| 08:05 | Coburn | Where do I find the latest kernel for the ClearFog? |
| 09:12 | Ke | why is this channel not called solid-run |
| 09:12 | Ke | about half of the discussion is not about cubox |
| 09:14 | Ke | Coburn: would mainline kernel work on the board? |
| 09:15 | vpeter | Ke: Probably Armbian is using mainline + some patches. |
| 09:16 | Coburn | ok |
| 09:16 | Coburn | I'm gonna vent a little and say that it seems that SR doesn't update the wiki |
| 09:17 | Ke | Coburn: not sure about Clearfog, but in general not having software support may be a good sign |
| 09:17 | Coburn | a lot of the stuff on that is outdated or has typos, like the kernel page for the a38x page is still 3.10.x |
| 09:18 | Ke | I would definitely hope that concept of vendor SDK would go away |
| 09:18 | Coburn | Not having software support is a good sign? |
| 09:18 | Coburn | What do you mean by that? |
| 09:18 | Ke | because it's just a sign of not committing to upstream support |
| 09:18 | Coburn | yeah |
| 09:18 | Coburn | I mean I run LEDE on my Clearfog right now |
| 09:18 | Coburn | but there's a bug in the 4.4.61 kernel that causes warnings to be emitted from the itlwifi driver |
| 09:19 | Coburn | then hostapd crashes when it tries to init the adapter to set up a wifi network |
| 09:19 | Ke | did you try 4.9.30? |
| 09:19 | Coburn | So I'm trying to roll my own image, but it's like "oh you gotta do it this way" |
| 09:19 | Coburn | I haven't yet, because I can't find the sources for 4.9.30 |
| 09:20 | Ke | https://cdn.kernel.org/pub/linux/kernel/v4.x/linux-4.9.30.tar.xz |
| 09:20 | Coburn | I did pull something from solid-runs lede git, dunno what that was |
| 09:20 | Coburn | does that mean i need to update the dtb blobs? |
| 09:20 | vpeter | Try armbian with 4.11? :) |
| 09:20 | Ke | dtbs should be backwards compatiblew |
| 09:20 | Coburn | This is for router use. The clearfog is acting as my router to my internet. |
| 09:20 | Ke | but updating dtb may be better option |
| 09:21 | Coburn | how do I do that? |
| 09:21 | Coburn | I brought it through my local (australian) contact because I wanted a board with 8 ports |
| 09:21 | Coburn | Almost makes me feel like GlobalScale all over again |
| 09:23 | Ke | hmm if you are not comfortable with mastering your u-boot, you should not probably listen to anything I say |
| 09:24 | Ke | I am assuming you can try out these things at almost 0 cost |
| 09:25 | Coburn | No, i've mastered uboot |
| 09:25 | Coburn | well |
| 09:25 | Coburn | i wrote my own bootcmd |
| 09:25 | Coburn | and know how to print, save and reboot :P |
| 09:34 | Ke | Coburn: you can eg. easily load kernels, dtb and initrds over tftp for testing |
| 09:34 | Ke | with almost 0 chance of bricking |
| 09:34 | Coburn | oh |
| 09:34 | Coburn | nope |
| 09:34 | Coburn | I just use microSD with the known good kernel backup as "emergency_kernel.img" |
| 12:13 | Artox | oh Coburn left :( |
| 12:13 | Artox | anyway, in case somebody still wonders |
| 12:13 | Artox | the Marvell Kernel for clearfog is 3.10. |
| 12:13 | Artox | But mainline works reasonably well today |
| 12:14 | Artox | and already on 4.9 |
| 12:14 | Artox | the solidrun lede is a fork where we add stuff when it is not yet in lede, but really important to get boards working |
| 12:35 | Bluerise | Hm |
| 12:35 | Bluerise | anyone with a macchiatobin and jtag? |
| 12:35 | Bluerise | rabeeh: are you using openocd? |
| 13:08 | Bluerise | afaerber: oh, you're that guy from openocd? |
| 13:21 | afaerber | Bluerise, I'm one contributor, yes |
| 13:22 | Bluerise | heh. you don't happen to know if I explicitly need to give the cti base or dbg base to openocd? |
| 13:22 | Bluerise | I'm trying to attach to a cortex a72.. |
| 13:23 | afaerbe | 13:23 * afaerber hasn't debugged aarch64 yet since it got merged |
| 13:23 | afaerber | in general if there is more than one CPU I think the bases need to be specified |
| 13:24 | afaerber | dap info X should hopefully give you some hints |
| 13:24 | afaerber | MacchiatoBin? |
| 13:24 | Bluerise | yes |
| 13:25 | Bluerise | currently getting.. |
| 13:25 | Bluerise | Error: Invalid ACK (5) in DAP response |
| 13:25 | Bluerise | Error: Debug regions are unpowered, an unexpected reset might have happened |
| 13:25 | Bluerise | Error: JTAG-DP STICKY ERROR |
| 13:25 | Bluerise | :/ |
| 13:25 | Bluerise | openocd -c "adapter_khz 1000" --file interface/jlink.cfg -c "transport select jtag" -c "jtag newtap auto0 tap -irlen 24 -expected-id 0x18040357" -c "target create mvl aarch64 -chain-position auto0.tap" |
| 13:26 | Bluerise | I'm a bit surprised about the irlen, ususally I only see values from 2 to 4, but those then produce "unexpected..." |
| 13:26 | Bluerise | Info : JTAG tap: auto0.tap tap/device found: 0x18040357 (mfg: 0x1ab (Purple Ray), part: 0x8040, ver: 0x1) |
| 13:26 | Bluerise | Error: IR capture error at bit 4, saw 0x05 not 0x...3 |
| 13:26 | Bluerise | like this, but with 24 there's no IR capture error anymore |
| 13:26 | afaerber | I need to update my firmware, so maybe I can play with it the weekend |
| 13:27 | afaerber | Purple Ray sounds weird :) |
| 13:27 | Bluerise | hah |
| 13:27 | Bluerise | I was surprised as well, but not sure if that's a gadget or a bug ;) |
| 13:37 | Bluerise | afaerber: trying to find out why the machine hangs. interrupts just... stop |
| 13:38 | Bluerise | like, clock doesn't tick anymore |
| 13:38 | Bluerise | of course, no interrupts, no console, no debugger, nothing |
| 13:38 | Bluerise | so I hoped jtag would help |
| 13:39 | afaerber | there were memory corruption issues, supposedly fixed by a recent firmware commit |
| 13:49 | vpeter | Artox: Why there is no announcements regarding such projects as LEDE is? There is some work done and user must finds out it. And be clever enough to know in what shape it is. |
| 13:50 | vpeter | Currently I'm using LEDE from your repository (build more than 1 year back). And now I see there is new repo available. Is solidrun makes things to find even harder? |
| 16:26 | Bluerise | afaerber: Oh? which firmware, atf? or IO? and how to update? :) |
| 16:30 | afaerber | Bluerise, https://github.com/MarvellEmbeddedProcessors/edk2-open-platform/commit/e9931e946ed8c257f7dbb86bfa56bae3f5886a5c |
| 18:35 | vpeter | Lightning yesterday killed my cable modem and seems wan port on my clearfog pro too. Even if was connected over belkin surge protector. |
| 18:35 | vpeter | Is there any way to identify how far damage come? There is no dmesg activity when modem/computer is connected to it. Lan ports and connected sata disk is working fine. |
| 18:55 | vpeter | Wondering if there is a possibility to remap this port. |
| 20:25 | vpeter | Maybe using sfp transceiver. |
| 20:30 | vpeter | rabeeh: Is there any very cheap one which would work? |
| 20:49 | Artox | vpeter, are we? |
| 20:50 | Artox | maybe |
| 20:50 | Artox | that old thing was based on a 4.4 kernel |
| 20:50 | Artox | adn a pain to maintain (which is why it never saw updates) |
| 20:50 | Artox | so for the past months we pushed to get the necessary pieces merged |
| 20:50 | Artox | (it really did take months) |
| 20:51 | Artox | I´ll ask around and see about publishing such news somewhere central |
| 20:51 | Artox | (the last lede stuff was merged may 30 btw) |
| 20:51 | Artox | so starting from that point lede master should give a good experience |
| 20:52 | Artox | our goal is the next stable release |
| 20:52 | Artox | that one might see some more love |
| 21:05 | vpeter | Artox: what I would like to tell is that such devices are usable only to people who really understands all the details about the hw and sw. And how to build it. Like switching pcie to msata. Image from Lede master probably doesn't cover this. Which means someone needs to figure out how to make final image to boot on his device. Which require at least build new u-boot. Not having good instructions means try & error. |
| 21:07 | vpeter | Hope I fix my clearfog somehow to actually make this new version. Current image served very well. |
| 22:20 | Bluerise | afaerber: ah, I got that one |
| 22:20 | Bluerise | do serdes SFPs work on the clearfog? |