**** BEGIN LOGGING AT Sun Oct 07 02:59:59 2018 Oct 07 06:04:06 Hello All Oct 07 06:08:45 These errors are normal? Oct 07 06:08:46 0000:01:00.0.bin failed with error -2 [ 12.776900] ath10k_pci 0000:01:00.0: Falling back to user helper [ 12.900449] firmware ath10k!pre-cal-pci-0000:01:00.0.bin: firmware_loading_st ore: map pages failed [ 12.913248] ath10k_pci 0000:01:00.0: Direct firmware load for ath10k/QCA988X/ hw2.0/firmware-6.bin failed with error -2 [ 12. Oct 07 06:38:37 xcom169: yes Oct 07 08:04:06 anyone know how the sysupgrade tool works Oct 07 08:05:02 can get my head around how to transfer the actual upgrade image to the device Oct 07 08:07:59 Michael76: copy it via scp from your computer to the device or download it directly to the device via wget Oct 07 08:08:14 (store it in /tmp, then run sysupgrade on the file in /tmp) Oct 07 08:08:45 ohh thank you guys it just dawned on me the very second you replied Oct 07 08:09:32 the instruction tells me to download the file directly to the device with wget Oct 07 08:10:22 ok great ill have a go at now hopefully i wont brick this thing Oct 07 08:29:31 Michael76: wget works too, just make sure it land in /tmp/ Oct 07 08:44:00 hmm will the backup option via the web gui due if you do the upgrade with the sysupgrade tool Oct 07 08:45:12 russell--, ohh yeah i read that part Oct 07 09:04:03 Michael76: there's a command line switch to keep the config, it's -c IIRC (just look it up in sysupgrade --help to confirm) Oct 07 09:06:51 xdarklight, yeah did see there where options for this with the tool Oct 07 09:09:12 xdarklight, was hoping to get the upgrade done asap and as i allready now how to do the backup with the ui id like to use it if possible Oct 07 09:11:14 xdarklight, you see the ui upgrade option seem broken at the moment so thats why im reverting to sysupgrade Oct 07 09:11:27 ah I see you really mean the backup option. give me a sec Oct 07 09:11:56 sysupgrade --create-backup /path/to/backup.tar.gz Oct 07 09:23:40 sorry for x-posting, but how is QCA9990 stability with ath10k (or -ct)? Oct 07 09:24:33 I can buy this Linksys LAPAC2600 with a discount, but the offer ends today. it's IPQ8065 with 2x QCA9990 Oct 07 09:24:57 quite tempted to buy it, as I have no Wave2 stuff yet Oct 07 09:28:12 stintel: you should buy it (i have no idea if it works ;-) Oct 07 09:28:41 russell--: it's not supported but could be relatively easy Oct 07 09:29:20 I'm seeing lowest price around EUR249, and can get it today for 140 Oct 07 09:30:14 dualcore 1.7GHz, 128MB flash, 512MB RAM, 802.3at support Oct 07 09:30:33 especially the last one became a hard requirement for me Oct 07 09:43:33 ah but there is already QCA9994, might explain the reason for dumping these linksys at low price Oct 07 10:13:48 okay strangely enough my settings seems to have been saved after upgrading even though i never instructed sysupgrade to save a thing Oct 07 10:15:25 just gave it the sysupgrade command pointed it to the file and nothing else Oct 07 10:16:40 weird Oct 07 10:21:13 if you dont want your settings saved use sysyupgrade -n Oct 07 10:21:20 ohh i see it apparently does this by default Oct 07 10:21:26 yes Oct 07 10:22:00 trine, yeah just read through the documentation again and saw that Oct 07 10:31:47 blogic: I'm going to test if I can reproduce the RPS crash on my spare ERL, if I can I'm going to push https://git.openwrt.org/cd74a1c5 Oct 07 10:33:33 feels weird though that nobody reported it already, the patch is in master for >1m Oct 07 10:52:43 stintel: as far as I understand it QCA9990 is just the industrial quality of QCA9980, so the mixture of ipq8065 and QCA9990 looks a bit strange (I would have expected QCA9984/ QCA9994 there), but it might still be a solid device (are you sure that QCA9990 really is wave2?) Oct 07 10:53:02 pkgadd: yes Oct 07 10:53:09 but 1st gen and 9994 2nd gen Oct 07 10:58:40 I would have expected QCA9990 to be wave1, like QCA9980 Oct 07 11:00:20 https://www.qualcomm.com/products/qca9990 Oct 07 11:02:32 ah, o.k. Oct 07 11:04:15 EUR140 feels like a lot to spend on a previous-gen thing, then again wave2 hardware isn't cheap Oct 07 11:05:02 stintel: another user recently reported repeated crashes on apu2 to me, we just couldn't narrow it down yet because he wasn't getting decent stack tracew Oct 07 11:05:09 stintel: but it might be the same issue Oct 07 11:05:13 and it's also been happening for a while now Oct 07 11:07:09 nbd: was this on irc, flyspray or forum? Oct 07 11:07:12 or ML ? Oct 07 11:07:21 xmpp Oct 07 11:07:22 ;) Oct 07 11:07:23 or directly to you, ah Oct 07 11:07:36 maybe you can point him to my staging tree and ask if that revert fixes it for him ? Oct 07 11:07:55 already gave him a link to that commit Oct 07 11:08:01 stintel: you can sometimes find the nbg6817 (QCA9984, so non-enterprise, non-PoE) for 115-125 EUR, but that might lack some features you're looking for Oct 07 11:08:48 I really need PoE Oct 07 11:08:53 power bricks suck :) Oct 07 11:09:42 one more reason I would consider getting the GW6404 but $800 is a bit too much Oct 07 11:10:05 to replace my APU2s and ERLs Oct 07 11:10:30 but I would like to get 4 eventually, and that's just too expensive :P Oct 07 11:10:49 nbd: alright :) Oct 07 13:11:30 * Monkeh tries puppy dog eyes on nbd Oct 07 14:53:43 can't reproduce it on ERL, but the stack trace in https://git.openwrt.org/cd74a1c5 shows ACPI which isn't used on that platform, maybe it's only a problem on platforms using ACPI Oct 07 15:04:02 Hi any now how to see what scaling governor openwrt is using? Oct 07 15:04:57 know* Oct 07 15:08:22 I found out you cant change it on mvebu any way. Oct 07 15:40:14 bah, seems I missed strongswan 5.7.1 Oct 07 16:21:19 updated openwrt/upstream, https://sdwalker.github.io/uscan/index.html Oct 07 20:02:27 https://bugs.openwrt.org/index.php?do=details&task_id=1749 Oct 07 20:03:29 I think thats what I'm hitting Oct 07 20:04:37 at least, part of it Oct 07 20:50:17 oh, TIL V=sc Oct 07 20:56:06 looks like -lpthread fixes it, might be worth committing Oct 07 21:14:42 oh, maybe not I'm blind, was looking at host portion... can't convince it to add anything to LDFLAGS Oct 07 22:08:22 finally started playing with my odroid c2 again: Linux OpenWrt 4.14.74 #0 SMP Sun Oct 7 14:49:47 2018 aarch64 GNU/Linux Oct 07 22:08:33 nice Oct 07 22:08:49 is mainline support good now/ Oct 07 22:08:50 ? Oct 07 22:09:52 well, not if you want hdmi Oct 07 22:10:07 oh, I know thats still broken Oct 07 22:10:14 but other stuff works well? Oct 07 22:11:42 so I figure out why I haven't hit this e2fsprogs issue before, I missed it from my new config generation stuff Oct 07 22:11:57 buuut I think its fixed Oct 07 22:12:02 tbh I haven't been following the developments recently Oct 07 22:12:09 ah Oct 07 22:12:10 but looks like after 4.14 there were still many changes Oct 07 22:12:14 yeah Oct 07 22:12:40 I've been following the linux-meson stuff Oct 07 22:12:50 oooh, 4.16 -> GX: HDMI and CEC support on Odroid-C2 Oct 07 22:13:04 I only have a c1, so S805 Oct 07 22:13:08 I hope that means with upstream u-boot :) Oct 07 22:13:28 exynos support looks to be in a pretty bad state Oct 07 22:13:36 but theres no proper changelog or anything like meson Oct 07 22:14:50 ooh Oct 07 22:14:51 Meson8b: Odroid-C1 Ethernet support Oct 07 22:14:53 for 4.17 Oct 07 22:15:32 what I still don't know is if I can use arm-trusted-firmware instead of https://git.openwrt.org/?p=openwrt/staging/stintel.git;a=commit;h=e89d16c2c66ea0181c3bcb49f673d69207a6d474 Oct 07 22:16:03 is odroid c1 armv8 ? Oct 07 22:17:13 v7 I think Oct 07 22:17:46 * Amlogic ARM® Cortex®-A5(ARMv7) 1.5Ghz quad core CPUs Oct 07 22:18:11 ah ok Oct 07 22:18:38 alright, one last build test Oct 07 22:18:51 well I called the target meson, with a subtarget gxbb, so it should be possible to add 8b as subtarget too Oct 07 22:18:54 if this builds I can probably deploy it (late night last minute maintenance ftw?) Oct 07 22:19:14 ah cool Oct 07 22:19:35 it looks like the support for most peripherals is there now at least, so it'll be usable with 4.19 lts Oct 07 22:19:48 but if the kernel configs differ too much it might make more sense to have 2 different main targets Oct 07 22:20:15 hm, yeah Oct 07 22:20:53 like brcm2708/brcm2710 ... it would be nice to have it separate because the kernel configs are difficult to keep clean Oct 07 22:21:12 but there are so many patches for brcm2708 that it's probably better to keep them together Oct 07 22:21:31 yeah Oct 07 22:21:36 yeah, 4.19 will be nice :) Oct 07 22:21:48 yeah Oct 07 22:21:57 whole bunch of nice stuff Oct 07 22:24:27 maybe I should just send my meson target to the ML for review Oct 07 22:24:55 heh ya Oct 07 22:26:20 ah no, I called it mesongx/gxbb Oct 07 22:26:51 oh Oct 07 22:28:09 it's been too long since I've actually done this, but then my C2 burnt itself and I couldn't continue Oct 07 22:28:20 so maybe I should get it in master before it burns again :D Oct 07 22:28:21 heh Oct 07 22:28:42 yeah Oct 07 22:30:05 guess I should fork openwrt on github so I can send a pull request Oct 07 22:30:15 not sending git patches via email anymore, I keep fucking it up :D Oct 07 22:35:11 god damn it Oct 07 22:38:55 this in.h/ether.h conflict nonsense is never ending Oct 07 22:41:46 what would be nice is if gcc could print out an include tree or something for things like that Oct 07 23:01:26 Now the VPU Power Domain has been fixed while boothing from Mainline U-Boot, Oct 07 23:01:27 VPU and HDMI nodes can finally be added to the Odroid-C2 DTS. Oct 07 23:01:35 well that looks promising :] Oct 07 23:01:43 nice Oct 07 23:01:53 I'm adding some patches :) Oct 07 23:04:07 argh Oct 07 23:04:14 why do we even have this problem it's ridiculous Oct 07 23:04:24 more importantly, doesn't present itself on the host Oct 07 23:06:19 http://ix.io/1oCe Oct 07 23:06:35 I don't really get it, are the kernel headers conflicting with each other?! Oct 07 23:13:11 ugh, this is just a retarded situation to be in Oct 07 23:34:28 sigh, I give up Oct 07 23:34:32 impossible to debug Oct 07 23:34:44 whatever openwrt is doing, it's just breaking it Oct 07 23:57:34 hey, everyone Oct 07 23:57:46 I was looking at buying a usb adapter Oct 07 23:58:29 i'm eyeing this comfast CF-WU782AC Oct 07 23:58:42 I mean, at no point is that even referenced Oct 07 23:58:54 seems like a cheap way to get a mediatek 7612 based adapter Oct 07 23:59:18 yeah I look here https://github.com/openwrt/mt76/blob/d533f1351834ba2899897b090aa56a5f5b922bd7/mt76x2/usb.c and it's not mentioned Oct 07 23:59:29 but maybe all it would take is an edit once i got it ? Oct 07 23:59:49 is there a difference between the 7612UN and the 7612U ? Oct 08 00:12:34 sigh Oct 08 00:13:01 at this point the file isn't being included yet its *still* causing redefinitions Oct 08 00:13:06 what the hell is the build system doing? Oct 08 00:22:14 whats the prob jwh ? Oct 08 00:23:11 whatever the build system does, it's causing redefintions of structs and others that aren't even referenced Oct 08 00:23:21 honestly it's a total joke of a situation Oct 08 00:23:44 between glibc doing dumb shit and musl refusing to be useful Oct 08 00:23:58 ah ok, what platform is this ? lantiq ? Oct 08 00:24:06 doesn't matter Oct 08 00:24:36 for lantiq glibc you gotta delete toolchain/gcc 850 patch Oct 08 00:25:54 it's not gcc Oct 08 00:26:01 nor platform specific Oct 08 00:26:19 it's header dumbassery, in a similar vein to how so many things need fixed to build on musl Oct 08 00:27:52 sometimes there's probs with lto for me , like with dropbear i disable it Oct 08 00:28:10 also for e2fsprogs i put back in --disable-threads Oct 08 00:28:35 and for midnight commander i revert the patch that moved away from librpc Oct 08 00:28:36 incidentally I also hit that Oct 08 00:28:46 put I just used -lpthread Oct 08 00:28:48 but* Oct 08 00:29:10 so I can't actually use this software on either libc Oct 08 00:29:24 doesn't work with musl because of lacking interfaces Oct 08 00:29:33 doesn't work with glibc because the openwrt build system does something Oct 08 00:29:37 (builds fine on the host) Oct 08 00:29:58 u tried like -lgcc ? Oct 08 00:30:03 what Oct 08 00:30:11 ah nevermind i forget heh Oct 08 00:30:37 -TARGET_LDFLAGS += -flto Oct 08 00:30:37 +TARGET_LDFLAGS += -flto -lpthread Oct 08 00:30:40 for e2fsprogs Oct 08 00:31:10 not sure if a) threads are even worth bothering with for openwrt, and b) why it's even built if theres no ext images Oct 08 00:31:29 yeah --disable-threads works too so, meh Oct 08 00:31:58 yeah Oct 08 00:32:29 glibc isn't tested enough it seems Oct 08 00:33:35 just try deleteing toolchain/gcc/ patch 850 if you haven't , it does mean rebuilding from scratch tho Oct 08 00:33:45 also don't use gcc 8 Oct 08 00:33:50 derp. no luck with backporting patches to get HDMI working on odroid C2 Oct 08 00:33:59 stintel: doh :( Oct 08 00:34:02 and it's 3:30AM, calling it a night ;) Oct 08 00:34:06 later Oct 08 00:34:08 good luck with glibc Oct 08 00:34:17 not a chance Oct 08 00:34:31 I'd actually just suggest dropping glibc support altogether at this point ;) Oct 08 00:34:49 since it doesn't really work Oct 08 00:35:04 i got glibc building working ok on lantiq , rasberry pi 2 and mt7621 Oct 08 00:35:13 iirc it's still required by arc* Oct 08 00:35:23 ah Oct 08 00:35:28 anyway Oct 08 00:35:30 * stintel & Oct 08 00:35:33 isn't that being dropped anyway? Oct 08 00:35:34 nn Oct 08 00:53:11 jwh: uh Oct 08 00:54:30 hi Oct 08 00:57:30 jwh: not sure I should continue ratling his chain Oct 08 00:57:52 you should not Oct 08 00:57:54 jwh: I was a bit unhappy that there wasn't much help for that 'fuxxy' person Oct 08 00:58:17 heh Oct 08 00:58:19 jwh: the SNR was too high Oct 08 00:58:24 lost in the noise now Oct 08 00:58:25 yup Oct 08 00:58:43 jwh: well, I thought this k-person was trolling, so I figured, let's see how serious he is Oct 08 00:59:11 check /names :D Oct 08 00:59:17 jwh: but hum, really, the U-Boot problem he has looks like a corrupted bootloader Oct 08 00:59:38 maybe reloading it with JTAG with ap152 U-Boot build might get him close-ish to a working result Oct 08 00:59:40 perhaps, maybe its also ram if its during relocation Oct 08 00:59:53 could be , yeah Oct 08 01:00:08 those things are dirt cheap, there's no nothing, but then it's SDRAM too iirc Oct 08 01:00:56 heh Oct 08 01:01:45 I'd rather like to steer this fuxxy person toward mainline U-Boot, if he could work on the qca955x port and upstream the patches, I'd like that :) Oct 08 01:02:44 yeah Oct 08 01:03:54 I'd just like to get this stuff working Oct 08 01:04:33 which stuff ? Oct 08 01:04:39 the DIR8xx ? Oct 08 01:04:42 no Oct 08 01:04:48 still wasting time on accel-ppp Oct 08 01:05:23 pretty close to not bothering, may have to replace some of my boxes with mainline linux distros instead but already well into negative return on investmet Oct 08 01:05:26 invstment Oct 08 01:05:27 kljkld Oct 08 01:06:44 the bt home hub 5a is still an awesome piece of equipment imho i just wish there was some way to get the wireless working faster, it seems like there's some kind of ceiling of 20 megabytes/s or so Oct 08 01:09:15 that plus dsl vectoring doesn't really work for me, the snr just slowly drops and drops until the speed drops, I don't have the same problem with broadcom based equipment, but Oct 08 01:09:27 I could ask my provider to try a different profile Oct 08 01:09:49 jwh: but ... it can have backdoors ? Oct 08 01:09:58 * Marex should probably stop :) Oct 08 01:10:19 the magican china chip ahaha Oct 08 01:10:21 its russian, so probably ;) Oct 08 01:10:41 jwh: MIPS Baikal ? Oct 08 01:10:51 jwh: that platform looked AWESOME Oct 08 01:10:55 size of a grain of rice and hacks everything Oct 08 01:14:43 it's definitely more convienient having an isp that doesn't use ppp Oct 08 01:17:01 no updates to the amd display code for 10 days, someone must be on holiday Oct 08 01:18:31 * Marex looks forward to CH330N Oct 08 01:18:37 ordered a 100, so let's see Oct 08 01:20:52 ugh Oct 08 01:20:56 bed time I guess Oct 08 01:21:06 cba dealing with this Oct 08 01:22:26 gn Oct 08 01:22:32 nn Oct 08 01:22:36 just remember, memes are illegal in the EU **** ENDING LOGGING AT Mon Oct 08 02:12:16 2018 **** BEGIN LOGGING AT Mon Oct 08 02:14:09 2018 **** ENDING LOGGING AT Mon Oct 08 02:14:35 2018 **** BEGIN LOGGING AT Mon Oct 08 02:16:09 2018 **** ENDING LOGGING AT Mon Oct 08 03:00:00 2018