| - chomwitt (QUIT: Remote host closed the connection) (~chomwitt@2a02:587:dc21:9e00:6dfd:37a:2bc3:ca7e) | 00:26 | |
| + chomwitt (~chomwitt@2a02:587:dc21:9e00:6dfd:37a:2bc3:ca7e) | 00:26 | |
| - GNUmoon2 (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 00:29 | |
| + GNUmoon2 (~GNUmoon@gateway/tor-sasl/gnumoon) | 00:30 | |
| - S0rin (QUIT: Ping timeout: 248 seconds) (~S0rin@user/s0rin) | 00:30 | |
| + S0rin (~S0rin@user/s0rin) | 00:33 | |
| - Christoph_ (QUIT: Remote host closed the connection) (~Christoph@p4fe73f87.dip0.t-ipconnect.de) | 01:02 | |
| - chomwitt (QUIT: Ping timeout: 248 seconds) (~chomwitt@2a02:587:dc21:9e00:6dfd:37a:2bc3:ca7e) | 01:34 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:40) | 02:41 | |
| - Guest447 (QUIT: Ping timeout: 248 seconds) (~nicolas@229-48-142-46.pool.kielnet.net) | 03:46 | |
| + nsc (~nicolas@244-99-142-46.pool.kielnet.net) | 03:47 | |
| * nsc -> Guest6749 | 03:48 | |
| - GNUmoon2 (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 05:18 | |
| + GNUmoon2 (~GNUmoon@gateway/tor-sasl/gnumoon) | 05:18 | |
| - Ar|stote|is (QUIT: Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) (~linx@149-210-16-210.mobile.nym.cosmote.net) | 05:38 | |
| + Ar|stote|is (~linx@149-210-16-210.mobile.nym.cosmote.net) | 05:38 | |
| - littlebobeep (QUIT: Ping timeout: 240 seconds) (~alMalsamo@gateway/tor-sasl/almalsamo) | 06:50 | |
| + littlebobeep (~alMalsamo@gateway/tor-sasl/almalsamo) | 06:51 | |
| - littlebobeep (QUIT: Remote host closed the connection) (~alMalsamo@gateway/tor-sasl/almalsamo) | 08:52 | |
| + littlebobeep (~alMalsamo@gateway/tor-sasl/almalsamo) | 08:58 | |
| + chomwitt (~chomwitt@2a02:587:dc21:9e00:d416:b93a:7a85:a330) | 09:18 | |
| - chomwitt (QUIT: Ping timeout: 248 seconds) (~chomwitt@2a02:587:dc21:9e00:d416:b93a:7a85:a330) | 10:42 | |
| + chomwitt (~chomwitt@2a02:587:dc21:9e00:e28b:dfc7:bbce:95de) | 11:48 | |
| - littlebobeep (QUIT: Remote host closed the connection) (~alMalsamo@gateway/tor-sasl/almalsamo) | 12:20 | |
| + littlebobeep (~alMalsamo@gateway/tor-sasl/almalsamo) | 12:30 | |
| - chomwitt (QUIT: Ping timeout: 252 seconds) (~chomwitt@2a02:587:dc21:9e00:e28b:dfc7:bbce:95de) | 12:55 | |
| - littlebobeep (QUIT: Ping timeout: 240 seconds) (~alMalsamo@gateway/tor-sasl/almalsamo) | 13:16 | |
| + littlebo1eep (~alMalsamo@gateway/tor-sasl/almalsamo) | 13:16 | |
| + chomwitt (~chomwitt@2a02:587:dc21:9e00:e28b:dfc7:bbce:95de) | 13:16 | |
| + Christoph_ (~Christoph@p4fe737fa.dip0.t-ipconnect.de) | 13:25 | |
| * littlebo1eep -> littlebobeep | 13:27 | |
| cinap_lenrek | ok, xhci works | 14:31 |
|---|---|---|
| cinap_lenrek | (when i run usb start in uboot before booting my kernel) | 14:32 |
| cinap_lenrek | time for the display | 14:34 |
| sl | yesss | 15:31 |
| cinap_lenrek | the mipi dsi spec seems to be closed | 15:33 |
| cinap_lenrek | i'd like to read it to at least get some overview | 15:33 |
| cinap_lenrek | there are so many moving parts in the whole display setup | 15:33 |
| cinap_lenrek | dcss -> nwl -> ti-sn65dsi86 | 15:34 |
| cinap_lenrek | the nwl controller has the capability of sending some kind of command packets | 15:35 |
| cinap_lenrek | i'm not sure if that is really needed | 15:35 |
| cinap_lenrek | but i cant know what is needed when i cant even read the dsi spec | 15:35 |
| cinap_lenrek | i bet half the code can be ignored | 15:36 |
| cinap_lenrek | i do not want a generic driver really | 15:36 |
| cinap_lenrek | just do the god damn modeset and get the fuck out | 15:36 |
| cinap_lenrek | i dont care about hdmi | 15:37 |
| cinap_lenrek | people who use hdmi can fuck themselfs | 15:37 |
| cinap_lenrek | yeah | 15:44 |
| cinap_lenrek | i think it is not needed | 15:44 |
| cinap_lenrek | as we'r talking to the bridge and that needs to be configured separately thru i2c | 15:48 |
| cinap_lenrek | and down the bridge it would be all standard vesa stuff | 15:48 |
| + stjohn (~stjohn@cpe-68-174-13-238.nyc.res.rr.com) | 16:16 | |
| sigrid | yay | 16:21 |
| sigrid | cinap_lenrek: you got ethernet working too? | 16:22 |
| cinap_lenrek | last weekend, yes | 16:22 |
| cinap_lenrek | its commited already | 16:22 |
| sigrid | oh nice | 16:22 |
| cinap_lenrek | also multicore | 16:22 |
| cinap_lenrek | its even more complicated | 16:24 |
| cinap_lenrek | we have dcss -> nwl -> dphy -> dsi2edp bridge -> panel | 16:25 |
| cinap_lenrek | poco a poco | 16:25 |
| sigrid | maybe the right way to do mipi dsi is to log which registers linux kernel sets to get shit working. and then do the same within u-boot, doesn't matter if it's completely unreadable bit-fuckery | 16:25 |
| cinap_lenrek | one sais like the software architecture is usually a reflection of the organizations structure | 16:26 |
| cinap_lenrek | this tells me i'll never want to work for nxp | 16:26 |
| cinap_lenrek | sigrid: exactly. | 16:26 |
| cinap_lenrek | sigrid: tho theres no way we can reprogram the dcss from uboot | 16:27 |
| cinap_lenrek | sigrid: but yes, this was the plan | 16:28 |
| cinap_lenrek | tho i bet linux will also be reconfiguring all the clocks and fuck around with gpios and who knows what too | 16:28 |
| cinap_lenrek | the whole display thing has like fuckin 7 different clock signals | 16:28 |
| cinap_lenrek | sigrid: the thing is we need to understand enougth of the relationships between these different parts to be able to troubleshoot it | 16:29 |
| cinap_lenrek | sigrid: because it will definitely not just work | 16:30 |
| cinap_lenrek | thats just completely impossible with so many moving parts | 16:30 |
| sigrid | I see :( | 16:31 |
| cinap_lenrek | plus i dont have a i2c driver nor a clock module driver ;) | 16:31 |
| cinap_lenrek | nor gpio driver | 16:32 |
| minute | cinap_lenrek: alternatively you can use lcdif instead of dcss, it's simpler | 16:35 |
| sigrid | crazy idea: linux, then kexec to 9front | 16:35 |
| minute | cinap_lenrek: there's a mux register that toggles between dcss and lcdif input to nwl-dsi IIRC | 16:35 |
| cinap_lenrek | oh | 16:36 |
| minute | i will also read the dsi spec soon even if it is closed, because i have lots of trouble with dsi and other chips, too | 16:36 |
| cinap_lenrek | minute: thanks for the hint! | 16:37 |
| minute | cinap_lenrek: bluerise also did some more exploratory code work on internal display in uboot, he has a branch somewhere | 16:37 |
| minute | i think sigrid has the link too | 16:37 |
| minute | bluerise theoretically set up everything correctly (using dcss) but it still didn't work (color bars from the sn chip ok, but black screen otherwise) | 16:38 |
| minute | i believe it could be related to delays or the order of things or something with LP vs HS mode on the dsi lines | 16:39 |
| bluerise | sup | 16:50 |
| - littlebobeep (QUIT: Ping timeout: 240 seconds) (~alMalsamo@gateway/tor-sasl/almalsamo) | 17:02 | |
| minute | bluerise: cinap_lenrek is attacking mipi dsi on the imx8m (for plan9) | 17:06 |
| cinap_lenrek | bluerise: do you need traces from linux for mipi dsi and the internal panel? | 17:06 |
| cinap_lenrek | i think thats where i'm going to start | 17:06 |
| cinap_lenrek | i found your repo | 17:07 |
| cinap_lenrek | :) | 17:07 |
| cinap_lenrek | the lcdif approach sounds nice | 17:07 |
| cinap_lenrek | the controller is alot simpler | 17:07 |
| cinap_lenrek | but nobody knows if it works at all | 17:07 |
| cinap_lenrek | maybe later when i'm understanding stuff more | 17:08 |
| cinap_lenrek | minute: my plan is to leave out all that mipidsi shit out of my kernel | 17:08 |
| cinap_lenrek | just give me a linear framebuffer and thats it | 17:08 |
| cinap_lenrek | the kernel is the wrong place for this stuff | 17:09 |
| + littlebobeep (~alMalsamo@gateway/tor-sasl/almalsamo) | 17:14 | |
| - V (QUIT: Ping timeout: 252 seconds) (~v@ircpuzzles/2022/april/winner/V) | 17:31 | |
| + V (~v@ircpuzzles/2022/april/winner/V) | 17:55 | |
| cinap_lenrek | wait a minute | 17:56 |
| cinap_lenrek | why's the linux kernel 486MB | 17:56 |
| erle | bloat | 18:00 |
| cinap_lenrek | ah | 18:00 |
| erle | what was it? | 18:00 |
| - V (QUIT: Quit: We're here. We're queer. Connection reset by peer) (~v@ircpuzzles/2022/april/winner/V) | 18:02 | |
| + V (~v@ircpuzzles/2022/april/winner/V) | 18:03 | |
| cinap_lenrek | wrong file | 18:15 |
| cinap_lenrek | the Image file is *ONLY* 32MB | 18:16 |
| cinap_lenrek | it is so rocket science to just load a kernel now | 18:17 |
| cinap_lenrek | i cant paste in the kernel as uboot looses the characters | 18:17 |
| cinap_lenrek | so i wanted to make a environment variable for the load lines | 18:18 |
| cinap_lenrek | just fetching the Image from tftp | 18:18 |
| cinap_lenrek | and the rest from mmc as the boot script | 18:18 |
| cinap_lenrek | but i cant remember how to escape the commands | 18:18 |
| cinap_lenrek | i tried: | 18:18 |
| cinap_lenrek | setenv bla ,.... '&&' .... | 18:18 |
| cinap_lenrek | printenv LOOKS good | 18:18 |
| cinap_lenrek | but apparently its not right | 18:18 |
| cinap_lenrek | anyone remmeber how to escape ; or && for setenv? | 18:19 |
| cinap_lenrek | got it | 18:22 |
| cinap_lenrek | run envname | 18:23 |
| cinap_lenrek | here we go | 18:24 |
| + mtm (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 18:57 | |
| cinap_lenrek | hm | 19:45 |
| cinap_lenrek | apparently theres already some infra for register tracing now | 19:45 |
| cinap_lenrek | with the regmap stuff | 19:45 |
| cinap_lenrek | but i cant get it to work | 19:45 |
| cinap_lenrek | defined my regmap_should_log function but nothing is logged | 19:45 |
| cinap_lenrek | i'm not really sure anymorw how you can pass linux arguments to the kernel | 19:46 |
| cinap_lenrek | i just set bootargs environment variable but it seems to have no effect | 19:46 |
| cinap_lenrek | i set loglevel=8 | 19:46 |
| cinap_lenrek | but no change | 19:46 |
| cinap_lenrek | any linux experts here? | 19:47 |
| _E | cinap_lenrek: depends on the definition of "expert" | 20:00 |
| cinap_lenrek | afaik, devicetree blob passed to the kernel would contain the boot command line, no? | 20:00 |
| cinap_lenrek | or maybe the whole concept of kernel command line doesnt exist in uboot anymore? | 20:01 |
| _E | aren't dtb's compiled in advance? | 20:02 |
| cinap_lenrek | they get patched | 20:02 |
| _E | ah | 20:02 |
| cinap_lenrek | it used to be all simple | 20:02 |
| cinap_lenrek | with atags on arm | 20:02 |
| cinap_lenrek | but my question is how i can pass the linux "loglevel=8" argument to the kernel | 20:03 |
| cinap_lenrek | iirc it would be just: | 20:03 |
| cinap_lenrek | setenv bootargs ..... | 20:03 |
| cinap_lenrek | before booti, no? | 20:03 |
| _E | well, that's how it worked for me | 20:04 |
| cinap_lenrek | maybe loglevel= is wrong | 20:04 |
| _E | I was doing setenv bootargs "$bootargs (whatever else)" | 20:04 |
| _E | but if you need to rewrite existing params, your option is what I would have ended up doing | 20:05 |
| cinap_lenrek | [?2004hroot@reform:/proc# cat cmdline | 20:12 |
| cinap_lenrek | [?2004lloglevel=8 | 20:12 |
| cinap_lenrek | it works | 20:12 |
| cinap_lenrek | its the regmap tracing that doesnt work | 20:12 |
| cinap_lenrek | i'll need to dig | 20:12 |
| cinap_lenrek | got it | 20:40 |
| cinap_lenrek | dev_name() is just some devietree argate | 20:40 |
| cinap_lenrek | garbage* | 20:40 |
| cinap_lenrek | i need dev_driver_string() to match | 20:40 |
| cinap_lenrek | http://felloff.net/usr/cinap_lenrek/mipi-dsi-trace1.txt | 20:46 |
| cinap_lenrek | the dcss is not using the regmap framework | 20:46 |
| cinap_lenrek | so its not appearing in the trace | 20:46 |
| cinap_lenrek | i'll hack it and provide logs | 20:46 |
| cinap_lenrek | http://felloff.net/usr/cinap_lenrek/mipi-dsi-trace2.txt | 21:16 |
| cinap_lenrek | its still virtual addresses | 21:16 |
| cinap_lenrek | theres a AT instruction to get physical addresses | 21:16 |
| cinap_lenrek | but i can't remember the gcc inline assembly syntax for it | 21:17 |
| cinap_lenrek | works. | 21:48 |
| cinap_lenrek | http://felloff.net/usr/cinap_lenrek/mipi-dsi-trace3.txt | 22:01 |
| cinap_lenrek | that should be all relevant registers | 22:01 |
| cinap_lenrek | and i checked that the display does indeed, turn on :) | 22:02 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 22:56 | |
| cinap_lenrek | http://felloff.net/usr/cinap_lenrek/mipi-dsi-trace4.txt | 23:11 |
| cinap_lenrek | tracing the registers for ctxld too | 23:11 |
| cinap_lenrek | syntax is <ctxid>/<addr> | 23:12 |
| cinap_lenrek | bluerise, sigrid: http://felloff.net/usr/cinap_lenrek/mipi-dsi-trace4.txt | 23:47 |
| cinap_lenrek | heres your traces from linux modeset | 23:47 |
| cinap_lenrek | when a register prints as x/ it means it gets queued to the ctx loader | 23:48 |
| cinap_lenrek | if anything, it helps seeing the order in which linux does stuff | 23:49 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!