**** BEGIN LOGGING AT Fri Feb 10 10:59:56 2006 Feb 10 14:24:03 ehh Feb 10 14:24:42 how do i get my kernel to use the "initrd.gz" located in /dev/mtdblock4 as a root-fs expanded to the ram? Feb 10 14:24:54 the rootfs is ready and packad to that initrd.gz Feb 10 14:25:12 but now i have to mess up with the kernelparams.. and dunno where to start Feb 10 14:26:15 first, he has to mount the /dev/mtdblock4 with jffs2 as filesystem, then he has to make that ramdisk in /dev/ram0 with ext3 as filesystem? Feb 10 14:26:24 are you trying to use a initrd with openslug? Feb 10 14:27:03 no, but i want to learn sth from openslug Feb 10 14:27:09 i built my own rootfs Feb 10 14:27:16 and my own kernel Feb 10 14:27:21 now, i wanna get it to boot Feb 10 14:27:35 hmm Feb 10 14:27:44 i have the initrd.gz ready Feb 10 14:27:47 why not just use the stock openslug setup and modify that? :) Feb 10 14:27:55 nah.. :) Feb 10 14:28:15 to big and (sorry) not good for my use Feb 10 14:28:48 have to use it with perl and libdeviceserial-port for controlling an rfid-reader (but without any nfs/usbstick/harddisk storage) Feb 10 14:29:01 Well, iirc, initrd (or initramfs) is appended to the kernel. Feb 10 14:29:25 appended? Feb 10 14:29:52 yes, but I think there's a problem with using an kernel+initrd that is larger than 1mb Feb 10 14:29:53 to fit in the 1mb big partition where the kernel has to get in? ;) lol! Feb 10 14:30:18 Have you tried ucslugc btw? Feb 10 14:30:26 nope Feb 10 14:30:39 With ucslugc you might get enough space to have a tarball on the flash that you extract to ram Feb 10 14:30:46 i built my own uclibc+busybox+perl system Feb 10 14:31:06 works fine... but has to be bootet ;) Feb 10 14:31:12 ucslugc uses uclibc+thumb+busybox Feb 10 14:31:19 ie, it's quite small Feb 10 14:31:20 thumb? Feb 10 14:31:28 yes, 16bit arm instruction set Feb 10 14:31:33 hm Feb 10 14:31:34 instead of 32 Feb 10 14:31:59 my initrd is 3,6mb big Feb 10 14:32:07 it's as small as we get it, I think. There's no filesystem support or anything in it by default. Feb 10 14:32:19 and holds a complete rootfs including perl Feb 10 14:32:28 my kernel is 987kb gib Feb 10 14:33:36 using thumb reduces the binary sizes with ~20% Feb 10 14:34:00 NAiL, morning. Feb 10 14:34:14 How did the builds go? Feb 10 14:34:15 I think what you want is basically to checkout the ucslugc distro, modify busybox to suit, add a perl RDEPENDS Feb 10 14:34:25 VoodooZ_Log: they're all finished :) Feb 10 14:34:57 That way, you don't need to duplicate busybox (which it seems you are doing now) Feb 10 14:35:00 good. no show-stoppers then. Feb 10 14:35:19 not on the compiling part, no. It won't be released until after the weekend though Feb 10 14:36:17 np Feb 10 14:36:37 I'll be keep an eye out for news on the web page. Feb 10 14:37:52 eehh.. that can't be true.. it has to be possible to use such an ramdisk wich is placed on the flash... Feb 10 14:38:13 i compete so far.. and now, it hangs at such a point.. Feb 10 14:38:21 rammy2: possible? Dunno. Easy, don't think so ;-) Feb 10 14:38:57 But why insist on a tarball? jffs2 has compression built-in. If you need more compression, and don't care about writing to the fs, you can switch to cramfs. Feb 10 14:41:15 hm because my rootfs is about 14mb large Feb 10 14:41:34 and it would get into the /dev/mtdblock wich is about 6mb big Feb 10 14:42:50 eek Feb 10 14:42:57 theres a kernelparam Feb 10 14:42:59 initrd=0xc0800000,4M Feb 10 14:43:06 but how do i get that memoryadress? Feb 10 14:43:56 look at the address used by RedBoot, then add a mb ;) Feb 10 14:44:15 "a mb"? Feb 10 14:44:24 (if your initrd is flashed instead of the jffs2) Feb 10 14:44:29 why 1? Feb 10 14:44:49 RedBoot loads the kernel from a specific offset. The kernel "partition" is 1mb Feb 10 14:45:28 actually, the kernel should print the redboot partition table on boot Feb 10 14:45:48 0x00060000-0x00160000 : "Kernel" Feb 10 14:45:48 0x00160000-0x00180000 : "Ramdisk" Feb 10 14:45:48 0x00180000-0x007e0000 : "Flashdisk" Feb 10 14:45:48 0x007e0000-0x00800000 : "FIS directory" Feb 10 14:46:35 though... I doubt the kernel tries to read the initrd from flash... Feb 10 14:58:26 hmm kk ill try it Feb 10 21:16:08 salut. hat allenfalls jemand binaries für einen proxy ? squid oder tinyproxy evtl ? Feb 10 21:20:45 ? Feb 10 21:23:09 t[cc]smar, do you want to know if there are any binaries for a proxy in OpenSlug? Feb 10 21:29:14 stein: yes Feb 10 21:30:14 squid is IMHO a *very* unlikely candidate for the slug Feb 10 21:30:26 maybe tinyproxy if someone cares to port it Feb 10 21:30:52 I don't think there there are any for openslug... Try ipkgfind.nslu2-linux.org Feb 10 21:30:54 somewhat yes. but since package DB on nslu2-linux mentionned it for unslung Feb 10 21:31:03 i thought someone should have had it going Feb 10 21:31:30 Nail, IMHO? Feb 10 21:31:41 In my humble opinion Feb 10 21:31:46 :-) Feb 10 21:32:01 The slug won't be able to run squid in any way I'd deem "useful" Feb 10 21:32:03 i did ipkgfind Feb 10 21:32:52 Either build from source or "upgrade" your openslug to opendebianslug (or similar)... Feb 10 21:33:23 could somebody try bb tinyprox for openslug for me ? i'm still decomputerized.... Feb 10 21:33:59 neither tinyproxy nor squid is in oe Feb 10 21:34:12 ? thought they should... lemme see. Feb 10 21:34:44 a right. nylon that is Feb 10 21:34:52 is nylon in ? Feb 10 21:35:23 tinyproxy is available as binary under opendebianslug... Feb 10 21:36:34 nylon doesn't appear to be a package. There appears to be packages *for* nylon, but not nylon itself Feb 10 21:37:47 argh Feb 10 21:38:11 debianslug gives me some creeps cause they ask for a swap partition right away. Feb 10 21:38:35 You could set up a swap file... Feb 10 21:38:37 sounds like it will be memory hungry, which is not for waste on slug Feb 10 21:39:07 currently, the slug targeted is on NFS. no good for that Feb 10 21:39:22 It's allways good to have a bit of swap... :-) Feb 10 21:40:16 the mailserver slug is allright without. but in principle you are right. execpt for real HW, since nowadays i'd just recommend to up some memory. cheap enough it is. Feb 10 21:41:27 True, btw, the memory on the slug is very limited. Even ctorrent crashed do to "out-of-memory"... (without swap).... Feb 10 21:42:31 that's the reaon i'm a bit sceptic aout debianslug. but i should't be until i checked it out. Feb 10 21:43:32 Yes, I believe opendebianslug packages are build a little fat... but I haven't noticed any difference from openslug... Feb 10 21:43:41 in performance Feb 10 21:44:41 is there a feed i could check ? maybe i can just peel the binaries out of the packages and have a try. Feb 10 21:45:18 www.debonaras.org I believe... They have a link to a feed. Feb 10 22:01:20 seems it will work Feb 10 22:17:49 well, seems it works but doesn't meet my needs. checking polipo no Feb 10 22:17:54 w Feb 10 22:54:16 polipo seems to be the right thing for slug. had very little issues with it. here's the link: http://www.pps.jussieu.fr/~jch/software/polipo/ - for whoever cares, this is a good addition to OE/openslug Feb 10 23:38:48 heh, polipo runs clean in 9M RAM Feb 10 23:40:06 n8 all Feb 10 23:40:26 nite Feb 11 03:00:07 beewoolie: I can send you those other things you wanted today Feb 11 10:27:04 effem ??? Feb 11 10:30:35 hi Feb 11 10:30:42 sorry new nick Feb 11 10:30:47 np Feb 11 10:30:48 query?? Feb 11 10:30:55 yeah Feb 11 10:33:01 do you got my messages? Feb 11 10:34:07 kami68 last msg from you was [11:22] mhh in german we call it : Vorführeffekt at this moment it runs??? Feb 11 10:34:47 for me it records, but no signal need to check the antenna cabling Feb 11 10:35:24 so back again : **** ENDING LOGGING AT Sat Feb 11 10:59:56 2006