**** BEGIN LOGGING AT Sun Nov 11 03:00:00 2018 Nov 11 04:44:21 Borromini: alright... thanks for explanation. Nov 11 04:59:18 Does anybody have suggestions for this problem (proper ways of switching between AP/IoT mode on mt7628)? Those undocumented magic numbers are quite annoying. https://github.com/openwrt/openwrt/pull/1500#issuecomment-437643835 Nov 11 05:38:05 gch981213[w]: Beyond somehow convincing Mediatek that actual documentation is required, not a few highlights? Nov 11 05:55:15 gch981213: I'm not sure you're correct that bit 15 of GPIO1_MODE is relevant - it seems as likely to be a typo as anything else. Nov 11 05:55:33 I'm sure all the magic is in the totally undocumented pin mapping registers. Nov 11 06:14:10 gch981213: I really can't quite puzzle out what they're doing with those registers, though - the fact they're shifting by the GPIO number is dying to tell me they have a pair of 32-bit registers per function to select which GPIO pin a function is routed to, but the three registers seen doesn't seem right. Nov 11 06:15:30 The first two sort of make sense, if 0x1350 is SD_CLK for the lower 32 GPIOs and 0x1360 is some other function, but then what's 0x1340 got to do with it? Nov 11 06:24:10 Monkeh: Thank you all the same :) Nov 11 06:27:32 gch981213: I'll bet it's an off-the-shelf option from Palmchip whatever it is. ;) Nov 11 06:31:39 mmk Nov 11 06:31:48 so I downgraded to 17 and its been fine since :/ Nov 11 06:31:56 I'll have to get another AP if I'm gonna debug this Nov 11 06:32:32 lynxis: ping - reproducibility question Nov 11 06:35:55 ldir: what on earth are you doing up so early on a Sunday? :) Nov 11 06:36:31 Tianjin Quanjian FC v Shanghai SIPG FC for Sky Nov 11 06:36:36 ahhh Nov 11 06:37:04 later will be Chelsea v Everton Nov 11 06:37:16 oh I've heard of those :D Nov 11 06:37:25 later than that will be Sporting Kansas City v Real Salt Lake Nov 11 06:37:44 busy busy huh Nov 11 06:37:48 and even later than that Atlanta United FC v New York City FC Nov 11 06:38:05 at that point I'll probably just collapse in a heap Nov 11 06:38:46 yup, that is a long day Nov 11 06:39:52 Home probably 0200 *tomorrow* after getting up 0420 this morning - urgh Nov 11 06:40:19 :( Nov 11 06:40:20 sucks Nov 11 06:41:01 anyway I was trying to find cases/proof of what I remember that some packages had parallel building disabled for reproducibility reasons and lynxis is the repro king :-) Nov 11 06:41:16 heh Nov 11 06:49:03 hey fix sucky ath9k pls :d Nov 11 06:49:27 the one bit of wireless equipment I use is busted on 4.14/18.06.1 Nov 11 06:49:33 should probably try master Nov 11 07:12:08 blogic: Any chance you can weigh in on the mystery registers? Nov 11 07:24:54 ? Nov 11 07:25:04 what register ? Nov 11 07:34:22 blogic: SYSCTL_BASE + 0x1340 (and 1350 and 1360) on MT7628 - I don't know what the source of the code is, but there are snippets around suggesting pin remapping (specifically for the SD interface) Nov 11 07:34:29 Wondered if you might have some inside knowledge on that Nov 11 07:39:24 Monkeh: http://community.onion.io/topic/1266/how-to-switch-agpio-gpio-18-19-omega2-fw-vs-lede-build/10 Nov 11 07:40:00 not exactly the same, but related Nov 11 07:40:10 DonkeyHotei: Yes, exactly not that. Nov 11 09:32:43 lol, finally figured out that rpi needs userspace actions to setup bluetooth Nov 11 10:32:28 man that bluetooth is a pita :) Nov 11 10:33:37 ? Nov 11 10:35:23 stintel: there's a non-working patch from Marcel Holtmann on the linux-bluetooth mailing list. depending on how much time you have and in which mood you are you might want to try that Nov 11 10:35:31 (try = try and fix) Nov 11 10:36:58 xdarklight: for rpi ? Nov 11 10:38:07 stintel: for rpi3 to be precise Nov 11 10:39:06 xdarklight: you happen to have a direct link, or a subject I can google for ? Nov 11 10:39:13 give me a sec Nov 11 10:39:54 thanks Nov 11 10:41:28 gonna finish the 4.14 support first, it's needed anyway iiuc (serdev) Nov 11 10:41:38 so I can send it out to the ML Nov 11 10:42:01 I can't find the mailing list link, but you can grab the patches here: https://github.com/jeremycline/linux/commits/rtl8723bs-3wire-working - best way is to check the history here: https://github.com/jeremycline/linux/commits/3eb8d755c63bae4dfdefedc12373b5a30726052b/drivers/bluetooth/bt3wire.c Nov 11 10:43:05 (you may want to look at Jeremy's patches since he's aiming for RTL8723BS support in the bt3wire driver which can be ignored for rpi3 since that's Broadcom based) Nov 11 10:44:12 oh it's a complete new driver Nov 11 10:44:40 that's pretty much over my head :( Nov 11 11:11:53 HDMI CEC module, does it make sense to add in package/kernel/linux/modules/video.mk ? Nov 11 11:18:36 or maybe it makes more sense to enable it in the kernel config of targets that support CEC Nov 11 11:39:05 hmmmz Nov 11 12:55:36 ldir: it should be in the git log if I disabled parallel for it Nov 11 13:14:52 CEC is going to be for another day Nov 11 13:18:25 and so is bluetooth :) Nov 11 14:11:25 I am starting to understand why brcm2708 never got 4.14 support :D Nov 11 14:53:30 Hi all. Got a question about an Archer A7. Accidentally bought this router thinking it was supported bc they reused the model number (AC1750) for both the C7 (supported) and A7 (not yet supported). Nov 11 14:53:52 At any rate I got this pull request: https://github.com/openwrt/openwrt/pull/1462 Nov 11 14:54:59 I set up a ubunutu VM, selected A7 as the target profile, and built it Nov 11 14:55:55 But I don't see lede-ar71xx-generic-archer-a7-v5-squashfs-factory.bin in bin/targets/ar71xx/generic/ Nov 11 14:56:10 Here are the contents of that folder: Nov 11 14:56:40 wasd@wasd-VirtualBox:~/openwrt$ ls -l bin/targets/ar71xx/generic/ Nov 11 14:56:40 total 17004 Nov 11 14:56:40 -rw-r--r-- 1 wasd wasd 105 Nov 10 16:58 config.seed Nov 11 14:56:40 -rw-r--r-- 1 wasd wasd 2182 Nov 10 17:16 openwrt-ar71xx-generic-device-archer-a7-v5.manifest Nov 11 14:56:40 -rw-r--r-- 1 wasd wasd 2490368 Nov 10 17:16 openwrt-ar71xx-generic-root.squashfs Nov 11 14:56:41 -rw-r--r-- 1 wasd wasd 1574771 Nov 10 17:16 openwrt-ar71xx-generic-uImage-lzma.bin Nov 11 14:56:42 -rwxr-xr-x 1 wasd wasd 5048020 Nov 10 17:16 openwrt-ar71xx-generic-vmlinux.bin Nov 11 14:56:44 -rwxr-xr-x 1 wasd wasd 5053144 Nov 10 17:16 openwrt-ar71xx-generic-vmlinux.elf Nov 11 14:56:46 -rw-r--r-- 1 wasd wasd 1638400 Nov 10 17:16 openwrt-ar71xx-generic-vmlinux.lzma Nov 11 14:56:48 -rwxr-xr-x 1 wasd wasd 1584892 Nov 10 17:16 openwrt-ar71xx-generic-vmlinux-lzma.elf Nov 11 14:56:50 drwxr-xr-x 2 wasd wasd 4096 Nov 10 17:16 packages Nov 11 14:56:52 -rw-r--r-- 1 wasd wasd 814 Nov 10 17:16 sha256sums Nov 11 14:56:54 wasd@wasd-VirtualBox:~/openwrt$ Nov 11 15:03:42 Guvnarrr: build with V=s and you'll see what goes wrong Nov 11 15:04:02 the build finished successfully though Nov 11 15:04:30 well it won't error out by default if your image is too big e.g. Nov 11 15:04:34 but the image won't be created Nov 11 15:05:37 Ok issuing another build Nov 11 15:11:29 perhaps you nailed it? Nov 11 15:11:30 WARNING: Image file /home/wasd/openwrt/build_dir/target-mips_24kc_musl/linux-ar71xx_generic/tmp/openwrt-ar71xx-generic-archer-a7-v5-squashfs-sysupgrade.bin is too big Nov 11 15:14:47 Borromini: I appreciate the tip but I gotta say that I'm not sure where to go from here Nov 11 15:16:48 Guvnarrr: You should disable some packages in menuconfig, and the image will be smaller Nov 11 15:17:37 That's where I was headed mentally. Any rather large ones that are rarely used you would recommend? Nov 11 15:18:14 If you did make V=s, you can find the size of the files in the output Nov 11 15:19:23 Guvnarrr: sounds like you're including everything and the kitchensink ;-) Nov 11 15:19:38 luci is the first one that come to my mind. Assuming you can configure the device without the web interface Nov 11 15:20:07 luaraneda: he has 16 MiB of flash if wikidevi is right, i think he has other issues than LuCI Nov 11 15:20:24 Guvnarrr: pastebin you ./scripts/diffconfig.sh output Nov 11 15:20:28 s/you/your/ Nov 11 15:23:11 https://pastebin.com/9fs3fWf8 Nov 11 15:23:15 Borromini: Oh, I thought the device has 4 MiB or 8 MiB Nov 11 15:24:03 luaraneda: yeah, that's what we tend to think when we see those errors pop up. hehe Nov 11 15:24:18 Guvnarrr: He meant the execution of that script Nov 11 15:24:28 Guvnarrr: don't pastebin the contents of that script please. Nov 11 15:24:34 pastebin the *output* of that script Nov 11 15:24:34 My only alteration to my openwrt checkout (aside from the pull request) was picking the Archer A7 Nov 11 15:24:44 ok sorry o it Nov 11 15:24:55 $ ./scripts/diffconfig.sh Nov 11 15:24:55 Guvnarrr: execute: Nov 11 15:24:56 make defconfig Nov 11 15:24:56 ./scripts/diffconfig.sh Nov 11 15:24:58 is how you run it Nov 11 15:25:04 yep Nov 11 15:25:19 https://pastebin.com/JvEP5P5u Nov 11 15:25:26 he probably won't even have luci if he's building off master Nov 11 15:25:40 nope nothing special there... weird Nov 11 15:25:44 is the default loaded with packages? Nov 11 15:25:56 Guvnarrr: how do you mean? Nov 11 15:26:04 the kitchen sink Nov 11 15:26:08 the default is pretty trimmed but does contain everything necessary :) Nov 11 15:26:23 no, the kitchen sink is what i thought you had added, because it broke Nov 11 15:26:23 hmmmm Nov 11 15:26:28 Yes, it should be enough Nov 11 15:26:39 it almost seems like the image size is set too low... Nov 11 15:27:06 any clues in what files were touched in the pull request? Nov 11 15:27:46 yes https://github.com/openwrt/openwrt/pull/1462/files Nov 11 15:28:15 Sorry, I mean was the size set lower in those changes Nov 11 15:28:29 From the github PR: Nov 11 15:28:29 xback (18 days ago): +1 for adding dynamic partitioning. I'm currently in the process of adapting all ar71xx targets to use this, as the os-image partition size of 0x180000 is too small and causes build errors using kernel 4.14 Nov 11 15:28:45 the PR sets IMAGE_SIZE := 15104k Nov 11 15:28:58 https://github.com/openwrt/openwrt/pull/1462/files Nov 11 15:29:03 so that sounds allright Nov 11 15:29:22 but maybe like luaraneda says it might be the 4.14 kernel playing up Nov 11 15:29:36 the rootfs probably isn't the problem. Nov 11 15:31:12 Guvnarrr: What if you try the patch sent to the ML instead the PR? Nov 11 15:31:12 http://lists.infradead.org/pipermail/openwrt-devel/2018-November/014559.html Nov 11 15:31:18 sorry that was the wrong link. https://github.com/openwrt/openwrt/pull/1462/files#diff-e2dd41c6c2f95fe5c5d79b932f3c33aa is the right one Nov 11 15:32:21 luaraneda: yeah that might be a better idea, that one is more recent by a month :) Nov 11 15:32:27 Deciphering... try the patch send to ML(?) instead of the PullRequest? Nov 11 15:32:40 no someone else sent a patch to the mailing list Nov 11 15:32:42 Mailing List Nov 11 15:33:11 Guvnarrr: sec, i'll link you to a patch you can easily apply. just make sure to reset your git checkout to defaults first. Nov 11 15:33:19 git reset --hard Nov 11 15:33:36 ok basically get off the branch with the PR and apply a diff Nov 11 15:34:09 sorry im an embedded sw engineer but we use SVN. Git is new to me as is the lingo Nov 11 15:34:21 did you just pull in vanilla openwrt and apply the PR? Nov 11 15:34:24 or something else? Nov 11 15:34:37 (git is in the horizon soon I hope :) ) Nov 11 15:34:42 hehe Nov 11 15:35:00 i like git, but it took me some getting used to as well (i know more about git now than ever I knew about SVN) Nov 11 15:35:30 download this: https://patchwork.ozlabs.org/patch/995935/mbox/ and rename it to a7-support.patch or sth Nov 11 15:35:33 https://pastebin.com/FEzR3qED Nov 11 15:35:43 ^ thats how i created my build Nov 11 15:35:57 ok Nov 11 15:36:18 what does git branch say Nov 11 15:36:27 does it show master? if so, then do $ git checkout master Nov 11 15:36:40 wasd@wasd-VirtualBox:~/openwrt$ git branch Nov 11 15:36:40 * (HEAD detached at FETCH_HEAD) Nov 11 15:36:40 master Nov 11 15:36:58 (sorry if that should go into pastebin, not sure what the rull of thumb is) Nov 11 15:37:04 no that's ok Nov 11 15:37:26 i havent run reset --hard Nov 11 15:37:38 that's no problem, since you weren't on master in the first place Nov 11 15:38:07 (fwiw, i didn't know you could just grab a PR with git fetch origin :P) Nov 11 15:38:42 when you're on master and have grabbed the patchwork patch, do $ git apply a7-support.patch Nov 11 15:38:55 that should apply cleanly, then you can rerun make menuconfig Nov 11 15:39:51 I would advice to delete the .config file first (just to be sure) Nov 11 15:40:02 and the .config.old Nov 11 15:40:52 To get back to master.... git checkout master Nov 11 15:41:04 ? Nov 11 15:43:46 yes Nov 11 15:45:02 where is .config.old? Nov 11 15:45:41 actually there are a few Nov 11 15:45:42 wasd@wasd-VirtualBox:~/openwrt$ find . -name .config.old Nov 11 15:45:42 ./tmp/.config.old Nov 11 15:45:42 ./build_dir/target-mips_24kc_musl/busybox-1.29.3/.config.old Nov 11 15:45:42 ./build_dir/target-mips_24kc_musl/linux-ar71xx_generic/backports-4.19-rc5-1/.config.old Nov 11 15:45:42 ./build_dir/target-mips_24kc_musl/linux-ar71xx_generic/linux-4.14.79/.config.old Nov 11 15:45:44 wasd@wasd-VirtualBox:~/openwrt$ Nov 11 15:46:51 the one that maters is on the top directory Nov 11 15:47:35 /home/wasd/openwrt/.config Nov 11 15:47:36 /home/wasd/openwrt/.config.old Nov 11 15:47:36 (in your case) Nov 11 15:47:55 don't worry if they don't exist Nov 11 15:47:55 I deleted .config in the base dir, no .config.old in the base Nov 11 15:48:01 applying patch Nov 11 15:49:10 and? Nov 11 15:49:22 no news is good news i assume :-D Nov 11 15:49:36 https://pastebin.com/ZvW1geyA Nov 11 15:49:59 Im just keeping you guys updated so I dont make a small disastrous error Nov 11 15:50:04 that's ok, not to worry. patch applied cleanly i presue Nov 11 15:50:08 * presume Nov 11 15:50:12 yes Nov 11 15:50:24 running make menuconfig Nov 11 15:50:25 good Nov 11 15:50:59 hmmm unexpectedly the target profile is already A7 Nov 11 15:51:35 should i delete the .config.old from tmp? Nov 11 15:52:11 no need. Just run a "make defconfig" after you close menuconfig Nov 11 15:52:15 ^^ Nov 11 15:52:21 oh nice! Nov 11 15:52:48 done reentered config Nov 11 15:53:09 default profile ---> changing to A7 Nov 11 15:54:20 hmmm there is no A7 for TP-LINK Nov 11 15:54:33 just Archer C7 Nov 11 15:56:09 According to the patch, the device should be named "TP-LINK Archer A7 v5" Nov 11 15:56:47 (from target/linux/ath79/image/generic-tp-link.mk) Nov 11 15:57:30 don't forget that ar71xx and ath79 are separate Nov 11 15:57:52 oh thats where i messed up i bet Nov 11 15:58:43 yep that was my issue (my bad) Nov 11 15:59:26 building Nov 11 16:13:24 wasd@wasd-VirtualBox:~/openwrt$ ll -al bin/targets/ath79/generic/ Nov 11 16:13:24 total 11676 Nov 11 16:13:24 drwxr-xr-x 3 wasd wasd 4096 Nov 11 09:12 ./ Nov 11 16:13:24 drwxr-xr-x 3 wasd wasd 4096 Nov 11 09:03 ../ Nov 11 16:13:24 -rw-r--r-- 1 wasd wasd 15592 Nov 11 09:03 config.seed Nov 11 16:13:25 -rw-r--r-- 1 wasd wasd 2305 Nov 11 09:12 openwrt-ath79-generic-device-tplink-archer-a7-v5.manifest Nov 11 16:13:28 -rw-r--r-- 1 wasd wasd 3813884 Nov 11 09:12 openwrt-ath79-generic-tplink_archer-a7-v5-initramfs-kernel.bin Nov 11 16:13:31 -rw-r--r-- 1 wasd wasd 4032870 Nov 11 09:12 openwrt-ath79-generic-tplink_archer-a7-v5-squashfs-factory.bin Nov 11 16:13:34 -rw-r--r-- 1 wasd wasd 4063948 Nov 11 09:12 openwrt-ath79-generic-tplink_archer-a7-v5-squashfs-sysupgrade.bin Nov 11 16:13:39 drwxr-xr-x 2 wasd wasd 4096 Nov 11 09:12 packages/ Nov 11 16:13:41 -rw-r--r-- 1 wasd wasd 592 Nov 11 09:12 sha256sums Nov 11 16:13:43 wasd@wasd-VirtualBox:~/openwrt$ Nov 11 16:13:50 go have a coffee and bailey's on me.... you guys have earned it Nov 11 16:22:41 hehe Nov 11 16:22:57 glad you got it sorted. it's better to have ath79 as well instead of ar71xx, since ar71xx will be phasedout. Nov 11 16:23:46 Guvnarrr: if it all works well, please e-mail the author with openwrt ml in CC (you need to be registered though) to say it works fine. always helps Nov 11 16:33:44 woooot. brcm2708/bcm27{0{8,9},10} with 4.14 succesfully compile-tested Nov 11 16:34:50 weren't you done for the day :P Nov 11 16:35:21 no, I decided to give up on bluetooth and HDMI CEC until after 4.14 was completely finished Nov 11 16:36:12 ah ok Nov 11 16:36:25 btw 4.19 was going to be the next LTS no? Nov 11 16:39:13 Borromini: yes Nov 11 16:39:18 :) Nov 11 16:40:00 it seems to have worked but no LUCI Nov 11 16:40:31 or at least i cant get to the GUI at 192.168.1.1 Nov 11 16:46:02 no, because you're on master Nov 11 16:46:50 do this: ./scripts/feeds update -a && ./scripts/feeds install -a Nov 11 16:46:54 then run make menuconfig again Nov 11 16:46:58 then enable luci in there Nov 11 16:47:00 and rebuild Nov 11 16:47:13 builds from master never have luci enabled by default Nov 11 16:56:54 Borromini: understood... logged into SSH can ping whatever from there, but opkg fails Nov 11 16:57:05 cant get package list Nov 11 16:57:22 ath79 isn't compiled yet afaik Nov 11 16:57:40 as in: by the buildbots. so no package repos etc Nov 11 16:57:58 if you need packages, better enable them in your buildroot and build them - or better yet, include them in your image Nov 11 16:58:21 yeah ok then push that as new firmware i guess? Nov 11 16:58:50 yeah Nov 11 16:59:06 you can use sysupgrade, you won't lose any settings like that Nov 11 16:59:40 sysupgrade is a utility to use from within SSH/linux Nov 11 16:59:42 ? Nov 11 17:05:29 yep Nov 11 17:05:38 just scp the new image to /tmp/ on your router Nov 11 17:05:45 then sysupgrade /tmp/new-image Nov 11 17:05:56 sysupgrade is neat, check out sysupgrade --help Nov 11 17:06:08 excellent... will do. I guess this is *weak* but i like the GUI Nov 11 17:06:41 just ran "./scripts/feeds update" and "./scripts/feeds install -a -p luci" Nov 11 17:06:59 looks like all i need to do now is enable luci in menuconfig Nov 11 17:07:18 i like LuCI for overview and reports, for config the CLI is often handier/more versatile Nov 11 17:12:17 I agree with that Nov 11 17:17:19 I suppose i should be SCPing the openwrt-ath79-generic-tplink_archer-a7-v5-squashfs-sysupgrade.bin file Nov 11 17:17:34 yeah Nov 11 17:17:40 factory is only to be used from stock to openwrt Nov 11 17:21:51 So since I have not really altered anything on this router (it's brand new), I probably can forgo using sysupgrade's "-c" option. Would you agree? Nov 11 17:23:52 sysupgrade will by default keep everything it can Nov 11 17:24:02 so just use it l ike i said Nov 11 17:25:36 i suggested you run sysupgrade --help to see what you could do with it, sorry if i gave you the idea you needed anything beyond just plain sysupgrade /path/to/file at this point Nov 11 17:26:36 WOOT Nov 11 17:26:42 up with luci now Nov 11 17:32:00 Borromini: Thank you so very much. I am up and now I am certain I didn't just throw $60 down the drain. I will def email the author with my success. Nov 11 17:36:24 yw. enjoy! Nov 11 17:36:39 bbl Nov 11 18:21:16 updated openwrt/upstream, https://sdwalker.github.io/uscan/index.html Nov 11 18:44:03 Can anyone check this patch? https://pastebin.com/HJ3LcXYG Its for unpriveledged LXC containers... Thanks Nov 11 18:44:51 I dont know what exactly needs to be mounted with RELATIME ... Nov 11 20:13:12 muhaha: you should make that a patch, describe properly, send to mailing list Nov 11 20:13:31 muhaha: that should be more reliable for getting a proper review Nov 11 20:13:48 muhaha: also, find out, *what* needs REALTIME first **** ENDING LOGGING AT Mon Nov 12 02:59:59 2018