**** BEGIN LOGGING AT Tue Oct 02 03:00:02 2018 Oct 02 05:12:08 russell--: placeholders Oct 02 05:12:17 some bits have undefined functions Oct 02 05:12:52 blogic: Exciting new features to be discovered Oct 02 05:13:19 (morning) Oct 02 05:25:09 blogic: why some "-" and others "rsvd Oct 02 05:25:11 ? Oct 02 05:25:35 fingers-- Oct 02 06:22:09 does anyone have experience with open-plc-utils (powerline)? I would like to know what firmware header I should be looking for :) Oct 02 06:22:12 russell--: for sake of consistency i guess Oct 02 06:22:22 russell--: i'll drop the pinctrl code soonish Oct 02 06:22:28 moving it to pure OF Oct 02 06:22:36 already have the dtsi files ready Oct 02 08:32:48 morning Oct 02 08:44:37 blogic:ping Oct 02 08:46:08 ldir: your patches worked =) Oct 02 08:47:00 yay! Oct 02 08:48:04 which suggests the built-in cake workaround doesn't for whatever reason. Oct 02 08:48:20 yeah, strangely enough Oct 02 08:48:22 and all the more reason to remove it. Oct 02 08:58:12 Hello, is Felix Fietkau online? Oct 02 08:59:21 nbd: ^ Oct 02 09:00:46 yes Oct 02 09:02:25 Hi, I am trying to add in the mac80211 package the compilation of the R8188EU driver, which is included in the stagind drivers. It supports RTL8188EUS USB 802.11n chipset that are not supported by rtl8xxx driver Oct 02 09:03:19 I need some help, because it is not clear to me how the package description in the mac80211 makefile is turned in correct CFGOPT options in the autoconf generated .config in the build dir of the backport package Oct 02 09:06:05 ah wait, I may have figured it out just now, it is a matter of package naming, it seems that now .config contains what I need: CPTCFG_BACKPORTED_STAGING=y CPTCFG_BACKPORTED_R8188EU=m CPTCFG_BACKPORTED_88EU_AP_MODE=y Oct 02 09:06:52 is it possible that the kernel package name must reflect the option? I named it rtl8188eu and it did not work, now I named it r8188eu and I see the correct options set in .config Oct 02 09:09:16 the name does not need to reflect the option Oct 02 09:09:25 you just need to use it consistently throughout the makefile Oct 02 09:13:46 well I did, now with: r8188eu added in PKG_DRIVERS and with a new define KernelPackage/r8188eu entry plus config-$(call config_package,r8188eu) += R8188EU and $(eval $(call KernelPackage,r8188eu)) it is compiling the driver Oct 02 09:14:14 before with all the same except that everything was named "rtl8188eu" the .config was not generated with the CPTCFG options set Oct 02 09:14:32 now it built: LD [M] /media/storage/das1/temp/openwrt_x86/build_dir/target-x86_64_musl/linux-x86_64/backports-4.19-rc5-1/drivers/staging/rtl8188eu/r8188eu.o Oct 02 09:20:01 did you run make oldconfig after making your change to the makefile? Oct 02 09:20:15 yes I did Oct 02 09:21:13 please show me the full patch with the rtl8188eu name that wasn't working Oct 02 09:24:52 I have the new "working" Makefile with r8188eu under test, the package built returned an error:Package kmod-r8188eu is missing dependencies for the following libraries: cfg80211.ko compat.ko lib80211.ko I think I have to add +kmod_mac80211 in the DEPENDS right? Oct 02 09:25:15 anyhow I will try to revert to the old naming Oct 02 09:31:40 +kmod-mac80211 Oct 02 09:31:47 or rather +kmod-cfg80211 Oct 02 09:32:03 and +kmod-lib80211 Oct 02 09:36:11 ok, I will do it. I have the patch with old naming, how can I send it to you? Oct 02 09:39:00 Menion: https://openwrt.org/submitting-patches ;) Oct 02 09:39:36 well, yes once it works I will submit a PR on GitHub, this is for trying to understand why the rtl8188eu package name does not work Oct 02 09:44:28 Menion: when you're done, let's try to the rename again Oct 02 09:55:42 mangix: it's probably worth asking the knower of all things mr nbd :-) Oct 02 10:09:41 dedeckeh: pong, patches ! Oct 02 10:15:55 ldir: time zones :) Oct 02 10:28:19 mangix: ahh, you need to leave stuff here so he trips over it ;-) Oct 02 10:38:07 anyone here with a bthomehub? Oct 02 10:51:44 jow: if you can wait until Sunday then yes Oct 02 10:51:53 well a hh5a Oct 02 10:51:54 nbd: I believe this was fixed upstream: https://github.com/openwrt/openwrt/blob/master/target/linux/generic/pending-4.14/100-MIPS-fix-cache-flushing-for-highmem-pages.patch Oct 02 11:08:15 g khal colors Oct 02 11:08:36 sorry about that, mouse focus mishap… Oct 02 14:36:48 Hello. Is there any support left for WEXT in OpenWRT? Oct 02 14:51:12 Menion: not really, at least not as far as the integration with the configuration system is concerned Oct 02 14:51:52 which driver? Oct 02 15:03:33 driver for rtl8188eu dongle, it is still not supported by rtl8xxx mac80211 driver Oct 02 15:04:48 wouldn't you only need the wext stuff if you were trying to use old vendor drivers? Oct 02 15:08:11 yes but wpad for instance is not built with wext enabled. apparently wpa_supplicat and hostapd is, I am going to try. But it is a pity that the support is just dropped, luci for example is still able to support it Oct 02 15:52:32 jow: Several, yes, if you're asking about the 5 Oct 02 16:03:44 Mon Oct 1 19:01:57 2018 user.notice mac80211: Failed command: iw phy phy0 set antenna all all Oct 02 16:03:52 I think I've seen this on different devices Oct 02 16:04:28 Driver doesn't support antenna configuration? Oct 02 17:34:48 bah, apparently the dropbear bug where it sometimes doesn't support EC algos is still there Oct 02 17:35:45 fucking heisenbugs Oct 02 18:15:42 jow: I ma a BTHH5a Oct 02 18:16:02 I have Oct 02 18:33:02 What is the compile option openwrt uses to build its packages? Oct 02 18:33:18 Will there be any use in my compiling them using LTO Oct 02 18:35:08 it might reduce size Oct 02 18:35:15 some packages are already built with LTO Oct 02 18:35:35 I would suspect enabling it globally will break some things Oct 02 18:36:30 Oh Oct 02 18:36:42 Yeah compatibility with LTO is a thing Oct 02 18:38:00 e.g. commit that enables it for a package: https://git.lede-project.org/154c0c40 Oct 02 18:45:24 maybe we can do something like enable LTO by default and disable it with PKG_LTO:=-0 Oct 02 18:48:16 How to resolve breakage? Oct 02 18:48:44 Linux kernel has being seeing LTO compatability for sometime, especially to support embedded applications Oct 02 18:50:46 I see you meant config options alone. Oct 02 19:03:26 * ldir slightly gnashes at cake, but nevermind. Oct 02 19:08:29 demonshreder: I made the kernle compile with LTO but this needed some patches Oct 02 19:08:41 I think I never got it fully booted up Oct 02 19:12:03 Oh Oct 02 19:54:52 Hauke: You are an HH5A? :P Oct 02 19:55:34 Uh-oh, OpenWRT has become sentient! Oct 02 19:55:49 It's okay, HH5As have reboot buttons. Oct 02 19:55:58 * Monkeh reboots Hauke Oct 02 20:03:15 * ldir does a nand backup of Hauke Oct 02 20:03:33 ;-) Oct 02 20:45:28 lol Oct 02 23:48:13 hello, what does 'SQUASHFS error: xz decompression failed, data probably corrupt' mean? Oct 02 23:48:55 i'm pretty sure that the nand flash is not corrupted nor the image i'm flashing **** ENDING LOGGING AT Wed Oct 03 03:00:01 2018