**** BEGIN LOGGING AT Fri Mar 02 03:00:00 2007 Mar 02 14:39:33 hi Mar 02 14:39:42 morning Mar 02 14:40:00 afternoon, 16:12 here xD Mar 02 14:40:50 i've put a serial port to my nslu and now i can debug ucslugc Mar 02 14:41:22 seems that the kernel is booting fine, but i have some problems with the rootfs Mar 02 14:41:24 Isn't it great to no longer work "blind" :) Mar 02 14:41:49 yep Mar 02 14:41:54 finnaly Mar 02 14:41:55 xDD Mar 02 14:41:57 Time: OSTS clocksource has been installed.^M Mar 02 14:41:57 x1205 0-006f: setting the system clock to 2007-03-02 15:44:05 (1172850245)^M Mar 02 14:41:57 VFS: Mounted root (jffs2 filesystem).^M Mar 02 14:41:57 Freeing init memory: 92K^M Mar 02 14:41:57 Segmentation fault^M Mar 02 14:41:59 ^MINIT: version 2.86 booting^M^M Mar 02 14:42:01 Segmentation fault^M Mar 02 14:42:26 how can i go forward with the debugging? Mar 02 14:43:06 segmentation fault? That's ... just weird Mar 02 14:43:21 (And btw, we're in the same TZ) Mar 02 14:43:32 init=/linuxrc on the kernel cmdline? Mar 02 14:43:53 where can i found that? Mar 02 14:44:28 If so, my bet is that the kernel is firing up /bin/sh to execute the /linuxrc script, and it's the /bin/sh that's segfaulting. (cmdline is printed early in the boot) Mar 02 14:44:30 as far as i know, ucslugc has been broken some months, so its not surprising that now that compiles does not work Mar 02 14:44:57 Kernel command line: rtc-x1205.probe=0,0x6f init=/linuxrc root=/dev/mtdblock4 rootfstype=jffs2 rw noirqdebug console=ttyS0,115200n8^M Mar 02 14:46:16 looks right Mar 02 14:46:51 Yep that's right. So I'm betting that its /bin/sh that's segfaulting when the kernel starts /linuxrc. Mar 02 14:47:03 /bin/sh is usually a link to /bin/busybox. Mar 02 14:48:11 i'll take a look to that rootfs Mar 02 14:48:55 Usually you would try booting a rootfs that would have /linuxrc be a simple C program (statically linked) that prints "Hello". Mar 02 14:49:26 ok Mar 02 14:49:29 i'll try that Mar 02 14:49:38 Once that works, you build up from there, with a dynamically linked app, until you get to the point where you can see if its a library problem, or a busybox problem, or a permissions issue, etc. Mar 02 14:57:22 mwester Mar 02 14:57:28 to compile i must use tmp/cross/armeb-linux-uclibc/bin/gcc ? Mar 02 14:59:44 im not sure about how must i compile something from my host arch for the arm arch Mar 02 15:00:01 i supose that i need the cross-compiler and the linker Mar 02 15:00:22 and the arm native libraries i need to link Mar 02 15:03:30 OE put all that stuff in the cross directory under tmp Mar 02 15:04:54 If you just compile and link with the tools in the .../tmp/.../cross/bin directory, I think it should select all the right stuff. Mar 02 15:07:58 yes Mar 02 15:08:01 i think i have it Mar 02 15:08:36 saladino@korn:/mnt/almacen/Arkaitz/bin$ file prueba Mar 02 15:08:36 prueba: ELF 32-bit MSB executable, ARM, version 1 (ARM), statically linked, not stripped Mar 02 15:09:02 if i put it in tmp/rootfs/ and do a make ucslugc-image it will put it in the final image? Mar 02 15:09:17 * mwester is not sure, but thinks that should work Mar 02 15:09:43 ok Mar 02 15:09:43 trying Mar 02 15:09:45 thanks mwester Mar 02 15:19:57 humm Mar 02 15:20:02 it doesn't work Mar 02 15:43:57 is there a way to mount the generated jffs2 image in a local dir? Mar 02 15:44:01 i cannot use loopback Mar 02 15:49:42 without loopback I doubt it's possible Mar 02 15:49:57 that's basically what loopback is for Mar 02 15:53:37 nop kolla Mar 02 15:53:37 as far as i know Mar 02 15:53:53 the jffs2 fs cannnot be mounted as loopback device Mar 02 15:54:05 did someone succeded mounting it as loopback? Mar 02 15:54:18 I dont remember if I ever did Mar 02 15:54:35 but I can imagine you're right Mar 02 15:54:43 i've been reading and no way Mar 02 15:55:01 it can be done using some virtual device that simulates a ramdisk Mar 02 15:55:06 im looking for a way Mar 02 15:55:09 :D Mar 02 21:04:39 * H4rvy is away: PUB **** BEGIN LOGGING AT Fri Mar 02 21:50:13 2007 Mar 02 23:37:29 * H4rvy is back **** BEGIN LOGGING AT Sat Mar 03 01:07:02 2007 **** ENDING LOGGING AT Sat Mar 03 02:59:59 2007