**** BEGIN LOGGING AT Fri Mar 16 02:59:58 2012 Mar 16 06:23:48 Hi I am using openwrt in 703N Mar 16 06:24:10 when I use trunk '29936' it works fine. Mar 16 06:24:25 but we I update my repo to 30834. Mar 16 06:24:37 it can't mount /dev/mtdblock3 anymore. it give: Mar 16 06:24:48 root@OpenWrt:/# mount /dev/mtdblock3 /mnt -t jffs2 Mar 16 06:24:49 [ 64.420000] jffs2: Too few erase blocks (4) Mar 16 06:24:49 mount: mounting /dev/mtdblock3 on /mnt failed: Invalid argument Mar 16 06:28:32 added comment at: https://dev.openwrt.org/ticket/10719#comment:3 Mar 16 06:41:47 by remove some packages. re-compile again. now it back to work. Mar 16 09:52:27 nbd * r30954 /trunk/target/linux/generic/ (2 files in 2 dirs): kernel: optimize out remaining netfilter hooks in the bridging code if bridge filtering is disabled Mar 16 13:53:30 has there been any work on multiuser support for luci ? Mar 16 13:53:52 no Mar 16 14:09:28 jow_laptop: some how those two words dont seem to belong in the same sentence Mar 16 14:10:35 well I think its about restricted access Mar 16 14:10:43 which is hard to implement right Mar 16 14:10:58 any tacked on solution will inevitably be full of weaknesses Mar 16 14:27:48 jow_laptop but someone really interested can develop a luci like interface with restricted access, right? Mar 16 14:28:36 sure Mar 16 14:28:48 lua, cgi-bin, wrapping stuff.. Mar 16 14:31:06 often the requirement is a read-only user, or allow users to control only specific addons, like for example the ftp server or something Mar 16 14:31:27 without granting access to firewall or dnsmasq or similar Mar 16 14:31:53 however the whole security is circumvented as soon as there is a way to do shell commands Mar 16 14:32:02 for example through the custom rules box in the firewall settings Mar 16 14:32:35 so security restrictions do not only apply per-page but also per-option in some cases Mar 16 14:32:40 thats where its getting complex Mar 16 14:33:12 implementing read-only users and per-page access restrictions is on my todo Mar 16 14:55:04 neat Mar 16 15:02:24 build #137 of ps3 is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/ps3/builds/137 Mar 16 15:35:12 build #122 of avr32 is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/122 Mar 16 16:05:31 juhosg * r30955 /trunk/target/linux/mpc52xx/ (Makefile config-2.6.30 config-3.2): mpc52xx: switch to 3.2.9 Mar 16 17:09:21 seeing http://fpaste.org/iOxE/ ... not sure why. Mar 16 17:09:58 earlier config-n.n versions have CONFIG_NF_CONNTRACK_MARK enabled... but not config-3.0 and later. Mar 16 18:30:53 juhosg * r30956 /trunk/target/linux/ar71xx/ (base-files/etc/diag.sh files/arch/mips/ath79/mach-ja76pf.c): ar71xx: change LED name prefix on the jjplus boards Mar 16 18:31:02 juhosg * r30957 /trunk/target/linux/ar71xx/ (7 files in 7 dirs): ar71xx: add support for the jjPlus JA76PF2 board Mar 16 18:31:07 juhosg * r30958 /trunk/target/linux/ar71xx/ (generic/profiles/jjplus.mk image/Makefile): ar71xx: add profile and generate image for the JA76PF2 board Mar 16 18:31:21 philipp64|laptop: did you run make defconfig ? Mar 16 18:39:23 philipp64|laptop: try this: http://luci.subsignal.org/~jow/xtables-require-conntrack-mark.patch Mar 16 20:21:42 jow_laptop: think it also needs CONFIG_NF_CONNTRACK_MARK Mar 16 20:22:21 besides CONFIG_NF_CONNTRACK_MARK ? ;) Mar 16 20:29:03 sorry I mean CONFIG_NF_CONNTRACK Mar 16 20:29:09 it wants both. Mar 16 20:29:51 if CONFIG_NF_CONNTRACK is off you got way bigger problems than xtables Mar 16 21:19:41 [OpenWrt-Devel] Build broken in uclibc++ http://dpaste.com/hold/717581/ <-- seems still broken Mar 16 21:22:26 jow_laptop: your patch needs CONFIG_NF_CONNTRACK_MARK=y since it's a bool. Mar 16 22:53:02 nbd * r30959 /trunk/package/mac80211/patches/565-ath9k_optimize_queue_wake.patch: ath9k: remove the queue wake optimization, it may be unreliable in some cases Mar 16 23:13:48 nbd: pung Mar 16 23:13:59 Chocks: poeng Mar 16 23:14:21 is it possible to change the machine type in the platform code dynamically, if you decide a certain CPU means a specific machine? Mar 16 23:15:22 no idea Mar 16 23:25:30 Chocks: you could set it via uboot machid variable Mar 16 23:25:56 uboot doesn't know the machine; it depends upon platform code to work that out Mar 16 23:29:32 I have a combined GH300N/N2/AG300H platform code. the same image works on all 3 machines. The minor issue is that it all appears as the same machine in /proc/cpuinfo, apart from the processor type. Mar 16 23:29:57 it'd be nice to know the exact machine I'm looking at, since there are minor practical differences between the 3 such as LED mapping Mar 16 23:30:12 and the switch configuration Mar 16 23:33:07 i see Mar 16 23:33:42 can you put somewhere your kernel file so i can take a look? Mar 16 23:34:26 what do you want to look at, exactly? Mar 16 23:35:02 i would like to see it then i might get idea how you could do it Mar 16 23:35:51 the kernel file where you have combined platform code Mar 16 23:39:09 you mean, the patch to the platform code, or the uImage? Mar 16 23:40:13 both? Mar 16 23:41:47 there's other stuff that goes with it, so the kernel alone probably isn't much use. but let me pastebin my platform code Mar 16 23:44:43 http://pastebin.com/ZdBHqj1K Mar 16 23:49:48 why dont you define multiple MIPS_MACHINE? Mar 16 23:50:49 look at target/linux/lantiq/files-3.1/arch/mips/lantiq/xway/mach-arv45xx.c for example Mar 16 23:51:53 lines 518 - 524 is wrong imho Mar 17 00:00:56 * Chocks looks Mar 17 00:09:08 that relies upon the correct machine name being passed on command line or by u-boot etc Mar 17 00:09:36 in that example, the MIPS_MACHINE decides the init code to call Mar 17 00:11:19 yes Mar 17 00:11:55 and you can do it in uboot using machid Mar 17 00:12:06 yes, but I think you're missing the point Mar 17 00:12:23 you dont want to mess with uboot? Mar 17 00:12:33 in openwrt, we build separate images for each machine, even though the kernel is 100% identical apart from the machine name we hard wire in Mar 17 00:12:54 the point in my case is to have a single image I can run on the 3 similar hardware platforms I have. Mar 17 00:13:04 this works, more of less, but I want to tidy the /proc/cpuinfo case Mar 17 00:13:39 I don't really want to mess with u-boot, no. but we ignore what that sets in any case Mar 17 00:14:14 ok, i'm not familiar with that target Mar 17 00:14:34 well, the target itself doesn't really matter if I can modify the machine name Mar 17 00:16:48 i think i see what you want to do... but lines 518 - 524 seem wrong Mar 17 00:17:03 if you want to do it that way you should put that into new function at least Mar 17 02:11:03 build #106 of rdc is complete: Failure [failed compile_7] Build details are at http://buildbot.openwrt.org:8010/builders/rdc/builds/106 **** ENDING LOGGING AT Sat Mar 17 02:59:58 2012