**** BEGIN LOGGING AT Mon Nov 05 03:00:00 2018
Nov 05 03:31:57 <ldir> stintel: I've poked the dnsmasq list for ideas.  Could you PM me your /var/etc/dnsmasq file please
Nov 05 07:02:55 <blogic> hi
Nov 05 07:24:58 <jow> morning
Nov 05 08:55:22 <rmilecki> hey
Nov 05 08:58:30 <stintel> ldir: https://www.linux-ipv6.be/OpenWrt/dnsmasq-static-unavailable.txt
Nov 05 08:58:30 <wigyori> morning
Nov 05 09:01:07 <stintel> morning o/
Nov 05 09:01:08 <stintel> ldir: sorry, I updated that file with what you asked :)
Nov 05 10:01:13 <karlp> this circleci thing won't let me look at logs of a failure without authorizing the app to have complete read/write access to every single one of my public and private repos on github....
Nov 05 10:01:36 <karlp> that's... not an imporvement over travis-ci.
Nov 05 12:51:49 <rmilecki> does it make sense to get some good SSD for building OpenWrt?
Nov 05 12:52:24 <rmilecki> i can get Samsung 970 PRO for $200, wondering if that is worth it
Nov 05 12:55:20 <Borromini> rmilecki: i can only compare NVMe against S-ATA SSDs, difference was minimal. I think HDD -> SSD should be more tangible I/O-wise
Nov 05 13:07:14 <karlp> hdd->ssd was fantastic for me, yes.
Nov 05 13:09:39 <stintel> using bcache here, 1TB HDD with 64GB Intel 520 backed SSD cache
Nov 05 13:09:49 <stintel> fast enough for me too
Nov 05 13:09:59 <stintel> if you have plenty of RAM, it doesn't matter all that much
Nov 05 13:13:16 <Borromini> :D
Nov 05 13:13:32 <Borromini> stintel: are you talking about building on a ramdisk?
Nov 05 13:14:36 <stintel> no
Nov 05 13:14:39 <stintel> just fscache
Nov 05 13:16:19 <xback> judging an IO bottleneck during build is easyier to see using htop
Nov 05 13:16:37 <Borromini> hi xback :)
Nov 05 13:16:45 <xback> 'lo
Nov 05 13:17:07 <Borromini> there's iotop as well, wouldn't that be more reliable for that?
Nov 05 13:17:36 <xback> I mostly check using htop,  and looking if I see any grey bars in CPU load graph
Nov 05 13:17:47 <xback> where Grey means "waiting on IO"
Nov 05 13:17:53 <Borromini> ok
Nov 05 13:29:57 <karlp> stintel: "plenty of ram" is subjective
Nov 05 13:30:07 <karlp> I had 8gig and going from hdd to ssd was very noticeable.
Nov 05 13:32:16 <stintel> karlp: 8GIG isn't plenty of RAM by far
Nov 05 13:34:04 <stintel> xback: I'm mostly using vmstat 1 to check IOwait
Nov 05 13:34:15 <karlp> sure, but 8gig was enough for plenty of things, and switching from hdd to ssd made a very marked improvement.
Nov 05 13:34:36 <karlp> 8gig is also maxed out on that laptop unfortunately.
Nov 05 13:34:42 <stintel> same problem here :(
Nov 05 13:35:03 <stintel> a month after I bought my xps13 9343, the 9350 was released with max 16GB
Nov 05 13:35:06 <stintel> fml
Nov 05 13:35:47 <stintel> and now I decided to wait until someone releases an ultrabook with max 32GB
Nov 05 13:37:08 <stintel> upgraded my workstation from 32GB to 64GB couple of months ago, should have done that way earlier
Nov 05 13:41:59 <xback> stintel: thanks.  didnt know that one :)
Nov 05 13:42:52 <stintel> iostat -x 1 works too
Nov 05 13:43:31 <karlp> rmilecki: aiui, there's ~0 benefit for the pro over the evo.
Nov 05 13:43:55 <karlp> I've had both an evo sata ssd and an evo nvme and have been very happy iwth them both. (n=1 of course)
Nov 05 13:44:01 <rmilecki> karlp: pro is MLC and evo is TLC
Nov 05 13:44:08 <rmilecki> i'm not sure if I want to trust TLC ;)
Nov 05 13:44:24 <karlp> you're trusting tlc, you're trusting the samsung controller
Nov 05 13:44:40 <karlp> if you're not writing forever all day every day, I don't see why you'd want to go pro
Nov 05 13:45:03 <karlp> this isn't ubifs directly on them ;)
Nov 05 13:45:39 * Borromini builds on his desktop, and he has 16 GB
Nov 05 13:45:55 <Borromini> too bad 32 GB was actually affordable a year *before* Ryzen hit the market :( :(
Nov 05 13:46:23 <xback> I bought it right at the perfect time :p
Nov 05 13:46:42 <xback> Corsair 32GB kit 2400Mc
Nov 05 13:47:09 <xback> bought it 2nd handed for 150 euro's :p
Nov 05 13:47:15 <stintel> https://www.alternate.de/G-Skill/DIMM-64-GB-DDR3-1866-Octo-Kit-Arbeitsspeicher/html/product/1102597 -> only 64GB kit in the memory compat list for my mobo
Nov 05 13:48:26 <Borromini> o_O
Nov 05 13:53:11 * ldir hides his wallet
Nov 05 13:54:18 <stintel> and apparently those memory compat lists do matter
Nov 05 13:56:15 * nitroshift_ gives stintel an old 386 pc
Nov 05 13:57:38 * mamarley gives stintel a 400MB ATA hard drive to go in it.  (Yes, mamarley has actually has one of those…)
Nov 05 13:59:07 <stintel> nitroshift_: cool. I ditched everything < 586 a long time ago
Nov 05 13:59:22 <stintel> maybe even everything < 686
Nov 05 13:59:25 <ldir> hands stintel his ST157A 'cos 40MB has to be enough for anyone.  And feels generous so adds a 387 co-pro
Nov 05 14:00:15 <stintel> the company I work for builds a virtual appliance that requires now 24GB RAM, so 32GB wasn't enough anymore
Nov 05 14:01:06 <stintel> when I'm going to build my next workstation I'll probably go for 128GB right away
Nov 05 14:03:09 <stintel> and seriously considering https://www.raptorcs.com/content/TL2WK2/intro.html as next workstation
Nov 05 14:03:40 * mamarley , with only 16GB, is feeling somewhat inadequate.  Two of the systems are maxed out there, one one can hold 32GB…
Nov 05 14:15:08 * ldir runs a central heating system instead
Nov 05 14:20:32 <stintel> ldir: 😂
Nov 05 14:21:09 <stintel> ldir: the municipality does that for me
Nov 05 14:21:23 <stintel> commie style
Nov 05 14:22:15 <mamarley> District heating?
Nov 05 14:22:16 <karlp> stintel: gonna run it in BE mode too, really test things out? :)
Nov 05 14:23:30 <KanjiMonster> stintel: euro/GiB even seems to be on the cheaper side
Nov 05 14:37:33 <stintel> mamarley: yes
Nov 05 14:37:39 <stintel> karlp: who knows ;)
Nov 05 14:37:47 <stintel> KanjiMonster: that 64GB kit?
Nov 05 14:37:55 <KanjiMonster> yeah
Nov 05 14:38:27 <mamarley> stintel: Cool, I wish they would do that (and district cooling) here too.  It would be more efficient and it would make using those fancy VAV systems easier/cheaper in residential applications.
Nov 05 14:38:32 <stintel> KanjiMonster: it's DDR3 ;)
Nov 05 14:39:12 <stintel> looks like similar DDR4 kit is 20-25% more expensive indeed
Nov 05 14:39:20 <KanjiMonster> ah, oh. didn't notice hat
Nov 05 14:39:24 <KanjiMonster> *that
Nov 05 14:40:07 <KanjiMonster> btw, saw that in the headers of your email from yesterday: X-Greylist: delayed 472 seconds by postgrey-1.27 at vger.kernel.org; Sun, 04 Nov 2018 23:28:40 EST
Nov 05 14:46:58 <KanjiMonster> stintel: also note that vger is in the future, 23 EST would have been 5 am of today here ;)
Nov 05 14:48:21 <stintel> KanjiMonster: huh?
Nov 05 14:48:42 <stintel> I saw the greylisting in my mailserver log
Nov 05 14:49:20 <stintel> that in the future thing is weird. I sent it 21:04 EET
Nov 05 14:51:00 <dedeckeh> ldir:ping
Nov 05 14:51:21 <ldir> dedeckeh: pong - hello!
Nov 05 14:51:49 <dedeckeh> ldir:hello !
Nov 05 14:52:32 <dedeckeh> ldir:I've updated iproute2 to 4.19 in my staging tree
Nov 05 14:52:54 <ldir> uh oh - this is gonna be cake related
Nov 05 14:53:03 <dedeckeh> ldir:could you give it a try as cake is supported now in 4.19
Nov 05 14:53:15 <dedeckeh> indeed ...
Nov 05 14:55:09 <ldir> sure.  IIRC there's nothing >4.14 #ifdef'd so it should just work(tm)
Nov 05 14:55:25 <dedeckeh> ldir:I looked into the cake patch we have; it seems to align with the cake functionality in 4.19.0
Nov 05 15:00:09 <ldir> dedeckeh: I tried hard to keep that patch inline with upstream so when we went to 4.19 it would reverse-apply and just be dropped.
Nov 05 15:00:25 <ldir> anyway - dog needs walking, so a bit later :-)
Nov 05 17:10:02 <stintel> nbd: what's your opinion on https://git.openwrt.org/?p=openwrt/staging/stintel.git;a=commitdiff;h=c48e0f49e1f34ef4c54183ae151b7eee2b58f914 ?
Nov 05 17:10:25 <stintel> reported at https://www.spinics.net/lists/linux-wireless/msg179703.html
Nov 05 17:10:29 <stintel> that should have been in the commit
Nov 05 17:10:50 <stintel> ah it's in the revert patch commit message
Nov 05 17:12:06 <stintel> looks like Arend's reply didn't make it to the list yet
Nov 05 17:14:38 <nbd> a better solution would be to set brcmf_cfg80211_ops.update_connect_params = NULL on roamoff=1
Nov 05 17:14:48 <nbd> then you wouldn't have to revert the nl80211 commit
Nov 05 17:14:55 <nbd> the WARN_ON indicates a driver bug
Nov 05 17:14:57 <stintel> is that the correct solution?
Nov 05 17:15:16 <nbd> i think so
Nov 05 17:15:50 <stintel> ok. I can probably fix that but wouldn't know what to write in the commit message, too unfamiliar with this code
Nov 05 17:15:54 <nbd> the update_connect_params callback is for updating parameters related to roaming
Nov 05 17:16:10 <nbd> there are places in nl80211 that only check for the presence of that callback
Nov 05 17:16:18 <nbd> instead of the roaming flag
Nov 05 17:16:41 <nbd> so it seems to me the WARN_ON was added to complain in case setting roaming parameters is allowed, but roaming itself isn't
Nov 05 17:17:00 <stintel> ok
Nov 05 17:17:40 <stintel> nbd: thanks!
Nov 05 17:18:46 <stintel> I'll stare at the code some more until I understand it better
Nov 05 17:24:03 <mangix> KanjiMonster: Torvalds had a complaint that gmail was flagging mailing list stuff as spam
Nov 05 17:26:56 <KanjiMonster> mangix: I know, I fish a few out of spam every day or so ;) The issue this time was the email not even being in spam (and taking almost 2h to appear)
Nov 05 17:29:54 <stintel> gmail is junk!
Nov 05 17:30:20 <stintel> pun intended
Nov 05 17:50:02 <ldir> stintel: jmail surely?
Nov 05 17:51:52 <stintel> ldir: haha good one
Nov 05 17:53:01 <ldir> I feel I'm doing quite well for someone who's had 3 hours sleep.
Nov 05 17:53:21 <stintel> haha, I had some more but a liter of wine and some wodka before that
Nov 05 17:55:04 <Borromini> i thought programmers only drank cafeine :P
Nov 05 17:55:09 <Borromini> and energy drinks
Nov 05 17:55:19 <Borromini> but this is just an AA meeting it seems.
Nov 05 17:55:44 <ldir> when does the guy with the pizza turn up??
Nov 05 17:55:58 <ldir> and what is the cross-eyed emoji?
Nov 05 17:56:14 <stintel> Borromini: Attitude Adjustment? :>
Nov 05 17:56:39 <Borromini> also ;)
Nov 05 17:56:44 <Borromini> >_>
Nov 05 17:59:06 <stintel> got some goodies delivered today, among which a couple of these: https://erelement.com/programmers-usb-modules/cable-pl2303?zenid=ecqppshfvmnh9gg8c8lgrjr903
Nov 05 18:02:40 <Borromini> your cyrillic must be fluent
Nov 05 18:02:52 <Borromini> all i can read is 'konektor' lol
Nov 05 18:03:08 <stintel> Borromini: I use chromium for websites that don't have an "EN" button ;)
Nov 05 18:03:22 <Borromini> ah ;)
Nov 05 18:03:38 <ldir> ah a genuine knock off usb->serial adapter ;-)
Nov 05 18:07:22 <stintel> ldir: as long as it works ;)
Nov 05 18:07:37 <ldir> lol
Nov 05 18:08:32 * Borromini grabs some popcorn and waits till stintel's routers get fried
Nov 05 18:27:30 <stintel> Borromini: 😂
Nov 05 18:35:09 <ldir> dedeckeh: iproute2-4.19 looks good to me from a cake viewpoint.
Nov 05 18:42:45 <mangix> KanjiMonster: oh I've had that happen as well
Nov 05 18:43:23 <mangix> if gmail is junk, how good of an idea is it to run a mail server on the router?
Nov 05 18:44:11 <Borromini> o_O
Nov 05 18:45:07 <Borromini> mangix: why not some small ARM thingy like a RPi or an odroid?
Nov 05 18:46:17 <mangix> well, the router is an ARM thingy :)
Nov 05 18:54:53 <dedeckeh> ldir:thx for testing!
Nov 05 18:57:29 <ldir> all we need now is the kernel and we can lose the out of tree kmod as well :-) yay!
Nov 05 19:07:20 <Hauke> ldir: I am working on that ;-)
Nov 05 19:10:24 <mangix> hrm
Nov 05 19:10:42 <mangix> been wondering with the new kernel setup how openwrt patches are going to be eliminated
Nov 05 19:10:53 <mangix> that is, superceeded
Nov 05 19:14:00 <Borromini> mangix: i wouldn't feel at ease with stuff like that running on my router, even with lower privileges and whatnot
Nov 05 19:17:36 <mangix> Borromini: how about in an LXC container?
Nov 05 19:21:05 <Borromini> mangix: yeah, i'm not familiar with that but that sounds like it's more contained than just running on your router :)
Nov 05 19:21:21 <Borromini> not familiar with any of the container technology i mean.
Nov 05 20:04:42 <mangix> Borromini: AFAIK it's basically a VM that runs the same kernel
Nov 05 20:20:54 <Latrina> Guys, anyone knows where the imgpproxy package can be found for 18.06?
Nov 05 20:21:10 <Latrina> I cannot seem to find it in the repos
Nov 05 20:21:18 <Latrina> perhaps has it been renamed?
Nov 05 20:23:12 <Borromini> isn't it 'igmp'?
Nov 05 20:23:17 <Borromini> and not 'imgp'
Nov 05 20:24:03 <Latrina> as far as I know it is called imgpproxy
Nov 05 20:24:08 <Latrina> I had it before on lede
Nov 05 20:24:19 <Latrina> this is the source code
Nov 05 20:24:21 <Latrina> https://github.com/openwrt/openwrt/tree/openwrt-18.06/package/network/services/igmpproxy
Nov 05 20:24:32 <Latrina> https://openwrt.org/packages/pkgdata/igmpproxy
Nov 05 20:24:37 <Borromini> double check what you just wrote please.
Nov 05 20:24:56 <Latrina> what you mean double check?
Nov 05 20:25:09 <Latrina> the name of the package is imgpproxy
Nov 05 20:25:18 <Latrina> unless if it has been renamed and I am not aware of it
Nov 05 20:25:25 <wigyori> you'll need to triple-check this
Nov 05 20:25:26 <Borromini> no it's not read what you write, and link to, and what i wrote
Nov 05 20:25:33 <Borromini> Latrina: ^^
Nov 05 20:26:17 <Latrina> I honestly cannot understand the question
Nov 05 20:31:14 <Latrina> never mind
Nov 05 20:44:14 <ldir> root@Router:~# opkg list | grep igmp
Nov 05 20:44:14 <ldir> igmpproxy - 0.2.1-4
Nov 05 20:47:10 <ldir> https://downloads.openwrt.org/releases/packages-18.06/mips_24kc/base/igmpproxy_0.2.1-2_mips_24kc.ipk
Nov 05 21:03:02 <greearb_> bisecting for 'poweroff' failure on r7800 is too much trouble...it was broken at least 1000 commits ago, and at 2000 commits ago, sysupgrade failed.
Nov 05 21:03:36 <greearb_> and with incremental build (ie, not make distclean between bisects), then it won't build, so that makes it take forever.
Nov 05 21:06:03 <russell--> make dirclean is what you want
Nov 05 21:06:26 <greearb_> ok, will try to remember that if I try again
Nov 05 21:06:42 <russell--> but yeah, my from-scratch builds take 20 minutes
Nov 05 21:14:03 <russell--> greearb_: my first guess would be a kernel transition, looks suspiciously at commit 18e9ed2482d from back in may
Nov 05 21:15:27 <greearb_> well, even if I managed to bisect to that, it doesn't really tell me much since the kernel would have changed so significantly :(
Nov 05 21:22:51 <Monkeh> greearb_: But it's only a very small piece of kernel you're concerned with.
Nov 05 21:23:01 <russell--> greearb_: do you know if poweroff ever worked?
Nov 05 21:23:28 <greearb_> someone said it worked, I have not been using r7800 for long and it has never worked for me.
Nov 05 21:23:50 <russell--> it isn't a super common thing to do, in my experience
Nov 05 21:23:54 <greearb_> Monkeh, well, *maybe*, but it is not always the obvious change
Nov 05 21:31:11 <mangix> greearb_: another device that has broken poweroff :)
Nov 05 21:31:45 <mangix> does it also apply to halt?
Nov 05 21:33:07 <greearb_> halt reboots
Nov 05 21:33:32 <mangix> ooo wait a minute. the other device had broken reboot
Nov 05 21:33:33 <mangix> nvm
Nov 05 22:38:36 <stintel> wtf. I have a microSD card that shows different content when booting it on my RPi vs when mounting it on my desktop
Nov 05 22:40:54 <stintel> dmes
Nov 05 22:42:06 <stintel> ah wait, no, it's just the overlay that isn't erased
Nov 05 22:42:28 <stintel> thought I was going mad here for a minute
**** ENDING LOGGING AT Tue Nov 06 03:00:00 2018