**** BEGIN LOGGING AT Sun Mar 10 02:59:56 2019 Mar 10 11:17:27 can someone point me at examples of interacting with the luci jsonrpc with curl or similar? Mar 10 11:17:37 hello, latest trunk fails with tplink wdr4900 Mar 10 11:17:54 I mean image geos right in, but doesn't boot, bootloader says bad CRC Mar 10 11:19:54 Then TFTP:d in new factory of same openwrt mage and flashed that, got in again but same end result, so something doesn't generate the image properly Mar 10 11:20:47 So.. I don't know how or where to fix that, but help :P Mar 10 11:22:16 whole tftp process: http://p.kapsi.fi/?235ddea3ee6a93b2#dk5cwHMWXrUlSOLO+Ggp67zMzK2qPB4CQoHlvS4jUHo= Mar 10 11:23:33 olmari: i don't have a wdr4900, is it a ppc? Mar 10 11:23:56 yeah it is Mar 10 11:24:46 From somewhere between... hmmh... maybe about 1 year and this date something changed that breaks I guess the fimnal image build and CRC'ing Mar 10 11:25:01 when was the last working version? Mar 10 11:25:40 russell--: there's examples on the wiki, are they not working for you? Mar 10 11:25:43 well the thing is that I can't remember offhand now.. somewhere like year ago was the old image inside, but no way of checking that now Mar 10 11:26:07 karlp: i haven't found them on the wiki, do you have a link? Mar 10 11:26:17 flashing recovery instructions work, but new images still odesn't work... old one propably does Mar 10 11:26:47 not what I was looking for, but https://github.com/openwrt/luci/wiki/JsonRpcHowTo has some too. Mar 10 11:26:49 https://oldwiki.archive.openwrt.org/toh/tp-link/tl-wdr4900 here is perfectly working tftp rcovery instructions.,.. https://oldwiki.archive.openwrt.org/toh/tp-link/tl-wdr4900 Mar 10 11:27:16 ..but new images build on trunk seems to have broken CRC :) Mar 10 11:27:59 I'll test latest stable from site in a sec Mar 10 11:29:42 olmari: git bisect maybe Mar 10 11:31:32 russell--: https://openwrt.org/docs/techref/ubus?s[]=curl&s[]=session has some notes o ngetting sessions with curl too, I think that's what I was remembering Mar 10 11:33:36 18.06.2 boots just fine Mar 10 11:35:51 russell--: well "I know of the command" but no idea how to use it or even more generally figure out where and what would be any meaningfull thing :P Mar 10 11:39:52 olmari: you do know how to build your own images? Mar 10 11:40:28 yes, that part I have gotten used to Mar 10 11:40:59 ..To confirm by the way, snapshot from site doesn't boot either, same deal there Mar 10 11:41:41 So definately something bad happened somewhere between 18.06.2 and trunk Mar 10 11:42:18 bisecting is about letting git checkout commits and then marking them good or bad and letting git narrow in on the single commit that went from good to bad Mar 10 11:42:56 there are some decent howtos available online Mar 10 11:43:30 git diff'ing between the release and trunk might give some ideas too Mar 10 11:44:40 well first I'd need to figure out where whole CRC is set to be calculated as in image creation time Mar 10 11:48:03 this commit looks like it might (wild speculation) be related: 576c69938bb38c7edfa23f220eb30cb5c1d0a726 Mar 10 11:51:06 git log target/linux/mpc85xx/ Mar 10 11:51:28 or, tig target/linux/mpc85xx/ Mar 10 11:52:45 could even :) Mar 10 11:52:56 at least sounds like exactly a thing that would break the stuff :P Mar 10 11:56:31 check out the immediately previous commit and building that and see if it works, then checkout the maybe-bad commit, build that and see if works Mar 10 11:56:54 with a make dirclean before each one Mar 10 11:57:24 soon enough, or I'll let buildsystem to finish off what it is doing now and then :) Mar 10 14:26:28 russell--: okay it wasn't that revision, nor immediate after that which is the problem, let the digging around continue :) Mar 10 14:34:16 mailin list suugest that it might not be any specific commit that breaks it, but kernel too big for stock bootloader or so Mar 10 14:35:26 Apparently https://github.com/openwrt/openwrt/pull/1773 Mar 10 15:07:13 blogic: I see that the ipq8064 uses stmmac, is this MAC between the switch and the PHY and how is this modeled with DSA? Mar 10 16:17:45 \o/ Mar 10 18:09:31 nbd: on 18.06 head with your latest mt76 bump, no hiccups so far thank you. Mar 10 18:17:53 dwmw2_gone: I had a look into the LCP echo problem, my current assumption is that the echo repsones were dropped in the switch becasue some queue was full and we do not have QoS activated Mar 10 18:20:35 dwmw2_gone: do you have a dump of the ptm0 interface? Mar 10 18:21:20 we can probably configure some rules into the switch to give the more important packages higher priority Mar 10 23:15:11 \o/ Mar 11 02:17:07 Hauke: stmmac runs the gmacs Mar 11 02:17:32 2(?) macs are wired to the internal qca8337n switch Mar 11 02:17:46 and qca8k runs the 8337 **** ENDING LOGGING AT Mon Mar 11 02:59:57 2019