**** BEGIN LOGGING AT Thu Aug 11 02:59:56 2011 Aug 11 06:28:00 Hello, any comments on https://lists.openwrt.org/pipermail/openwrt-devel/2011-August/011837.html ?? Aug 11 06:56:05 sekon_: use uci-defaults ? Aug 11 07:08:41 blogic: https://lists.openwrt.org/pipermail/openwrt-devel/2011-August/011859.html (the OT's response) Aug 11 07:10:36 s/OT/OP Aug 11 07:19:00 sekon_: use uci-defaults Aug 11 07:19:09 everything you describe can be covered by uci-.defaults Aug 11 09:16:48 build #45 of mpc52xx is complete: Failure [failed compile_10] Build details are at http://buildbot.openwrt.org:8010/builders/mpc52xx/builds/45 Aug 11 09:48:18 jogo * r27952 /trunk/target/linux/generic/patches-2.6.30/971-ocf_20110720.patch: Aug 11 09:48:18 kernel: 2.6.30: remove bogus additional find_task_by_vpid export from OCF Aug 11 09:48:18 find_task_by_vpid is already exported, thus breaking builds. Aug 11 09:49:23 hauke * r27953 /trunk/target/linux/generic/patches-3.0/941-ocf_20110720.patch: Aug 11 09:49:23 kernel: fix ocf patch for kernel 2.6.30 Aug 11 09:49:23 find_task_by_vpid is already exported by the kernel without any patches needed. Aug 11 09:55:49 hauke * r27954 /trunk/target/linux/generic/patches-3.0/941-ocf_20110720.patch: Aug 11 09:55:49 kernel: revert r27953 Aug 11 09:55:49 KanjiMonster did the right patch, I was just confused because of linux-3.0 and linux-2.6.30 Aug 11 11:20:26 florian * r27955 /trunk/target/linux/realview/config-default: [realview] disable hardware perf support to work in qemu Aug 11 11:20:28 florian * r27956 /trunk/target/linux/generic/ (config-2.6.39 config-3.0 config-3.1): [kernel] add more missing touchscren symbols (3.0+) Aug 11 11:20:32 florian * r27957 /trunk/target/linux/ep93xx/modules.mk: Aug 11 11:20:32 [ep93xx] package more board specific modules: Aug 11 11:20:32 - keyboard matrix driver Aug 11 11:20:32 - SoC sound support Aug 11 11:20:33 based on patch by KanjiMonster Aug 11 12:48:10 build #67 of brcm47xx is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/brcm47xx/builds/67 Aug 11 13:54:17 nbd * r27958 /trunk/package/mac80211/ (67 files in 2 dirs): mac80211: update to 2011-08-10 Aug 11 13:54:28 nbd * r27959 /trunk/target/linux/ar71xx/files/arch/mips/ar71xx/ (mach-dir-615-c1.c mach-tew-632brp.c): ar71xx: adjust the mtd layout of tew-632brp and dir-615c to match the image layout (fixes #9922) Aug 11 14:27:34 KanjiMonster: Are you around? Aug 11 14:47:30 I see my problem. The new Buffalo WZR has two flash banks, and in trying to keep the partitions as similar as possible, I've crossed the bank boundary. Aug 11 14:52:04 Is there a way to join two mtd partitions into a single "root_fs", or do I need to have two distinct filesystems? Aug 11 14:53:08 on the wzr with the two flash banks, the kernel first concatenates both chips into one mtd device and then sets up partitions on top of that Aug 11 14:53:21 so partitions can cross flash bank boundary Aug 11 14:58:57 nbd: well that makes me feel better.. Aug 11 14:59:23 nbd: unfortunately, the kernel only detects a 16MB flash chip, when it is, in fact, 32. Aug 11 14:59:31 I think it only sees bank 0 Aug 11 14:59:36 which wzr are you talking about? Aug 11 14:59:41 wzr-hp-ag300h? Aug 11 14:59:49 no.. Aug 11 14:59:59 it is a WZR-HP-G300NH2 Aug 11 15:00:02 oh Aug 11 15:00:07 haven't worked on that one yet Aug 11 15:00:21 you can probably copy some of the initialization code from the wzr-hp-ag300h setup Aug 11 15:00:23 marked WZR-HPG302HV2 on the PCB Aug 11 15:00:28 it will likely work pretty much the same way Aug 11 15:00:35 I'll check the ag300 Aug 11 15:00:42 the SOC is different than the HP300 Aug 11 15:00:51 what kind of soc is it? Aug 11 15:00:58 ar7241? Aug 11 15:01:17 but it's still SPI flash, right? Aug 11 15:01:22 Yes, spi flash Aug 11 15:01:23 http://pastebin.com/2gnNDrTf Aug 11 15:01:28 that's the dd-wrt boot log Aug 11 15:01:37 and.. I have the chip specifics here.. one moment Aug 11 15:01:40 ok, if it's spi flash, then you can use the same code for registering the flash chips Aug 11 15:01:51 dd-wrt implements it differently Aug 11 15:01:58 they have a hacked up flash driver that treats both as one chip Aug 11 15:02:13 Atheros AR7242 rev 1.1 (0x1101) Aug 11 15:02:17 ok Aug 11 15:02:58 i should probably refactor the wzr-hp-ag300h init to move out the mtd concat init Aug 11 15:03:08 so that it doesn't have to be duplicated for the 300nh2 Aug 11 15:03:30 if you just want to test it, then copy&paste of those parts should work Aug 11 15:03:36 I'm back in backfire, but I feel that I will have to move to trunk shortly. Aug 11 15:03:45 I don't think that backfire even has the ag300h.. ;-) Aug 11 15:03:55 yeah, you probably need trunk for that Aug 11 15:03:57 Let me find it in my trunk copy Aug 11 15:04:34 and backport. I have been monkeying with the rtl stuff since it appears that the eth chipset is somewhat random now. Aug 11 15:05:16 In general, pretty much every hardware address starting with 4C is a rtl8366rb Aug 11 15:05:49 That's not a solid rule, though. I found one with a rtl8366rb that started with 00. Aug 11 15:27:26 nbd: building now. Aug 11 15:28:18 The board definitions are not very consistent. A lot of them are done quite differently. Aug 11 15:28:45 but, perhaps it appears more that way because I am not all that familiar with them. Aug 11 15:42:49 at some point they need to be replaced with device tree files anyway Aug 11 15:43:03 but that'll be quite a bit of work Aug 11 15:43:44 mtd_concat_create -- any reason why this would not be built on backfire? Aug 11 15:43:56 I failed at link time Aug 11 15:43:59 build #66 of atheros is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/atheros/builds/66 Aug 11 15:44:35 mtdconcat is not being built. Let me check something. Aug 11 15:57:23 hmm wrt image with samba/ntp/luci/collectd crap no stripping in 8M :) -> hmm i wonder where uclibc is needed for.. -> http://pastebin.com/YG62gMXy Aug 11 15:57:28 where do we still need uclibc for ? Aug 11 15:57:32 :p Aug 11 15:58:10 glibc based.. using gentoo lastweek as toolchain. Aug 11 15:59:35 one point i still need to fix run chmod +x on /lib/*.s* and /usr/lib/*.s* Aug 11 16:01:31 big diff i dare to say now.. uclibc/eglibc -> more cpu/sysload 0.10~0.50 while idle. compared to (r27947) Load: 0.01 0.02 0.05 Aug 11 16:03:10 I can't really make sense of that. Can you start over, try to think before you type something. Thanks. Aug 11 16:03:37 regarding size, glibc is still too big for many systems. Aug 11 16:04:09 ChanServ: to start with i do not wanna start bashing on uclibc since thats the main openwrt project target nomather wat result i have on x86 non embedded. Aug 11 16:04:11 :S Aug 11 16:04:14 Chocky: Aug 11 16:04:43 other point wat do you not understand instaid of expecting me every time retype everything i say specialy for you Chocky ? Aug 11 16:05:16 Chocky: if 8M is to big then i dunno Aug 11 16:06:28 Chocky: and think about the point "nostripping" and still this small. Aug 11 16:06:53 Chocky: try that with eglibc good luck :) Aug 11 16:07:03 * WillieNL still remember +400M Aug 11 16:16:37 Chocky: toolchain is gcc 4.4.5 + binutils 2.20.1 works with less issues for me. fixed mosth problems at the first place. (just match versions with upstream distros) Aug 11 16:42:30 build #70 of at91 is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/at91/builds/70 Aug 11 16:55:55 willie: I still can't make sense of this. the point is, a "stream of consciousness" feedback is not use to me. You will save us both time by being clearer. Aug 11 16:56:38 fullstop: you need to select mtd concat support in the kernel config Aug 11 17:02:52 nbd: I made it that far. thanks! Now it Oopses during boot. Aug 11 17:03:23 I need to get a better stack trace. Aug 11 17:27:21 build #64 of orion is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/64 Aug 11 17:29:34 build #68 of ubicom32 is complete: Failure [failed compile_3] Build details are at http://buildbot.openwrt.org:8010/builders/ubicom32/builds/68 Aug 11 17:31:23 The boot parameters which uboot passes to the kernel.. are they mostly ignored / replaced by the mtd partitions in the board definition? Aug 11 17:44:16 half-booted! \o/ Aug 11 17:54:19 build #63 of ps3 is complete: Failure [failed compile_6] Build details are at http://buildbot.openwrt.org:8010/builders/ps3/builds/63 Aug 11 19:45:41 build #66 of lantiq is complete: Failure [failed compile_6] Build details are at http://buildbot.openwrt.org:8010/builders/lantiq/builds/66 Aug 11 20:34:49 nbd * r27960 /trunk/target/linux/generic/base-files/init: initramfs: allow preinit to enable switching from initramfs to tmpfs to allow overlayfs to be mounted on top of it Aug 11 21:23:34 nbd, do u have a link to the mac80211 changelog? Aug 11 21:26:12 RealOpty: https://git.kernel.org/?p=linux/kernel/git/linville/wireless-testing.git;a=history;f=net/mac80211;hb=HEAD Aug 11 21:26:20 ty Aug 11 21:28:22 build #65 of cobalt is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/cobalt/builds/65 Aug 11 21:28:34 jow * r27961 /trunk/package/iptables/Makefile: [package] iptables: drop iptables-utils and ip6tables-utils, ship needed symlinks in iptables and ip6tables packages (#9691) Aug 11 21:36:46 jow * r27962 /branches/backfire/package/iptables/Makefile: [backfire] backport r27961, deprecate iptables-utils and ip6tables-utils but leave them in place to not break backwards compatibility Aug 11 21:43:19 /win 23 Aug 11 21:43:23 \o/ Aug 11 22:11:03 nbd * r27963 /trunk/package/ (jshn/ libubox/Makefile ubus/Makefile): update libubox and ubus to latest, libubox now includes jshn Aug 11 22:29:41 build #60 of ppc44x is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/60 Aug 11 23:13:55 nbd: what do you know about what Linux does with x86 "high memory"? Aug 11 23:14:30 no Aug 11 23:14:39 nothing Aug 11 23:17:28 ok. pulling my hair out trying to figure out where that region gets remapped to. Aug 12 00:54:49 is there a good example of a "virtual" package that is only a list of dependent packages? That is, i want to select one thing and have it pull in a set of packages without itself actually containing anything. Aug 12 01:22:58 russell--: create as usual and define the Package/.../install section with "true" as only command Aug 12 02:42:06 build #70 of s3c24xx is complete: Failure [failed compile_10] Build details are at http://buildbot.openwrt.org:8010/builders/s3c24xx/builds/70 Aug 12 02:54:36 jow_laptop: awesome, thanks! **** ENDING LOGGING AT Fri Aug 12 02:59:59 2011