**** BEGIN LOGGING AT Sun Aug 23 02:59:56 2009 Aug 23 03:08:05 rektide: have you looked at the threads at open-rd.org regarding building the marvell kernels for open-rd Aug 23 03:35:42 tinker-f595: every thread i've seen is all "i want to use these patches, but i dont know how to" Aug 23 03:36:36 actually, its just rc3 going "i dont get it" Aug 23 03:42:46 ha Aug 23 03:44:19 I see you made a comment in the thread http://groups.google.com/group/openrd/browse_thread/thread/640eab6a8a0091ba# Aug 23 03:47:46 in the thread Dhaval talks about hosting a git repository but does not give the location Aug 23 04:04:52 ah the thread for the open-rd base kernell patches http://groups.google.com/group/openrd/browse_thread/thread/68e7fb1eb2ea3e0f# refer to http://computingplugs.com/index.php/Building_a_custom_kernel Aug 23 17:34:05 anyone here using ubuntu with his/her plug computer? Aug 23 17:40:42 yeah, but I'm a total nube, having just gotten talking to it. Aug 23 17:42:46 * SDuensin runs Ubuntu on his. Aug 23 17:44:27 just wondering whether there are security updates for arm architecture Aug 23 17:44:40 security.ubuntu.org/ubuntu seems to host only i386 and amd64 Aug 23 17:44:56 It gets updates pretty regularly through the other repos. Aug 23 17:45:20 http://ports.ubuntu.com/ Aug 23 17:45:52 hah. Thanks. Aug 23 18:18:52 Weird, one of my sheevaplugs has started showing up with the mac of 00:62:81:A0:00:00 Aug 23 18:47:09 just tried booting a sheevaplug with 230.5 kernel without specifying mtdparts. Worked fine, managed to find 3 parts. Now... do I really need that parameter then? :) Aug 23 18:50:17 yacoob: do you have the bootlog somewhere? Aug 23 18:50:28 it probably found U-Boot's partition table through the use of magic Aug 23 18:50:48 and does anyone here use the Gentoo image? Aug 23 18:51:04 I just untarred it to my external hard drive and got Gentoo booting Aug 23 18:51:10 but I can't login XD Aug 23 18:51:21 there's root and operator for users Aug 23 18:51:30 but I can't find the password for either Aug 23 18:53:03 debio264, no, but I can get it, hang on. Aug 23 18:53:54 and I just found /etc/rc.local. Weird stuff. Aug 23 18:57:11 debio264, http://pastebin.com/d6515cf1f Aug 23 19:01:35 odd Aug 23 19:01:47 and you're sure the partitions weren't specified on the command line? Aug 23 19:01:53 have you checked /proc/cmdline? Aug 23 19:02:23 if not, it's silently reading the partition table from U-Boot Aug 23 19:02:46 which is possible, but on another ARM box I work with, it explicitly logs that it's reading Redboot's partition table Aug 23 19:03:13 # Aug 23 19:03:14 Kernel command line: console=ttyS0,115200 rw root=/dev/mmcblk0p5 rootdelay=3 Aug 23 19:04:09 and it isn't verbose about the source of the information Aug 23 19:04:24 # Aug 23 19:04:26 NAND device: Manufacturer ID: 0xad, Chip ID: 0xdc (Hynix NAND 512MiB 3,3V 8-bit) Aug 23 19:04:29 # Aug 23 19:04:31 Scanning device for bad blocks Aug 23 19:04:34 # Aug 23 19:04:36 Bad eraseblock 134 at 0x0000010c0000 Aug 23 19:04:37 (gah, sorry about ^#$s) Aug 23 19:04:39 # Aug 23 19:04:42 Creating 3 MTD partitions on "orion_nand": Aug 23 19:07:21 I'm looking at moving the root file system to SD, as described inhttp://www.openplug.org/plugwiki/index.php/SD_Card_As_Root_File_System, but there's this note: Aug 23 19:07:30 WARNING: At least one user using a stock (late Aug 23 19:08:10 August 2009) plug has found this procedure to work fine right up to the "saveenv" part. Subsequent boots complain of a CRC error in the initial kernel load. No resolution yet. Aug 23 19:08:54 rbgray, I got mine plug last week and I'm running with rootfs on sd card Aug 23 19:09:20 So you did this procedure successfully? Aug 23 19:09:31 tbh, I wasn't following that procedure Aug 23 19:09:44 but it looks ok Aug 23 19:10:07 you need to partition sd card, copy the fs there, then boot as usual, with root= pointing to your new filesystem Aug 23 19:10:19 rootdelay=N might be needed Aug 23 19:10:54 yeah, I don't want to mess with the onboard flash until I'm a little more up to speed on all this u-boot and stuff :/ Aug 23 19:11:18 this procedure looks ok, and doesn't do anything to flash Aug 23 19:11:19 yacoob: do you have it set to fallback to internal if no card? Aug 23 19:11:30 nope Aug 23 19:11:58 I think I've seen that described, I just don't understand all the parms yet. Aug 23 19:12:01 btw, why is /'s device listed as 'rootfs' instead of... well, whatever the device is? Aug 23 19:12:24 rbgray, yeah, http://www.openplug.org/plugwiki/index.php/Multi-Boot Aug 23 19:12:51 it's quite easy - upon power up, uboot will try to run commands from bootcmd Aug 23 19:13:09 by default you have there something like 'nand read.e ...; bootm ...' Aug 23 19:13:21 which means 'read kernel from NAND flash to RAM; run it' Aug 23 19:13:43 hey, I think I"m even beginning to understand :) Aug 23 19:13:50 the kernel gets value of bootargs as parameters Aug 23 19:13:55 so bootargs point to various things Aug 23 19:14:01 like 'console is on ttyS0' Aug 23 19:14:14 or 'root device is on /dev/mtdblock1' Aug 23 19:14:48 so now, if you want to have your plug use another device as root filesystem, you create it, then change 'root=' parameter in bootargs Aug 23 19:15:14 brb Aug 23 19:15:21 now, the multi-boot wiki page describes how to stash both bootargs and bootcmd into separate variables Aug 23 19:15:28 so you can later switch between them Aug 23 19:15:47 and if you have: Aug 23 19:15:48 bootcmd=run bootcmd_sd; run bootcmd_nand Aug 23 19:16:16 it means 'first run commands from bootcmd_sd - this should be sufficient to boot. If you won't boot from that and still are in uboot, run commands from bootcmd_nand' :) Aug 23 19:20:38 yeah, makes sense, so I guess I'd best prep my SD card :) Aug 23 19:20:53 * rbgray thinks he's gonna have fun with this :) :) Aug 23 19:20:57 I'm the author of the Warning you read rbgray Aug 23 19:21:10 hi davidthings Aug 23 19:21:26 I'm currently trying to figure out what went wrong Aug 23 19:22:00 I thought I'd put the warning up to maybe save others from some extra work! Aug 23 19:22:52 I followed the procedure pretty tightly (I think), and it worked great. But then I got a CRC in my boot ... Aug 23 19:22:59 I guess I'll see if mine goes wrong too. Do you have a particular (u-boot?) version involved? (Yes, the warning is appreciated!) Aug 23 19:23:25 Marvell>> boot Aug 23 19:23:27 NAND read: device 0 offset 0x100000, size 0x400000 Aug 23 19:23:28 Reading data from 0x4ff800 -- 100% complete. Aug 23 19:23:30 4194304 bytes read: OK Aug 23 19:23:31 ## Booting image at 00800000 ... Aug 23 19:23:33 Image Name: Linux-2.6.22.18 Aug 23 19:23:34 Created: 2009-03-19 9:18:16 UTC Aug 23 19:23:36 Image Type: ARM Linux Kernel Image (uncompressed) Aug 23 19:23:37 Data Size: 2106696 Bytes = 2 MB Aug 23 19:23:39 Load Address: 00008000 Aug 23 19:23:40 Entry Point: 00008000 Aug 23 19:23:42 Verifying Checksum ... Bad Data CRC Aug 23 19:23:43 Marvell>> Aug 23 19:24:19 dang, to bad it didn't display the checksum and there isn't another tool to verify it. Aug 23 19:25:04 I was going to repair the boot partition, then try again. Aug 23 19:25:41 I guess I'll forge ahead and see what happens. :| Aug 23 19:26:25 I think it happened when I wrote the environment... saveenv... Aug 23 19:26:36 But could have been something else. Aug 23 19:26:59 can't see how that would screw up an image crc.... (but I'm not smart about this stuff yet..) Aug 23 19:27:55 I was wondering where the environment was actually saved... if it's saved in NAND somewhere I guess it could be all confused and write over the boot images Aug 23 19:28:08 * davidthings is not that smart about this either Aug 23 19:28:09 I would interpret this as the Kernal image itself having a bad crc, either on the card or after being loaded into memory Aug 23 19:29:22 I'm thinking (please correct) that the procedure in question doesn't copy the kernel... certainly it doesn't make the same partition structure that the NAND had Aug 23 19:32:07 * rbgray is getting confused again... Aug 23 19:33:26 davidthings, 'nand read.e' copies kernel. From flash to ram. Aug 23 19:33:33 only that :) Aug 23 19:33:59 Yeah. I'm thinking about the "SD_Card_As_Root_File_System" procedure Aug 23 19:35:07 Doesn't change where the kernel image is coming from, just where the root fs is, if I'm right Aug 23 19:35:22 yes Aug 23 19:35:49 Can you give me a pointer as to how I could repair my bootimage... Aug 23 19:35:56 so the gain here is a bit of speed, plus making sure that flash won't get worn out Aug 23 19:36:27 davidthings, the doc that came together with sheevaplug says how to boot kernel from tftp Aug 23 19:37:09 Yeah. Just looking at that. Seems sad to redo the whole of the NAND when all I need is to repair the boot Aug 23 19:37:14 once you're booted, you can use flash_eraseall and nand_write to write new kernel into your nand flash Aug 23 19:37:22 what do you mean by 'whole NAND'? Aug 23 19:37:35 you don't need to touch uboot nor root filesystem Aug 23 19:37:43 (unless they're broken too :) Aug 23 19:37:47 Oh! Aug 23 19:38:00 So flash_eraseall doesn't erase all flash Aug 23 19:38:12 no, you point it at mtdN Aug 23 19:38:18 (AFAIR :) Aug 23 19:38:26 Oh. Great. I'm on it. Thanks for the pointers. Aug 23 19:39:04 the difference seems to be that _eraseall erases whole "device", while with _erase you can specify the range to wipe Aug 23 19:39:29 and "device" is /dev/mtdN - run cat /proc/mtd to see what do you see Aug 23 19:39:44 Sadly, I can't run anything right now. Aug 23 19:39:50 (Dead in the water) Aug 23 19:40:05 IIRC, the pdf from CD was describing that you should see 3 devices, while by default you get only 2 (kernel image and rootfs) Aug 23 19:40:27 davidthings, as long as you have working uboot you should be able to recover, nuked uboot is harder to recover from Aug 23 19:40:33 (needs extra hw I think...) Aug 23 19:40:46 Yeah. UBoot is good, so far Aug 23 19:41:05 davidthings, also, the procedure in the pdf talks about pointing root to nfs Aug 23 19:41:29 which is probably not needed if your rootfs in flash is fine Aug 23 19:41:42 OK Aug 23 19:42:37 Will I need NFS etc. for tftpboot? Aug 23 19:43:16 you need nfs if you want to use root fs from nfs Aug 23 19:43:34 if the only thing you want to do is get kernel, tftpd should be enough Aug 23 19:43:47 Great. Thanks again. Aug 23 19:44:33 good luck, I was in similar situation yesterday :) Aug 23 19:44:56 Oh boy. Aug 23 19:47:00 * rbgray has to be careful, because he's on a Mac and is not sure of the utilities to put things back (w/jtag and such) Aug 23 19:48:59 you mean, to bring back borked uboot? Aug 23 19:49:18 rbgray: is it an Intel Mac? (Either way, you can download the Ubuntu Desktop ISO, burn that to disk, and boot from that for recovery when MacOS can't) Aug 23 19:49:42 standby, wife is talking. Aug 23 19:49:56 high priority interrupt! :) Aug 23 19:54:05 hah. Found excellent page about small, yet annoying problem: http://www.shallowsky.com/linux/noaltscreen.html Aug 23 19:55:45 Funny. That's annoyed me for 10 years. Aug 23 19:57:56 less -X also works, but obviously only for less Aug 23 19:58:16 yacoob: Any reason why you choose TFTP boot rather that USB stick boot? Aug 23 19:58:51 yes, didn't have usb stick on hand :) Aug 23 19:58:56 OK Aug 23 19:59:52 whew! Aug 23 20:00:11 * rbgray made the mistake of trying to block that signal this morning..... Bad Idea! :o Aug 23 20:01:14 so, here's another question. What the hell is discoverd in /root ? Aug 23 20:02:02 Q_Continuum: yes, but I have VMWare, so I that's probably how I would go if I need to, assuming the USB serial would not be a problem. Aug 23 20:02:15 I hope to set up a native tool chain eventually... Aug 23 20:02:42 or just use pre-built stuff, loading would be the only problem. Aug 23 20:02:45 Eh, I don't know how well virtual machines have access to USB-serial connections...IF it does, great. If not, just reboot off the Live Distro and fix it, then go back to mac :-D Aug 23 20:03:25 Q_Continuum: supposedly usb should work ok. We'll see. Aug 23 20:03:40 Either way, that's your fix-path. Ubuntu in a VM or off a disc :-D Aug 23 20:03:40 * rbgray returns to SD card building from two levels of interrupt... Aug 23 20:03:50 wow, you're so reentrant! Aug 23 20:03:52 8) Aug 23 20:04:18 less so than I used to be... Aug 23 20:05:37 hmm, I think I'm done with the plug. Now I can get to the part where I set up a lot of stuff on it and we all live happily everafter. Aug 23 20:05:53 Then you order a second one because you want to putz with it more :-P Aug 23 20:06:04 Also, I think I convinced the local MicroCenter to order Plugs to sell... Aug 23 20:06:36 oooh, MicroCenter... Aug 23 20:07:01 I have one about 40 minutes away in Cincinnati Aug 23 20:07:03 Showed it to one of the com sales reps I know, and I know he uses Linux Aug 23 20:07:16 I've waited ~2months for my order to arrive... although the delay was vendor's fault, not customs :) Aug 23 20:07:17 Was right at closing time, so they're all crowding around. Aug 23 20:08:48 Pulled up the page, described it...coworkers walked up, he describes it...sales manager comes by one of them goes "We gotta get these to sell!" "Ok, email me the link and I'll get it to the ordering dept" Aug 23 20:08:56 I'm kind of tempted to have a go at preparing a kernel with initrd for lvm, but tbh I don't really need lvm... :) Aug 23 20:09:27 what's initrd? Aug 23 20:09:45 initial ramdisk Aug 23 20:09:57 * rbgray swims through a sea of new acronyms. Ah. Aug 23 20:10:04 minifilesystem consisting of scripts that will run as kernel boots Aug 23 20:10:07 Heh. Aug 23 20:10:23 you need it when you need to do special stuff before you access / for example Aug 23 20:10:24 Q_Continuum: which MC was this? Aug 23 20:10:27 rbgray: I've found it's easier to just focus on what you need, learn as you go :-D (But learning is never bad) Aug 23 20:10:29 (like, lvm or raid...) Aug 23 20:10:36 Minnesota. Aug 23 20:15:26 Sometimes I think Radio Shack would be a good place for some widgets which could be programmed by end users. So many products these days are just crappily designed as far as their user interfaces. My last answering machine was TERRIBLE. I'd love to have a port on the back where I could plug in and reprogram it. Or turn it into different things. Imagine if they got user programmable... Aug 23 20:15:28 ...devices and paired them with open source projects? Most consumers couldn't program the devices, but the open source community could and then there'd be a rich source of options available to all. Aug 23 20:16:06 * rbgray stops dreaming and gets back to starting on that SD card again... Aug 23 20:18:12 What? Radio Shack has things other than cell phones now? :-P Aug 23 20:18:51 'The Shack' still has some parts you will not find at big box stores... Aug 23 20:19:14 The local 'Shack' guy figures 1/3 of their sales is from the DIY parts corner. Aug 23 20:19:40 breadboards, resistos, etc Aug 23 20:20:49 the good old days :) Although this morning I realized that my Sheeva Plug is 2009's version of the KIM-1. ;) Aug 23 20:21:04 That's what it all used to be. Poor Radio Shack. Aug 23 20:21:13 Never had a KIM-1. :-( Aug 23 20:21:42 I did build a computer out of that microprocessor kit 'Shack used to sell. That was probably close to the same level of annoying as the KIM. :-D Aug 23 20:23:18 heh, whole 1K RAM 2K ROM, 20 ma current loop TTY interface, audio tape storage - what a difference 3 decades makes! http://oldcomputers.net/kim1.html Aug 23 20:24:00 Yea, storage would have been a nice feature on mine. Had to key things back in every time. Aug 23 20:26:36 Look at it now. The SheevaPlug is equivalent to a 1999 $2000 PC. That fits in a jacket pocket. For $99. Aug 23 20:27:23 A bit harder to read the screen, but yea. Aug 23 20:28:00 It's almost literally milliions of times better! Aug 23 20:28:09 Smaller, cheaper...no moving parts... Aug 23 20:28:11 why the sheeva doesnt have eSATA is beyond me Aug 23 20:28:16 then it would be a PC Aug 23 20:28:22 rektide: OpenRD if you need more I/O Aug 23 20:28:32 i bought openrd, but its frakking huge Aug 23 20:28:47 I wish they'd stack the OpenRD with more RAM and max out the CPU. Aug 23 20:28:59 1.5GHz/2GB is the max the chipset/board can handle. Aug 23 20:29:07 imagine 2GB of RAM. Aug 23 20:29:14 Marvell.com advertises 1.2GHz tops Aug 23 20:29:22 The specs for the board list 1.5 Aug 23 20:29:29 Marvell's specs. Aug 23 20:29:35 as do the docs on the CD with the plug Aug 23 20:29:41 hrm Aug 23 20:29:54 http://www.marvell.com/products/embedded_processors/kirkwood/index.jsp Aug 23 20:29:57 that says 1.2GHz Aug 23 20:30:28 That's what they have now. Aug 23 20:31:43 I'd love 2GB, but it's just too expensive. Aug 23 20:31:45 Reedy: the MAC address is the default if you are using Sheevaplug install alpha-6. You need to edit uboot-custom.txt if you want to change that Aug 23 20:31:54 ram is too expensive? really? Aug 23 20:32:01 Yeah, i had already found and sorted that :) Aug 23 20:32:06 Reedy: or you can change 'ethaddr' in u-boot Aug 23 20:32:23 rektide: yes, SDuensin is our resident "building something off the Sheeva design" guru (by virtue of simply doing that) Aug 23 20:32:29 Reedy: something like - 'setenv ethaddr ' Aug 23 20:32:31 My routers being a PITA, so i cant work out if its the plug being funny or just the router (most likely the router atm) Aug 23 20:32:43 rabeeh, yeah, i fixed it by editing the txt and reflashing :) Aug 23 20:32:45 and apparently more RAM for the Sheeva IS expensive. (Single supplier, they don't like doing multiple models) Aug 23 20:32:56 rektide - I asked about it. What Q_Continuum said. Aug 23 20:33:06 You can change the MAC Addr from the Marvell command line (UBoot?) Aug 23 20:33:27 The first 512M is like $3. The next 512M is $29. Aug 23 20:33:50 Which I would've paid for. Going to 2GB, I doubt it. Aug 23 20:34:02 SDuensin: what did you build from the Sheeva design? Aug 23 20:34:18 rektide - www.getrosie.com Aug 23 20:35:03 Is it the processor itself thats locked in to using this single suppliers ram? Aug 23 20:35:13 or is it just that using more conventional ram would require board redesign? Aug 23 20:35:29 It's part of the board, so getting more requires a board assembled with it. Aug 23 20:35:56 At least, currently. Aug 23 20:36:09 Someone had the idea of clip-on RAM but I don't know how well that'd work. I'd buy it if someone made it, though :-D Aug 23 20:36:18 Just add a bit more cost ;) Aug 23 20:36:40 If you could buy a single plug with extra RAM, it wouldn't be so bad. However, you need to order an entire run to do it. So multiply that extra $29 by a few thousand units. :-) Aug 23 20:36:46 I need to do an Ubuntu install to an external HD later today. Aug 23 20:36:58 First gotta find the HD... lol Aug 23 20:37:23 SDuensin: What is their base run count? 5k, 10k? Aug 23 20:37:30 Or do you not really know? Aug 23 20:37:52 Q_Continuum - I don't really know. I think it depends on how big a change you're making and what the "sweet spot" is price-wise. Aug 23 20:38:01 Figured. Heh. Aug 23 20:38:08 For the right price, they'll build 20. Aug 23 20:38:10 I mean, you *CAN* get one plug with 2 gig. It'd probably cost you a few thousand dollars though. :-D Aug 23 20:38:14 Exactly. Aug 23 20:39:02 Like the black case Rosie uses. I managed to get a small run, but they were expensive. However, it was better than getting thousands of them and then deciding I didn't like it. Aug 23 20:39:41 Which page on the wiki documents directing an install to say, an external HD? Aug 23 20:39:46 I'm not finding it :-/ Aug 23 20:40:18 Found the doco on network-booting Aug 23 20:40:40 http://plugcomputer.org/plugwiki/index.php/External_Hard_Drives Aug 23 20:40:46 Thanks Aug 23 20:41:07 SDuensin: you noted that Rosie comes with a 110 US plug, but failed to note that it can operate on, what is it, 100-240 volts 50-60 Hz. Aug 23 20:41:27 Right. I need to update the site. BADLY. Aug 23 20:44:37 It's going to be interesting to see what comes of all this - some interesting appliances, I bet. Aug 23 20:45:16 All I want to NOT see is bankruptcy. Aug 23 20:54:18 ;) Aug 23 20:54:36 * rbgray thinks writing the ext3 journal is taking a long time... Aug 23 20:54:43 I really need to get my plugs into production use ;) Aug 23 20:55:46 do these things have any sort of power failure detection such that the system might be able to do a Really Fast shutdown if it gets unplugged? Aug 23 20:56:09 i suspect not Aug 23 20:56:16 Nope. Aug 23 20:56:22 it'll only have whatever is in the capacitors keeping it alive for seconds Aug 23 20:56:50 but that's no good if it doesn't know that power has been lost... Aug 23 20:56:59 And the caps aren't going to be much help spinning discs. Aug 23 20:57:05 ah, SD card is initialized. Aug 23 20:57:28 SDuensin: right, those would suck the PS dry very quickly too. Aug 23 20:58:23 assuming it was USB powered... Aug 23 21:09:54 :( Aug 23 21:09:56 I'm blaming my plugs being funny on the router. It needs replacing Aug 23 21:11:41 yacoob: booting from tftp... thanks. That's fun. Aug 23 21:13:12 Makes development a lot easier, that's for sure. Aug 23 21:17:02 my openrd's clock doesnt move. :( Aug 23 21:17:24 http://groups.google.com/group/openrd/browse_thread/thread/a5233afb33708a94 Aug 23 21:24:05 * rbgray boots off his 8G SD card :) Aug 23 21:32:57 Hello Aug 23 21:33:13 Greetings. Aug 23 21:35:02 Does someone know what "FTL header not found" means ? Aug 23 21:35:13 (It seems related to MTD partitions...) Aug 23 21:38:05 Faster Than Light? Cool. Aug 23 21:41:24 I had always known that Linux was ahead of it's time. It's looking for a quantum computer Aug 23 21:43:11 Seriously though, I don't know what that is. Aug 23 21:48:51 In my opinion, Linux says I detected a partition here but there is no correct header Aug 23 21:53:37 ah, Flash Translation Layer Aug 23 21:54:50 got multiple Google hits for your message "FTL header not found", but don't understand them... :/ Aug 23 21:55:47 So do I Aug 23 21:59:59 gargh Aug 23 22:00:22 what username/password should I use to login to the Ubuntu install after running recovery? Aug 23 22:01:56 looks like there's just a root account Aug 23 22:01:59 but... Aug 23 22:02:00 oh Aug 23 22:02:15 silly me Aug 23 22:02:19 it was the original password Aug 23 22:07:18 rbgray: No issues with the SD card, I hope? Aug 23 22:24:22 seems ok, but I've been dealing with various "emergencies" like my daughter's Mac Mini dying - but wait, it was the monitor that died.... Just gettinb back to it to set up multiboot between SD & NAND. Aug 23 22:25:02 * rbgray must now move another monitor back to the basement and maybe even start the charcoal too. ;) Aug 23 22:25:50 aww, if we loose another CRT, we'll have to replace it with a flat screen... ;) Aug 23 22:27:17 I can't get my rootfs mounted, I think there is something wrong with my mtdparts parameter Aug 23 22:27:59 Where is rootfs located when you used marvell rootfs ? 0x300000 ? 0x500000 ? Aug 23 22:28:37 *you used marvell rootfs to flash tour nand Aug 23 22:29:06 I want to get the multi-boot working, then dive in on the SD card, doing all the fixes and updates, etc. Then maybe put something interesting on it, like Sendmail or maybe gcc so it can build itself or the $DAYJOB product. Aug 23 22:31:24 Doc says 0x500000 but kernel says 0x300000 and none works Aug 23 22:42:27 Both works if I specify rootfstype=jffs2 Aug 23 22:42:32 Strange... Aug 23 22:45:53 The last works (and it's the one which make sense), but, Debian says no space left on root Aug 23 22:46:01 Which is logical Aug 23 22:53:40 http://www.openplug.org/plugwiki/index.php/Multi-Boot sez mmcinit is not in the current flash? Guess I'm off to Linux land after all. Or is that needed if one just moves the root fs? Aug 23 23:45:45 Good bye everyone Aug 24 00:17:05 arg ethernet isnt working on the git marvell tree for me Aug 24 00:20:03 rektide: what's not working? Aug 24 00:20:48 the driver gets intialized, but it doesnt negotiate the link correctly (its still set at 10mbps) and its not send/receiving traffic Aug 24 00:45:03 ADDRCONF(NETDEV_UP): eth0: link is not ready Aug 24 00:45:28 rektide: what's the last version that worked for you? Aug 24 00:45:40 the bundled 2.6.22.18 Aug 24 00:47:24 http://pastebin.com/mbc2ded4 is everything ethernet related i can compile Aug 24 00:54:11 honestly i have no idea whether MV643XX is meant for the kirkwood platform or not Aug 24 00:59:32 crap Aug 24 00:59:41 http://www.cyrius.com/debian/kirkwood/qnap/ts-219/status.html says it should be supported via the mv643xx module Aug 24 01:48:23 rektide: you still having problems? Aug 24 01:48:32 yeah. Aug 24 01:52:08 rektide: so it never worked for you on a kernel.org kernel? Aug 24 01:52:17 rektide: I see in the defconfig CONFIG_NET_DSA_MV88E6123_61_65=y Aug 24 01:52:33 tinker-f595: that's for switch chips, which that box doesn't have Aug 24 01:53:06 then why is in the open-rd defconfig ? Aug 24 01:53:28 and I see CONFIG_NET_DSA_MV88E6XXX=y Aug 24 01:53:33 lennert: this is a marvell git kernel, and its the only kernel i've attempted, outside the bundled 2.6.22.18 Aug 24 01:53:47 tinker-f595: 88e6xxx is a line of switch chips Aug 24 01:53:52 tinker-f595: and i don't know which openrd defconfig you're referring to Aug 24 01:54:03 the one Dhavel posted Aug 24 01:54:34 i neglected to mention i was running OpenRD Aug 24 01:54:38 lennert: look in this thread http://groups.google.com/group/openrd/browse_thread/thread/68e7fb1eb2ea3e0f Aug 24 01:54:40 where? Aug 24 01:55:14 or look here http://openrd.googlegroups.com/attach/e1ced388a2c3eba9/openrd_defconfig?view=1&part=2 Aug 24 01:55:19 kirkwood boards should just use the kirkwood defconfig Aug 24 01:55:26 no need to have a different defconfig for each board Aug 24 01:55:41 and the kirkwood defconfig has dsa support because e.g. the 6281 rd board has a dsa switch chip Aug 24 01:55:53 so i assume he just took the kirkwood defconfig and changed a couple of lines Aug 24 01:56:02 but either way there shouldn't be a separate defconfig for it Aug 24 01:56:22 http://pastebin.com/m7260d4c3 Aug 24 01:56:36 i seem to have all the relevant config's compiled in Aug 24 01:56:53 lennert: openrd-base and openrd-client are different from the sheevaplug Aug 24 01:57:00 tinker-f595: i know Aug 24 01:57:35 I am wondering if Dhavel posted the wrong file? Aug 24 01:57:38 tinker-f595: kirkwood_defconfig currently already supports multiple different kirkwood boards Aug 24 01:57:45 tinker-f595: i.e. all the boards currently supported in the kernel tree Aug 24 01:58:19 well according to dhavel he had to make a number of changes and supply patches Aug 24 01:58:38 either that or he's used to the old pxa style way of doing things which was to have one defconfig per board Aug 24 01:58:45 tinker-f595: well, sure, you'd have to add support for the board Aug 24 01:58:51 tinker-f595: that still doesn't mean you need a separate defconfig Aug 24 01:59:34 but it would need certain things commented out or uncommented that are not active on the sheevaplug Aug 24 01:59:55 .... Aug 24 02:00:13 ? Aug 24 02:00:48 hmm im going to have to switch from usb boot to sd boot Aug 24 02:00:49 http://pastebin.com/m3866f920 Aug 24 02:01:27 for some reason u-boot is failing to boot usb when i have a not-insignificant number of usb devices plugged in. Aug 24 02:01:54 lennert: Dhaval also had this thread on patches for the Open-RD Client Aug 24 02:02:34 lennert: both rektide and I want to get an up to date kernel working on our open-rd devices Aug 24 02:03:02 it may be worthwhile to diff Dhaval's patches against the orion versions of the files Aug 24 02:03:09 but man that'd be a lot of work Aug 24 02:03:34 I want to get fed 11 booting off a sata/esata device attached to my OpenRD -Client Aug 24 02:03:43 if dhaval's patches never gave you working ethernet, then the ethernet phy address is probably wrong or so Aug 24 02:04:04 rektide you asked dhaval about that? Aug 24 02:04:32 I havent asked him. The idea of asking for that public git server is smart. Aug 24 02:04:53 lennert: any idea if he sent his patches up stream? Aug 24 02:05:21 tinker-f595: there's quite a few OpenRD itmes on the marvell tree (which is what i've been using) Aug 24 02:05:29 lennert: I am assuming you are the same person listed on the fedora arm page Aug 24 02:06:04 tinker-f595: don't know if he sent his patches upstream, i don't real linux-arm-kernel@ actively anymore Aug 24 02:06:07 tinker-f595: and yep Aug 24 02:07:58 lennert: do you still stuff with the fedora arm port project? Aug 24 02:08:09 not much Aug 24 02:08:17 kedar/jitesh do pretty much all of the work these days Aug 24 02:08:30 * lennert is back to kernel things Aug 24 02:09:23 Followup "please make public your git" post: http://groups.google.com/group/openrd/browse_thread/thread/b4981ecb40628a00 Aug 24 02:09:35 lennert: so it is being actively support still. That is good Aug 24 02:10:07 rektide: thanks for that post Aug 24 02:10:09 tinker-f595: yep Aug 24 02:10:43 lennert: I am doing a proof of concept of porting the OLPC XS to ARM. Aug 24 02:10:45 you should just discuss this on linux-arm-kernel@ Aug 24 02:11:07 email list server? Aug 24 02:11:38 http://lists.infradead.org/mailman/listinfo/linux-arm-kernel i think it moved to Aug 24 02:11:38 tinker-f595: i figued since he's in india, he's probably getting to work sometime soon-ish, and he could use a little push. Aug 24 02:11:57 the people that can help you with these issues are likely not subscribed to your group Aug 24 02:13:55 lennert: thanks Aug 24 02:14:15 wonder if we should start an OpenRD irc channel Aug 24 02:15:48 i joined. :) Aug 24 02:15:52 we can see if anyone shows up Aug 24 02:16:09 but i still plan mostly on using this channel Aug 24 02:16:57 man i'm kind of freaked out uboot dies so poorly once i have multiple usb devices plugged in Aug 24 02:17:42 i've done a considerable amount of mix & matching, trying various hubs, cables, and sets of components. it seems pretty reliably to fail if i have any more than a hub and two devices. Aug 24 02:18:06 just a pity, cause sdhc is so ridiculously slow compared to these flash usb drives Aug 24 02:18:09 Ok, for setting up an install of Ubuntu off of an external drive...do I do an actual install, or do I copy the files onto the drive, then set uboot to boot from it? (The wiki seems to only include the info on setting uboot up to accept it as an option) Aug 24 02:18:34 rektide: I joined linux-arm-kernel Aug 24 02:19:25 Q_Continuum: just copy the files Aug 24 02:21:14 Q_Continuum: I worte that page just to show how to boot. There are a number of other wiki pages that describe how to build the file system. I guess I could put a link to one of those. The main purpose of the page I created was to get compatibility information Aug 24 02:21:34 I guess I should post to the forum that it exists and the purpose of the page Aug 24 02:22:55 Yeah. I'm setting it up off a USB HD, with fallback to NAND. Which the multiboot page will tell me how to do, now just to get an install on the HD :-D Aug 24 02:23:33 (Also, do I want to use the same image as in the alpha-6 installer? It can write to the HD all it wants, so I'd rather it not boot the OS to RAM and waste that space...just boot normally) Aug 24 02:23:36 let me remember were I got that info... Aug 24 02:23:52 I used the amse stuff as the alpha 6 Aug 24 02:24:21 Take your time, I'm not in a huge rush - just want to get the box started at some point :-D Aug 24 02:25:18 what drive and enclosure you planning on using? Aug 24 02:26:29 Toshiba 250GB SATA, in a NexStar-SX enclosure. Aug 24 02:26:33 I can find links to both if needed. Aug 24 02:28:27 Had the drive for awhile, just picked up the enclosure from MicroCenter yesterday. Aug 24 02:32:16 2.5 " or 3.5" Aug 24 02:32:51 Q_Continuum: be interesting to see how it performs Aug 24 02:40:30 ok, u-boot reconfigured & refactored for sd support Aug 24 02:40:50 crap, no mmcinit in the stock uboot? Aug 24 02:41:35 frell Aug 24 02:44:51 use the one with the alpha6 installer Aug 24 02:47:00 2.5" SATA, so don't need a power source either :-D Aug 24 02:49:07 Damn, I can't locate my spindle of blank DVDs. I think I might have to actually go buy some again... Aug 24 02:49:29 Gotta put Win7 final on the old laptop, and it won't boot from USB, so can't use that method that I used for the desktops. Aug 24 02:49:37 final/RTM Aug 24 02:58:12 2.5" sata have lower inertial mass generally and use less energy to start moving Aug 24 02:58:13 Win7 why? **** ENDING LOGGING AT Mon Aug 24 02:59:57 2009