| 00:00.09 | *** join/#nslu2-linux gouki (n=gouki@bl11-50-232.dsl.telepac.pt) |
| 00:16.48 | VoodooZ_ | rwhitby: ok. so the default redboot uses a blank? or skips that region of the flash? |
| 00:17.42 | rwhitby | VoodooZ_: what default redboot? |
| 00:18.44 | rwhitby | oh, the one from linksys images? yes. the mac address is inserted during production. |
| 00:19.16 | VoodooZ_ | is that one of the skipped blocks in the FIS? |
| 00:19.45 | VoodooZ_ | I'm back to apex so I'll need to figure out the address and do it manually.. |
| 00:21.46 | *** part/#nslu2-linux treitmayr (n=thomas@85.127.14.184) |
| 00:23.26 | rwhitby | no, mac address is inside the redboot fis partition |
| 00:27.43 | *** join/#nslu2-linux yvasilev (n=yvasilev@189.140.141.23) |
| 00:29.59 | VoodooZ_ | would you happen to remember the address? I didn't see it mentionned on the wiki.. |
| 00:32.08 | mwester | 0x3FFB0 |
| 00:34.53 | rwhitby | http://www.nslu2-linux.org/wiki/Info/BootFlash |
| 00:35.13 | rwhitby | third link in the "RedBoot/Flash" area of the information pages ... |
| 00:38.11 | VoodooZ_ | ok. thanks. I'll add this to my log book... (it's about time I started one!) |
| 00:50.39 | VoodooZ_ | woot! I think I might have finally succeeded.. flashed an LE apex... now I need to figure out how to get a complete slug image for angstrom.. |
| 00:57.21 | rwhitby | VoodooZ_: how did you build the LE apex? |
| 00:58.18 | VoodooZ_ | straight build and used the slugosle built cross compiler... used devio to swap the binary and voila! Strange as I tried similar things yesterday... |
| 00:58.50 | VoodooZ_ | I think the jtag tool was messing things up... so I flashed a proven BE apex and did some RAM tests instead.. |
| 00:59.41 | VoodooZ_ | How do I inherit that slugos-image to get angstrom to build me one? the sample local.conf is empty... unless I have to mess with angstrom itself? |
| 01:00.40 | VoodooZ_ | right now, I can only get the angstrom kernel to decompress... it stalls after... I need to either manually flash all the parts or get that slugos-image inheritance thing working.. |
| 01:05.10 | rwhitby | pINHERIT += "nslu2-image" |
| 01:05.16 | rwhitby | oops, remove the initial 'p' |
| 01:06.59 | VoodooZ_ | and that's in the local.conf under angstrom right? |
| 01:07.05 | rwhitby | yep |
| 01:07.24 | VoodooZ_ | ok. thanks. I'm trying the manual way for now but I'll start a fresh build before going to bed.. |
| 01:07.41 | CIA-22 | 03rwhitby * r7681 10optware/trunk/sources/optware-bootstrap/ (target-specific.mk vt4/ vt4/bootstrap.sh vt4/postinst): bootstrap for vt4, untested |
| 01:08.20 | VoodooZ_ | btw, I see the latest FIS dir has a "loader" block... I assume this is what use to be the ramdisk... is this where you store the second stage apex now? |
| 01:10.01 | rwhitby | yep |
| 01:10.23 | rwhitby | actually loader replaces kernel, kernel moves up |
| 01:10.39 | rwhitby | (since redboot does not support FIS dir on nslu2) |
| 01:10.46 | VoodooZ_ | ah ok... good. Just trying to get situated as things have change quite a bit since 2006.. |
| 01:11.24 | VoodooZ_ | I noticed that the kernel block is now 0x120000 which is nice... |
| 01:11.33 | rwhitby | slugimage adjusts it to whatever size the kernel you give it is |
| 01:11.43 | VoodooZ_ | nice... |
| 01:12.01 | VoodooZ_ | I like apex's fis driver... |
| 01:12.16 | rwhitby | you will probably want to build a custom kernel to give you more rootfs space. you won't need the network or sata disk drivers ... |
| 01:12.56 | VoodooZ_ | yeah, definitely... will go custom later... but I have a robot competition in May so I need to rampup fast... |
| 01:13.23 | VoodooZ_ | I need gpsd which appears broken... It's a magellan type of thing (GPS nav robot) |
| 01:13.47 | VoodooZ_ | The extra FP boost will come in handy too. |
| 01:14.19 | VoodooZ_ | you rarely get something for free... on that side, the slug is a real joy... first the de-underclocking then this... sweet! |
| 01:15.05 | VoodooZ_ | btw, do you know what those two FIS skips are in the kernel fis entry? |
| 01:16.19 | rwhitby | yes, first is for the kernel size header, and second is for the ramdisk size header |
| 01:16.35 | rwhitby | (this is how the kernel can be bigger than 1MB) |
| 01:17.06 | VoodooZ_ | ramdisk? I thought we weren't using one anymore.. |
| 01:17.14 | CIA-22 | 03bzhou * r7682 10optware/trunk/platforms/toolchain-vt4.mk: toolchain-vt4: copy libdl.so* to STAGING_LIB_DIR |
| 01:17.22 | rwhitby | correct, but redboot still loads whatever those 16 bytes point to. |
| 01:18.19 | VoodooZ_ | ok. so should I just flash my kernel image to that block without worrying? |
| 01:18.42 | VoodooZ_ | I wonder what those skip affect in apex? copy function too? |
| 01:20.16 | rwhitby | no, you can't flash your kernel to that block. apex will skip 16 bytes from the middle of your kernel. |
| 01:20.22 | rwhitby | this is why we have the slugimage tool. |
| 01:20.43 | VoodooZ_ | ok.. I guess that's one good reason not to do it manually... |
| 01:21.22 | VoodooZ_ | Unless of course Idecide to override those skips and go linear as I use apex... |
| 01:22.48 | VoodooZ_ | so that means that the individual parts (kernel, jffs, etc..) in the deploy/images/ folder don't yet contain the skips.. Only once the full 8M image... |
| 01:24.10 | rwhitby | correct. |
| 01:24.55 | VoodooZ_ | ok.. I'm finally on the same page as people in 2006! :) I made sure to enable my irssi log so I won't have to ask twice... |
| 01:24.55 | rwhitby | you can't do anything manually on the nslu2 without risking problems due to all this stuff |
| 01:25.03 | VoodooZ_ | yes.. |
| 01:25.31 | VoodooZ_ | and even if I did I would mess up the ability to have a 1Meg+ kernel right? |
| 01:26.43 | rwhitby | well, if you remove redboot and use apex as first stage loader, then you don't need the skips. |
| 01:27.12 | CIA-22 | 03rwhitby * r7683 10optware/trunk/platforms/packages-vt4.mk: Enable vt4-optware-bootstrap |
| 01:27.14 | VoodooZ_ | ah.. that's what I was wondering.. |
| 01:27.57 | VoodooZ_ | so the > 1meg kernel thing wouldn't affect a pure apex setup? |
| 01:28.34 | VoodooZ_ | I'd have to figure out how to change the fis dir i gues... |
| 01:29.23 | CIA-22 | 03bzhou * r7684 10optware/trunk/platforms/toolchain-vt4.mk: toolchain-vt4: copy symlinks of libdl*so* |
| 01:29.33 | VoodooZ_ | I'll start simple (too late!) and then I'll cleanup the historical stuff later for fun.. |
| 01:30.11 | VoodooZ_ | I'm no longer afraid now that I've brick the thing and came back.. I'm JTAG empowered now! :) |
| 01:35.27 | rwhitby | with apex there are no constraints - you can make the partitions as large as you like. |
| 01:35.57 | rwhitby | I would suggest you look into modifying slugimage to suit your needs. |
| 01:37.13 | VoodooZ_ | ok.. |
| 01:37.34 | VoodooZ_ | rebuilding angstromle-image with inherit thing... |
| 01:38.11 | VoodooZ_ | as for modding the fis dir: it's probably best done within linux right? |
| 01:38.31 | VoodooZ_ | But then again, slugimage could probably do it for me.. |
| 01:39.37 | *** join/#nslu2-linux jacques (n=jacques@nslu2-linux/jacques) |
| 01:53.17 | VoodooZ_ | rwhitby: Should I be overriding the kernel command line in apex or just let the kernel use whatever...? |
| 01:56.11 | rwhitby | kernel has no default now - you need to set it in the bootloader |
| 02:02.40 | VoodooZ_ | ok. that must still be valid right? root=/dev/mtdblock4 rw rootfstype=jffs2 mem=32M@0x00000000 init=/linuxrc noirqdebug console=ttyS0,115200n8 |
| 02:10.51 | VoodooZ_ | which package creates the fis directory that slugimage packages? |
| 02:16.23 | rwhitby | slugimage creates the fis directory itself |
| 02:16.28 | rwhitby | it's a perl script - have a read |
| 02:25.21 | *** join/#nslu2-linux eiffel_ (n=eiffel@dslb-084-061-130-062.pools.arcor-ip.net) |
| 02:26.13 | VoodooZ_ | good I understand perl better than python... that's on my todo list.. |
| 02:27.08 | VoodooZ_ | I noticed that apex's default setup for the slug is outdated... I updated my .config to use the FIS entries instead of hard-coded address to avoid this int he future... (ex: fis://Kernel) |
| 02:27.17 | VoodooZ_ | I'll email the changes to Marc if they work... |
| 02:29.21 | VoodooZ_ | thanks to that I was able to reach the "Uncompressing Linux..." stage... but ends fails with "bad gzip magic numbers" for obvious reasons.. |
| 02:30.03 | *** join/#nslu2-linux aadamson (n=adamson_@c-76-97-53-219.hsd1.ga.comcast.net) |
| 02:31.05 | *** join/#nslu2-linux mithro (n=tim@169.222.10.205) |
| 02:59.29 | rwhitby | VoodooZ_: the default nslu setup is designed for stock firmware |
| 02:59.46 | rwhitby | use either the slugos or debian setups from the nslu2-linux kernel svn repo |
| 03:12.06 | VoodooZ_ | ok... that's a total unknown to me... I'll have to explore that area... |
| 03:12.43 | VoodooZ_ | but while I'm waiting for my angstrom build to finish I manually copied the kernel to the ram and booted and voila! my first angstrom boot! |
| 03:13.09 | VoodooZ_ | wow. 23.3M RAM free! |
| 03:14.16 | VoodooZ_ | oh crap! I just realised I don't have turnup now though.. :( more learning... |
| 03:25.44 | VoodooZ_ | Is it possible that my partitioning on my usb flash disk is no longer valid because of the byte ordering change??? doesn't make sense... |
| 03:35.20 | VoodooZ_ | wow. that angstrom build is definitely not ready for primetime... It recognizes usb devices but won't let me mount them... that and lots of alignment traps... |
| 03:40.49 | VoodooZ_ | oh well. I've done enough today.. good night rwhitby and thanks for all your help... |
| 03:41.34 | rwhitby | VoodooZ_: you did want to be bleeding edge ... |
| 03:43.54 | VoodooZ_ | yep.. I guess I'll pay dearly for that small fp perf. boost... |
| 03:44.53 | rwhitby | indeed you will. |
| 03:45.27 | rwhitby | VoodooZ_: if you want to cut down boot even more, you could load rootfs directly from flash by compiling modules into the kernel |
| 03:45.34 | rwhitby | (like we do for the fsg3) |
| 03:46.25 | VoodooZ_ | you mean for slugos... yes... but I was refering to EABI's Floating point speed improvement... |
| 03:46.49 | VoodooZ_ | If slugos was to become EABI compliant I'd be right back in flash... |
| 03:55.09 | VoodooZ_ | later.. |
| 04:05.39 | rwhitby | VoodooZ_: there is a good chance that SlugOS 5 will be EABI |
| 04:36.00 | VoodooZ_ | rwhitby: relly? And what is best guess that will happen? |
| 04:39.02 | *** join/#nslu2-linux Annirak (n=chatzill@S0106000129d15bac.vc.shawcable.net) |
| 04:39.05 | *** join/#nslu2-linux Annirak[W] (n=bmoran@h2-cpe102.insinc.com) |
| 04:44.15 | Annirak | So, I'm trying to get my NSLU2 going with an updated kernel image that fits with Debian. I was using slugos/le + debian before. I downloaded and installed kernel-image-2.6.23-1-ixp4xx, and created the corresponding ramdisk. I downloaded the apex binary at the bottom of this wiki page: http://www.nslu2-linux.org/wiki/Debian/CompileApex. Then I took the debian installer image, and unpacked... |
| 04:44.17 | Annirak | ...it with slugimage. I took my new apex, vmlinuz, and initramfs, replaced the ones in the debian installer with mine, and rebuilt the image with slugimage. I then flashed the new image onto my NSLU2 with upslug2. When I was done that, I halted apex while it booted, and reset the environment variables to the values that slugimage listed as it built the image. Now, my NSLU2 seems to... |
| 04:44.18 | Annirak | ...continuously load and reload the kernel image. When it gets to loading the initramfs, it stops, goes back, and loads the kernel image again. I can't interrupt the process. What have I done wrong? |
| 04:50.07 | *** join/#nslu2-linux mithro (n=tim@169.222.10.205) |
| 05:30.03 | *** join/#nslu2-linux hillct (n=H@cpe-024-211-242-232.nc.res.rr.com) |
| 05:48.58 | *** join/#nslu2-linux DrZimmerman (n=theo@2001:5c0:8fff:fffe:0:0:0:6af1) |
| 07:11.30 | *** join/#nslu2-linux drath_ (i=vmaster@p5B07F12F.dip.t-dialin.net) |
| 07:36.18 | *** join/#nslu2-linux ka6sox-laptop (n=tom@nslu2-linux/ka6sox) |
| 08:16.25 | *** join/#nslu2-linux jbot (i=ibot@pdpc/supporter/active/TimRiker/bot/apt) |
| 08:16.25 | *** topic/#nslu2-linux is nslu2-linux developer channel - use #nslu2-general for end-user questions |
| 08:30.18 | *** join/#nslu2-linux erikja (n=erikja@0x57314066.abnxx10.adsl-dhcp.tele.dk) |
| 08:31.39 | *** join/#nslu2-linux erikja (n=erikja@0x57314066.abnxx10.adsl-dhcp.tele.dk) |
| 09:20.40 | *** join/#nslu2-linux kolla (n=kolla@drift1.uninett.no) |
| 10:57.08 | *** join/#nslu2-linux erikja (n=erikja@0x57314066.abnxx10.adsl-dhcp.tele.dk) |
| 11:07.37 | *** join/#nslu2-linux mithro (n=tim@169.222.15.33) |
| 13:28.44 | VoodooZ_ | rwhitby: the inherit= change didn't seem to work on angstrom... Still slug image.. |
| 13:34.21 | *** join/#nslu2-linux lyakh (n=lyakh@p57BD288D.dip0.t-ipconnect.de) |
| 13:50.38 | *** join/#nslu2-linux GPSFan (n=kenm@12.10.255.246) |
| 14:08.07 | *** join/#nslu2-linux ByronT_ (n=byron-po@nslu2-linux/ByronT) |
| 14:08.12 | VoodooZ_ | rwhitby: I was thinking that I might just go back to good old slugosbe for the time being and deal with EABI whenever it comes... |
| 14:08.36 | VoodooZ_ | That said, if you want a guinea pig to test it on slugos, I'm your man! |
| 14:26.40 | *** join/#nslu2-linux Ragnor (n=Ragnor@dslb-084-060-141-145.pools.arcor-ip.net) |
| 16:07.45 | *** join/#nslu2-linux HopsNBarley (n=hops@nslu2-linux/HopsNBarley) |
| 18:49.20 | *** join/#nslu2-linux thraxisp (n=thraxisp@ottgate.precidia.com) |
| 20:03.44 | CIA-22 | 03engy * r7685 10optware/trunk/ (3 files in 2 dirs): nmap:4.50->4.53 |
| 20:09.58 | *** join/#nslu2-linux yvasilev (n=yvasilev@dune.math.cinvestav.mx) |
| 20:13.49 | CIA-22 | 03bzhou * r7686 10optware/trunk/ (4 files in 2 dirs): uemacs: fix strncpy and promoted on cs05q3armel, syno-x07 |
| 21:17.48 | CIA-22 | 03engy * r7687 10optware/trunk/ (make/tcpdump.mk sources/tcpdump/): tcpdump:3.9.7->3.9.8, script cleanup |
| 22:14.44 | *** join/#nslu2-linux gouki (n=gouki@ubuntu/member/gouki) |
| 22:45.31 | *** join/#nslu2-linux yvasilev (n=yvasilev@189.140.141.23) |
| 22:51.36 | *** join/#nslu2-linux Reedy_Boy (n=Reedy_bo@wikimedia/Reedy-Boy) |
| 23:03.35 | *** join/#nslu2-linux ka6sox-work (n=tking@nslu2-linux/ka6sox) |
| 23:37.11 | *** join/#nslu2-linux kolla (n=kolla@ip-14-202-241-92.dialup.nmt.net) |
| 23:39.51 | *** join/#nslu2-linux mithro (n=tim@169.222.10.205) |
| 23:56.00 | *** join/#nslu2-linux thraxisp (n=thraxisp@24.139.16.154) |