if so, sometimes that's hard because the drivers are based on code names which may or may not be in datasheets and other documentation |
21:56.38 | alan_o | I mean the driver names |
21:56.39 | Forty-Bot | at the moment, I have been going through headers for different drivers looking for appropriate registers |
21:56.50 | alan_o | Forty-Bot: sometimes it's like that |
21:56.53 | Forty-Bot | yeah, there are no names in the official docs |
21:57.11 | Forty-Bot | though they are clearly using some hardware interfaces from related hardware |
21:57.17 | alan_o | sometimes insiders from $company know "these all use the $thing IP block" |
21:57.45 | alan_o | Forty-Bot: also look at device trees for related hardware |
21:57.52 | alan_o | see if anything matches up |
21:58.47 | alan_o | also unrelated hardware. A lot of times IP comes from third parties and different manufacturers will have the same IP blocks, with some level of similarity between the addresses of the registers |
21:59.31 | Forty-Bot | ok, thanks for the tips |
21:59.32 | alan_o | but for a uart, maybe it's not that hard to write one, but depending on your use case, you may have to use DMA to ensure that you don't lose data |
21:59.46 | Forty-Bot | I think I found the appropriate uart driver |
22:00.15 | Forty-Bot | but it was targeted at only one board, so there may be some implicit assumptions I'm going to have to work out |
22:00.42 | alan_o | could be |
22:05.20 | L29Ah | 00:55:26]<alan_o> L29Ah: meta-debian will give us the best of both worlds |
22:05.21 | L29Ah | does it have use flags, /etc/portage/patches/ and -9999 packages? |
22:06.08 | alan_o | It has more than -9999 packages. I don't now the number exactly, but I'm sure it's a positive number. |
22:07.16 | L29Ah | "-9999" means "fetched from vcs and rebuilt if it's updated" |
22:07.43 | L29Ah | s/use/USE/ |
23:20.49 | *** join/#elinux vstehle (~vstehle@88.121.149.49) |