**** BEGIN LOGGING AT Sun Dec 13 02:59:56 2020 Dec 13 06:40:17 rsalvaterra: ping Dec 13 06:42:28 neoraider: can you rebase https://patchwork.ozlabs.org/project/openwrt/patch/0f48abf2ac872957d6a4a150ead39564053f2afc.1589716209.git.mschiffer@universe-factory.net/ ? Dec 13 08:46:34 * aparcar[m] sent a long message: < https://matrix.org/_matrix/media/r0/download/matrix.org/KszvMgLcdcZiqZTWkgZIUcLj/message.txt > Dec 13 08:52:02 aparcar[m]: that's expected when you're trying to use kmods that were not build along with the kernel... Dec 13 08:52:36 Or do you mean the kmod had wrong dep inside? Dec 13 08:52:36 PaulFertser: yea but opkg even download the package from the special kmod archive Dec 13 08:52:57 aparcar[m]: btw, you hilighting jow didn't work as your long message was sent as a URL to a web site. Dec 13 08:53:22 oh bummer Dec 13 08:54:37 dangole: https://paste2.org/XNEGmeyy please check this opkg output Dec 13 08:54:40 aparcar[m]: hah, Depends: kernel (=5.4.81-1-19448ecbdd32381bda1125c6da8fe07e) inside that kmod. So opkg is showing something else, weird. Dec 13 08:54:42 something is still bad Dec 13 08:55:04 Did you download the file just recently? Dec 13 08:55:10 PaulFertser: I'm guessing it crashes at some point inside, the crash log sent to dango is related Dec 13 08:55:20 yea like 4 minutes ago Dec 13 08:56:35 aparcar[m]: and what kernel version are you running there atm? Dec 13 08:58:12 the same as the arhcive Dec 13 08:58:14 *archive Dec 13 08:58:42 aparcar[m]: I wonder where 5.4.82 can come from then, as I downloaded the ipk unpacked everything manually, and it's not inside. Dec 13 08:59:31 Does "opkg info kernel" show 5.4.81 too? Dec 13 09:01:33 Version: 5.4.81-1-19448ecbdd32381bda1125c6da8fe07e Dec 13 09:04:17 aparcar[m]: is it possible opkg is trying to install some other ipk, not the one it just downloaded? Dec 13 09:04:52 I thought the error would be fixed by a series of patches from dangole and me, but now it seem to come again in some corner cases Dec 13 10:47:34 aparcar[m]: pong, but I guess now it's a bit late for you. :/ Dec 13 13:38:14 stintel: testing lldpd on 19.07, it seems the "interface" section of the config doesn't work for me. On APs with a single ethernet (eth0) and a single "lan" device ("br-lan"), lldpd doesn't listen on lan. It only listens on lo. Expected? Dec 13 13:41:31 oh. This looks like a race. Restarting the service "fixes" the issue Dec 13 14:10:27 olmari: any chance you can test ujail on the wdr4900 any time soon? Dec 13 14:29:54 https://lore.kernel.org/netdev/20201213140710.1198050-1-vladimir.oltean@nxp.com/T/#t Dec 13 17:39:42 build #752 of sunxi/cortexa53 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org/master/images/builders/sunxi%2Fcortexa53/builds/752 Dec 13 17:57:07 build #549 of ath79/tiny is complete: Success [build successful] Build details are at http://buildbot.openwrt.org/master/images/builders/ath79%2Ftiny/builds/549 Dec 13 18:03:26 build #543 of ipq40xx/generic is complete: Success [build successful] Build details are at http://buildbot.openwrt.org/master/images/builders/ipq40xx%2Fgeneric/builds/543 Dec 13 18:40:46 congrats on the new wiki look :) Dec 13 18:53:14 updated openwrt/upstream, https://sdwalker.github.io/uscan/index.html Dec 13 19:11:03 rsalvaterra: ping2 Dec 13 19:11:52 oh cool wiki Dec 13 19:12:54 aparcar[m]: 'morning! :) Dec 13 19:13:19 I remeber you had some kmod stuff that looked like it wouldn't be to hard to review Dec 13 19:13:44 98581ec478 zram-swap: remove the compression streams settings Dec 13 19:13:57 this is entirely obsolete because it's done automatically in kernel? Dec 13 19:14:25 Exactly. The kernel uses per-CPU compression streams. That's a no-op. Dec 13 19:14:48 (There's a link to the relevant kernel commit in my patch explaining the whole deal.) Dec 13 19:15:26 okay sounds legit Dec 13 19:20:04 I also sent a patch adding the lzo-rle.ko to the kmod-lib-lzo package, because zram (and probably other crypto API users) expect it to be there. Upstream is a mess in that regard, I don't know if I'll be able to untangle it without breaking anything. At least I fixed the zram dependencies for 5.11, I think (i.e. linux-next hasn't screamed at me yet). Dec 13 19:28:47 do you have both patches at hand? Dec 13 19:30:24 Sure, they're both on Patchwork. Dec 13 19:30:39 https://patchwork.ozlabs.org/project/openwrt/patch/20201019182113.958966-1-rsalvaterra@gmail.com/ Dec 13 19:30:46 https://patchwork.ozlabs.org/project/openwrt/patch/20201210132825.2342-1-rsalvaterra@gmail.com/ Dec 13 19:31:46 okay I'm currently trying to fix an opkg issue but try to review it afterwards Dec 13 19:32:26 Ok, thanks! Dec 13 19:32:51 can't we go full on zstd instead of lzo? Dec 13 19:33:47 aparcar[m]: Not before Linux 5.11. ;) Dec 13 19:34:32 That's the dependency issue I fixed. Until Linux 5.11, zram will always have a hard dependency on lzo. Dec 13 19:35:39 But going full zstd (or as much as possible, at least) is my objective, yes. Dec 13 19:38:17 Well… Come to think of it, I could always backport the kernel patch to our tree, though… Dec 13 19:39:43 rsalvaterra: yea people tend to do quite some backporting... Dec 13 19:40:26 I guess this is simple enough, but I'll wait until we move to 5.10. Dec 13 19:40:45 rsalvaterra: change of plan. I go surf but will review your two patches afterwars ;) Dec 13 19:41:15 Ahahah! I would too, if I lived in Hawaii. XD Dec 13 19:41:19 Go for it! ;) Dec 13 20:32:17 just read that the Archer AX50 has a Intel CPU °° Dec 13 20:33:19 but actually, it does not look like x86 Dec 13 20:33:23 https://deviwiki.com/wiki/TP-LINK_Archer_AX50 Dec 13 21:13:18 zorun: they bought lantiq Dec 13 21:13:26 it's like lots of those Fritz!Boxes Dec 13 21:14:08 Intel sold their Xscale, only to buy into the non-x86 space again a few years later. Money to burn eh. Dec 13 21:15:00 apparently they sold it of this year again though Dec 13 21:27:36 website is revamped? https://openwrt.org/docs/techref/ubus has TRX vs TRX2 on top, strange Dec 13 21:29:36 so does : https://openwrt.org/docs/techref/filesystems.network, layout has some bug for sure Dec 13 21:32:18 all techref namespace seems to be affected Dec 13 21:32:23 already reported here apparently https://forum.openwrt.org/t/implementation-of-the-new-styling-on-openwrt-websites/81672/8 Dec 13 21:33:01 tabbing is broken as well Dec 13 21:33:02 Borromini: ah, they re-sold Lantiq? Dec 13 21:33:07 or did you mean Xscale? Dec 13 21:33:56 zorun: no apparently lantiq got resold as well. Dec 13 21:34:09 https://en.wikipedia.org/wiki/Lantiq Dec 13 21:34:17 see last line of 'corporate history' Dec 13 21:40:37 oh well Dec 13 21:40:54 never even heard about MaxLinear Dec 13 21:41:43 i run daily master build, somethin went wrong as my new build did not come up at all Dec 13 21:45:23 ath79 target, i saw libubox, selinux, fstools and procd new changes added Dec 13 21:48:49 zorun: me neither. Dec 13 21:48:57 rr123: and if you roll them back? Dec 13 21:49:47 Borromini: reboot did not help, need get it near my PC to reset to factory then see which commit broke it Dec 13 21:50:42 a few weeks back similar brick-age happened due to procd changes Dec 13 21:51:17 yes i remember stintel talking about that Dec 13 22:00:58 IIRC that had to do with procd no longer running as root, but with the new user missing in /etc/passwd Dec 13 22:01:23 caused me to have a brick too, after I restored an old config on a build that expected the new user Dec 13 22:48:17 svanheule[m]: indeed. I tried bisecting it, but as I was forced to write factory image after each attempt, it did not expose the bug as it was in sysupgrade Dec 13 22:49:04 I guess it's my own fault as I warned about users being created in packages not working with sysupgrade but never fixed it myself Dec 14 02:35:37 mangix: hey, does this look right to you? https://github.com/openwrt/packages/compare/master...tofurky:fwknop_gpgme_deps i saw you'd recently switched around the gpgme options which is why i ask. if so i can open a PR. thanks Dec 14 02:49:36 m4t: I don't remember touching gpgme. I would replace ln -sf with $(LN) Dec 14 02:51:43 m4t: both those packages also require PKG_RELEASE bumps **** ENDING LOGGING AT Mon Dec 14 03:00:00 2020