**** BEGIN LOGGING AT Fri Apr 09 03:00:09 2021 Apr 09 06:26:33 mangix, that linux support might be "vendor" and not "official" - the year at https://www.kernel.org/category/releases.html is different from CIP ex: https://wiki.linuxfoundation.org/civilinfrastructureplatform/start Apr 09 06:38:32 plntyk: that kernel tree lives on kernel.org: https://git.kernel.org/pub/scm/linux/kernel/git/cip/linux-cip.git/ Apr 09 06:39:11 official enough for me Apr 09 06:47:06 yeah i saw that too - that is better compared to "hidden" LTS distro kernel sources Apr 09 10:08:29 How could I use Meson in a package in my own feed? It says that I should have "include ../../devel/meson/meson.mk" in my Makefile, but that of course won't work if the package is in a different feed. Apr 09 10:10:29 wb9688: well, in your case, topdir could be a solution Apr 09 10:11:45 Ah, like $(TOPDIR)/feeds/packages/devel/meson/meson.mk? Apr 09 10:13:17 That seems to work indeed, thanks! Though it's of course not ideal because e.g. someone could name that feed differently Apr 09 10:22:23 nbd: do we need #define _POSIX_C_SOURCE for MacOS? Apr 09 10:22:30 Hi, how do I add a cfe RAM file to the bcm63xx-cfe package? I can't seem to locate it in the OpenWRT repo I've cloned... Apr 09 10:22:33 it's used for getopt() https://git.openwrt.org/?p=openwrt/openwrt.git;a=blob;f=tools/firmware-utils/src/zyimage.c;h=6aacf2ff1bd94fa55258f30a6d4aa9e93312b072;hb=HEAD#l11 Apr 09 10:22:53 Festivenari: check Makefile for URL Apr 09 10:22:54 wb9688: currently a solution like that is needed. Eventually, meson will move to base Apr 09 10:22:54 stintel: Speaking of module unloading issues… :P https://marc.info/?l=linux-block&m=161793727423976&w=4 Apr 09 10:23:10 rmilecki: i don't think so. i'll give it a try Apr 09 10:23:13 rmilecki: no, no, I want to do that locally Apr 09 10:23:17 nbd: for me it causes Apr 09 10:23:18 src/zyimage.c:10:0: warning: "_POSIX_SOURCE" redefined Apr 09 10:23:21 src/zyimage.c:11:0: warning: "_POSIX_C_SOURCE" redefined Apr 09 10:23:37 Festivenari: ah, i missed the point of question Apr 09 10:23:55 running that make command results in one makefile that doesn't work; missing /package.mk Apr 09 10:23:59 Festivenari: so you may fork bcm63xx-cfe repo and replace URL Apr 09 10:24:02 Festivenari: or add a patch Apr 09 10:24:18 Festivenari: or hack Makefile to copy your cferam to the build dir Apr 09 10:24:28 I guess I will try that out, then Apr 09 10:24:54 rmilecki: still builds after removing both Apr 09 10:24:58 nbd: thanks Apr 09 10:25:32 rmilecki: try compiling with clang's -Weverything Apr 09 10:26:56 Festivenari: add define Build/Prepare step Apr 09 10:27:28 Festivenari: with something like Apr 09 10:27:29 $(CP) /home/festivenari/foo $(PKG_BUILD_DIR)/ Apr 09 10:27:39 rmilecki: yeah, I was going to try that Apr 09 10:28:11 Festivenari: you may also be able to use $(TOPDIR) Apr 09 10:28:18 $(CP) $(TOPDIR)/foo Apr 09 10:35:55 FWIW: I don't have clang, but I tried -Wextra and it's ok Apr 09 10:37:44 rmilecki: sorry, swapped routers. just reconnected. Apr 09 10:38:08 rmilecki: thanks. It's no longer complaining about missing the cferam image Apr 09 10:38:11 *file Apr 09 10:38:33 mangix: I googled -Weverything, seems something development only, I don't have clang, so I tried Wextra Apr 09 10:38:45 Festivenari: np Apr 09 10:39:05 right. -Wall and -Wextra don't really enable all warnings. Apr 09 10:39:14 there are useful ones left out Apr 09 10:39:47 meson tries to abstract it away by offering warning levels. Apr 09 10:40:01 and adding a bunch based on what compiler is being used. Apr 09 10:40:24 gcc unfortunately has nothing like -Weverything Apr 09 10:46:29 nbd: are you planning on pushing a more recent mt76 to 21.02? The March 26 bump does seem to fix the MCU timeouts for me on MT7613BE. Apr 09 10:50:40 Borromini: And the broadcast issues on MT7615. ;) Apr 09 10:51:55 speaking of mt76, i should really figure out how to get this router running on OpenWrt... Apr 09 10:52:22 mangix: The MT7621 you ordered? Apr 09 10:52:53 rsalvaterra: good =) Apr 09 10:53:23 my WDS performance seems down though. can't have it all i guess Apr 09 10:53:33 rsalvaterra: yeah. i can't figure out how to get properly formatted firmware. everything i try results in the bootloader giving me an LZMA decompression error Apr 09 10:55:50 mangix: Header shenanigans? Apr 09 10:56:29 who knows Apr 09 10:57:07 I also ordered a MT7621-based board :D Apr 09 10:57:44 mangix: which model? Apr 09 10:58:13 linksys e7350 Apr 09 10:59:33 Borromini, rsalvaterra: how much have you tested the latest mt76, did you encounter any regressions? Apr 09 10:59:56 I have an update for 21.02 prepared, just wanted to make sure that it was tested enough Apr 09 11:00:45 nbd: All fine up to now. Apr 09 11:05:05 nbd: daily driver here for over a week Apr 09 11:05:44 my 2,4 GHz WDS (MT7615) seems to perform worse than before, i should re-run some iperf tests Apr 09 11:05:56 In other words, my dad doesn't complain, so that means "ship it" in my book. :P Apr 09 11:06:13 :P Apr 09 11:10:48 What's the best way to analyse a broadcom-based firmware image? Apr 09 11:11:05 binwalk doesn't work since it's all in one single jffs2 partition Apr 09 11:12:24 Festivenari: I came across this, the other day: https://github.com/sviehb/jefferson (never used it). Apr 09 11:12:58 Don't know if it helps or not. Apr 09 11:14:17 Festivenari: use jefferson to extract JFFS2 Apr 09 11:14:25 rmilecki: yeah, doing that right now Apr 09 11:14:45 getting the original FW image was not exactly simple; had to setup port mirroring on a switch and then wireshark for the URL Apr 09 11:15:02 Festivenari: you can also pastebin "hexdump -C -n 512 foo.bin" and "hexdump -C foo.bin | tail -n 10" and link hee Apr 09 11:16:01 Festivenari: was device / SoC is that? Apr 09 11:16:29 rmilecki: uh, a Netgear V7610 (BCM63168) Apr 09 11:16:44 luckily, the CFE web server is enabled Apr 09 11:17:01 Festivenari: well, take a look at target/linux/bcm63xx/image/Makefile Apr 09 11:17:25 Festivenari: that JFFS2 is most likely a bootfs partition with cferam and vmlinux Apr 09 11:17:51 rmilecki: yeah, it's a NAND one, so bcm63xx_nand.mk, to be exact Apr 09 11:17:59 I believe I just need to get the layout right and it should accept the image Apr 09 11:18:09 guess I will hexdump it for you Apr 09 11:18:32 Festivenari: well, probaably no need now Apr 09 11:18:39 it's a bcm63xx image with tail Apr 09 11:18:43 ah Apr 09 11:19:01 tail contains chekcsum, device info, flash, flags Apr 09 11:19:08 bcm63xx supports all of that I believe Apr 09 11:20:05 >Voice Board Configuration (0-19) : SI32260_3050 Apr 09 11:20:08 hmm... is that why? Apr 09 11:20:17 I see it's in the original firmware Apr 09 11:20:33 if I use a hex editor to search for it Apr 09 11:22:23 oops... just killed something related to dbus while trying to kill mono Apr 09 11:22:46 Festivenari: what's your question, again/ Apr 09 11:22:47 ? Apr 09 11:23:22 rmilecki: uh, well, when I input "print" into the CFE console, I can see that there's a "voice board" configuration option Apr 09 11:23:42 The original firmware happens to have that particular value in it Apr 09 11:23:52 but not the OpenWRT one; I guess it needs to be added? Apr 09 11:24:16 *"p" into the CFE console Apr 09 11:24:57 i don't know, do more debugging Apr 09 11:25:02 see where that string comes from Apr 09 11:25:08 cferom? cferam? Linux kernel? Apr 09 11:25:19 yeah, gonna see Apr 09 11:26:40 yeah, doesn't seem like that matters Apr 09 11:49:11 Where does the value for CFE_WFI_VERSION come from? Apr 09 11:53:05 Festivenari: how does it matter? Apr 09 11:53:20 Festivenari: someone took it from existing image or found in GPL package Apr 09 11:53:28 ah Apr 09 11:53:45 I do have the GPL sources, however Apr 09 15:39:45 aparcar[m]: hi Apr 09 15:40:02 did you had a chance to check the issue regarding keys in the imagebuilder? Apr 09 15:40:11 https://git.openwrt.org/?p=openwrt/openwrt.git;a=commit;h=418362b1cc106b9aca3905150199f60548906fff Apr 09 15:40:28 it's pretty simple to simulate: Apr 09 15:40:49 - Extract the imagebuilder. go into the folder. try to run make package_index Apr 09 15:41:06 Generating index for package ./wpa-supplicant-basic_2020-06-08-5a8b3662-32_mips_24kc.ipk Apr 09 15:41:08 Generating index for package ./zlib_1.2.11-3_mips_24kc.ipk Apr 09 15:41:09 Cannot open file '/home/koen/firmware/cus_16626/openwrt-imagebuilder-ath79-mikrotik.Linux-x86_64/key-build' for reading Apr 09 15:41:11 Makefile:136: recipe for target 'package_index' failed Apr 09 16:32:59 any hint why this build has failed? https://buildbot.openwrt.org/master/images/#/builders/71/builds/1 Apr 09 16:33:03 make: *** [/builder/shared-workdir/build/include/toplevel.mk:230: package/compile] Error 2 Apr 09 18:07:31 xback: thanks I'll look into that Apr 09 21:31:13 Build [#44](https://buildbot.openwrt.org/master/images/#builders/51/builds/44) of `ramips/rt3883` failed. Apr 10 00:06:27 hrm dango's not here Apr 10 00:08:04 dangole_: ? Apr 10 00:12:35 xback: ping Apr 10 00:13:54 aparcar[m]: not showing up for me Apr 10 00:15:28 xback: I just tried to reproduce this but it work just fine with x86/64 and it doesn't look arch specific... Apr 10 00:15:37 mangix: I guess he's afk Apr 10 00:24:14 no i mean in the user list Apr 10 00:30:42 /home/mangix/devstuff/test-openwrt/staging_dir/target-mips_24kc_musl/usr/lib/libnfnetlink.so: undefined reference to `__time64' . funny. guess this happens with a dirty build_dir Apr 10 00:56:49 wow found an amazing bug **** ENDING LOGGING AT Sat Apr 10 02:59:57 2021