**** BEGIN LOGGING AT Mon Jun 24 02:59:57 2019 Jun 24 03:23:37 ynezz: urngd is indeed faster than haveged. With 2 core devices it seems to get rid of all uninitialized urandom read warnings. More Info: https://gitlab.labs.nic.cz/turris/turris-build/issues/40 Jun 24 07:02:55 uerm Jun 24 07:02:58 http://downloads.openwrt.org/releases/18.06.2/targets/ Jun 24 07:03:08 seems that 18.06.2 images are being erplaced in many cases Jun 24 07:03:17 rather than going into 18.06.3 directory [?!??!] Jun 24 07:09:16 yes. during mirror server migration, the 18.06.2 downloads got corrupted, so they're being rebuilt fro mtheir respective git tags now Jun 24 07:09:28 18.06.3 builds are quired but not due yet Jun 24 07:10:31 huerm, queued? Jun 24 07:11:16 ok, OOo what 'is expected' for 18.06.3, may take a week or what? Jun 24 07:12:36 2-3 days Jun 24 07:13:18 jow: ok, thaknyou =) Jun 24 07:36:10 jow: hi, as discussed in hamburg, I've reworked the padding of the images https://git.openwrt.org/?p=openwrt/staging/ynezz.git;a=shortlog;h=refs/heads/upstream/fix-x86-images Jun 24 07:37:43 as the image padding is not optional, I've decided to decrease the default 256MiB rootfs size to 64MiB in order to save some space https://git.openwrt.org/b3db380bc81f92a17a5894b1269c03453280ee9c Jun 24 07:38:08 I'm not sure about this one as the images are gzipped so the size difference is neglible Jun 24 07:46:43 ynezz: fine with me Jun 24 07:54:09 oh, we have rasperry pi 4 now :) Jun 24 07:54:44 with emmc/nand finally? Jun 24 07:56:05 but 4GB of RAM is certainly nice Jun 24 08:00:36 https://www.raspberrypi.org/blog/raspberry-pi-4-on-sale-now-from-35/ Jun 24 08:00:47 still uses microSD, i hate this little detail Jun 24 08:01:19 my microSD cards keep dying in raspberry :( Jun 24 08:01:54 real Gigabit Ethernet, 802.11ac dual band, 2 x USB 3.0, HEVC support - that sounds really nice Jun 24 08:11:30 use Industrial SD cards with SLC NAND Jun 24 08:11:37 broadcom wireless? Jun 24 08:14:19 ds_shadof: I think, that the problem is firmware on this SD cards, not the storage medium Jun 24 08:14:49 ds_shadof: industrial usually doesn't imply more stable firmware Jun 24 08:14:56 jow: i guess so Jun 24 08:15:06 jow: FullMAC SDIO i guess Jun 24 08:16:24 ds_shadof: in other words, my SD cards usually die after sudden power-off(removal of the card) during write Jun 24 08:17:45 jow: https://www.raspberrypi.org/blog/raspberry-pi-4-on-sale-now-from-35/ mentions CYW43455, but that chipset supports Bluetooth 4.2 only while Raspberry Pi 4 is said to have Bluetooth 5 Jun 24 08:18:17 can't you just add BT5 with firmware upgrade? Jun 24 08:20:35 rmilecki: Cypress... never heard of it Jun 24 08:20:39 ;) Jun 24 08:20:42 or is that the new Broadcom after the merger? Jun 24 08:21:04 not sure if you're joking or not Jun 24 08:21:20 Cypress bough some part of Broadcom's wireless Jun 24 08:21:26 not all of it though Jun 24 08:21:46 ah yes, was just skimming the wikipedia page Jun 24 08:22:01 no I was not joking actually, its not that common to see "new" wireless vendor names these days Jun 24 08:22:27 i think Cypress now owns wirelss parts used for some small devices (IOT, Raspberry, etc.), all SDIO based (not sure about USB) Jun 24 08:22:38 I think Broadcom still has all PCIe devices (not sure about USB) Jun 24 08:25:44 root@Tappers-router:~# fw_printen Jun 24 08:30:34 https://libreelec.tv/2019/06/libreelec-9-2-alpha1-rpi4b/ "The 4B now uses SPI flash for the bootloader. Current firmware supports SD card boot only – Network and USB booting are still on the Pi Foundation to-do list." Jun 24 08:31:18 apart from HEVC I need so much, another good news: "Also on the list is HBR audio (current audio capabilities are the same as the 3B) and 3D video. The 4B hardware is HDR capable but software support has a dependency on the new Linux kernel frameworks merged by Intel developers (with help from Team LibreELEC/Kodi) in Linux 5.2 os a kernel bump will be needed to use them." Jun 24 08:36:29 Hi people I made a forum post about my probs with my wrt3200acm Jun 24 08:36:32 https://forum.openwrt.org/t/wrt3200acm-will-not-flash-to-second-partition/39399 Jun 24 08:58:13 jow: rmilecki: no need to find out what cypress is, infineon already bought cypress recently ;) Jun 24 08:58:34 oh, great ;) Jun 24 08:58:59 and nxp took over wireless from marvell Jun 24 09:03:34 wow infineon is buying everything now huh. they bought IR recently too. Jun 24 09:04:42 so everybody is ditching wireless because too difficult and let the new owners find out the hard way? :P Jun 24 09:06:28 gweentea: i wouldn't call 2014 "recently" Jun 24 09:08:12 oh my bad, I'm getting old. Jun 24 09:08:21 aren't we all :-) Jun 24 09:08:36 felt like yesterday to me Jun 24 10:06:06 jow: BTW what about the python3 switch? I just got Acked-by: Yousong Zhou but I remember some discussion here, that we probably need to install python3 on the buildbots in order to proceed, right? Jun 24 10:06:34 so probably I should postpone pushing it Jun 24 10:06:51 KanjiMonster: ping Jun 24 10:06:59 KanjiMonster: regarding the feeds.conf thing Jun 24 10:07:03 we agreed that Jun 24 10:07:05 pong Jun 24 10:07:20 1) we add a src-include directive to not have stale .default files Jun 24 10:07:34 2) keep the parameters for building the feeds.conf Jun 24 10:07:40 3) fix the dshes etc Jun 24 10:07:44 right ? Jun 24 10:07:50 right Jun 24 10:07:51 i would code that now Jun 24 10:07:55 thanks ! Jun 24 10:08:38 basically your command would be to create a new feeds.conf and populate that, and commands to edit/modify it we can add at any later date (the main bikeshedding ;) Jun 24 10:09:43 gonna get an rpi4. was planning to get a khadas vim3 because rpi3b+ has no 4k support, but the 4 does. might still get a vim3 though. also, coincidence the vim3 sale starts on the same day as the pi4? I think not Jun 24 10:17:19 KanjiMonster: well, lets keep the bikeshedders busy Jun 24 10:21:14 guys, would you check my "sysupgrade: extending firmware validation" e-mail? Jun 24 10:21:27 jow: that possibly affects LuCI Jun 24 10:21:45 blogic: you like designing things, maybe check that one? Jun 24 10:23:53 rmilecki: lets have a look right now Jun 24 10:25:31 rmilecki: sysupgrade is already triggered from procd Jun 24 10:26:04 so the sysupgrade shell foo could gain a helper that you pass a result which gets sent out to procd via ubus Jun 24 10:26:20 the flash process would need to be split into check and flash Jun 24 10:26:28 ad procd would call them separatley Jun 24 10:26:49 * rmilecki is thinking Jun 24 10:26:57 and if the first return !0 then it just does not start 2/ramfs and waits for the result json Jun 24 10:27:11 and when the result json is there we could provide it on ubus Jun 24 10:27:47 that way you can throw anything into the json, procd would just memcpy the blob and have a sysupgrade.status or similar ubus method Jun 24 10:27:53 as for signing .... Jun 24 10:27:59 doesn't sysupgrade already have a check only command/option? Jun 24 10:28:12 KanjiMonster: it does but not via ubus/procd i think Jun 24 10:28:12 sysupgrade -T /tmp/foo.img Jun 24 10:28:21 it's 0/1 result Jun 24 10:28:22 KanjiMonster: it requires the luci-helper thingy Jun 24 10:31:21 sorry, i'll need more explaning there, starting with the: "so the sysupgrade shell foo could gain a helper that you pass a result which gets sent out to procd via ubus" Jun 24 10:31:23 do you want a a new call like "/sbin/sysupgrade foo /tmp/file" that calls "ubus call procd something"? Jun 24 10:35:41 /sbin/sysupgrade needs a new function inside that gets called from the platform_check_image() Jun 24 10:37:11 and that new function should call ubus? or print JSON? Jun 24 10:39:52 sysupgrae_check_result() or whatever that send the reason via ubus to procd Jun 24 10:39:52 so that each target can be upgraded to this new image check code Jun 24 10:39:52 if the check_image fails the target just calls that helper Jun 24 10:39:52 and procd puts the result into ubus as a call the webui can do Jun 24 10:41:54 ok, I think I got the idea, thanks! Jun 24 10:42:13 i'll try to implement it piece by piece and see how it goes Jun 24 11:51:00 new day .. new bumps .. Jun 24 11:51:20 ldir: you might want to refresh the bumps. just did some alterations to finish them off Jun 24 11:52:06 refresh, as in, re-apply :) Jun 24 11:52:46 gotcha :-) Jun 24 12:22:40 xback: I have 2 * Archer C7 v2 running with your bumps Jun 24 12:47:23 ldir: thank you! Jun 24 13:21:00 ynezz: purely fyi, on my cns3xxx board, i'm seeing this one: [ 9.620000] urngd: jent-rng init failed, err: 2 Jun 24 13:24:40 probably will happen on rpi1/0 too. it cannot load the jitterentropy kernel module Jun 24 13:29:50 jow: any objections to https://git.openwrt.org/?p=openwrt/staging/jogo.git;a=commitdiff;h=76971461b77d8ace7c4ff208de0510301e75ecc1 apart from the missing commit message? Jun 24 13:34:16 rmilecki: why did opcodes.o have to be added to LUAC_O? it seems unrelated to changing the name of the installed 5.1 files Jun 24 13:34:45 karlp: did i do that? let me check Jun 24 13:35:42 it's in v2, patch 1, near the very bottom Jun 24 13:37:02 karlp: i didn't change that line, see a colored diff: https://patchwork.ozlabs.org/patch/1120630/ Jun 24 13:37:18 (diff for a diff can be confusing in reading) Jun 24 13:38:34 cool. Jun 24 13:38:45 karlp: see https://git.openwrt.org/?p=openwrt/openwrt.git;a=blob;f=package/utils/lua/patches/020-shared_liblua.patch;h=64f683ddbe004891930aebfd7d153cb977833615;hb=HEAD Jun 24 13:39:00 it's existing patch 020-shared_liblua.patch that modifies LUAC_O Jun 24 14:20:34 ping nbd Jun 24 14:20:42 cotequeiroz: pong Jun 24 14:22:18 I was looking at my patches, and found this long forgotten one: https://patchwork.ozlabs.org/patch/952204/, assigned to you. It adds chacha20-poly1305 suites to ustream-ssl (at the top of the order). Jun 24 14:26:13 cotequeiroz: looks good to me. i'll merge it Jun 24 14:26:17 thanks Jun 24 14:28:26 I can send a v2, with an up-to-date commit message, if necessary, to get rid of the DMARC header and the old 'our current version' comment. Jun 24 14:29:40 no need Jun 24 14:32:00 applied Jun 24 14:39:43 Thanks. Jun 24 15:17:08 rmilecki: https://www.cypress.com/documentation/datasheets/cyw43455-single-chip-5g-wifi-ieee-80211nac-macbaseband-radio-integrated - the link says BT 4.1, but the pdf itself says BT 5.0 Jun 24 15:30:05 hi, i am going to report an issue: using the imagebuilder to make a new image with luasec package, the building process will fail with error: Unknown package 'luasec'. Jun 24 15:31:51 i was used openwrt-imagebuilder-18.06.2-ar71xx-generic.Linux-x86_64 to make an new image. Jun 24 15:31:51 have you uses the image builder to make other packages? Jun 24 15:33:48 i was made another image by using the same imagebuilder one month ago, everything went well. Jun 24 15:36:09 several days ago i found the opkg didn't work properly, i was going to install iperf3 package but opkg told me it cannot find that package. Jun 24 15:36:28 i had run opkg update already. Jun 24 16:01:50 KanjiMonster: not at all, I've been bitten by this one many times as well Jun 24 16:03:25 xback: ping Jun 24 16:18:34 ldir: pong Jun 24 16:19:23 dude! Jun 24 16:19:53 ? Jun 24 16:20:26 to say I'm seeing funkiness with ath10k-ct as well Jun 24 16:20:45 see email of day/2 days ago Jun 24 16:21:16 don't know if kernel bump or ath10k-ct firmware bump (or both) related Jun 24 16:21:32 I'm currently rebuilding my custom fw (which use ibss by default for mesh) Jun 24 16:21:54 was going to first revert the fw, then the driver, and then apply the fw again Jun 24 16:22:24 cool - I'm trapped in work for a week so not really able to do too much Jun 24 16:24:23 funny-strange-bnw Jun 24 16:27:32 xback: interesting, it means, that this platform probably doesn't have high resolution timer Jun 24 16:28:12 ynezz: could very well be :) would need to check with Tim from Gateworks who has the soc doc Jun 24 16:30:21 stintel: fyi, urngd don't use the kernel module, it has jitterng library(providing the same functionality as the kernel module) compiled in Jun 24 17:17:31 oh noes. I broke 18.06 Jun 24 17:18:03 * mangix tries to figure out why Jun 24 17:18:47 ynezz: no idea if you noticed my comment Jun 24 17:20:48 mangix: you mean the binary builds? Jun 24 17:21:45 talking about this: https://downloads.openwrt.org/releases/faillogs/mips_24kc/packages/ Jun 24 17:22:29 for some reason, deprecated OpenSSL APIs are now disabled, even though they default to y Jun 24 17:23:31 as a result there are long range build failures Jun 24 17:24:43 ok Jun 24 17:29:25 I killed my wrt3200acm lol It will not boot now! Jun 24 17:29:43 Going to dig out the ttl cable later Jun 24 17:30:26 you're a badass :P Jun 24 17:31:40 Yeah fam dem routers best watch out, when badman like tapper rolls tru! Jun 24 17:31:56 * Tapper Points his gun fingers Jun 24 17:32:19 Brap! lol Jun 24 17:32:41 lol Jun 24 17:32:51 wicked Jun 24 17:35:16 The breakage is caused by (at least) openssl 1.1.1c (which is in master/19.07, but not in 18.06) being built in 18.06 ... https://downloads.openwrt.org/releases/packages-18.06/mips_24kc/base/ Jun 24 17:35:43 ^-- mangix Jun 24 17:36:02 ldir: I have some issues with ath10k-ct too. https://pastebin.com/p0Pe2iwq Jun 24 17:36:08 why is that happening? Jun 24 17:40:57 https://git.openwrt.org/?p=feed/packages.git;a=commitdiff;h=1784615d361ca05b6fc074e6870e945deec3a427 Jun 24 17:41:05 haproxy was updated and removed select CONFIG_OPENSSL_WITH_DEPRECATED Jun 24 17:41:07 maybe related Jun 24 17:41:09 bbl Jun 24 17:47:49 Any one know how to get the case off the wrt3200acm off? Jun 24 17:48:39 I have the front off and undue the 4 scroos from inside the feet. Jun 24 17:50:03 ynezz: oh interesting Jun 24 17:51:46 jow: I suggest that there's some confusion in the build bots regarding 19.07 and 18.06. There's no way that haproxy commit is causing openssl 1.1.1 to be built for 18.06 Jun 24 17:58:01 jow: mangix: for some reason the builder used master for the base packages feed -> http://release-builds.openwrt.org/18.06/packages/builders/mips_24kc/builds/53/steps/updatefeeds/logs/stdio Jun 24 18:02:24 hi, i am one of the haproxy maintainers... did i do something wrong? Jun 24 18:02:27 O grate! one of the antenna comes off the board. Jun 24 18:02:33 gladiac1337: np Jun 24 18:02:36 gladiac1337: no Jun 24 18:03:31 yea... my commits do not touch openssl Jun 24 18:04:00 builtbot problem. Jun 24 18:10:28 gch981213: did the ath79 eth interfaces change? It seems to have stopped working on my qca9531 device Jun 24 18:31:41 When finding out what the pins are on the TTL port witch way up should I have the board for my wrt3200acm Jun 24 18:32:04 Remember I cant see the picks on the web Jun 24 19:12:59 Tapper: did you get an answer? Jun 24 19:19:37 if looking down on board from above, where LEDS are to the top, network ports to the bottom, then serial header pins are a white connector at the top left of the board. Jun 24 19:20:52 These pins, from Right to left are - Ground, TX, unknown, RX, unknown, unknown. Jun 24 19:24:21 ldir: do they still use the annoying 2.0mm pin spacing? Jun 24 19:25:14 I don't know, I don't have the board - I'm looking at the pics on the hardware page and trying to translate for our sightless friend. Jun 24 19:27:52 googled, and yes they do. so standard connectors don't fit (which use 2.54mm/0.1") Jun 24 19:31:31 that's a bit of a pain Jun 24 19:33:18 correct, still non-standard connectors Jun 24 19:38:15 They're to a standard, just not the one we'd prefer :-) Jun 24 19:42:28 obligatory https://xkcd.com/927/ Jun 24 19:46:39 KanjiMonster: yeah that one never gets old Jun 24 19:47:57 extra fun if the device doesn't boot while the connector is attached Jun 24 19:51:05 Hi ldir sorry mate was eating. Thanks, but my ttl cable will not fit on the pins in the port. Jun 24 19:51:52 ldir do you want a wrt3200acm? I will post it to you if you like Jun 24 19:52:47 One of the antenna cables has come away from the PCB so there is no hope of me fixing it. Jun 24 19:53:34 find someone who can handle a soldering iron Jun 24 19:54:21 I don't know any one that can do that for me mate. Jun 24 19:55:18 I have a wrt1900ac v2 as a back up and I got the wrt3200acm for free so I don't mind giving it to some one that can fix it. Jun 24 19:56:36 nbd: I've submitted a patch to apply the ustream-ssl change to openwrt: https://patchwork.ozlabs.org/patch/1121498/ Jun 24 19:57:10 very kind but honestly don't have a use for either - happy to have a go at repairing for you Jun 24 19:57:47 Do you have the cable? Jun 24 19:58:05 That would fit on these tinny pins Jun 24 19:58:20 tiny* Jun 24 20:00:40 no. I'd probably get the soldering iron out and wire to the back of the board and add a 3.5mm 'headphone' style socket Jun 24 20:00:57 ldir oo you flash bugger lol Jun 24 20:01:47 I will email you and you can give me details if that's ok? Jun 24 20:01:51 this is the correct cable: https://www.amazon.de/gp/product/B019XLMT8W Jun 24 20:01:52 oh yes, that's me - not - lol - anyway have to go, had about 12 hrs sleep in past 72 hours and not feeling well Jun 24 20:02:16 get better soon mate. later. Jun 24 20:02:46 I will by one of thos cables and send it in the box if that would be better. Jun 24 20:04:26 Tapper: you can buy adapter cables for the smaller pins, e.g. https://www.amazon.com/2-54mm-2-0mm-pitch-Header-Adapter/dp/B00HFZK652/ Jun 24 20:04:28 i've become more of a fan of headphone jacks for removable serial recently Jun 24 20:05:06 O KanjiMonster cool Jun 24 20:08:40 Tapper: it was linked by a linksys community technical support person in their forums, so I assume that is the right one Jun 24 20:08:57 I cant get it in the UK Jun 24 20:09:08 thanks anyway mate Jun 24 20:40:35 so how much time do i have left to get my final package updates into 19.07? Jun 24 20:44:03 gladiac1337: depends on if they are new versions/features or fixes Jun 24 20:44:55 i want to get the new haproxy version (2.0) in 19.07 Jun 24 20:46:09 the new version is already in master Jun 24 20:55:31 gladiac1337: I think you can just open a PR. See hnyman's comment in https://github.com/openwrt/packages/pull/9241#issuecomment-502755356 Jun 24 20:57:41 alright, thank you Jun 24 20:59:56 gladiac1337: in general after branching only fixes should go into the branch Jun 24 21:02:14 when was the branching time? i kinda missed it... Jun 24 21:06:20 about two weeks ago Jun 24 21:37:09 i see, thank you Jun 25 01:14:10 as a learning experience... Jun 25 01:14:32 i'm fiddling with the 4.19 kernel Jun 25 01:15:27 i'd like to trace the history of this file: drivers/cpuidle/cpuidle-arm.c Jun 25 01:16:23 from what I see in the current openwrt code to this change: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/diff/drivers/cpuidle/cpuidle-arm.c?id=dbb3c26034fa5336de45e667e2faf80a78531cd0 Jun 25 01:16:32 appologies for the long link... Jun 25 01:17:01 what are some ways to do that? Jun 25 01:33:17 perhaps bad timming, bad question, or both. I hang out for a bit and post it in the forums Jun 25 02:37:40 got to got now... **** ENDING LOGGING AT Tue Jun 25 03:00:38 2019