**** BEGIN LOGGING AT Mon Dec 07 02:59:57 2009 Dec 07 10:13:34 lars * r18673 /trunk/target/linux/s3c24xx/ (config-2.6.31 files-2.6.31/ patches-2.6.31/): [s3c24xx] Remove 2.6.31 support. Dec 07 12:49:17 Hi, I'm trying to build a ramdisk image for ALIX boards with latest trunk version of OpenWRT (revision 18672), but my image doesn't start and I don't known why Dec 07 12:49:45 last lines of log are: Dec 07 12:49:59 VFS: Cannot open root device "ram" or unknown-block(1,0) Dec 07 12:50:15 Please append a correct "root=" boot option; here are the available partitions: Dec 07 12:50:22 ... Dec 07 12:50:53 follwed by sda, sda1, etc... Dec 07 12:51:09 but I don't want to boot on disk, only from ramdisk.... Dec 07 12:51:58 <[florian]> hi cedric_ricard Dec 07 12:52:20 <[florian]> cedric_ricard: are you doing a "ramdisk" Target Image ? Dec 07 12:52:42 yes Dec 07 12:52:56 full log here: http://openwrt.pastebin.com/m4ce87242 Dec 07 12:52:58 <[florian]> ok, so it should be able to find the ramdisk by itself Dec 07 12:53:22 <[florian]> cedric_ricard: can you remove the root=/dev/ram part ? Dec 07 12:53:41 [florian]: ok, I'll try Dec 07 12:54:51 ok, problem is different now Dec 07 12:55:01 the same as friday Dec 07 12:55:47 <[florian]> cedric_ricard: can you show me the different log ? Dec 07 12:57:23 I have updated the previous log : http://openwrt.pastebin.com/m309538b8 Dec 07 12:59:41 <[florian]> from this line "VFS: Mounted root (ext2 filesystem) readonly on device 8:1." Dec 07 12:59:55 <[florian]> it mounted the compact flash, and I assume this is not what you want right? Dec 07 13:00:33 [florian]: yes, I just want to test this image, booting it from ramdisk, and only ramdisk Dec 07 13:01:57 <[florian]> cedric_ricard: can you try without the rootwait option? I want to check something Dec 07 13:02:09 [florian]: ok Dec 07 13:02:58 [florian]: I should be working on usbroot next week and the updated preinit (with the changes Felix wanted) Dec 07 13:03:12 <[florian]> cshore1: cool Dec 07 13:04:28 [florian]: same problem :( Dec 07 13:04:40 <[florian]> cedric_ricard: ok, let me look for a solution to this Dec 07 13:05:27 I have removed the hard drive to test, but the problem is the same, except the last line of log : Dec 07 13:06:08 VFS: Cannot open root device "" or unknown-block(8,1) Dec 07 13:06:42 <[florian]> cedric_ricard: oh, can you tell me which compression method do you have when doing make menuconfig -> Target Images -> Ramdisk compression? Dec 07 13:06:47 <[florian]> cedric_ricard: you might want to try with lzma Dec 07 13:07:14 [florian]: compression was 'None' Dec 07 13:07:22 <[florian]> cedric_ricard: ok Dec 07 13:07:36 <[florian]> cedric_ricard: can you try with lzma? Dec 07 13:08:27 [florian]: ok, give me few minutes... Dec 07 13:09:07 <[florian]> cedric_ricard: no problem Dec 07 13:25:14 [florian]: same problem with lzma Dec 07 13:25:33 last line is "Kernel panic - not syncing: junk in compressed archive" Dec 07 13:27:39 <[florian]> cedric_ricard: ok, so it cannot properly find the initramfs Dec 07 13:29:37 And can it be a problem that comes from my build ? or a new bug in OpenWRT ? Dec 07 13:29:59 <[florian]> cedric_ricard: I do not think the problem comes from your build Dec 07 13:30:12 <[florian]> cedric_ricard: but it probably comes from one or several patch that we have added in openwrt Dec 07 13:33:19 [florian]: Can you give me some lead to help me finding (and I hope fixing) this problem, please? As you known, i'm not an Linux Kernel expert... Dec 07 13:35:15 <[florian]> cedric_ricard: can you try removing that patch: target/linux/generic-2.6/patches-2.6.31/053-lzo_compression_for_initramfs.patch? Dec 07 13:35:34 [florian]: ok, thanks Dec 07 13:36:48 [florian]: if it can't find preinit, I doubt that's the issue Dec 07 13:37:38 <[florian]> Kaloz: I am wondering it we are not having an issue here with the kernel not able to find initramfs at all Dec 07 13:38:06 VFS: Mounted root (ext2 filesystem) readonly on device 8:1. Dec 07 13:38:37 <[florian]> I am also having that issue with uml here Dec 07 13:38:56 uml was broken since day 1 or 2 ;) Dec 07 13:39:08 <[florian]> na na, I had it working a couple of times :p Dec 07 13:39:32 every time you fixed it temporary ;) Dec 07 13:39:57 [florian]: should I keep the lzma compression option? Dec 07 13:40:10 <[florian]> cedric_ricard: no you can restore your original seting Dec 07 13:41:09 [florian]: ok, and I just need to do a 'make package/kernel-clean' followed by a make, isn't it ? Dec 07 13:41:52 <[florian]> cedric_ricard: make target/linux/clean, make should be sufficient Dec 07 13:42:27 [florian]: ok, thanks Dec 07 14:13:14 [florian]: same thing. Kernel panic - not syncing: No init found. Dec 07 14:14:52 <[florian]> cedric_ricard: it was not meant to change anything Dec 07 14:14:59 <[florian]> cedric_ricard: let me dig this problem Dec 07 14:15:11 Maybe this is not related to ramdisk, because my first tests, friday, were done with ext2 image, directly copied to hardrive. And I had the same problem... Dec 07 14:19:01 <[florian]> cedric_ricard: something is still wrong as it cannot mount your ramdisk Dec 07 16:48:16 florian * r18674 /packages/net/mdnsresponder/ (7 files in 3 dirs): [package] add mDNSReponder (#6102) Dec 07 16:48:19 florian * r18675 /packages/libs/libsoup/ (. Makefile): [package] add libsoup (#6297) Dec 07 16:48:22 florian * r18676 /packages/libs/libdmapsharing/ (. Makefile): [package] add libdmapsharing (#6298) Dec 07 16:48:24 florian * r18677 /packages/net/dmapd/ (. Makefile files/ files/dmapd.init): [package] add dmapd (#6299) Dec 07 16:48:26 florian * r18678 /packages/multimedia/gst-plugins-ugly/ (5 files in 2 dirs): [package] add gst-plugins-ugly (#6300) Dec 07 16:48:27 florian * r18679 /packages/multimedia/gstreamer/ (3 files in 2 dirs): [package] update gstreamer to 0.10.25 (#6301) Dec 07 16:48:29 florian * r18680 /packages/multimedia/gst-plugins-base/ (4 files in 2 dirs): Dec 07 16:48:29 [package] update gst-plugins-base to 0.10.25, modularize plugin building Dec 07 16:48:30 (#6302) Dec 07 16:48:34 florian * r18681 /packages/multimedia/gst-plugins-good/ (Makefile patches/ patches/030-no_shave-make.patch): [package] update gst-plugins-good to 0.10.17, modularize build (#6303) Dec 07 16:55:35 florian * r18682 /trunk/target/linux/rdc/image/mkimg_sitecom.pl: [rdc] make mkimg_sitecom.pl executable Dec 07 16:56:02 florian * r18683 /trunk/target/linux/uml/README: [uml] update README after r18624 Dec 07 16:56:07 florian * r18684 /trunk/package/busybox/ (35 files in 19 dirs): [package] update busybox to 1.15.2 (#5926) Dec 07 16:58:32 hmmm.. any ideas why i got problems with ar5k in master mode (wpa2) on Dec 07 16:58:34 Kamikaze (r18640)? Dec 07 17:03:32 details please Dec 07 17:03:43 i cannot see the network Dec 07 17:04:08 its a wl500gP v1 with winston cm-9 atheros minipci Dec 07 17:04:41 so either my notebook doesnt do 5ghz wifi anymore and or the driver is broken, or the atheros does something strange Dec 07 17:04:51 it 'looks fine' so far. Dec 07 17:05:11 most likely some configuration issue leading to hostapd problems Dec 07 17:05:15 like an invalid channel etc. Dec 07 17:05:37 well.. its not like ar5k does leave me much choices Dec 07 17:05:44 kill hostapd, start it with -dd and the generated config in /var/etc/ Dec 07 17:06:07 e.g. it would be helpful to be able to set the region to europe Dec 07 17:06:13 us channels suck Dec 07 17:06:16 is crda installed? Dec 07 17:06:35 ? Dec 07 17:06:42 iw reg get/set is what i use Dec 07 17:06:50 does not work w/o crda Dec 07 17:06:58 whats crda? Dec 07 17:07:07 the regulatory database used by cfg80211 Dec 07 17:07:46 argh... fuckers, cant we just add a compile-time option to ignore that and leave it to me(the user) to do the right thing? Dec 07 17:08:02 in the end i'Ăll override it anyhow, so its for no gain Dec 07 17:08:23 well if you're able to override it anyhow you'll sure find out whats the actual problem Dec 07 17:08:30 sorry.. didnt mean you or so.. just annoyed by some coders blockwart-madness Dec 07 17:08:59 jow_laptop i dont't think its the channel. tried multiple ones, like 36, 40, 48 Dec 07 17:09:16 please do what I said and start hostapd with debugging options Dec 07 17:09:28 throw the log in pastebin and post it here Dec 07 17:09:36 else it is just guessign around Dec 07 17:10:08 will do in a moment Dec 07 17:15:36 btw... is the default fw-config the same on svn like on 'releases' ? Dec 07 17:17:08 wtf... hostapd isnt running at all... its installed, but not running Dec 07 17:17:46 try to run it manually with the generated /var/etc/hostapd.conf Dec 07 17:17:57 doesnt exist Dec 07 17:18:10 what's your wireless config? Dec 07 17:18:11 it doesnt generate the config Dec 07 17:18:47 oh sorry Dec 07 17:18:50 http://pastebin.ca/1705742 Dec 07 17:18:58 proper patch is /var/run/hostapd-$ifname.conf Dec 07 17:19:05 -t Dec 07 17:19:15 patch? Dec 07 17:19:35 *path Dec 07 17:20:48 http://pastebin.ca/1705747 Dec 07 17:21:10 iw list says otherwise Dec 07 17:21:39 http://pastebin.ca/1705749 Dec 07 17:23:19 to make it short.. i repllaced a wrt54 classic with 2.4brmc kernel with a wl500gP with atheros wifi and a fonera (for 2.4ghz) since the broadcom wifi is slow, instable and does not perform at all when using wpa2 and doing 16mbit dsl pppoe on it Dec 07 17:24:13 the fonera just bridges the 2.4ghz wifi to a port on the wl500g. if now the 5ghz wifi would work i would be done... and hopefully be done with performance issues also Dec 07 17:25:04 check the generated /var/run/hostapd-wlan0.conf and maybe google for "IEEE 802.11 Configured channel (36) not found from the channel list of current mode (2) IEEE 802.11a" Dec 07 17:25:10 and retry with crda installed Dec 07 17:25:16 I have to run now. bbl Dec 07 17:25:24 jow_laptop thanky... Dec 07 17:25:27 irgh Dec 07 17:53:41 hm.. installing crda doesnt help Dec 07 17:59:33 nbd ping Dec 07 18:01:04 will changeset 18662-18665 fix my problem? Dec 07 18:05:30 pong Dec 07 18:06:04 probably won't Dec 07 18:06:39 uh.. how so? Dec 07 18:06:42 but i have an idea what you can try Dec 07 18:07:06 there is a regulatory related config option in openwrt make menuconfig near the kmod-ath module Dec 07 18:07:09 enable that Dec 07 18:07:21 and then copy http://nbd.name/regulatory.bin to /usr/lib/crda/ Dec 07 18:07:32 ? Dec 07 18:07:47 i have installed crda now. and it allow setting DE and lists correct channels now Dec 07 18:07:53 yes Dec 07 18:07:54 only hostapd seems stupid Dec 07 18:07:58 but the channels are marked as DFS Dec 07 18:08:04 so it probably refuses to use them Dec 07 18:08:09 since ath9k does not have DFS support Dec 07 18:08:15 i use ar5k Dec 07 18:08:20 same Dec 07 18:08:22 winston cm9 Dec 07 18:08:27 both share the same regulatory domain handling code Dec 07 18:10:14 only replacing the bin doesnt help Dec 07 18:13:20 i know Dec 07 18:13:26 you need to recompile the kmod-ath package Dec 07 18:17:38 nbd after doing what exactly? (config-change?) Dec 07 18:19:48 config change, then make package/mac80211/compile Dec 07 18:19:58 well.. chat config change Dec 07 18:20:22 fsck.. i did a svn up in that tree Dec 07 18:21:18 CONFIG_ATH_USER_REGD ? Dec 07 18:22:38 fsck.. the svn up broke only rebuilding that Dec 07 18:24:56 nbd perhaps it was naive to think one can just compile, and use atheros based stuff for as 5ghz ap ;) Dec 07 18:25:26 CONFIG_ATH_USER_REGD, yes Dec 07 19:22:02 hm.. doesnt build anymore... seems a make clean is in order Dec 07 19:22:24 is there a openwrt developer good practices guide somewhere? Dec 07 19:37:59 not yet Dec 07 19:40:05 hm.. it doesnt want to build... need a distclean perhaps? Dec 07 19:40:17 show me the error, please Dec 07 19:41:31 http://pastebin.ca/1705926 , but i did make distclean since then Dec 07 19:42:21 i don't think distclean helps there Dec 07 19:42:25 no idea what's causing that error, though Dec 07 19:42:34 did you try make oldconfig? Dec 07 19:42:44 nope Dec 07 19:42:46 that's always a good idea after svn up Dec 07 19:42:52 just like with the kernel ;) Dec 07 19:43:05 i copied out my .config, did make distclean, copied it back, saved config, make Dec 07 22:03:15 fsck. Dec 07 22:03:23 i think i broke my install. Dec 07 22:03:46 doesnt fork any shell anymore Dec 07 22:07:25 hangs in the shell which replaces busybox Dec 07 23:39:23 florian * r18685 /packages/libs/libpng/ (Makefile patches/100-config_fix.patch): [package] update libpng to 1.2.41 (#6312) Dec 07 23:39:25 florian * r18686 /packages/libs/tiff/ (5 files in 2 dirs): [package] update tiff to 3.9.2 (#6316) Dec 07 23:39:27 florian * r18687 /packages/multimedia/gst-plugins-good/Makefile: [package] fix gst-plugins-good ipkg directory Dec 07 23:39:28 florian * r18688 /packages/net/hiawatha/Makefile: [package] update hiawatha to 6.19 (#6317) Dec 07 23:47:49 florian * r18689 /packages/sound/scmpc/Makefile: [package] add missing libpthread dependency to scmpc (#6290) Dec 07 23:59:08 florian * r18690 /trunk/package/kernel/modules/fs.mk: [package] package nls cp866 support (#6281) Dec 08 00:29:08 nbd ping Dec 08 00:46:19 found another bug. clicking on routes in luci admin mode brings up this: http://pastebin.ca/1706265 and ff barfs about invalid xml Dec 08 00:48:19 yeah.. wifi seems to work with your db on channel 36 Dec 08 00:48:28 roh: will look into it tomorrow Dec 08 00:48:44 xMff the luci thing? thanks Dec 08 00:49:37 i guess something should happen with hostapd too. it just exiting and no feedback to the webinterface when it doesn like some regulatory domain setting Dec 08 00:50:03 I know Dec 08 00:50:15 but I consider this a responsibility of the base system :) Dec 08 00:50:51 *sigh* Dec 08 00:51:08 we're working on bringing more wifi feedback to the ui but there are certain corner cases that can't be reliably checked Dec 08 00:51:22 well.. i consider myself a 'avanced poweruser#' .. i know about linux and all its dirty and niceness Dec 08 00:51:26 like the error you had, even iw aggreed on the channel Dec 08 00:51:52 the ui can't know it better then the underlying tools Dec 08 00:51:55 i basically use openwrt because its more work to do it myself, and somebody needs to maintain it to provide updates Dec 08 00:52:41 and i am too annoyed by the 'progess' of wireless on linux to follow all details myself.. you guys have that seemingly much better under control **** ENDING LOGGING AT Tue Dec 08 02:59:56 2009