**** BEGIN LOGGING AT Sat Apr 24 03:00:29 2021
Apr 24 03:01:14 <Grommish> just make -C <path_to> from the buildroot?
Apr 24 03:11:08 <Grommish> I only realize afterwards how silly it sounds :D
Apr 24 04:29:04 <thagabe> Afternoon, gents. I was wondering if there was going to be improvements on mvebu in terms of performance. Just yesterday I fired up a 21.02-rc1 build and tested it on my gigabit fiber network and I reckon I lost about 20% of my total broadband speed. I know there was going to be a performance penalty but nearly 20% is quite steep.
Apr 24 06:12:24 <guidosarducci> thagabe: FYI, I replied to your apcups msg but you kept disconnecting. Just tried 21.02 and can confirm the bug. Problem is in luci-app-statistics, introduced by jow during js client changes, and never tested it seems. Raise an issue on GH/openwrt/openwrt-luci and mention him.
Apr 24 06:15:16 <guidosarducci> thagabe: root cause is "Host" parameter arbitrarily changed to list type. Workaround: edit /etc/config/luci_statistics and change "list Host 'localhost'" back to "option Host 'localhost'". then reload luci_statistics."
Apr 24 06:22:47 <thagabe> guidosarducci Thank you, Yes 21.02 (Kernel 5.4 in general) has been very buggy on me. I've had many issues in general but i guess that's why it's called rc1. I'm glad you have brought to the attention of the devs the regressions with luci statistics!
Apr 24 06:52:14 <guidosarducci> thagabe: The issue won't get any attention unless you file that bug report, with the logs you already have and extra details I just offered. IRC isn't really a substitute and you'd be lucky if jow noticed otherwise. Thanks.
Apr 24 06:55:23 <thagabe> Oh, I'm sorry I misread that. I thought you had already raised the issue to jow. I'll get on it tonight.
Apr 24 07:47:55 <Grommish> If a luci-app has an issue with a display element working properly under a given theme...  Is that an app package issue or an issue for the theme package?
Apr 24 07:48:41 <Grommish> the app displays fine under bootstrap, for example, but not under 2020
Apr 24 07:49:27 <Grommish> but in 2020 theme, it only fails to work properly in that apps page..  So.. app or theme?
Apr 24 10:17:33 <dorf_> Grommish: theme most likely.
Apr 24 10:18:03 <dorf_> if it's working fine in the baseline theme, then it's a theme issue..
Apr 24 10:44:54 <aparcar[m]> is the kmod-amd-xgbe package missing in x86?
Apr 24 11:03:08 <ldir> aparcar[m]: I think I saw it added recently to master
Apr 24 11:45:19 <guidosarducci> aparcar[m]: any idea where things stand with my BTF PR #3855? I've had no response to queries on GH or IRC; would be nice to know one way or the other.
Apr 24 11:59:05 <aparcar[m]> *hide*
Apr 24 12:23:44 <guidosarducci> aparcar[m]: you know we can still see you even with the blanket on your head, right? :^) Could you ping ynezz or jow or whoever you chatted with before, since all I get is unparseable silence. Another option is just waiting until someone <cough>ldir</cough> is in a helpful mood. Thanks in advance.
Apr 24 12:36:07 * ldir wonders how those commits got there and tries to look innocent
Apr 24 13:18:42 <russell--> i'm getting CPLD driver for Mikrotik RB4xx series boards (MFD_RB4XX_CPLD) [N/m/y/?] (NEW) on ath79
Apr 24 13:56:05 <stintel> russell--: are you running with 5.10 kernel ?
Apr 24 13:57:51 <stintel> russell--: if yes, try https://git.openwrt.org/c8341ca5
Apr 24 14:32:15 <kab-el> nbd: do you know if mt76 will support MT7975AN ? I have an AsiaRF AW7915-NP1, which according to wikidevi contains MT7915AN + MT7975AN
Apr 24 14:33:25 <kab-el> nbd: iw phy reports only '#channels <= 1', I understand this as no DBDC, but according to wikidevi this card should support DBDC
Apr 24 14:37:59 <aparcar[m]> ldir: do you have an opinion on the bpf stuff?
Apr 24 14:47:45 <nbd> kab-el: DBDC means a single chip will show up as 2 PHYs
Apr 24 14:47:52 <nbd> this is only supported on 7915D
Apr 24 14:47:55 <nbd> not the AN variants
Apr 24 14:48:21 <nbd> i heard that 7975 is the same chip type as 7915, so it should be supported
Apr 24 14:53:48 <blocktrron> nbd: do you have any information what is required for MT7915AN + MT7975AN combination to do 160MHz channels?
Apr 24 14:55:17 <blocktrron> This is the combination found on the U6 Lite, where VHT/HE160 are not supported, Ubiquiti also does not list it as a supported operation mode
Apr 24 15:08:41 <nbd> no idea
Apr 24 15:09:21 <Namidairo> the page for that asiarf card only seems to have up to HE80 in the specs
Apr 24 15:10:49 <Namidairo> 2402mbit so HE80 with a 0.8us guard interval
Apr 24 15:12:46 <Namidairo> mentions 80+80 though
Apr 24 15:22:57 <kab-el> hmm
Apr 24 15:23:22 <kab-el> wikidevi also says Configurable 4×4/3×3 or 2×2n+2×2ax DBDC
Apr 24 15:23:40 <Namidairo> that's probably copy pasted from the product page
Apr 24 15:24:48 <kab-el> it is
Apr 24 15:25:25 <kab-el> but what is this MT7975AN ?
Apr 24 15:25:33 <kab-el> should it be visible as another phy? because it isn't
Apr 24 16:04:19 <nbd> please show me the output of cat /proc/bus/pci/devices
Apr 24 16:43:03 <kab-el> nbd: https://blackhole.sk/~kabel/tmp/proc_pci_bus_devices.txt
Apr 24 16:57:21 <PaulFertser> kab-el: do you have a link to some block diagram for that AsiaRF module?
Apr 24 16:57:49 <PaulFertser> kab-el: there's only one PCI device that looks like mediatek-something, 14c37915
Apr 24 17:00:28 <kab-el> PaulFertser: no, I only have the information available on the wikidevi page and the product page
Apr 24 17:00:45 <PaulFertser> kab-el: https://www.asiarf.com/wp-content/uploads/2020/12/MT7915-11ax-WiFi-6-block-diagram-compact.jpg implies MT7975 is a PHY, so no wonder it doesn't need its own driver.
Apr 24 17:01:14 <kab-el> PaulFertser: ok that makes sense
Apr 24 17:01:35 <PaulFertser> kab-el: and https://www.asiarf.com/shop/hot-sales/wifi6-4t4r-dual-bands-selectable-mpcie-card-ieee802-11ax-ac-a-b-g-n-2-4g-5ghz-aw7915-np1/ says "Dual Bands Selectable", not DBDC!
Apr 24 17:02:06 <kab-el> PaulFertser: but it also says Configurable 4×4/3×3 or 2x2n+2x2ax DBDC
Apr 24 17:10:12 <nbd> kab-el: ah, so you only have a single card
Apr 24 17:11:12 <nbd> i don't think 7915 actually supports DBDC being configurable
Apr 24 17:15:14 <blogic_> th mt7915 is the digital wmac
Apr 24 17:15:20 <blogic_> and then has an analog frontend
Apr 24 17:15:21 <blogic_> MT7975A: A band 4T4R 5GHz BW160
Apr 24 17:15:22 <blogic_> MT7975G: G band 4T4R 2.4GHz BW40
Apr 24 17:15:22 <blogic_> MT7976A: A band 4T4R 5GHz BW160 (ePA)
Apr 24 17:15:37 <blogic_> MT7976D: DBDC 2T2R 2.4GHz BW40 & 3T3R 5GHz BW160
Apr 24 17:16:34 <blogic_> nbd: been running the latest mt76 + usteer for ~4 days now with "always answer probe-req" flag
Apr 24 17:17:02 <blogic_> had one fail which was a misconfigured roaming domain after a reflash
Apr 24 17:17:23 <blogic_> also could you extend the sample uci with
Apr 24 17:17:30 <blogic_> # this option does fo
Apr 24 17:17:33 <blogic_> # this option does foo
Apr 24 17:17:43 <blogic_> #option foo default val used in the code
Apr 24 17:17:48 <blogic_> that would be awesome
Apr 24 17:53:49 <kab-el> blogic_: what is usteer?
Apr 24 17:54:12 <kab-el> blogic_: it does not have a README on github...
Apr 24 17:54:29 <kab-el> ah, client steering
Apr 24 17:54:32 <kab-el> what does that mean?
Apr 24 17:55:20 <hexa-> kab-el: pushing stations towards other aps
Apr 24 17:55:32 <hexa-> or in fact the same ap but a different frequency band
Apr 24 17:57:24 <kab-el> hexa-: do these other APs need to have same SSID and encryption? or can I push a client to a completely different wifi? :D
Apr 24 17:58:34 <hexa-> they all need to run usteer and I believe those need to be the same
Apr 24 18:18:21 <blogic_> it needs to be same ssid but if your sta knows different psk / bssid it will also work
Apr 24 19:50:09 <enyc> Hrrm.... *some* older clients which are  WPA2-compatile  fail to connect to WPA2+WPA3 dual-mode APs
Apr 24 19:50:23 <mangix> this is true
Apr 24 19:50:44 <mangix> cough*broadcom*cough
Apr 24 19:50:51 <enyc> at least, with  Openwrt 21.02-SNAPSHOT on lantiq xrx200 and atheros, candelatech suspect firmware etc.
Apr 24 19:51:00 <enyc> mangix: coo whats' going on, what workarounds, fixes, ....... ?
Apr 24 19:51:09 <mangix> don't use wpa3
Apr 24 19:51:10 <mangix> lol
Apr 24 19:51:24 <enyc> mangix: or, as I do, have extra SSID for compatibility mode or so
Apr 24 19:51:31 <mangix> sure
Apr 24 19:52:09 <enyc> mangix: but, if this is whdespread orpoblem, do some linux distros work around in driver / wpa_supplicant, patch bcm firmware, whataver it is?
Apr 24 19:54:40 <mangix> doubt it
Apr 24 19:55:25 <mangix> I've mostly seen this issue with phones
Apr 24 19:55:39 <mangix> i imagine printers would have issues as well
Apr 24 20:20:59 <Borromini> enyc: it's not easy to work around issues with code that is essentially a black hole.
Apr 24 20:22:15 <Hauke> I have sen this problem with an oldr HP printer with wifi and an iPhone 5
Apr 24 20:22:28 <Hauke> they do not connect to WPA2+WPA3 networks
Apr 24 20:27:16 <Hauke> The SAE part of WPA3 is normally completly done by hostapd in Linux, hardware or firmware supprot is only needed for PMF
Apr 24 20:29:40 <Borromini> i have two SSIDs in the air, way easier.
Apr 24 20:33:45 <mangix> Hauke: iOS 9?
Apr 24 20:34:26 <mangix> I remember on an iPhone 3GS WiFi stopped working when I enabled 802.11w on my router. Just optional.
Apr 24 20:34:50 <mangix> actually I remember the issue. I sent a patch for it that got rejected.
Apr 24 20:34:53 <Hauke> iOS 10
Apr 24 20:35:00 <Hauke> optional 802.11w works
Apr 24 20:35:33 <mangix> right. The 3GS was running iOS 6.
Apr 24 20:36:47 <karlp> I mean, my wife's iphone doesn't work on wifi  _at all_ at home.
Apr 24 20:37:12 <karlp> or, to rephrease it might work, but it might mean she doesn't have cell reception anymore, or vice versa.
Apr 24 20:37:31 <karlp> absolute trash fire of sofware, and everyone just accepts it, or throws out everything andbuys nly apple.
Apr 24 20:37:42 <mangix> yeah iphones have shish kebab reception compared to android phones. no idea why.
Apr 24 20:38:26 <mangix> even broadcom android phones have better reception
Apr 24 20:38:28 <Hauke> it probably depends what you comapre them with
Apr 24 20:38:29 <karlp> and it's not openwrt issues either, current home is running stock mikrotik, but it's busted at her sisters house with teh ISP router as well.
Apr 24 20:38:44 <karlp> I'm not really talking about range, just a straight "works, not works"
Apr 24 20:38:56 <karlp> she litereally can't receive calls if wifi is turned on at home.
Apr 24 20:39:28 <Hauke> does the phone try to route the calls over wifi and this fails?
Apr 24 20:39:35 <karlp> anyway, just an anecode that it hasn't even improved since 3gs on ios6...
Apr 24 20:39:36 <Hauke> this wifi calling
Apr 24 20:40:10 <karlp> Hauke: I've no idea, I'v eno way of diagnosigin it, and it happens with at least three different home routers in three different houses
Apr 24 20:40:21 <karlp> so it's effectively unfixable for me.
Apr 24 20:40:30 <mangix> wifi calling is very strange
Apr 24 20:40:44 <mangix> it fails with my archer c7v2 but works with my turris omnia
Apr 24 20:41:09 <mangix> both using qca9880
Apr 24 20:41:21 <Hauke> I had no problems with wifi calling, but my phone doesn't do it at home beacsue I have good cell reception
Apr 24 20:41:31 <Hauke> mangix: do you block ipsec?
Apr 24 20:41:50 <Hauke> the phone builds up an ipsec tunnel to the ISP and then does SIP on it
Apr 24 20:42:35 <mangix> i do not
Apr 24 20:43:05 <mangix> the openwrt config can be identical. still fails.
Apr 24 20:43:20 <Hauke> and the ISP checks if you are in your home country, if you are outside, you should use cell because it is more expensive
Apr 24 20:43:49 <Hauke> I read that it does not work when you use the google DNS server instead of the provider DNS server
Apr 24 20:44:25 <mangix> I'm talking about the same place for the router. I'm using ISP DNS.
Apr 24 20:44:40 <Hauke> ok, I am not an expert in this
Apr 24 20:44:55 <mangix> hmmmmm there may be one possibility
Apr 24 20:45:02 <mangix> actually nvm
Apr 24 20:45:12 <mangix> both run dnsmasq
Apr 24 21:03:35 <Borromini> guys, is the kernel/rootfs split dynamic for all openwrt targets? or does it depend on the target
Apr 24 21:14:38 <owrt-snap-builds> Build [#66](https://buildbot.openwrt.org/master/images/#builders/64/builds/66) of `realtek/generic` failed.
Apr 24 21:25:19 <blocktrron> mangix: are you on android?
Apr 24 21:28:28 <blocktrron> If so, feel free to read https://issuetracker.google.com/issues/186147999
Apr 24 21:29:39 <Hauke> blocktrron: why is there a login required to view a bug report
Apr 24 21:31:19 <mangix> what the...
Apr 24 21:31:58 <mangix> this implies it's some configuration issue
Apr 24 21:32:39 <mangix> strange...
Apr 24 21:34:50 <Borromini> Hauke: i got this stacktrace on 21.02 head: https://paste.debian.net/1195038/
Apr 24 21:34:56 <Borromini> MT7613 radio
Apr 24 21:40:15 <blocktrron> mangix: what do you mean with configuration issue?
Apr 24 21:40:41 <blocktrron> It's jsut always sending DNS queries for the IMS to the default gateway, regardless whether or not it has to pass this host
Apr 24 21:40:57 <blocktrron> Might explain your problem, not sure who is at fault there
Apr 24 21:41:49 <Hauke> Borromini: I think I saw this also some time ago, for me wifi continued to work
Apr 24 21:42:27 <Borromini> Hauke: same here, it seems.
Apr 24 21:42:53 <Borromini> i don't know to read it, and since it's 21.02 and that's close i thought it might help if i shared
Apr 24 21:42:54 <Hauke> Borromini: no, I think I saw this on my Raspberry Pi runing Debian testing  (kernel 5.10)
Apr 24 21:43:01 <Borromini> ok.
Apr 24 21:43:47 <svanheule> Borromini: mtdsplit handles the automatic splitting, AFAIK there's nothing platform specific about it
Apr 24 21:44:04 <Borromini> svanheule: ok thanks, so for all targets then. thanks!
Apr 24 22:08:09 <Borromini> svanheule: does mtdsplit only take care of the kernel/rootfs split or does it also handle the rootfs_data partition?
Apr 24 22:08:43 <svanheule> normally both, but by different parts of the driver
Apr 24 22:09:04 <Borromini> ok, thanks
Apr 24 22:09:10 <svanheule> kernel/rootfs has quite a few types of splitters
**** ENDING LOGGING AT Sun Apr 25 02:59:56 2021