**** BEGIN LOGGING AT Tue Sep 03 03:00:09 2019 Sep 03 04:16:06 jow: I can find the feeds.buildinfo nor commit.buildinfo at the targets folders https://downloads.openwrt.org/snapshots/targets/ath79/generic/. However this should be there since https://github.com/openwrt/openwrt/pull/2121 was merged some time ago. I tried it myself by selecting a target and setting the BUILDBOT variable, the files are created as exspected. could you help me out here? Sep 03 07:09:33 good morning people! Sep 03 07:17:37 Tapper: hey Sep 03 07:18:13 Hi there. Sep 03 07:18:42 one to make some people smile. Staffordshire man pranks bailiff by paying £1.2k fine – in pennies https://www.stokesentinel.co.uk/news/stoke-on-trent-news/staffordshire-man-pranks-bailiff-paying-3277858 Sep 03 07:58:22 Hauke: rmilecki could you please merge this as well? https://patchwork.ozlabs.org/patch/1152044/ Sep 03 08:02:22 i wonder how powerfull i9 really is Sep 03 08:02:39 i am considering to buy a thinkpad p53/s Sep 03 08:02:48 they come with i5/7/9 Sep 03 08:04:40 what comes next? i11? Sep 03 08:04:44 blogic: powerful for a few minutes, until throttling kicks in Sep 03 08:05:30 aparcar[m]: no idea, about that. YOu developed the code, so if you don't know how should I? :) Sep 03 08:06:03 aparcar[m]: maybe the problem is that the buildinfo stuff is not part of sha256sums Sep 03 08:06:10 in this case it won't get uploaded Sep 03 08:06:36 jow fair point, I just wanted to be sure there is no rsync script ignoring the files actively Sep 03 08:06:45 that's an interesting point, will look into that Sep 03 08:06:54 i cant find iproute2 in menuconfig anymore. was it renamed? Sep 03 08:07:01 stintel: really ? Sep 03 08:08:04 aep: man grep Sep 03 08:08:08 package/network/utils/iproute2/Makefile Sep 03 08:08:36 aep: should be in network -> routing & redirection iirc Sep 03 08:08:42 odly enough not visible in menuconfig. i wonder what broke Sep 03 08:08:56 unsatisifed dep Sep 03 08:09:00 ooohh. *facepalm* Sep 03 08:09:22 sorry for the noise. its in routing & redirection. search didnt work because its not named iproute Sep 03 08:09:36 jow: could you look at the opkg license patch we worked on in hamburg? I'd be somewhat sad if it just gets never merged Sep 03 08:11:18 jow: running the buildroot in buildbot mode locally creates a sha256sum containing the files Sep 03 08:11:42 blogic: it's usually like that on laptops. might be a bit better on the larger ones due to better / more airflow, but I wouldn't bet that it will never throttle Sep 03 08:11:44 I'll do some more debugging Sep 03 08:12:03 stintel: i am worried that the same is true for i7 Sep 03 08:12:14 stintel: maybe i should just go back to getting a tower Sep 03 08:12:22 blogic: I wouldn't be surprised if the even the i5 does so Sep 03 08:12:37 these things are not made to run 100% CPU on all cores/threads for an extended amount of time Sep 03 08:12:40 stintel: my current i5 sometimes throttels on hot days :-D Sep 03 08:12:51 stintel: but .. but ... Sep 03 08:13:00 mine run full blast always :-) Sep 03 08:17:34 well even the i7 in my mega tower with lots of fans and hydro cooling throttles sometimes Sep 03 08:17:46 even with AC in the room Sep 03 08:18:18 it improved a bit after applying new thermal paste Sep 03 08:18:44 When I rebuild I am going amd all the way Sep 03 08:18:48 might get a ryzen 3950x based system this month Sep 03 08:19:04 Nice! Sep 03 08:19:10 stintel: hmmm Sep 03 08:19:17 stintel: i am a bit scared of that still Sep 03 08:19:21 heard some weird stories Sep 03 08:19:45 I'm not 100% convinced either. I burnt a few AMDs in the past Sep 03 08:19:53 blogic: do you have news regarding the build servers from prpl? Sep 03 08:19:54 this i7 3930k is running 7 years now Sep 03 08:20:05 I'm not entirely convinced the AMD would last that long Sep 03 08:20:13 but only 1 way to find out I guess Sep 03 08:20:13 Some of thoes amd mobos are stuped expensive! I have seen one for £720 Sep 03 08:20:30 Tapper: I was looking at Gigabyte x570 aorus extreme or something Sep 03 08:20:35 because 10gigE Sep 03 08:20:50 stintel: i need 10g aswell Sep 03 08:20:52 then I have a reason to buy a new switch, and upgrade my fileserver, ... Sep 03 08:20:54 Yeah that's the one damb! Sep 03 08:20:59 this is going to be expensive Sep 03 08:21:03 i hate being new x86 HW Sep 03 08:21:22 blogic: well I don't really *need* it, but if I buy something now it has to be future proof Sep 03 08:21:56 stintel: ASRock X470 Sep 03 08:22:17 although now I'm doing some VMs with disks on iSCSI and you easily hit fill up gigE with that Sep 03 08:22:39 so yeah, it's preferred to have 10gigE whenever I buy something new Sep 03 08:23:47 stintel: https://www.asrock.com/MB/AMD/X470%20Taichi%20Ultimate/index.asp Sep 03 08:24:15 aep: if I would buy now I would get the latest chipset Sep 03 08:24:25 also because I have a fancy new PCIe 4.0 GPU Sep 03 08:24:39 *shrug* ok. well latest is expensive Sep 03 08:25:24 it is, but if it survives 7 years like my current workstation it's not that crazy Sep 03 08:25:42 mainboard in this one was also EUR500 iirc Sep 03 08:26:00 hi all, i need in openwrt .config pass somehow which modules to build for that device and which completely disable, which make built in. it is for x86_64 i need to add module which is in upstream kernel since 3.12 in staging and since 3.18 in drivers - xillybus to kernel modules - other Sep 03 08:26:43 indy__: you mean you want to change the kernel .config ? Sep 03 08:27:02 yes but not directly Sep 03 08:27:27 make kernel_menuconfig ? or you mean you want to add an option to the buildroot menuconfig? Sep 03 08:27:51 exactly - add option to buildroot menuconfig Sep 03 08:28:19 target/linux/*/modules.mk Sep 03 08:29:16 openwrt's kconfig parses that file? Sep 03 08:30:18 yup Sep 03 08:33:46 aep: thanks Sep 03 08:34:07 good to know :) Sep 03 08:50:53 on full load I sometimes see a difference of 10C between the hottest and the least hot core Sep 03 08:50:54 aep: 'grep -r KCONFIG:=' revealed additional places for modification Sep 03 08:51:21 maybe I should reapply thermal paste again and this time spread it instead of going with just a drop in the middle and let the pressure spread it out Sep 03 08:52:06 * russell-- would expect that spreading first risks bubbles Sep 03 08:53:18 jow: I added the config.buildinfo to a builder and run it, here is the output: http://5.9.6.233/targets/ath79/generic/ It contains the files and also the buildinfo files are in the sha256sums Sep 03 08:53:28 stintel, indy__ : there's a good video from linus tech tips on doing it properly Sep 03 08:53:42 aep: I'll have a look, thanks :) Sep 03 08:54:16 ordered a tube of thermal grizzly kryonaut yesterday, should arrive tomorrow or thursday Sep 03 08:54:52 I could always delid the CPU but is it worth the risk Sep 03 09:02:25 stintel, o/ Sep 03 09:02:34 nitroshift: yo. how was montenegro ? Sep 03 09:02:45 stintel, too short Sep 03 09:02:54 but *very* nice Sep 03 09:02:58 isn't it always :( Sep 03 09:03:14 yeah, true Sep 03 09:04:33 aep: i haven't found any video on kconfig in linus tech tips Sep 03 09:07:17 indy__: I think that was meant for russell-- Sep 03 09:16:35 stintel O MG I don't have the balls or money to delad a new CPU! lol Sep 03 09:31:29 aep: i found one that compared thermal paste methods and found they were all exactly the same with the exception of "not enough" Sep 03 09:39:51 aparcar[m]: well you need to test using the buildbot Sep 03 09:40:13 aparcar[m]: no idea why these files are missing. maybe they're triggered by a make target which buildbot does not invoke Sep 03 09:41:43 "$ grep -r buildinfo rules.mk include/ scripts/ package/Makefile" is empty in my git clone, so no idea where to look Sep 03 09:44:26 apparenty you rely on "make prepare" getting called. The buildbot does not do that Sep 03 09:45:45 is there anything special to get failsafe support beyond adding the led-failsafe alias in dts? if I ". /etc/diag.sh && echo $failsafe" it correctly replies with a led name, but I get no blinking pattern in boot, though I do see the "press f now" in the serial console Sep 03 09:47:41 jow: yea seems like I do, sorry did not look deep enough into the buildbots yet. I though activating the menuconfig entry would be enough to imitate the behaviour Sep 03 09:51:11 https://git.openwrt.org/?p=buildbot.git;a=blob;f=phase1/master.cfg;h=a91864aaabf5bcad87dd20f2769e4b597797f216;hb=de1247ce78808cbcbabb5c56f652f545b0d5f513#l920 Sep 03 09:51:54 either add further steps for buildversion and feedsversion or change the existing diffconfig step to a prepare on instead Sep 03 09:52:11 yea just found it Sep 03 09:52:31 will do, I think prepare would be fine to use instead Sep 03 09:52:58 or could .config $(tools/stamp-compile) $(toolchain/stamp-compile) cause trouble there? Sep 03 10:00:19 jow: patch sent Sep 03 10:07:17 jow: last question for today: is make tools/install specific for each target or are these tools (like sed, xz etc) share between all build? Sep 03 10:07:56 aka is that the "lowest common denominator" (kleinster gemeinsamer nenner) between all targets? Sep 03 10:09:48 i saw mike has done some work on kernel 5.2 under generic, anyone working to port mvebu platform forward? Sep 03 10:09:58 * russell-- is looking at meraki mr24, its define in target/linux/apm821xx/image/Makefile has a BLOCKSIZE of 63k (used to pad out the kernel and dtb while concatenating), what is the significance of the 63k? Sep 03 10:12:16 aparcar[m]: some tools are only built for specific targets Sep 03 10:13:56 all right thank you! Sep 03 10:14:00 and thanks for the merge Sep 03 10:14:14 the reproducible-builds people will be happy! Sep 03 10:16:54 oh wow jow you're on a merging spree Sep 03 10:23:24 nitroshift not that I know of. I don't know if the 5 ghz driver for the wrt.xx routers will run under k5.2. Sep 03 10:24:43 O it's k5.3 https://github.com/kaloz/mwlwifi/pull/369 Sep 03 10:24:55 There is a patch in openwrt for now. Sep 03 10:27:20 * yanosz is confused again Sep 03 10:27:37 afair 802.11s (some variantes at least) use wcett as a metric Sep 03 10:27:49 is it correct? Is it possible to get this value using ssh on openwrt? Sep 03 10:28:07 do you have any idea? thx in adv Sep 03 10:43:45 Hi, I have a small question. I just got a mail from Patchwork that my patch was superseded, but I can not see by which patch nor any message or commit about that. Sep 03 10:52:13 Pepe: the bzip2 patch? Sep 03 10:52:19 yes Sep 03 10:52:52 bzip2 was updated to 1.0.8 which supposedly fixes the CVE Sep 03 10:53:23 I noticed while trying to merge your patch, hence I set it to superseded (superseded by the version bump of bzip2 to 1.0.8 to be specific) Sep 03 10:53:26 Oh, know I see the point, where we don't understand each other. It was meant for 18.06 branch. :-/ Sep 03 10:53:45 oh sorry, didn't notice. Will merge it there Sep 03 10:53:53 My fault. Sep 03 11:16:41 so I have this mesongx target sitting in my staging tree for a few years already. it's for amlogic series 9 devices. now looking at http://www.linux-meson.com/doku.php#supported_wip_soc_families I'm thing of renaming it to just mesong Sep 03 11:16:49 s/thing/thinking/ Sep 03 11:17:26 as GXBB, GXL, GXM, AXG, G12A and G12B look like they should be supportable by a single target Sep 03 11:18:04 I could also just call it meson, I don't think anybody is going to add support for the 8 or 6 series at this point Sep 03 11:18:57 any thoughts ? Sep 03 11:19:45 got my Khadas Vim3 recently, so I have a 2nd device now, I'd like to finish up my work and get it merged finally Sep 03 11:20:31 turned out you need to define led-boot if you want to get a pattern to enter failsafe. led-failsafe is for blinking when you get _into_ failsafe. not sure what led-upgrade is for, it's nto doing anything while sysupgrade is running at least. Sep 03 11:25:11 is it possible to use kdump with OpenWRT? Sep 03 11:32:49 hmmmz, forgot the make target for u-boot again (e.g. uboot-imx6) Sep 03 11:35:33 make package/uboot-mesongx/{clean,compile} Sep 03 11:37:28 * russell-- understands the 63k thing now. it comes from the layout of the kernel.bin (a 1k header + 63k of dtb + vmlinux) Sep 03 11:44:14 blah, editing basefiles/etc/board.d/01_leds and "make" doesn't include the changes :| Sep 03 11:58:08 blah, make package/base-files/{clean,compile} hasn't picked it up either :| Sep 03 12:13:54 lol: UNRECOGNIZED COMMAND LOGGED TO CLOUD SERVERS. Sep 03 12:57:13 karlp: are you sure the PLATFORM 01_leds is not overwriting yours? If you look at the Package/base-files/install, they are copied after './files' Sep 03 12:58:57 well, I thin it might have been an interattion with upgrade vs first install? so I'd removed something from it, but it then left what was already there from an earlier install. Sep 03 13:03:45 I just use $(TOPDIR)/files, as it will trump everything else in the image, but then it won't be part of the .ipk. Sep 03 13:09:06 well, I'm trying to fix the 01_leds file :) Sep 03 13:14:21 But it will work around your problem, if you're just flashing to test it. Sep 03 13:14:45 karlp: are you sure the image file you flash is the newly built one? On more than one occation I spend an hour trying to figure out why my changes didn't do anything, only to notice I was flashing an old image over and over again instead of the correct one Sep 03 13:15:21 well, it worked when I sysupgraded -n so, yeah, I think so. Sep 03 13:15:27 ah, okay Sep 03 13:15:33 I think my final problem is actually that I don't want a "led-runnign" alias, Sep 03 13:16:48 I still haven't seen led-upgrade do anything though, I thought it would blink something while upgrading, but it just seems to turn it off. Sep 03 13:19:47 is there any reason we still have all the "led_netdev" stuff rather than just updating all the dts files to have trigger-sources? Sep 03 13:19:58 it that just a "haven't gottend around to it" or is there some other reason for it? Sep 03 14:24:56 karlp: have not gotten round yet i would assume Sep 03 14:31:54 np, just curious if there was something else. Sep 03 14:32:42 I just tried moving a oneshot definition into dts, but dts only supports the trigger name, not the timers (in this kernel) so yeah, all just slow migrations. Sep 03 14:32:45 karlp: i doubt it Sep 03 14:33:28 * karlp is gonna assume someone like adrian schmutzler will do it :) Sep 03 14:35:56 ok, upgrade-led works now. wonder why it didn't before.... Sep 03 14:36:27 because no one was asking for it, the alias node is probably the wrong place for led assignment, and it should go into the chosen node (with an appropriate prefix, like "openwrt,led-foo = <&led_bar>;") Sep 03 15:03:38 where does /var/run/led.state come from? I've got a dts that sets a led up as a oneshot, and a ucidef that sets it up as a oneshot, but there's a period when /etc/init.d/led runs that it gets set to "none" again, and my process starts up with no oneshot trigger file. /etc/init.d/led continues and sets it up as per /etc/config/system eventually, but... why did it interfere at all? Sep 03 16:31:06 what can i do about not having a sysupgrade file? this is sunxi Sep 03 16:31:20 default appears to be a rw ext4 Sep 03 21:21:02 My Pixel 3a phone does not want to conenct to my Wifi any more after the upgrade to Android Q Sep 03 21:21:19 I am using wpa3-mixed mode, I think someone else already reported this some time ago Sep 03 21:48:05 WPA3 works with mt76 on master, but not with my ath10k/ath9k on 19.07 from 2 months ago Sep 03 22:13:21 ethernet cable! **** BEGIN LOGGING AT Wed Sep 04 01:07:28 2019 **** ENDING LOGGING AT Wed Sep 04 03:00:01 2019