| agneli | gnarface: hello :) are u still here? | 08:14 |
|---|---|---|
| gnarface | agneli: still alive, only partially here | 09:57 |
| agneli | i have issue on devuan with chees | 14:46 |
| agneli | e | 14:46 |
| agneli | my board is armv7 | 14:46 |
| agneli | without neon | 14:46 |
| agneli | which is the baseline for armhf on devuan(debian) | 14:46 |
| agneli | and I am getting "illegel instruction" | 14:46 |
| agneli | :( | 14:46 |
| gnarface | from firefox? | 15:02 |
| gnarface | i seem to recall someone complaining a while back about certain things requiring neon when they weren't supposed to, specifically firefox but maybe some other stuff too | 15:02 |
| agneli | cheese is a video camera app | 15:20 |
| agneli | gnarface: that complainig person was probably me :) :) :) | 15:20 |
| gnarface | agneli: say, could you possibly confirm something for me about the dts code for a "pine64 plus" (AllWinner A64) board? someone has told me i can just add "sd-uhs-sdr104;" to the right spot in the dts file and it will make the SDIO bus go faster. | 16:57 |
| gnarface | do you know if that sounds plausible and safe? | 16:58 |
| gnarface | hmm... nope. doesn't seem to change anything. :( | 17:31 |
| agneli | i guess that depends on the hw capabilities | 18:44 |
| agneli | gnarface: do you know what the the ic that manages sd card interface? | 18:44 |
| agneli | if yes do you have this ic datasheet? | 18:45 |
| gnarface | no :( | 18:45 |
| agneli | perhaps it is the soc directly? | 18:45 |
| gnarface | not sure | 18:45 |
| agneli | for pine all schematics are there | 18:45 |
| agneli | so you can very easily check | 18:45 |
| agneli | no need to guess... like on windows you always do :) | 18:46 |
| agneli | gnarface: which one is yours: https://pine64.org/devices, please? | 18:51 |
| agneli | https://pine64.org/documentation/Pine_A64/Further_information/Schematics_and_certifications/ | 18:52 |
| agneli | here you have everything you can dream of :) | 18:53 |
| agneli | seems there are 2 revisions of the Pine A64+ boards | 18:53 |
| agneli | changing things in dts without careful examination of the schematics seems a bit careless | 18:54 |
| agneli | idk what is t-card but that looks like 4bit sd card interface | 18:59 |
| gnarface | agneli: assuming it's any of these, it would have to be the A64+ 1GB with the Wifi/BT module attached, but i wouldn't claim to be sure how to check | 19:09 |
| gnarface | agneli: it's this one https://pine64.com/product/pine-a64-1gb-board/ | 19:13 |
| gnarface | ...with this attached https://pine64.com/product/pine64-wifi-802-11bgn-bluetooth-4-0-module/ | 19:13 |
| gnarface | the hardware may already be going at max speed, i don't actually know, but it's definitely bottlenecking the wifi | 19:14 |
| agneli | SD physical layer specification V2.0, SDIO card specification V3.0 | 19:15 |
| agneli | this is what the SoC supports | 19:15 |
| agneli | this is SDIO v3.0 http://applelogic.org/files/SDIO.pdf | 19:16 |
| gnarface | does that mean it can go faster than 25Mbps or not? | 19:19 |
| agneli | 90 pages to read :) | 19:20 |
| gnarface | :( | 19:22 |
| agneli | hehe OK gnarface I think I know the person that recommended you the SDR-104 change | 19:23 |
| agneli | so I have another question: how exactly you implemented that change? | 19:23 |
| gnarface | well i had an existing patch i was using that was working but clearly slower than the wifi device itself | 19:24 |
| gnarface | so i just added that one line to it | 19:24 |
| gnarface | but nothing changed | 19:25 |
| agneli | well seems A64 does support SDIO v3 | 19:25 |
| gnarface | literally just this patch: http://paste.debian.net/1307642/ | 19:25 |
| agneli | and acoording to the SDIOv3 docs UHS-I is in the specs | 19:26 |
| gnarface | ... and this was my modified attempt http://paste.debian.net/1307643/ | 19:26 |
| agneli | ok you applied the patch and then what please? | 19:26 |
| agneli | kernel rebuild? reboot? | 19:26 |
| gnarface | yes of course, this is already a kernel i've built several times so far successfully | 19:26 |
| agneli | and you have dtb build into the kernel? | 19:27 |
| gnarface | no, just loading it from the same boot partition, same directory, from the extlinux.conf, but i know it is using my changes because the wifi device wouldn't have worked without it in the first place | 19:28 |
| gnarface | and i checked the byte length of the files when i changed them | 19:29 |
| gnarface | i was pedantic as possible about that part | 19:29 |
| gnarface | the issue at hand though, is this is a patch that i got from a old mailing list archive and it didn't apply to this kernel exactly as i found it. i had to actually edit it by hand to update it for the kernel i'm using, and i don't know if i even did that part exactly right, ... but like i said it worked | 19:30 |
| gnarface | i couldn't get anyone to check it for me | 19:30 |
| gnarface | i found some stackexchange post or something like that that suggested if i enabled debugging, it would let me manually set a speed in /sys, but i didn't try that yet | 19:32 |
| agneli | that last statement is probably correct | 19:33 |
| agneli | but that might depend | 19:34 |
| gnarface | yes, i assume that if my dts is still wrong it just wouldn't work | 19:34 |
| agneli | some clocks can be blocked by some processes | 19:34 |
| gnarface | or may be unstable | 19:34 |
| agneli | I had to write my own procedures to get the wifi working correctly etc | 19:34 |
| agneli | it was also clock related issue | 19:35 |
| agneli | I am not familiar with that board/ chip | 19:36 |
| agneli | I believe you were given good info | 19:36 |
| agneli | iiuc you need to change the voltage to sth lower, no? | 19:38 |
| agneli | and to get your dts checked... well probably somebody working on that board as we speak shoudl do it | 19:39 |
| agneli | otherwise it is a challenge | 19:39 |
| gnarface | heh, yea like i said i asked around but couldn't get anyone who was sure to answer me | 19:40 |
| agneli | there is no group where pople are using this board? | 19:41 |
| agneli | and there is no kernel from pine64 that is already fully enabled and working? | 19:41 |
| agneli | what kernel version are you using, please? | 19:41 |
| gnarface | yea i'm in their irc channels too | 19:50 |
| gnarface | literally nobody would answer me except one guy who said he might look at it then got sick and never got back to me | 19:51 |
| gnarface | i don't think anyone knows to be honest | 19:51 |
| agneli | probably nobody is interested in this board :( | 19:52 |
| gnarface | i think this board was discontinued quickly, the patch never got upstreamed, and everyone moved on to the later revision or the rockchip ones | 19:52 |
| agneli | did you hare your patch upstream for review? | 19:52 |
| agneli | *share | 19:52 |
| gnarface | no, the original creator did, and i found the mailing list archive of it getting refused | 19:53 |
| gnarface | i forget which kernel it was originally for, but i used basic pattern matching skills to update it by hand for kernel 5.18 | 19:54 |
| agneli | so u have kernel 5.10 | 19:54 |
| agneli | ok | 19:54 |
| agneli | pine says mainilne kernel should be ok | 19:54 |
| gnarface | it absolutely was not as of 5.18 | 19:55 |
| gnarface | you think this is fixed in 6.1? | 19:55 |
| agneli | https://pine64.org/devices/pine_a64/ | 19:55 |
| agneli | " Currently the majority of Linux distributions for the PINE A64 are using mainline Linux " | 19:55 |
| agneli | we are now @6.7.4 | 19:55 |
| agneli | http://kernel.org/ | 19:56 |
| agneli | idk maybe just download the sources and analyze the tree? | 19:56 |
| gnarface | yea, but to be honest i don't think any of them were using this SDIO wifi module as of that statement, | 19:56 |
| agneli | before compiling :) | 19:56 |
| agneli | iiuw you were told also that uhs-1 reuires going down to 1.8 V | 19:57 |
| agneli | you need to change proper voltage regulator | 19:57 |
| agneli | u have regulators tree in the schematics | 19:57 |
| gnarface | this board has been booting for much longer than just since 5.18 you see, but without this patch the SDIO port just never worked | 19:58 |
| * agneli recomends you analyse the docs I shared today, understand what needs to be done, then do it :) | 19:58 | |
| gnarface | i was actually told it should as recently as that, and when i demonstrated that was wrong, i found out i was the only one who had actually tried it | 19:58 |
| gnarface | but i do suspect possibly they were just holding out on me | 19:59 |
| agneli | the only thing I can tell you is that I am used to ask questions nbody cares to answer | 19:59 |
| gnarface | noted | 20:00 |
| agneli | so I just go to the docs, source code | 20:00 |
| gnarface | that's going to take some time... | 20:00 |
| agneli | we can always discuss the progress once you start working on it | 20:00 |
| agneli | of course... even years... but well :) | 20:00 |
| agneli | we are not in a hurry are we? | 20:01 |
| gnarface | eh, no not really, i was just hoping to be able to use Steam Remote Play over it, but it's bottlenecking at 2MB/s, which is only almost enough to maintain a stable stream | 20:01 |
| agneli | IMHO you need to understand first to which of the SoCs three controllers your SDIO wifi is connected | 20:02 |
| gnarface | i may just put a USB wifi device in it and be done with it, if i can find one better than this thing | 20:02 |
| agneli | well... I avoid using any wireless connection for any gameplay | 20:02 |
| agneli | this board has 1GB eth | 20:03 |
| agneli | why using crap like wifi? | 20:03 |
| gnarface | Steam Deck has no ethernet port | 20:03 |
| agneli | that is the reason I do not have it :) | 20:03 |
| gnarface | getting a usb-c ethernet device to try on it did also occur to me | 20:04 |
| agneli | I have to go... whatever I can help you I will surely do | 20:04 |
| gnarface | thanks, have a good one | 20:04 |
| agneli | I use wireless connecitons only when I really cannot use cables | 20:04 |
| agneli | like children getting teeth broken because of the cable in the middle of the room... etc ;) | 20:05 |
| agneli | good luck! | 20:05 |
| agneli | and wih usb-c you can go up to 5 GB eth probably if you have the switch | 20:05 |
| gnarface | the switch is only 1GB, but Steam Remote Play doesn't even need 50 Mbit | 20:06 |
| gnarface | it literally just needs a bit more than 25 Mbit, and i'm getting 20 | 20:07 |
| agneli | 20Mbit/s | 20:08 |
| agneli | Man | 20:08 |
| agneli | 20 Mbit/s real throughput? | 20:08 |
| agneli | and what is your AP? | 20:09 |
| agneli | what is the speed you are using on other devices? | 20:09 |
| gnarface | this is the AP | 20:15 |
| gnarface | this thing is what i'm using for the AP | 20:15 |
| gnarface | station dump from the pinebook shows it at like 58/72 Mbit down/up steady during transfers | 20:18 |
| gnarface | but dmesg confirms "high speed" sdio only, which is 25 Mbit/s | 20:18 |
| gnarface | real throughput 16-20 Mbit/s | 20:18 |
| gnarface | sometimes spikes to 25 Mbit/s for a couple seconds at a time | 20:19 |
| gnarface | only if i stand real close though | 20:19 |
| gnarface | so, diagnosis has been that hostapd is configured correctly and the clients are seeing the right link speeds | 20:20 |
| gnarface | something must be wrong with my kernel patch or config | 20:21 |
| gnarface | i've been told that dmesg would change and show that the SDIO was at UHS speed first if it was going to work | 20:24 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!