**** BEGIN LOGGING AT Thu Aug 08 02:59:57 2019 Aug 08 06:50:25 xdarklight: That was a RFC pull request on GitHub: https://github.com/openwrt/openwrt/pull/2299 Aug 08 07:32:45 wow, 1Gbit/s on ipq4019. Might be old news, but I never stop being amazed by good performance on relatively low-cost devices Aug 08 08:35:39 gch981213: ping Aug 08 08:36:55 gch981213: looks like ag71xx.c got upstreamed since some time. Checking the code shows quite a few changes and fixes compared to our latest state in ath79. https://elixir.bootlin.com/linux/v5.3-rc3/source/drivers/net/ethernet/atheros/ag71xx.c Aug 08 08:37:38 gch981213: would you agree to send our latest fixes upstream and adapt ath79 to use the upstreamed version then? This way we get fixes for free and are closer to mainline Aug 08 08:38:58 blogic: ^^ (as you were the original author of the ath79 flavour) Aug 08 08:44:05 xback: Well... The author of that driver chose to drop all mii interface setup code and I'm too lazy to compare two drivers and patch everything back now :P Aug 08 08:54:47 gch981213: If I find some sime, i'll give it a try :-) Aug 08 08:54:52 time* Aug 08 09:21:35 gch981213: olerem is right here if you want to talk to them.... Aug 08 09:29:07 gch981213: can you argue that it was wrong choice? Aug 08 09:29:25 karlp: We have talked about this here when that driver was in linux-next and I agree that our implementation isn't quite suitable for upstream. Aug 08 09:34:22 gch981213: sorry, I actualyl meant to tag xback, not you. Aug 08 11:32:49 xback: oh the pengutronix guy Aug 08 11:32:54 oh well Aug 08 11:33:07 works on my machine (TM) let me slap mt (c) on it Aug 08 11:33:13 i remember Aug 08 12:38:15 blogic: so what exactly is your blame? Aug 08 12:38:55 blogic: it was public driver review. you can add your commets everywhere Aug 08 12:39:23 point me a line of code what is wrong... if you can ;) Aug 08 12:41:21 olerem: were all the quirks added ? Aug 08 12:41:27 i did comment that in the pub;lic review Aug 08 12:41:37 and i believe the reply was "lets add them later" Aug 08 12:42:01 * xback seems to accidentally touched a sore spot .. Aug 08 12:42:04 blogic: what exact quirks do you mean? mii code insight of ethernet driver? Aug 08 12:42:40 olerem: does the upstream driver have the same feature set and supports all the different SoCs as the owrt code base that it was derived from ? Aug 08 12:42:57 blogic: don't ask questions. read the code! Aug 08 12:43:22 same attitude that i remembered :-D Aug 08 12:43:59 xback: regardless, we need to figure out whats missing and then retest Aug 08 12:44:19 I've already got a bunch of changes staged in my tree. currently testing them .. Aug 08 12:44:33 xback: right now i'd vote for using the downstream driver until its clear what feature set was dropped Aug 08 12:44:44 unfortunatley we cannot ask theperson that upstreamed the work Aug 08 12:44:54 xback: cool, can you point me at the tree ? Aug 08 12:45:01 i can also test with the HW i have here Aug 08 12:45:16 https://git.openwrt.org/?p=openwrt/staging/xback.git;a=summary Aug 08 12:45:20 cheers Aug 08 12:45:34 blogic: thanks for your 2 cents. appreciated Aug 08 12:45:52 I am dreading the testing Aug 08 12:46:10 it took ages to get it stable across the several 100 Socs/hw configs Aug 08 12:47:01 I can imagine :-) i;ve already spend 3 days trying to clean it up without breaking anything Aug 08 12:47:12 main reason is 2 issues: Aug 08 12:47:48 on my rb922 board, after some time .. ethernet just stops. checking via serial shows the ethernet port suddenly died without any details in logs Aug 08 12:47:52 there are no 100 SoCs. All board variants should not be supported by C-based code. Only devicetree. The mainline driver is supporting alls SoC as the old driver Aug 08 12:48:21 olerem: no need to get involved, we'll read the code Aug 08 12:48:21 thanks Aug 08 12:48:34 xback: yup, i recall there being funky stuff in regards to phys Aug 08 12:48:35 on my rb912, running ath9k and a USB LTE modem together results in a bunch of splats each time ath9k executes a scan. until the LTE modem dies Aug 08 12:48:40 blogic: so the atitude is your side more then my sideß Aug 08 12:48:43 need to see if i can find my notes, been a while Aug 08 12:48:51 the curious part is that it only happens on a rb922 not on a rb912 Aug 08 12:49:10 xback: had the same on some netgear units Aug 08 12:49:40 xback: i probably have the git tree from back then somewhere incl the whole histroy prior to folding it all Aug 08 12:50:20 as the ethernet issue is quite critical here at the company, I've got loads of time to work on this until it gets fixed Aug 08 12:50:36 xback: goodie :-) Aug 08 12:50:55 xback: i am currently fixing a mediatek ethernet issue Aug 08 12:51:16 i'll prbably have some time tomorrow to so how different the drivers are and then we'll see whats next Aug 08 12:51:20 hi all! I just reported on the forum a problem I had using VLAN 802.1ad on cable on a YouHua WR1200JS (MediaTek): https://forum.openwrt.org/t/mediatek-and-vlan-802-1ad-on-ethernet/42346 Aug 08 13:05:04 hi, I like to upload images to the wiki Aug 08 13:05:23 But it says that I do not have enough rights.. can someone help me out here? Aug 08 13:07:37 use the media namepsace Aug 08 13:10:30 ok, using ":media:tplink:tl-cpe210" name space I see a "Select Files.." button Aug 08 13:10:54 but when I click it, nothing happens. No other popup window. Aug 08 13:12:44 site reload does not help Aug 08 13:17:30 If someone feels bored, feel free to upload them https://mwarning.de/misc/tplink_cpe210_v3.tgz Aug 08 13:17:51 hm, works as expected here Aug 08 13:17:56 clicking the button will open a file browser Aug 08 13:18:04 a native one Aug 08 14:22:57 hm Aug 08 14:23:11 jow: thanks for testing Aug 08 14:26:57 hm, I cannot get it to work Aug 08 14:27:42 FF 68.0 Aug 08 15:31:10 maybe some javascriptwitchery? :p Aug 08 15:42:48 ynezz: any news on enabling ath79 on 19.07? Aug 08 16:01:25 there are probably few developers interested in providing ath79 images in 19.07 (and hopefully supporting it as well), so it's possible that there would be images for ath79 in 19.07 Aug 08 17:00:57 ynezz: so why don't you revert your source-only commit. I am supporting it and already having trouble because users can't get images for new devices Aug 08 17:42:02 xback: last I tested ath79 (maybe two weeks ago), ethernet kept disconnecting when I run iperf Aug 08 17:46:31 mangix: you better report a bug and mention the device. I am using lots of devices and run iperf with no problems Aug 08 20:52:28 does anyone know if there's any specific reason why kernel has complete upstream dtc package in the scripts/dtc folder, but uses only dtc? Aug 08 20:52:51 I mean, why it's not possible to build fdtget for example? Aug 08 20:53:47 the patch to make it build in the kernel is not trivial http://paste.ubuntu.com/p/VstrndMk3g/ Aug 08 22:17:48 * ynezz wonders where is ldir when someone wants to buildtest something on the otherOS :p **** ENDING LOGGING AT Fri Aug 09 02:59:57 2019