**** BEGIN LOGGING AT Thu Jun 04 02:59:58 2009 Jun 04 06:35:52 hmm. having trouble with a recent build booting on wgt634u. r16322, gets as far as: Please be patient, while OpenWrt loads ... Jun 04 06:35:52 Algorithmics/MIPS FPU Emulator v1.5 Jun 04 06:36:02 then hangs indefinitely Jun 04 06:39:38 juhosg * r16323 /trunk/package/ (3 files in 2 dirs): [package] kernel/modules: move SPI drivers into a new submenu Jun 04 06:39:50 * russell_ tries an older build Jun 04 06:44:00 my build of r15577 boots (i think that's the last one i actually flashed anywhere) Jun 04 07:02:05 wtf? Jun 04 07:02:41 after the first boot, i'm seeing this: Jun 04 07:03:17 et0: chipattach: NVRAM_GET(et0phyaddr) not found Jun 04 07:03:17 et0: chipattach error Jun 04 07:03:17 et0: etc_attach failed Jun 04 07:03:17 Could not attach bcm4710 ethernet: -2 Jun 04 07:10:37 russell_: < loswillios > weird. I just flashed my wgt634u with a fresh build and after the second reboot it wouldn't go further than CFE, printing this message: et0: chipattach: NVRAM_GET(et0phyaddr) not found Jun 04 07:10:53 yeah, like that! Jun 04 07:11:22 * russell_ had to read twice before i realized that i didn't write that ... so familiar!! Jun 04 07:11:32 yes Jun 04 07:11:40 I haven't had time to investigate further Jun 04 07:11:53 did you include unvram in your image? Jun 04 07:12:08 not that i know of Jun 04 07:12:21 it's per default [*] Jun 04 07:12:29 oh, then probably Jun 04 07:12:34 * russell_ checks my .config Jun 04 07:12:45 I wouldn't know of anything else messing with nvram Jun 04 07:13:22 CONFIG_DEFAULT_nvram=y Jun 04 07:13:33 CONFIG_PACKAGE_nvram=y Jun 04 07:14:29 printenv looks abbreviated compared to this: http://forum.openwrt.org/viewtopic.php?pid=89144 Jun 04 07:22:53 the macaddrs had been reset too Jun 04 07:23:07 * russell_ restores them from the ones printed on the case Jun 04 07:23:56 booting again Jun 04 07:24:55 rebooting again seems to work too Jun 04 07:25:28 very weird, never seen that before Jun 04 07:29:14 me neither. could you post a short summary how you got it fixed in the above thread? Jun 04 07:33:30 i just replicated the values provided by KillaB, fixing the macaddr's to match the ones i originally had (according to the case) and remembering to quote long strings with embedded spaces Jun 04 07:34:34 * russell_ posting a short summary as well Jun 04 07:40:01 fwiw, i have nocatauth running on a wgt now Jun 04 07:51:30 juhosg * r16324 /trunk/target/linux/ar71xx/ (4 files in 4 dirs): [ar71xx] add SPI driver for the Vitesse VSC7385 switch Jun 04 07:54:09 russell_: nice Jun 04 07:54:11 juhosg * r16325 /trunk/package/kernel/modules/spi.mk: [package] kernel/modules: package the VSC7385 driver Jun 04 07:55:00 russell_: isn't nocatsplash the successor? Jun 04 07:56:51 juhosg * r16326 /trunk/target/linux/ar71xx/files/arch/mips/ar71xx/mach-ap83.c: [ar71xx] pass ucode name to the VSC7385 driver on the AP83 board Jun 04 07:58:09 loswillios: not nocatsplash, nocatauth ... the perl dependent one Jun 04 07:58:36 oh, yeah, (sorry, /me reads now) Jun 04 07:58:51 nocatsplash is in kind of a horrid state atm Jun 04 07:59:33 we have been using nocatauth forever at our CWN, so the fact that it: a) fits; and b) seems to work is kind of a welcome surprise. Jun 04 08:00:31 getting some complaints from iptables about DROP rules in the nat table Jun 04 08:00:46 "not intended for filtering" or something Jun 04 08:07:44 heh, ok Jun 04 08:16:51 we tried to switch to wifidog, but users hated it, it was misleading in 'splash-only mode' and was broken in various ways and the developers weren't very enthusiastic about diagnosing and fixing. Jun 04 08:18:19 kind of soured me on wifidog, retreat to stability ensued from there. Jun 04 09:57:48 florian * r16327 /packages/net/ (5 files in 3 dirs): [package] merge vnstati to vnstat package, use tarball's configuration file, remove osbolete cron script (#4989) Jun 04 09:59:29 florian * r16328 /trunk/target/linux/brcm63xx/files/ (3 files in 2 dirs): [brcm63xx] register gpiochip earlier, allowing gpio-based runtime detection to be performed in board_init callback Jun 04 10:01:04 florian * r16329 /trunk/target/linux/brcm63xx/files/ (3 files in 2 dirs): [brcm63xx] use the correct number of GPIOs for bcm6358 Jun 04 10:01:57 florian * r16330 /trunk/target/linux/brcm63xx/files/arch/mips/bcm63xx/gpio.c: [brcm63xx] cache gpio values in gpio_set Jun 04 10:02:07 hehe [florian], i'm seeing good stuff coming... ;) Jun 04 10:02:31 <[florian]> danage: eh eh Jun 04 10:02:37 florian * r16331 /trunk/target/linux/brcm63xx/files/arch/mips/bcm63xx/gpio.c: [brcm63xx] be SMP safe even with gpiolib Jun 04 10:03:05 florian * r16332 /trunk/target/linux/brcm63xx/files/drivers/pcmcia/bcm63xx_pcmcia.c: [brcm63xx] request gpio pins and set direction correctly for pcmcia Jun 04 10:03:43 florian * r16333 /trunk/target/linux/brcm63xx/files/drivers/pcmcia/bcm63xx_pcmcia.c: [brcm63xx] add missing null entry in pcmcia driver Jun 04 10:04:12 florian * r16334 /trunk/target/linux/brcm63xx/files/drivers/pcmcia/bcm63xx_pcmcia.c: [brcm63xx] fix pcmcia device removal Jun 04 10:04:55 florian * r16335 /trunk/target/linux/brcm63xx/files/drivers/net/bcm63xx_enet.c: [brcm63xx] fix ethernet driver removal Jun 04 13:14:02 florian * r16336 /trunk/package/iptables/Makefile: [package] do not apply iptables patches when building with CONFIG_EXTERNAL_KERNEL_TREE Jun 04 13:43:37 hello every body Jun 04 13:47:10 I'm building brcm-2.6 and from 10 minutes ago I get this message "Waiting for string-inst.o.lock to be removed", still waiting or stop it and start again? Some one have any idea? Jun 04 13:51:03 hello Jun 04 13:55:47 anyone can help me to disassemble cfe.bin? Jun 04 14:53:51 nbd: ping Jun 04 14:59:22 <[florian]> ftp21: can you show me again your bootlog ? Jun 04 14:59:40 http://openwrt.ftp21.eu/good/log.txt Jun 04 14:59:44 acinonyx * r16337 /packages/net/ntpclient/files/ntpclient.hotplug: Jun 04 14:59:44 [packages] ntpclient: Allow ntpclient to run without default route installed Jun 04 14:59:44 This allows the use of ntp server inside a LAN where there isn't any internet access Jun 04 14:59:44 <[florian]> ftp21: I want to check that the flash driver does not access beyond the 8MB Jun 04 14:59:47 <[florian]> ftp21: thanks Jun 04 15:00:04 <[florian]> ftp21: you had to define your own board structure for this to work, right ? Jun 04 15:00:15 my board boot veery well Jun 04 15:00:28 but after first reboot crc is corrupted Jun 04 15:00:33 <[florian]> ftp21: I mean, you have to modify arch/mips/bcm63xx/boards/board_bcm963xx.c Jun 04 15:02:16 <[florian]> ftp21: "Erasing tag at partition [1] imageptr address 0xbe010100!" that's weird Jun 04 15:02:36 yes because the crc is invalid Jun 04 15:03:05 <[florian]> yes, but I do not really understand how it could be valid once and not later, unless the mtd driver in linux erases parts of the flash which contains such informations Jun 04 15:04:05 <[florian]> ftp21: can you try to reserve 128KB for your CFE partition instead of 64KB Jun 04 15:05:14 my cfe is a modified cfe by the isp Jun 04 15:06:25 now my board is bricked with some trouble during cfe flash :D Jun 04 15:07:01 <[florian]> ftp21: can you try the following patch: http://openwrt.pastebin.ca/1447655 ? Jun 04 15:07:48 madwifi crashes on trunk : http://openwrt.pastebin.com/m5395f96f Jun 04 15:07:58 ok but now i must reflash the cfe :( becaus is corrupted :D Jun 04 15:08:13 <[florian]> ftp21: use jtag Jun 04 15:08:19 yes Jun 04 15:23:51 mirko * r16338 /packages/Xorg/xorg/xserver/xorg-server/files/x.init: get rid of the grey grid when Xorg starts up Jun 04 16:09:08 hi.. I'm building a simple app to test case a problem with sqlite3. This simple program http://paste.pocoo.org/show/120971/ get a segmentation fault on the very first call to sqlite3_open .. :-/ (compiled with http://paste.pocoo.org/show/120972/ .. any clue? :) Jun 04 16:48:12 xxiao: please turn on 'kernel symbol table information' under 'global build settings', recompile and run again Jun 04 17:05:07 mirko * r16339 /trunk/target/linux/s3c24xx/Makefile: switch back to 2.6.29 until some target specific bugs are fixed in 2.6.30 Jun 04 18:07:43 nbd: will compile it in now will send result out tonight, i'm in US CST time Jun 04 18:08:35 actually, anyone knows a good remote cheap power-on reset solution? i want to remotely reset the board(PoE passive) when it's locked up Jun 04 18:09:22 a relay that attach to a linux PC will be good Jun 04 18:18:38 juhosg * r16340 /trunk/target/linux/ar71xx/ (2 files in 2 dirs): [ar71xx] improve MZK-W300NH support (thanks to Kazuki Shimada for the board) Jun 04 18:47:58 zandbelt * r16341 /packages/net/asterisk-1.6.x/ (6 files in 2 dirs): update asterisk-1.6.x to 1.6.2.0-beta2 Jun 04 19:04:11 back Jun 04 19:14:48 juhosg * r16342 /trunk/ (2 files in 2 dirs): [ar71xx] build firmware image for the MZK-W300NH board Jun 04 20:40:02 zandbelt * r16343 /packages/net/asterisk-1.4.x/ (9 files in 2 dirs): update asterisk-1.4.x to 1.4.25 Jun 04 20:40:33 zandbelt * r16344 /packages/net/asterisk-addons-1.4.x/ (4 files in 2 dirs): update asterisk-addons-1.4.x to 1.4.8 Jun 04 21:02:00 hi Jun 04 21:44:02 nite all, sweet dreams Jun 04 22:53:18 nbd: i just had a look at the wprobe source: sweet stuff! Jun 04 22:54:42 :) Jun 04 22:55:25 what is it Jun 04 22:58:15 it's an interface to gather measurement data from wifi drivers (currently only madwifi, but that'll change) Jun 04 23:00:10 cool Jun 04 23:28:54 * russell_ going to try bisecting the thing that caused wgt634u to not boot anymore Jun 04 23:29:35 r16322 bad, r15577 good-ish Jun 04 23:34:48 any details about wgt not booting? Jun 04 23:34:58 kernel crash? lockup at boot time? Jun 04 23:35:01 what's the issue? Jun 04 23:35:18 it locks up ... uh, let me get the last message Jun 04 23:35:36 Freeing unused kernel memory: 136k freed Jun 04 23:35:36 Please be patient, while OpenWrt loads ... Jun 04 23:35:36 Algorithmics/MIPS FPU Emulator v1.5 Jun 04 23:35:41 and then nothing Jun 04 23:35:47 odd Jun 04 23:35:51 yeah Jun 04 23:35:57 both revs using the same kernel version? Jun 04 23:36:07 not sure Jun 04 23:36:31 r15577 was built in early may Jun 04 23:36:46 i was out of town and hadn't actually flashed anything since Jun 04 23:37:00 i'm going to confirm and then go from there Jun 05 00:05:39 interesting. on a fresh checkout, make doesn't create any images. Jun 05 00:14:03 hmm. didn't like the -j17 i guess Jun 05 00:19:39 hmm. must be accumulated build tree cruft Jun 05 00:19:55 fresh checkout is booting Jun 05 00:22:13 aha! Jun 05 00:22:25 seeing the same nvram corruption Jun 05 00:22:30 same as last night Jun 05 00:22:38 only after a reboot Jun 05 00:22:50 that's not good Jun 05 01:06:06 * russell_ wondering if the partition table is right, letting the jffs2 stomp on it or something Jun 05 01:06:25 on it == on the nvram data Jun 05 01:17:57 can i force firstboot to run again? Jun 05 01:39:57 xMff: ping Jun 05 01:41:20 nbd: ping Jun 05 01:44:46 for this lib, http://www.inf.puc-rio.br/~roberto/struct/#pack I need do a package or is possible embed it directly into posix Jun 05 02:15:20 nbd: i compiled with the -g for kernel but the dump looked the same http://openwrt.pastebin.com/m494c2743 Jun 05 02:15:37 nbd: it's quite obvious to me that madwifi is not working on trunk Jun 05 02:15:42 it hangs Jun 05 02:15:50 my card is CM9 Jun 05 02:17:39 well, i asked you to enable symbol table information in make menuconfig Jun 05 02:17:43 not do anything with -g Jun 05 02:18:01 nbd: symbol table is also enabled Jun 05 02:18:07 ah, ok Jun 05 02:18:21 so after the part you pasted it hangs? Jun 05 02:19:02 Compile packages with debugging info & Compile the kernel with symbol table information Jun 05 02:19:10 yes it hangs Jun 05 02:20:46 ok, i'm about to go to sleep Jun 05 02:20:49 i'll check it out tomorrow Jun 05 02:21:20 ok. replaced with a different cm9, http://openwrt.pastebin.com/m3d443716 Jun 05 02:21:34 still hangs. however ath9k/SR71A did not hang and worked under 11g well Jun 05 02:21:39 using the same slot Jun 05 02:22:16 there are some nasty incompatibility issues between the ar71xx pci controller and atheros 11a/g cards Jun 05 02:22:17 ath5k did not do well either. 80211s will crash it Jun 05 02:22:51 nbd: could be, i was told routerboard RB433AH has issues with some atheros cards as well Jun 05 02:23:17 atheros's ar71xx does not like its ar5xxx/ar9xxx Jun 05 02:23:23 these issues are also the reason why those cards can only do dma transactions of 4 byte, not mroe Jun 05 02:23:44 newer ar9xxx cards are not affected by the same bug Jun 05 02:23:45 imho Jun 05 02:23:50 what...? then why need dma at all Jun 05 02:24:02 pci core bugs ... Jun 05 02:24:15 nbd: yes...ath9k/ar9160 ran fine on the same routerstation board Jun 05 02:24:29 it's those ath5k/madwifi cards...that have issues Jun 05 02:24:51 if it's pci-core, then it's troublesome,geez Jun 05 02:25:02 yes, indeed Jun 05 02:25:06 workaround will impact performance, to say the least Jun 05 02:25:12 it does Jun 05 02:25:20 but there is no known way around it Jun 05 02:25:31 atheros uses the same workaround internally Jun 05 02:25:34 iirc Jun 05 02:25:45 does openwrt contain those workaround? Jun 05 02:25:50 yes Jun 05 02:25:58 but i don't know yet why it's not enough Jun 05 02:26:07 need to do some testing myself Jun 05 02:26:10 then the workaround need be workaround... Jun 05 02:26:18 ed Jun 05 02:26:26 :0 Jun 05 02:26:27 :) Jun 05 02:26:42 ;) Jun 05 02:26:46 i'll let you know when i find something Jun 05 02:27:03 don't know what to do ... Jun 05 02:27:57 i do need both madwifi-or-ath5k and ath9k on the same board Jun 05 02:27:59 sigh Jun 05 02:28:14 ok..good nite..and thanks! **** ENDING LOGGING AT Fri Jun 05 02:59:57 2009