**** BEGIN LOGGING AT Mon Mar 18 02:59:58 2013 Mar 18 07:25:22 cyrus r36083 trunk/package/network/ipv6/odhcp6c/Makefile Mar 18 07:25:22 odhcp6c: fix preference handling, custom client-ID Mar 18 07:25:22 * thanks to mbidewell for investigating Mar 18 07:57:41 hello, how do i compile a single kernel module? i forgot to include kmod-nls-cp437 and make package/kernel/compile doesn't seem to generate the package i need.. Mar 18 08:21:35 dape_x61: make target/linux/compile package/kernel/compile - note that you will have to reflash since you changed the kernel config and you can't just install the kmod afterwards Mar 18 08:22:08 ah, got it, so reflash anyways, i cant just do a separate module.. Mar 18 08:22:25 dape_x61: for explanation: target/linux/compile compiles the .ko; package/kernel/compile creates the ipk Mar 18 08:22:39 thanks! Mar 18 08:24:22 dape_x61: reasoning is that many kernel modules implicitly enable subsystems (/support code) in the kernel, which make the new compiled module incompatible with the kernel, so we force that the kernel config used for building the image must match the kernel config for the modules Mar 18 08:24:45 so basically recheck and build everything to make sure its in sync Mar 18 08:24:53 yes Mar 18 08:26:00 (so if you want to avoid to rebuild because you forgot something, select all modules as =m - with a bit of sed magic this should be easily doable automated - but don't ask me how ;) Mar 18 08:27:29 okay, thanks for detailed info, if you have time to look over https://dev.openwrt.org/ticket/13193 , i dont recall that alsa-lib pulling evinput and snd module dependencies few hours ago Mar 18 08:30:05 kmod-sound-core seem to be selected as module now Mar 18 08:57:46 why does alix2 get CONFIG_PCI_SUPPORT, but x86-generic doesn't ? Mar 18 08:58:52 it did on 3.3, but not on 3.7 Mar 18 09:07:24 was okay on r35472 Mar 18 09:08:34 and on r35754 Mar 18 09:17:34 wtf, now i'm seeing it again??? Mar 18 09:17:41 heisenbug! Mar 18 09:18:15 oh, duh, no. Mar 18 09:24:16 looks like something had gone wonky in my build tree, somehow in the act of bisecting it cured itself? Mar 18 09:30:02 well, i got some stuff selected after syncing svn today, this morning, some unwanted sound modules Mar 18 09:40:30 stintel: HEAD is working for me on x86-generic-net48xx (it wasn't last night) Mar 18 09:40:38 including openvpn Mar 18 09:40:49 /dev/net shows up, etc Mar 18 09:41:05 ok, I'll retest when back @ home Mar 18 09:41:14 or I hope to Mar 18 09:41:21 new macbook arrives today Mar 18 09:41:28 I might be doing other stuff Mar 18 09:41:28 ;) Mar 18 10:10:31 build #197 of rb532 is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/197 Mar 18 10:14:36 I'm packaging a app that only have one binary file and one html file, but they be located in the same folder Mar 18 10:14:38 shall I place them both in /usr/bin/myapp/? Mar 18 10:14:57 *they must be located in the same folder Mar 18 10:15:50 build #197 of ppc44x is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/197 Mar 18 11:18:56 jow r36084 packages/net/net-tools/ patches/001-config.patch Makefile * net-tools: since we ship "slattach" it makes sense to enable SLIP support Mar 18 11:41:08 build #187 of sibyte is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/187 Mar 18 11:51:47 KanjiMonster: sed 's/^# \(CONFIG_PACKAGE_kmod-.*\) is not set/\1=m/' .config Mar 18 14:23:25 build #217 of orion is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/217 Mar 18 14:46:11 build #192 of xburst is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/192 Mar 18 16:12:03 nbd r36085 trunk/package/network/ services/hostapd/Makefile services/hostapd/patches/700-ubus_support.patch * hostapd: fix build errors on supplicant-only builds Mar 18 16:59:35 <_trine> can someone include a link to this in openwrt Mar 18 16:59:43 <_trine> ERROR: module '/home/draco/docknew/trunk/build_dir/target-arm_v5te_uClibc-0.9.33.2_eabi/linux-kirkwood/linux-3.3.8/sound/core/snd-compress.ko' is missing. Mar 18 17:05:02 <[florian]> _trine: will fix that Mar 18 17:05:13 <_trine> thanks Mar 18 17:05:51 <_trine> the dockstar code is still running in 3.3.8 most of the other platforms have moved up Mar 18 17:54:25 _trine: Exactly. Mar 18 18:14:20 <_trine> [florian], could you give me a shout when its fixed please? Mar 18 18:14:45 <[florian]> _trine: within 20 minutes or so Mar 18 18:14:55 <_trine> thanks again Mar 18 18:39:45 <[florian]> _trine: can't push the fix just now as I have to move Mar 18 18:39:57 what would happen if fs_128 was used when fs_64 was required? Mar 18 18:39:58 <[florian]> _trine: if you want to fix it locally just edit sound/core/Kconfig Mar 18 18:39:59 <_trine> ok np Mar 18 18:40:15 <[florian]> _trine: and add a string description for SND_COMPRESS_OFFLOAD Mar 18 18:40:26 I'm just wondering if what I am seeing is a result of that. I'm rebuilding with 64 just to test. Mar 18 19:03:17 nunojpg r36086 packages/utils/ dump1090/Makefile dump1090 * [packages] add dump1090, Mode S decoder for the Realtek RTL2832U Mar 18 19:07:01 nunojpg r36087 packages/net/sshtunnel/files/sshtunnel.init Mar 18 19:07:01 [packages] sshtunnel: port value in server section is not read Signed-off-by: Ingo Schmidt Mar 18 19:32:28 There are plans to add realtek wifi N support? Mar 18 19:33:10 rtl8188 and rtl8192 Mar 18 19:33:44 there are some linux upstream drivers for it, but i don't know how good AP mode support and stability are Mar 18 19:34:03 i have yet to see any realtek hardware that doesn't suck :) Mar 18 19:34:14 xD Mar 18 19:34:35 yes, the rtl8187 is very very bad Mar 18 19:36:40 their ethernet MACs are bad, their SoC designs are bad, they have many extremely bad wifi chipsets Mar 18 19:36:58 they seem to suck at everything except making things extremely cheap Mar 18 19:36:59 ;) Mar 18 19:37:18 i have a rtl8188 and, with debian works, but not tested it much Mar 18 19:37:41 yes, realtek make the things in his own way Mar 18 19:38:26 so while i'm ok with accepting patches to add the drivers, i will not do any testing or fixing work on them Mar 18 19:38:37 i prefer to work on decent hardware ;) Mar 18 19:38:48 of course Mar 18 19:38:49 nbd: touche Mar 18 19:38:56 JAJAJA Mar 18 19:42:27 i want ralink rt3883 wlan support :( Mar 18 19:42:44 I want ponies! Mar 18 19:43:04 found this: http://patchwork.openwrt.org/patch/1960/ If you want i can test it Mar 18 19:43:14 ponies? Mar 18 19:43:17 KanjiMonster: ponicorns! Mar 18 19:43:27 xD Mar 18 19:43:55 i don't think wanting wlan support on a wireless router is such an outlandish wish :P Mar 18 19:44:32 Interesting admin name, "[no address given]" ... Mar 18 19:45:34 Pteridium: but I think I know which patch this is, it applies, but is missing most actual dependencies Mar 18 19:46:23 KanjiMonster: i want ponies aswell Mar 18 19:46:25 pink Mar 18 19:46:27 \o/ Mar 18 19:46:29 the drivers have some ugly implicit ones; I haven't found a way of expressing that correctly in Kconfig speak Mar 18 19:46:45 blogic: just like /. a few years ago ;D Mar 18 19:47:26 awesome Mar 18 19:47:28 i supose it, but i don't know how to "repair" the dependencies Mar 18 19:47:46 sadly i am not programmer Mar 18 19:48:01 only PC asm Mar 18 19:48:07 PIC* Mar 18 19:48:25 Pteridium: I tried, and I failed, then omg squirrel and I forgot to work on it more Mar 18 19:48:59 no problem, it was an idea Mar 18 19:49:38 Pteridium: maybe I should just commit it in the current almost working state Mar 18 19:49:38 Can be enabled in the kernel? Mar 18 19:50:18 sorry, my english skills are no t good Mar 18 19:50:23 not* Mar 18 19:50:37 staging drivers Mar 18 19:52:25 thanks KanjiMonster Mar 18 19:52:56 maybe is time to use my brain harder Mar 18 19:53:28 nbd: is there any way of cleanly expressing that kmod-a depends on kmod-b if kmod-b is selected in menuconfig KCONFIG speak? I couldn't get it to properly propage m=>y transitions Mar 18 19:54:40 openwrt menuconfig or kernel menuconfig? Mar 18 19:54:46 openwrt menucofnig Mar 18 19:54:50 hm Mar 18 19:55:02 kernel menuconfig dependencies are already broken Mar 18 19:55:06 not sure if it can be done without tripping onto the recursive dep check Mar 18 19:55:23 blogic: with the latest spi-gpio fixes for the DGN3500 (rev 35989), can this device now be flashed with openwrt? Mar 18 19:55:26 my guess would be +PACKAGE_kmod-b:kmod-b Mar 18 19:56:12 nbd: IIRC that one's satisfied with kmod-a=y and kmod-b=m Mar 18 19:56:28 it shouldn't be Mar 18 19:56:47 because it should emit a select PACKAGE_kmod-b if PACKAGE_kmod-b Mar 18 19:57:06 lemme check what I have Mar 18 19:57:10 and select should enforce that the symbol being selected has at least the value of the symbol containing the select Mar 18 19:57:17 sungami: yes Mar 18 19:57:24 if that line is satisfied with a=y and b=m, check the generated code Mar 18 19:57:31 the firstboot takes 5 minutes as flash writing is really slow Mar 18 19:57:33 maybe it's something that needs to be fixed in metadata.pl Mar 18 19:57:41 sungami: but ti works really apart from that Mar 18 19:58:55 nbd: yepp; if I select kmod-usb as m and rtlwifi as y kmod-usb stays m Mar 18 19:59:15 ok, check the generated code Mar 18 20:03:45 nbd: it doesn't generate any selects, only a "depends !(PACKAGE_kmod-usb-core) || USB_SUPPORT" Mar 18 20:04:14 ok, needs a fix in metadata.pl then Mar 18 20:07:22 nbd: I guess it's the transitive depends that cause issues; this is also causing recursive dependencies if anything +depending on kmod-rtlwifi +depends on kmod-usb directly Mar 18 20:09:42 nbd: if you want I can give you my WIP patch to play around Mar 18 20:09:48 blogic: the instructions on the wiki still say to boot via tftpboot from uboot. Are there any updated instructions available somewhere? Mar 18 20:11:48 sungami: yes sort of Mar 18 20:12:07 there is a hack you need to apply to bypass the bootloader checksum foo Mar 18 20:12:25 the bootloader will only boot signed images Mar 18 20:12:32 but i found a way to bypass it Mar 18 20:12:45 PKG_RELEASE should be counted from 0 or 1? Mar 18 20:13:23 1 Mar 18 20:13:42 Thanks. Mar 18 20:16:06 sungami: i should update the wiki tomorrow Mar 18 20:19:07 blogic: great, thanks. Mar 18 20:32:21 Why is not included ltq-hcd in the kernel? Compatibility problem? Mar 18 20:33:05 It would be useful for ext4 images. Mar 18 21:06:19 KanjiMonster: testing Realtek wifi N with this "patch": https://lists.openwrt.org/pipermail/openwrt-devel/2012-March/014424.html Hope there will not be compilation problems. Mar 18 21:06:44 If not tomorrow i will test the driver. Mar 18 21:06:57 If not, tomorrow i will test the driver.* Mar 18 21:07:35 Pteridium: it should work, it just misses some dependencies (so e.g. make sure to have e.g. ath9k selected as =m or else hostapd will be missing 11n support Mar 18 21:08:43 the first step is to test if it is stable Mar 18 21:09:40 Why depends of ath9k? Mar 18 21:13:19 Pteridium: because the realtek drivers should select DRIVER_11N_SUPPORT, but they don't - so you have to have a different driver selecting DRIVER_11N_SUPPORT enabled - and you only need to have it enabled in menuconfig as =m, you don't need to have it installed Mar 18 21:13:48 ok, thanks Mar 18 21:14:00 compiling... **** ENDING LOGGING AT Tue Mar 19 02:59:58 2013