**** BEGIN LOGGING AT Thu Jul 26 03:00:01 2018 Jul 26 03:54:00 jwh: https://openwrt.org/contact#irc_channels shows only 2 channels. #openwrt isn't one of them Jul 26 03:55:13 it should probbly be on there Jul 26 03:55:53 but either way theres more people in there and you may well get a better response to most questions Jul 26 03:56:21 that true Jul 26 03:57:20 I'll try to add #openwrt there Jul 26 04:00:49 jwh: There it is. Added. Jul 26 04:00:52 o/ Jul 26 04:01:29 I mean I'm just an observer but it kinda makes sense to keep the noise in one place heh Jul 26 04:01:31 ups, maybe bump from 2 to 3 channels :D Jul 26 07:42:25 * ldir is back after Felix broke the internets :-) Jul 26 07:42:47 so i'm trying to get this ar8316 detected for rspro using ath79 target. i keep getting an unaligned access in ag71xx_mdio_reset -> regmap_write Jul 26 07:43:05 added some printk: [ 2.119188] ag71xx_mdio_reset: regmap_write(ffffffea, 20, 80000007) Jul 26 07:44:04 https://paste.ee/p/vEMD0#M1tf9Px2TGOCbyFz9ambITw6A6dcnddX is the current iteration of the .dts i'm working with Jul 26 07:44:58 https://paste.ee/p/BZAwT#zIk8ptrhU8XcfqvBy2SLpCg8RU3R8fkw is dmesg w/stack trace **** BEGIN LOGGING AT Thu Jul 26 08:30:46 2018 Jul 26 08:41:33 m4t: 0xffffffea is -22, which is -EINVAL - somewhere an error return is not properly checked Jul 26 08:42:11 KanjiMonster: thanks pulling my hair out over here :( Jul 26 08:43:06 im trying to figure out the dev so i can dev_name() and enable LOG_DEVICE for it Jul 26 08:43:24 KanjiMonster: do you still have your rspro? Jul 26 08:43:38 yes Jul 26 08:44:11 it's my home router though, so can't use it for testing Jul 26 08:44:40 hah, my family gave me an old wndr3700v2 over christmas, so i just merged my configs onto that and swapped it out Jul 26 08:45:02 did that earlier. so i've got the rspro free Jul 26 08:45:38 https://photos.app.goo.gl/crQoxz3oPWi8TYN87 Jul 26 08:47:45 so anyway, you should follow where the first argument to regmap_write came from, and then why it is -22 / -EINVAL - might be some wrong value(s) in the dts Jul 26 08:48:04 i'm sure there are some wrong values there heh Jul 26 08:48:07 thanks, i'll do that Jul 26 08:49:05 do you still netboot? i was wondering, if i or someone else gets ath79 going on it, if it'd be possible to "multiplex" spi between flash+rtc and get the rtc going, finally Jul 26 08:49:13 idk how descriptive dts can be in that aregard Jul 26 08:51:06 not sure how much worth it is with the rtc missing a battery Jul 26 08:52:23 true Jul 26 08:52:36 especially with hundreds of days uptime at a time and a local gps ntp server... Jul 26 08:52:51 still, it'd be nice to have everything working Jul 26 08:54:12 [ 1.141592] libphy: Fixed MDIO Bus: probed Jul 26 08:54:14 [ 1.771889] ag71xx_mdio_init: is -EINVAL! Jul 26 08:54:16 heh Jul 26 08:54:18 yep.. Jul 26 08:58:02 m4t: there was some short discussion for spi multiplexing support when someone tried to add support for a hybrid spi flash chip (both nand and nor, with a special "select chip" command to switch; but the actual nand/nor chips did behave like normal nand/nor chips), but it never resulted in any actual code changes Jul 26 09:01:04 huh something like this? https://patents.google.com/patent/US20120072647A1/en Jul 26 09:01:11 kinda strange Jul 26 09:01:34 guess it could be good for real estate and e.g. keeping some stuff ro the entire stuff Jul 26 09:02:48 m4t: http://www.winbond.com/hq/product/code-storage-flash-memory/spistack-flash/?__locale=uk&partNo=W25M161AW Jul 26 09:03:57 er, -the entire stuff, butyeah, cool Jul 26 09:04:01 (I like how the two chips require different voltages) Jul 26 09:04:08 hahah Jul 26 09:06:24 *clicks on serial nand* so nand chips with the same voltage as the nor chip do exist (and they even have them), they just opted to make integration more complicated Jul 26 09:07:35 also they do seem to have a hybrid one with both taking ~3.3V Jul 26 09:10:08 do you know if anyone ever ported u-boot to rspro? i looked around but couldn't find anything Jul 26 09:11:29 i ran into some issues with .elf + appended dts, and the .appended_dts mips thing didn't work. so i'm using cat linux.bin rspro.dtb > appended.bin atm, and manually specifying entry point. well, tftp'ing the .bin to the entry point. Jul 26 09:11:53 .appended_dtb i mean Jul 26 09:13:05 m4t: if you have a linux.bin, then you likely don't have an elf, but a raw binary Jul 26 09:13:25 yeah, cat linux.elf rspro.dtb > appended.elf does not work Jul 26 09:13:42 nor does using objcopy and the kconfig setting to add the .appended_dtb section to the elf Jul 26 09:13:57 thats why i'm using appended .bin Jul 26 09:14:00 I never tried the elf thingy Jul 26 09:14:33 i'm just thinking in terms automatically generating images in buildroot Jul 26 09:15:23 do we have any targets/devices that use elf kernels? Jul 26 09:16:51 oh, does rspro use .bin on ar71xx? Jul 26 09:17:38 it does ".. fis load -d -e kernel" which is elf Jul 26 09:17:48 i think...to boot .bin i'm having to add '-r' Jul 26 09:18:31 er wait a sec...lemme check redboot docs Jul 26 09:19:11 m4t: from flash it does load a raw binary -> https://github.com/openwrt/openwrt/blob/master/target/linux/ar71xx/image/generic-ubnt.mk#L225 Jul 26 09:19:55 how old is rspro? it uses redboot still? Jul 26 09:20:12 the rspro is quite old Jul 26 09:20:35 still uses mini-pci as well, and not mini-pcie Jul 26 09:20:37 Ubiquiti Networks certified release, version 0.9.00318M.0905121200 - built 12:01:38, May 12 2009 Jul 26 09:21:00 maybe what i added to generic-ubnt.mk in ath79 is busted. it spits out a uImage for initramfs. Jul 26 09:22:16 egh, i'll look at that more later. trying to figure out why syscon_regmap_lookup_by_phandle is returning einval Jul 26 09:26:06 uImage is for uboot, not redboot, so sounds quite wrong Jul 26 09:26:58 yep Jul 26 09:27:30 swapping out compatible = "qca,ar7100-eth"; (from ar7100.dtsi) with compatible = "qca,ar7100-eth", "syscon"; seemed to make it do more Jul 26 09:27:31 (likely just a default that needs to be overridden somewhere) Jul 26 09:28:53 [ 2.190995] switch0: Atheros AR8316 rev. 1 switch registered on mdio-bus.0 Jul 26 09:28:55 :-) Jul 26 09:32:50 m4t: in general if you see addresses that start with 0xffffff.., you most likely have an error pointer, not an actual address ;) Jul 26 09:33:04 yeah it looked funny to me ;) Jul 26 09:41:42 i need to figure out how to read the mac addresses from redboot or w/e Jul 26 09:41:49 i have them hard coded atm Jul 26 09:44:58 64 bytes from 192.168.1.1: icmp_seq=2 ttl=64 time=0.216 ms Jul 26 09:45:01 nice! Jul 26 09:45:17 i'm surprised my devicetree was correct, aside for the crap with missing syscon compat Jul 26 09:49:00 Hey anyone know a way to renormalize line endings for existing files on github, without delete/re-add all files? We try to change a repo from CRLF to LF endings. Jul 26 09:49:59 maybe something like find repo/ -type f -exec dos2unix {} \; ? Jul 26 09:50:13 but that'd be a massive commit... Jul 26 09:51:11 mhh i was hoping for a "silent" git command to-do the job, but i guess without re-adding the files they wont change in the database? Jul 26 09:51:54 yeah there's not any way around it, checksums change Jul 26 09:52:34 but if you do a git blame, you can ignore whitespace using -w flag Jul 26 09:53:58 we just try to change the source files, so patches created on unix don't have mixed line endings in them. Jul 26 09:54:41 for some reason they put the cross-platform sources with crlf endings up, which is kinda odd for github Jul 26 09:55:49 yuck Jul 26 10:17:37 andy2244: that is not possible without rewriting the history. Maybe git's client side crlf handling could be an option Jul 26 10:19:56 yeah but want to use this for openwrt, so it needs to-be checked in with LF, openwrt does no special handling as far as i'm aware or is there some hidden packge define to handle this? Jul 26 10:20:43 andy2244: include/toplevel.mk:export GIT_CONFIG_PARAMETERS='core.autocrlf=false' Jul 26 10:21:05 thats all special handling I am aware of Jul 26 10:22:40 mhh i could set this to true, but thats still ugly and similar to just do dos2unix on package/prepare, should have been fixed on the database. They also have the problem that after using .gitattributes now all new files are LF, while old are CRLF... Jul 26 10:25:03 btw did something break yesterday regarding dns in the master? Had to revert my FW from yesterday, since dns did not work anymore, port was unreachable? Jul 26 10:25:47 the firewall was broken and didn't start - not sure if this was your issue Jul 26 10:26:29 will find out later and check if a fresh build still has this problem Jul 26 11:02:42 Hello! Jul 26 11:03:36 I observed uci starts doing invalid memory accesses in valgrind, when duplicate sections with the same name are in a file. I was looking around in the code but some help would be apreciated Jul 26 11:04:41 m4t: here You have working image preparation code for routerstation should be similar to pro: https://github.com/tmn505/source/commit/c1efafd6cf881932104b3354cb36d227fe838eeb. You also can see in dts how I set up mac address. Jul 26 11:05:52 m4t: BTW do you get valid irq's for minipci cards? Jul 26 11:17:43 nbd: may I disturb you? Jul 26 11:18:16 nbd: regarding an uci problem I am not able to diagnose in the C source code Jul 26 14:36:48 connection to my ZNC instance too problematic. Exiting, sorry. Jul 26 16:18:22 adblock.sh, illegal instruction Jul 26 16:18:24 rofl Jul 26 16:32:04 at least not payload.sh Jul 26 20:00:45 Hi, I have a problem flashing and booting Mikrotik RB450G with openwrt-18.06-rc2 Jul 26 20:01:21 logs from sysupgrade -> https://pastebin.com/X3EGTgPg Jul 26 20:01:38 log from boot -> https://pastebin.com/Bj4wf5LN Jul 26 20:02:52 should I open new bug report? Or is it open already? Jul 26 20:10:57 I think I found it -> https://bugs.openwrt.org/index.php?do=details&task_id=1516&string=mikrotik&type%5B0%5D=&sev%5B0%5D=&pri%5B0%5D=&due%5B0%5D=&reported%5B0%5D=&cat%5B0%5D=&status%5B0%5D=open&percent%5B0%5D=&opened=&dev=&closed=&duedatefrom=&duedateto=&changedfrom=&changedto=&openedfrom=&openedto=&closedfrom=&closedto= Jul 26 22:28:25 I see nightly builds of openwrt Jul 26 22:28:44 This is now confusing what bins to download Jul 26 22:28:50 openwrt or lede ? Jul 26 22:29:08 want to do an upgrade on an old NEXX wt2030 Jul 26 22:29:13 I see images for both Jul 26 22:30:32 LEDE is the most current release, 17.01.5, until the new OpenWrt branded 18.06 goes golden (but there is 18.06-rc2 already) Jul 26 22:30:55 unless you dabble into the distant past, the only difference is just the naming and branding Jul 26 22:31:11 so use the newest one you're comfortable with Jul 26 22:31:32 snapshot images (contrary to release or rc images) don't have the webinterface preinstalled Jul 26 22:33:54 MT7620n in general is a chipset where you wouldn't want a too old version, so I'd rather suggest 18.06-rc2 than 17.01.5, as there has been work on the rt2x00 wlan drivers recently (which still won't win any prices, but should at least be better than in 17.01.x) Jul 26 22:34:16 ...and I hope you have the 8 MB flash variant, otherwise https://openwrt.org/supported_devices/432_warning applies Jul 26 22:34:29 pkgadd: have you got a silicon errata for 7620? Jul 26 22:34:44 or were you talking about older versions of openwrt/lede? Jul 26 22:35:00 pkgadd: yeah, I think I have 4MB, not 8. Jul 26 22:35:07 but thanks for the info on 18 Jul 26 22:36:36 how can I check from the CLI if I have 4 or 8? Jul 26 22:39:34 * javi404 digs around for 18.06-rc2 Jul 26 22:41:35 karlp: no, I don't have any specs about it - and I was exclusively referring to the rt2x00 driver situation Jul 26 22:42:16 last generation that's usually (always) coupled with rt2x00 wlan hardware Jul 26 22:42:20 pkgadd: so snapshot here, is this as new as 18.xx ? https://downloads.openwrt.org/snapshots/targets/ramips/mt7620/openwrt-ramips-mt7620-wt3020-4M-squashfs-sysupgrade.bin Jul 26 22:42:30 or is there a better link Jul 26 22:42:44 I don't mind bleeding edge for this device. Jul 26 22:42:51 snapshot is newer than 18.06-rc2 Jul 26 22:43:01 pkgadd: thanks for confirming Jul 26 22:48:39 stupid question, why doesn't wget work, server returns a 401 Jul 26 22:48:47 sorry, 400 Jul 26 22:49:01 use http, not https Jul 26 22:49:06 ah Jul 26 22:49:07 ok Jul 26 22:49:15 https support in wget (ustream) is optional Jul 26 22:49:15 then let me do a sysupgrade -v http://etc Jul 26 22:49:47 ...exactly for 4 MB devices like yours(?), as ssl support requires quite a chunk of flash Jul 26 22:50:54 yeah, these were purchased years ago Jul 26 22:50:59 great little things so small Jul 26 22:51:06 i doubt the one I have is 8MB Jul 26 22:51:17 flashed and rebooted Jul 26 22:51:19 let's see Jul 26 22:51:26 I did manage to get Luci on it Jul 26 22:51:47 last time back in the 15.0.x nightly it was just currently running. Jul 26 22:51:57 that's quite unlikely to succeed again, with 18.06 or 17.01 Jul 26 22:52:14 maybe if you use the imagebuilder Jul 26 22:52:23 but not with opkg on the fly Jul 26 22:53:08 (imagebuilder stores luci in the squashfs component, which has significantly better compression than jffs2 as used for the overlay - but it's extremely tight either way) Jul 26 22:53:26 got it Jul 26 22:53:39 well lets see, it rebooted Jul 26 22:53:41 maybe Jul 26 22:53:42 haha Jul 26 22:54:58 you have at least enough RAM, sad they they coupled it with that little amounts of flash Jul 26 22:59:57 pkgadd: yeah. Jul 26 23:00:01 flash success Jul 26 23:00:17 they make great hidden low power devices. Jul 26 23:00:51 ok Jul 26 23:00:56 time to reconfigure it Jul 26 23:01:12 thanks for your help pkgadd, looks like the led is fixed now, not dark. Jul 26 23:01:26 let me know if you need it for testing. Jul 26 23:01:32 or for me to report anything Jul 26 23:20:01 tmn505: cool. my wifi cards are detected and i can view their e.g. channels with 'iw' but i haven't tried bringing them up. i have to remove an antenna from one of the cards to connect the db9 port :/ Jul 26 23:28:43 question, why do you guys enable ULA by default? Jul 26 23:56:43 javi404: counter question, why not? hint, ULA prefixes make a lot of sense, if the WAN uplink isn't 100% reconnects more than just occassionally (and comes with additional niceties like dynamic prefixes) Jul 26 23:59:18 hi all Jul 26 23:59:23 is there anyone here Jul 26 23:59:26 pls help Jul 27 00:00:26 with? **** ENDING LOGGING AT Fri Jul 27 03:00:01 2018