**** BEGIN LOGGING AT Sun Jul 08 03:00:01 2018
Jul 08 03:22:09 <luaraneda> Lantis: Check the "include/image-commands.mk" file
Jul 08 03:22:24 <luaraneda> and look for "Build/check-size"
Jul 08 03:23:23 <luaraneda> It will usually (not sure if always), check for the image size
Jul 08 03:23:57 <luaraneda> you can change it temporary to generate the images
Jul 08 03:28:06 <vitus13> Quick question. I have a Linksys EA8500 that I cracked open and sucessfully flashed with ddwrt over serial. I'd rather have OpenWRT on it however. I'm curious if I should be flashing 'EA8500-squashfs-factory.bin' or 'EA8500-squashfs-sysupgrade.tar'. I'm not flashing from factory, but neither am I upgrading OpenWRT.
Jul 08 03:28:54 <vitus13> I guess the safer route would be to put the factory Linksys firmware back on, but if I could go straight to OpenWRT that'd be more convenient
Jul 08 03:31:36 <Fishman> vitus13: yes, that is the general guideline and what the instructions on the wiki have
Jul 08 03:32:48 <vitus13> OK, will do. Thanks
Jul 08 03:33:31 <Fishman> to clarify, recommended is putting factory oem firmware on it, then installing openwrt
Jul 08 03:34:33 <vitus13> Understood.
Jul 08 03:34:52 <vitus13> Just looking around for a sha256sum of the factory firmware before I flash it
Jul 08 03:35:29 <vitus13> mfg website doesn't have, but I downloaded it twice and got the same sum so good enough
Jul 08 03:35:41 <DonkeyHotei> vitus13: there is an easier way
Jul 08 03:36:07 <DonkeyHotei> if you have serial access, you have access to the bootloader
Jul 08 03:36:48 <vitus13> I do
Jul 08 03:37:14 <DonkeyHotei> many models have "initramfs" openwrt images as an option. feed that into the device's ram via tftp and boot to ram
Jul 08 03:37:43 <DonkeyHotei> then use the running system to flash the "squashfs" openwrt image
Jul 08 03:38:04 <vitus13> And I get the initramfs image from that sysupgrade.tar?
Jul 08 03:38:12 <vitus13> I see it has a root and a kernel
Jul 08 03:38:33 <DonkeyHotei> no, it's a separate download
Jul 08 03:38:43 <vitus13> I don't see any explictly labeled initramfs for this model, so I'll have to go via the factory image
Jul 08 03:45:06 <luaraneda> The ipq806x target, which the ea8500 belongs to, doesn't have the "ramdisk" feature. That means no initramfs images are build
Jul 08 03:45:54 <DonkeyHotei> ahh
Jul 08 03:46:39 <DonkeyHotei> btw, apparently, that model supports "advanced reboot" so you will need to reboot it after installling the oem and then install the oem again
Jul 08 03:46:49 <DonkeyHotei> then openwrt
Jul 08 03:47:33 <DonkeyHotei> http://cache-www.belkin.com/support/dl/FW_EA8500_1.1.4.169978_prod.img
Jul 08 03:48:17 <pie> hah, fun
Jul 08 03:54:50 <vitus13> Yeah, it has two flash partitions and you have to play with the power switch during boot to ensure you've flipped to the other. Or you can use the uboot to force it to boot to one or the other, which is more reliable
Jul 08 03:58:00 <DonkeyHotei> openwrt has an optional gui module for advanced reboot to manage that
Jul 08 04:34:44 <Lantis> Why does the build halt here and not continue? https://pastebin.com/dJwPBKvT
Jul 08 04:36:20 <Lantis> I'll clarify. Why does the oversized image halt the entire process rather than just skipping it, and can anything be done about it simply?
Jul 08 04:38:29 <vitus13> So after I flashed the factory image I remembered that the factory image doesn't have the ability to upload a new firmware. It can only auto-update from Linksys servers. So I figured I'd go for broke and just flash the EA8500-squashfs-factory.bin onto just one of the two partitions. It booted fine.
Jul 08 09:13:20 <Tohin> Hello. Is there any information about bpi-R2 support?
Jul 08 09:24:49 <russell--> dir860l-b1 has been OOM'ing on an image from a few days ago
Jul 08 09:25:12 <russell--> r7360-e15565a01c
Jul 08 09:45:04 <Hauke> Tohin: there is already an image for the bpi-r2: https://downloads.lede-project.org/snapshots/targets/mediatek/mt7623/
Jul 08 09:45:11 <Hauke> but I haven't tested it I do not have this device
Jul 08 10:21:29 <Borromini> ldir: not to nag but i think you missed some punctuation in 5781fc6b3f4fade6229390c364c7d7cca514ec76
Jul 08 10:21:59 <ldir> ha - probably - where ?
Jul 08 10:22:51 <Borromini> 'sensitive filesystem required)  Cygwin is unsupported because of the lack' < a point after the bracket?
Jul 08 10:23:05 <Borromini> and you have double spacing instead of one space
Jul 08 10:23:47 <Borromini> at the end of point 2: 'packages into package/feeds/' you're missing a point as well
Jul 08 10:23:59 <ldir> now for some reason I think that a full stop after a close bracket isn't required
Jul 08 10:24:01 <f00b4r0> ldir: ping?
Jul 08 10:24:08 <ldir> pong
Jul 08 10:24:13 <Borromini> and '4. Run "make" to build your firmware.  This will download all sources [...]' has the same double spacing
Jul 08 10:24:28 <ldir> I like double spacing
Jul 08 10:24:37 <Borromini> ldir: if you don't intend to make Cygwin the start of a new sentence I'd add a colon
Jul 08 10:24:59 <Borromini> double spacing looks weird. Like my glasses are off.
Jul 08 10:25:25 * ldir vows never to touch a readme again.
Jul 08 10:25:43 <Borromini> i like pink elephants, but they're no part of accepted punctuation convention unfortunately :P
Jul 08 10:25:49 <Borromini> sorry ;)
Jul 08 10:25:51 <ldir> and blames mkresin for pointing me at a PR
Jul 08 10:27:47 <Borromini> blame the germans, always works.
Jul 08 10:28:07 <ldir> That seems unfair.
Jul 08 10:28:24 <Borromini> depends. you can blame the french, but they can't help themselves
Jul 08 10:28:46 <f00b4r0> meh ;3
Jul 08 10:28:52 <Borromini> hehehe
Jul 08 10:38:00 <Hauke> nbd: your macos fix for nasm brekas the Linux build
Jul 08 10:38:22 <Hauke> the build only works for me when it uses gcc-ranlib instead of ranlib
Jul 08 10:38:59 <Hauke> is gcc-ranlib a standard thing and can we also use that on mac os?
Jul 08 10:39:07 <Hauke> probably we should then also use gcc-ar
Jul 08 10:40:53 <Tohin> Hauke, I have R2-board and I'm ready to test this image. What should I do with the downloaded files? And who should I report the result to?
Jul 08 10:40:56 <f00b4r0> there's nearly no gcc-anything on macos. it's all llvm
Jul 08 10:41:57 <Hauke> The openwrt build anyway needs some specail settings for mac os
Jul 08 10:42:22 <Hauke> Tohin: I do nmot know how to flash these files onto the R2
Jul 08 10:43:25 <ldir> don't break the builds on macos or I'm outta here :-)
Jul 08 10:43:48 <Hauke> Tohin: there is also a pull request doing some additional changes for the BPi-R2: https://github.com/openwrt/openwrt/pull/1135/files
Jul 08 10:43:53 <Hauke> I do not know if they are needed
Jul 08 10:44:10 <Hauke> ldir: do you build OpenWrt on mac os?
Jul 08 10:44:22 <ldir> yep
Jul 08 10:44:31 <Hauke> ldir: do you have gcc-ranlib and gcc-ar?
Jul 08 10:44:41 <f00b4r0> Hauke: no. I'm on 10.13
Jul 08 10:44:54 <ldir> nope
Jul 08 10:48:33 <Tohin> It seems I need to find "kristrev" or "blogic"..
Jul 08 10:53:19 <Hauke> f00b4r0: ldir: could one of you please test this patch on macOSX: https://patchwork.ozlabs.org/patch/941010/
Jul 08 10:53:30 <Hauke> with an x86 build
Jul 08 10:57:02 <ldir> Hauke: I can see what you've done and it looks good to me Mr Incredible, will test later today :-)
Jul 08 10:59:17 <ldir> oi yew Borromini https://git.openwrt.org/?p=openwrt/staging/ldir.git;a=commit;h=a365cae93837f6b6287f7a9a11672881317c4378 satisfactory?
Jul 08 11:00:44 <Borromini> Excellent, thanks for endulging my pedantry
Jul 08 11:02:00 <ldir> indulging
Jul 08 11:02:51 <f00b4r0> lol
Jul 08 11:02:52 * ldir decides to shut up 'cos his Flemish is just shocking these days :-)
Jul 08 11:03:07 <nbd> Hauke: your patch works for me
Jul 08 11:03:40 <ldir> Hauke: ^^^^^ ldir points at a real expert :-)
Jul 08 11:11:19 <Borromini> ldir: thanks again ;)
Jul 08 11:11:40 <Borromini> i always vacillate between endulge and indulge, now i know.
Jul 08 11:14:36 <Borromini> ok... so i just read 'stdin' as 'stalin'
Jul 08 11:14:41 <Borromini> must be the heat.
Jul 08 11:38:50 * ldir comes back inside from mowing the lawn
Jul 08 11:39:00 <DonkeyHotei> Tohin: i do not have this board, but a simple web search finds this http://forum.banana-pi.org/t/bpi-r2-openwrt-lede-support/4810
Jul 08 12:20:20 <Tohin> DonkeyHotei, yes? i know about this forum. But in their version does not work opkg fnd I do not have the ability to compile packages myself.
Jul 08 12:20:38 <Tohin> *yes, i know*
Jul 08 12:20:47 <DonkeyHotei> you were looking for instructions on how to use the images
Jul 08 12:20:56 <DonkeyHotei> that page lists them
Jul 08 12:22:38 <DonkeyHotei> you might want to use those instructions with the sources in the pull request you linked, but note that the pull request removes support for booting from emmc for now
Jul 08 12:22:40 <Hauke> nbd: thanks for testing
Jul 08 13:59:55 <ldir> chaps - https://patchwork.ozlabs.org/patch/916121/ appears in patchwork yet doesn't appear when I use the git pw utility - I wonder if that's why this patch has been left/ignored for a while?
Jul 08 16:09:12 <Hauke> is there a framework in Linux to dynamically switch the PHY between USB and PCIe if your hardware supports this?
Jul 08 16:09:33 <Hauke> I read something like this here: git.openwrt.org/4def81f30f2e89eeecc6ec06e3091f1dd6b3c3d9
Jul 08 16:10:45 <Hauke> I know this is possible with different device tree files, but I am looking for a solution to do this dynamically at runtime
Jul 08 16:25:50 <luaraneda> If you need to modify a device-tree at runtime, It might be possible to do it with device-tree overlays, but I'm not sure if all the infrastructure to support them is upstreamed. Some patches might be still in discussion
Jul 08 16:26:23 <KanjiMonster> Hauke: I think the idea is to dynamically load device-tree overlays, but the last time I looked there is no upstream (accepted) interface for that (the kernel does support runtime modification of the device tree though)
Jul 08 17:05:23 <Hauke> KanjiMonster: thanks for the information
Jul 08 17:06:27 <KanjiMonster> Hauke: userspace interface that is - IIRC there are kernel functions to load overlays, but nothing that exposes it to userspace
Jul 08 17:06:54 <Hauke> ok so if I have a driver which can detect what to use I can switch
Jul 08 17:11:00 <Borromini> ldir: your github exp branch is gone? :'(
Jul 08 17:17:58 <ultito> morning
Jul 08 17:21:05 <Bahgin> Any issue in making my APs (mix match of models) under the same SSID? This is not a unified system.
Jul 08 17:23:51 <lynxis> Bahgin: client usually doesn't like it, when you have 2 different layer2 network behind it. otherwise, it's no problem.
Jul 08 17:24:01 <ryao> Bahgin: What lynxis said.
Jul 08 17:47:31 <pkgadd> ldir: iirc, there was a concern about the size requirements on 4 MB flash devices - and an idea to add a new wpad-tiny for those tiny targets, but afaik no real plan
Jul 08 17:48:23 <Borromini> what would that wpad variety be lacking then? compared to wpad-mini?
Jul 08 17:49:23 <pkgadd> Borromini: that was referring to the patch adding ieee 802.11r to the wpad-mini package by default (so obviously 802.11r)
Jul 08 17:49:33 <Borromini> oh, sorry
Jul 08 17:52:34 <pkgadd> as in ldır's question from 13:59:54 UTC
Jul 08 17:53:46 <Borromini> :)
Jul 08 18:00:21 * ldir coughs https://git.openwrt.org/?p=openwrt/staging/ldir.git;a=shortlog;h=refs/heads/exp
Jul 08 18:01:31 <Borromini> i had noticed that one oh master of git branches.
Jul 08 18:01:38 <Borromini> but that's not for them kernalz
Jul 08 18:02:00 <Borromini> that's ok though i am ploughing through 4.14.54 atm
Jul 08 18:02:15 <Borromini> just one netfilter hiccup rest seems fine.
Jul 08 18:10:11 <huaracheguarache> now that the archer c7 v2 has been ported to ath79 can I just go ahead and sysupgrade to the new target?
Jul 08 18:10:53 <pkgadd> yes, but you may need to use -F (--force) and want to use -n (to omit the existing config)
Jul 08 18:11:16 <huaracheguarache> ok, thanks!
Jul 08 18:11:32 <pkgadd> my tl-wdr3600 and tl-wdr4300 are working great with ath79
Jul 08 18:12:27 <DonkeyHotei> note that ath79 is not in the 18.06 branch
Jul 08 18:12:43 <DonkeyHotei> only master
Jul 08 18:12:55 <huaracheguarache> I only run master
Jul 08 18:12:59 <pkgadd> and source-only so far (which makes sense)
Jul 08 18:13:03 <huaracheguarache> but thanks for the heads up
Jul 08 18:13:57 <pkgadd> e.g. the tl-wr1043ndv1 image doesn't work yet (switch problems, ethernet doesn't come up - or rather only rarely, so rarely that I haven't been lucky so far)
Jul 08 18:16:51 <pkgadd> and afaik AR7161 (like netgear wndr3700/ 3800) also still has problems with its PCI wlan
Jul 08 18:17:02 <DonkeyHotei> Slimey: speaking of ath79, did anything ever come of your gpl source code request?
Jul 08 18:17:09 <pkgadd> but the images that do work correctly, are full featured and work well
Jul 08 18:21:47 <Borromini> oh well my 1043 v1 is still lying belly up here :P
Jul 08 18:21:58 <pkgadd> :(
Jul 08 18:23:28 <Borromini> not because of ath79, because  of a botched uboot upgrade attempt
Jul 08 18:23:40 <Borromini> anyway... time to go. later gents!
Jul 08 18:24:08 <pkgadd> you probably need a CH341A for ~2.50 EUR and a SOIC8 test clip, which is sadly a bit more expensive - and to unsolder it, as flashing in-circuit is unlikely to succeed (as it partially powers up the SOC, which then interfers with eeprom access)
Jul 08 18:26:02 <DonkeyHotei> no jtag?
Jul 08 18:27:23 <pkgadd> wasn't jtag slightly difficult on ar71xx?  I don't really know, but flashing the spi-nor externally should be easier than trying to get jtag working
Jul 08 18:28:11 <pkgadd> e.g. just backup ART from the existing flash, reflash with a working firmware - boot and reflash ART at runtime#
Jul 08 18:28:45 <pkgadd> with unsoldering/ resoldering the flash as the only 'problem'
Jul 08 18:29:56 <Monkeh> If you're going to desolder it, skip the clip and just solder some wires on
Jul 08 18:30:56 <pkgadd> also an option (but slightly more risk to mess up the chip)
Jul 08 18:31:19 <Monkeh> If you can manage to desolder it and resolder it, you can solder some wires on..
Jul 08 18:32:02 <Monkeh> ldir: Please sir, can we have .11w as well? :)
Jul 08 19:05:18 <f00b4r0> Monkeh: sounds like wpad(-full)
Jul 08 19:07:31 <Monkeh> f00b4r0: No, uh, rather a lot smaller than that.
Jul 08 19:13:05 <Monkeh> f00b4r0: iirc .11w adds about 9K uncompressed to wpad-mini - wpad proper is almost 200K bigger compressed..
Jul 08 19:23:08 <ldir> Monkeh: Could you send in a patch? I'm aware that dwmw2_gone patch has been lurking/stalled for some time, so trying to inject a little energy behind it before we lose something useful 'forever'
Jul 08 19:24:29 <ldir> There's talk in the patch of building with -lto which reduces the size increase... but/and if there's an idea to have a wpad-notquitesomini then that's fine too, but I've personally no idea how to achieve that.
Jul 08 19:29:03 <mangix> ldir: about https://git.openwrt.org/?p=openwrt/staging/ldir.git;a=blobdiff;f=target/linux/ath79/Makefile;h=73b6122ac220d4925532151b9c84e556b56b518f;hp=2d8ca7efdbbbfbad25576fb0b2a31086951d88d4;hb=5704a10f9a8f274dc2630bec6eec8c1ff69ce5f6;hpb=a4dda081dfcadf024caf1d8ed382ba665b8a182d
Jul 08 19:29:12 <mangix> does that make any difference?
Jul 08 19:31:34 <mangix> I ask since the dsp2 stuff doesn't seem to work for me. DSPr2 ASM from libjpeg-tuurbo still fails to compile. Well, requires PKG_NO_MIPS16:=0 as well
Jul 08 19:31:45 <mangix> dsp2 is not supported with MIPS16
Jul 08 19:37:19 <mkresin> Hauke: KanjiMonster: it is possible to load dts overlays and the kernel provides a user interface to do so: https://gist.github.com/mkresin/b987b98c41793143c4b0a49da60b275b
Jul 08 19:37:58 <mkresin> f00b4r0: do you have by any chance a nand routerboard with bad blocks?
Jul 08 19:38:43 <mkresin> f00b4r0: I suspect a k4.9+ regression in the out of tree nand drivers
Jul 08 19:39:45 <mkresin> f00b4r0: it looks like no error code is returned if a mtd_read fails due to bad blocks
Jul 08 19:40:30 <mkresin> f00b4r0: or at least it would be the most feasible explaination for FS#1516
Jul 08 19:41:26 <f00b4r0> mkresin: you have me confused for someone else I'm afraid
Jul 08 19:42:30 <f00b4r0> i didn't open that FS and I have no board with NAND
Jul 08 19:44:57 <KanjiMonster> mkresin: can't find anything for that in the vanilla linux kernel, is this a local change?
Jul 08 19:46:20 <mkresin> f00b4r0: wasn't it you who added a bunch of mikrotik boards? or am I totally puzzled?
Jul 08 19:47:14 <mkresin> KanjiMonster: I've done it definitely with OpenWrt. but I can't remeber which target it was
Jul 08 19:48:03 <f00b4r0> i brought support for the NOR family
Jul 08 19:48:03 <KanjiMonster> mkresin: googling points to it being part of the raspi patches, but it's not part of the vanilla linux kernel
Jul 08 19:48:09 <f00b4r0> which are entirely different beasts
Jul 08 19:48:34 <f00b4r0> spi-nor, specifically
Jul 08 19:49:09 <mkresin> KanjiMonster: now that you mention rpi, it's quite likely I've played with one
Jul 08 19:49:38 <luaraneda> mkresin: KanjiMonster I was just reading about it, and the patch, which is not upstreamed is
Jul 08 19:49:39 <luaraneda> "OF: DT-Overlay configfs interface" by Pantelis Antoniou
Jul 08 19:49:58 <mkresin> f00b4r0: so I wasn't entirely wrong. just the wrong kind of flash. dangit! would have been to easy
Jul 08 19:50:01 <KanjiMonster> submitted once 2014, but then nothing
Jul 08 19:50:29 <mkresin> f00b4r0: to find someone with the board and the required debugging/coding skills
Jul 08 19:51:31 <luaraneda> The funny thing is it's required to support FPGA programming using device-tree overlays
Jul 08 19:52:20 <Hauke> mkresin: thanks for the inforamtion
Jul 08 19:52:42 <luaraneda> And FPGA vendor's kernels are using it
Jul 08 19:53:39 <Hauke> there is this documentation for device tree overlays: https://www.kernel.org/doc/Documentation/devicetree/overlay-notes.txt
Jul 08 19:54:03 <f00b4r0> i'm packing, bbl
Jul 08 19:54:09 <Hauke> ah it is only in kernel API
Jul 08 19:54:24 <luaraneda> I just found it on the bcm2708 target:
Jul 08 19:54:25 <luaraneda> target/linux/brcm2708/patches-4.9/950-0090-OF-DT-Overlay-configfs-interface.patch
Jul 08 19:57:05 <KanjiMonster> looking at the last submission, the latest comments were basically "ugh, world writable? that's a bad idea. And we need a security model for that.", which is probably why this hasn't been attempted again
Jul 08 19:57:35 <luaraneda> I'll test the patch on the zynq target in a couple of days
Jul 08 20:00:01 <luaraneda> Hauke: Can these kind of patches be applied to generic, or they belong to a target until they get upstreamed? It's highly likely that I will need it for the zynq target, and brcm2708 already has it
Jul 08 20:08:54 <Hauke> luaraneda: it would be nice if they will go into mainline
Jul 08 20:09:02 <Hauke> we already have a lot of patches in OpenWrt
Jul 08 20:09:22 <Hauke> when you need them for both targtes this shoiuld probably go to generic
Jul 08 21:19:15 <Fishman> what the heck is SpaceRat's deal?
Jul 08 21:19:45 <pkgadd> very weird, every 10 minutes joining/ parting for a second or less
Jul 08 21:19:56 <Fishman> yeah
Jul 08 21:20:11 <Fishman> seems like something a bot that harvests nicks would do
Jul 08 21:21:44 <pkgadd> that nick has been a human user in the past, participating in a few discussions, but the current join flood is extremely annoying
Jul 08 21:22:17 <Fishman> i've been seeing it for a couple months now
Jul 08 21:22:27 <pkgadd> weeks at least, yes
Jul 08 21:22:32 <Fishman> -SpaceRat- VERSION xchat 2.8.9 Windows NT 6.2 [i386/3.49GHz] :GemX/MSWin32 2.7.5 + EDK/X 0.01 + PornKing
Jul 08 21:23:00 <Fishman> i haven't looked up the gemx or edk/x stuff yet...
Jul 08 21:23:08 <pkgadd> at this point it would make sense to ban him for a day or two
Jul 08 21:23:28 <Fishman> or try to talk to him...
Jul 08 21:24:51 <pkgadd> ah, he seems to be logged into freenode - really doing just a join/ part dance every 10 minutes, I kind of assumed him logging out of freenode completely
Jul 08 21:25:45 <Fishman> i don't see indication of when he connected to the network, though, so no idea if he's also disconnecting or not
Jul 08 21:40:55 <Fishman> maybe i'll query him... when i'm not in need of a nap
Jul 08 21:41:07 <c|oneman> FISHMAN UNIT
Jul 08 21:41:12 <c|oneman> hmm, wrong network
Jul 08 21:41:35 <Fishman> O.o
Jul 08 21:41:42 <c|oneman> im guessing you're not the fishman who married a stipper.
Jul 08 21:41:54 <jwh> what
Jul 08 21:41:57 <Fishman> seeing as how i've never married, i am not :-)
Jul 08 21:45:50 <DonkeyHotei> i'm looking at the gpl tarball for the d-link dap-1860, which looks like a device that shouldn't be too hard to support
Jul 08 21:46:08 <DonkeyHotei> the factory firmware is a hacked up openwrt kamikaze
Jul 08 21:46:30 <DonkeyHotei> mt7621
Jul 08 21:47:56 <swalker> updated openwrt/upstream, https://sdwalker.github.io/uscan/index.html
Jul 08 21:48:21 <DonkeyHotei> appears to be 4x4 wifi
Jul 08 21:50:12 <DonkeyHotei> ah, no. i think it's 2x 5ghz and 2x 2.4
Jul 08 21:50:30 <pkgadd> be happy, 4x4 would have been mt7615?
Jul 08 22:02:49 <DonkeyHotei> it's sold as an ac2600 device though
Jul 08 22:17:11 <russell--> dir860l-b1 still OOM'ing
Jul 08 22:17:34 <russell--> r7440-6476148034
Jul 08 22:17:53 <russell--> after a make dirclean world
Jul 08 22:19:10 <russell--> [ 4985.767768] modprobe invoked oom-killer: gfp_mask=0x14201ca(GFP_HIGHUSER_MOVABLE|__GFP_COLD), nodemask=(null),  order=0, oom_score_adj=0
Jul 08 22:19:13 <russell--> [ 4985.792299] CPU: 1 PID: 13245 Comm: modprobe Not tainted 4.14.53 #0
Jul 08 22:41:12 <DonkeyHotei> maybe it is mt7615, i can't tell
Jul 08 22:57:48 <russell--> trying with 4fb27cc9d896f8b73689d1c6a85ca3cff4a9582d reverted
Jul 08 23:01:39 <DonkeyHotei> i am presently connected here via mir3g with that commit
Jul 08 23:06:56 <russell--> it's still early, but it looks like i'm still leaking with that reverted
Jul 08 23:07:31 <russell--> it seems like wireless because my ubnt on the same rev doesn't seem to leak
Jul 08 23:07:40 <russell--> ubnt-erx
Jul 08 23:54:14 <russell--> actually, with the revert, looks like it isn't leaking
Jul 08 23:54:29 <russell--> nbd: ^^^^
Jul 08 23:59:10 <DonkeyHotei> over the past 5 hours, i've gone from this:
Jul 08 23:59:10 <DonkeyHotei> Mem:        253668      42280     211388        400       2488       8120
Jul 08 23:59:11 <DonkeyHotei> to this
Jul 08 23:59:11 <DonkeyHotei> Mem:        253668      43020     210648        400       2488       8124
Jul 09 00:13:14 * russell-- goes back to the leaking version to collect illustrative data
Jul 09 00:57:03 <russell--> grr. not so promising now. seems to have stabilized. heisenbugs--
**** ENDING LOGGING AT Mon Jul 09 03:00:02 2018