**** BEGIN LOGGING AT Sun Jan 29 02:59:57 2006 Jan 29 03:20:29 ~seen beewoolie Jan 29 03:20:33 beewoolie was last seen on IRC in channel #openjtag, 1d 2h 13m 41s ago, saying: 'Ah, now I see. It's burried in the description.'. Jan 29 03:21:47 morning everyone Jan 29 03:22:03 philips any progress on audio? Jan 29 10:09:19 just managed to load and start apex.bin using tftp Jan 29 10:09:54 as I have no serial yet, (yeah, I know I should spend some time on it) I can only see that it starts (beep + led turns red) Jan 29 10:10:18 anyway, added some meat to the wiki page concerning this experiment: http://www.nslu2-linux.org/wiki/HowTo/RunApex Jan 29 10:10:34 http://www.nslu2-linux.org/wiki/HowTo/RunApex Jan 29 10:19:45 my slug is running with apex, but i have a serial... Jan 29 10:34:19 mental note: someone promised me a cable, need to check if it is already available Jan 29 10:34:55 <[g2]> sinclair73de did you flash APEX ? Jan 29 10:36:07 if I could telnet into it like I could do with apex, I think I would be able to find the courage to flash it (power outages are very rare here) Jan 29 10:39:20 yes, after testing it in sdram Jan 29 10:40:04 <[g2]> sinclair73de cool. Do you know about the APEX list ? Jan 29 10:40:26 <[g2]> and what are you using APEX for ? Jan 29 10:40:43 APEX list? Jan 29 10:41:08 Using apex for 128mb ram access Jan 29 10:41:18 <[g2]> one sec... Jan 29 10:41:24 <[g2]> search for the URL Jan 29 10:41:44 <[g2]> did you go with 32M16 parts ? Jan 29 10:41:53 sinclair73de, did you upgrade the slug yourself to 128M ? Jan 29 10:42:03 pretty cool Jan 29 10:42:35 <[g2]> sinclair73de I see you are #4 on the Alternate Bootloaders list :) Jan 29 10:42:42 yes 32M16 Jan 29 10:42:43 haven't really run into memory issues, but I feel sometimes I'm filling it up reasonably well Jan 29 10:43:09 cool, wish I could do that, but it is way beyond my soldering capabilities Jan 29 10:44:18 yes, i'm unsolded und soldered the two chips Jan 29 10:44:53 not easy this work Jan 29 10:44:57 nope Jan 29 10:44:58 <[g2]> heh Jan 29 10:45:07 I might've broken one of my slugs :-P Jan 29 10:45:29 <[g2]> sinclair73de I was going to sell memory modded slugs and have professional re-work done Jan 29 10:45:47 * [g2] wound up building hw Jan 29 10:46:05 i thought this too... but the problem was a bad soldered pin of the ram Jan 29 10:47:03 I managed to break a solder pad. Fortunately it was NC Jan 29 10:47:22 g2: whats the cost for these mod? Jan 29 10:48:26 <[g2]> sinclair73de to bump to 64MB memory and add serial was like and extra $75 at cost Jan 29 10:48:39 <[g2]> that was just the serial header Jan 29 10:48:44 the pads are no problem, i used hot air to unsolder Jan 29 10:49:02 <[g2]> sinclair73de are you in DE ? Jan 29 10:49:20 g2: yes Jan 29 10:50:00 why? Jan 29 10:50:45 btw, do you know this guy: http://www.sigmadesigns.com/pdf_docs/Envision8634_br.pdf Jan 29 10:51:08 300 mhz mips, additional high definition mpeg encoder etc Jan 29 10:51:15 looks like a very neat toy to me Jan 29 10:51:30 <[g2]> sinclair73de do you do much rework ? Jan 29 10:52:19 <[g2]> and sinclair73de which OS are you running unslung, openslug, or debian ? Jan 29 10:52:24 anyone tried to run Debianslug (LE) with apex? Jan 29 10:52:36 <[g2]> I'm running Debian LE on my hw Jan 29 10:52:40 g2: running openslug 2.7 Jan 29 10:53:05 <[g2]> have you run stressful HD access with > 64MB enabled ? Jan 29 10:53:25 g2: is it possible to test the Debianslug-image in ram with apex or must i flash it? Jan 29 10:54:22 <[g2]> sinclair73de are you JTAG enabled ? Jan 29 10:54:58 copy about 500mb is no problem, but nfs access to these (mpg)-files crashes the slug Jan 29 10:55:21 i dont know why... Jan 29 10:55:53 <[g2]> do you get errors ? Jan 29 10:56:10 g2: no, jtag not enabled, but i have a jtag adapter if i need one Jan 29 10:56:26 yes... but not logged Jan 29 10:56:38 <[g2]> and more importantly some decent soldering skills :) Jan 29 10:56:45 the next time i copy it Jan 29 10:58:33 i recorded 3GB mpeg data from my dbox2 with udrec (a mono application) on my slug, that works without any problems Jan 29 10:59:28 <[g2]> Is that running the 2.6.11.12 kernel ? Jan 29 11:00:04 what? Jan 29 11:01:10 <[g2]> sinclair73de I was wondering which kernel you were using with 2.7 Jan 29 11:02:47 <[g2]> sinclair73de I've got to run out right now Jan 29 11:03:07 <[g2]> could you possibly e-mail me ? Jan 29 11:03:23 <[g2]> The contact @ http://www.giantshoulderinc.com Jan 29 11:03:33 2.6.12.2 Jan 29 11:03:52 <[g2]> You may want to see the http://www.giantshoulderinc.com/hw-4533 pics Jan 29 11:04:09 <[g2]> or http://www.giantshoulderinc.com/ab3/case.jpg Jan 29 11:04:19 sinclair73de udrec? what system were you using? debianslug? Jan 29 11:04:39 <[g2]> EvilDevil openslug 2.7 Jan 29 11:04:41 ok, g2 i email to you Jan 29 11:04:46 <[g2]> THX Jan 29 11:05:05 <[g2]> I'll e-mail back in a bit (a couple hours) Jan 29 11:05:13 <[g2]> bbl Jan 29 11:05:35 EvilDevil: you know udrec? Jan 29 11:06:15 sinclair73de yep, i followed the discussion about porting it... :) would be nice because recording with nfs isn't that reliable here Jan 29 11:08:11 EvilDevil, sinclair73de, why record using nfs. That means you always need another system on while recording. I record on a local hd, then stream using UPnP or playback on pc using nfs & mplayer Jan 29 11:09:10 eFfeM: dbox2 doesn't support internal hd... Jan 29 11:09:25 eFfeM and my slug is up 24/7 anyway Jan 29 11:10:27 i still using a notebook with suse for recording with udrec_suite and want to do this on my slug (is possible) Jan 29 11:11:04 EvilDevil, I now understand. didn't know what a dbox2 is. I run a hauppauge pvrusb2 on my slug (the one with the hw mpeg2 encoder) Jan 29 11:11:25 so my recording happens on the slug (which is also on 24/7) Jan 29 11:12:24 eFfeM a dbox2 is a set top box for dvb-s with 10mbit ethernet interface Jan 29 11:12:43 and running linux :-) Jan 29 11:13:07 EvilDevil, is this similar to dreambox ? Jan 29 11:13:44 eFfeM yes.. the dreambox is kind of a successor to the dbox2 Jan 29 11:14:00 ah ok Jan 29 11:14:54 they're nice Jan 29 11:15:14 but pretty buggy... ;) Jan 29 11:17:21 anyway... sinclair73de how did you compile udrec? Jan 29 11:18:45 i compile udrec on my windows pc under vs2003, but compiling mono native on the slug was not simple Jan 29 11:19:26 i think not everything is working, but what i need works Jan 29 11:20:14 okay... was there any code-mangling ? Jan 29 11:21:41 i dont know what code-mangling meens Jan 29 11:22:12 changing makefiles / source-code? Jan 29 11:23:25 in mono? no, i loaded the precompiles runtime and compile it native... i got many warnings because of the BE, but it works Jan 29 11:23:51 okay... did you built mono from the source? Jan 29 11:24:25 no, thats not working... i tried it, but no success Jan 29 11:26:13 well.. i'll try it later... on debianslug... :-) thanks for the infos so far Jan 29 11:26:31 i want try this too Jan 29 11:27:00 i think LE is better for this Jan 29 11:27:19 good to hear that :) may need some expert advice on this Jan 29 11:27:36 EvilDevil: what is your slug running? OpenSlug? OpenDebianSlug? Jan 29 11:28:31 debianslug Linux slug 2.6.15 #1 Thu Jan 19 12:14:56 GMT 2006 armv5tel GNU/Linux Jan 29 11:29:26 and what bootloader? Jan 29 11:30:32 stock redboot Jan 29 11:31:47 no idea how to run debianslug without flashing? Jan 29 11:34:20 no, sorry Jan 29 11:34:51 maybe telnet into redboot and load firmware by tftp Jan 29 11:36:29 i tried these... but got many error, seams Endian specific problems Jan 29 11:47:49 no clue.. haven't tried it at all :) Jan 29 11:48:26 just did a straight flash from SerComm Jan 29 11:48:47 why don't you just back up your old flashimage and flash the new one? Jan 29 11:51:06 thats possible... changes flashing with sercomm anything on the bootloader? Jan 29 11:52:28 i'm not sure. it didn't change my bootloader, but if you're using another bootloader you may get other results Jan 29 11:54:26 thats why i'm ask, i dont know if it works with apex, and where in in boot process the cpu is switched to LE Jan 29 11:57:37 i think there were some mails on the nslu2-developer mailing list about this topic Jan 29 12:21:52 morning, jbowler Jan 29 12:24:03 morning... have you seen devlabel? Jan 29 12:24:27 not gotten around to looking at it, read the mail though Jan 29 12:25:24 looks like a good idea Jan 29 12:34:09 is there any reason not to do a release when devlabel works? Jan 29 12:34:40 ugh Jan 29 12:34:45 * NAiL rereads the mail Jan 29 12:38:40 devlabel, if it works, and a working 'beep' command - need to find the source for the Debian package... Jan 29 12:40:10 http://ftp.debian.org/pool/main/b/beep/ the .tar and the .diff Jan 29 12:40:29 plus the patch in http://bugs.debian.org/350214 Jan 29 12:42:42 The source is at http://ftp.debian.org/pool/main/b/beep/ Jan 29 12:42:50 Oops, bad paste... Jan 29 12:43:02 The source is at http://johnath.com/beep/ Jan 29 12:43:29 Now at 1.2.2 Jan 29 12:46:02 Debian seems to have loads of patches in the diff, btw Jan 29 12:56:42 n8 Jan 29 13:06:11 jbowler: what's the problem with devlabel? Jan 29 13:06:21 I'm not home, so I can't test it on a slug right now Jan 29 13:06:27 I don't think there is one yet - I just haven't looked at it (I've just built it...) Jan 29 13:06:34 ah, ok Jan 29 13:06:47 It sounds like it does the right thing, but I'm not sure how it interacts with udev Jan 29 13:07:46 I'll try it out later tonight Jan 29 13:07:47 I think turnup can drive it to add something like '/dev/root' and /linuxrc can mount using the -U option (so the fstab has /dev/root) Jan 29 13:08:27 sounds good Jan 29 13:08:55 beep 1.2.2 compiles, now to see if it works ;-) Jan 29 13:09:15 * NAiL sees a release closing in fast ;) Jan 29 13:11:59 I think maybe I should start stripping shell scripts, /sbin/sysconf is 24k and only compresses to 7.4k Jan 29 13:12:21 Need a shell script compiler... Jan 29 13:28:18 g2: back again? Jan 29 13:28:50 <[g2]> sinclair73de nod Jan 29 13:31:29 <[g2]> sinclair73de are you wondering about getting LE running ? Jan 29 13:32:29 very nice things on your homepage :-) Jan 29 13:35:02 <[g2]> heh Jan 29 13:35:52 i dont know so much about BE/LE... I only see the error messages when try to run debianslug from sdram Jan 29 13:36:27 g2: i think... not cheap this boards Jan 29 13:38:31 <[g2]> sinclair73de if you want to run debian LE you to a couple different ways you can go Jan 29 13:39:27 <[g2]> A) you can reflash the kernel in LE mode and boot APEX in LE and just boot the kernel normally in LE Jan 29 13:39:42 <[g2]> you'll have to build the kernel for this Jan 29 13:40:03 <[g2]> there's a 2.6.15 version in CVS to do something like this Jan 29 13:40:55 <[g2]> B) you can use the CVS version but prepends the LE commands to the vmlinuz Jan 29 13:41:48 <[g2]> c) you can use that debianslug plan as it being worked out Jan 29 13:46:14 i have used mastermakefile with make debianslug-image Jan 29 13:48:07 sinclair73de: that image (zImage-loftle) just works, but you need to byte swap the command line Jan 29 13:48:35 because i dont want to flash at this time, i unpack the image and tried to upload the ramdisk and kernel from apex Jan 29 13:48:49 [cc]smart: pb__ in #oe has a problem with a missing /etc/resolv.conf on non-openslug (well, non-slugos) apparently because populate-var.sh isn't creating it any more. Jan 29 13:49:10 <[cc]smart> just saw it, that's why i'm in. Jan 29 13:49:18 <[g2]> jbowler do you know how the mem param works he's got a 128MB unit Jan 29 13:49:34 sinclair73de: you can't do that, you need to run the zImage, if you uncompress it you will strip the switch-to-le code off the front. Jan 29 13:49:42 <[cc]smart> now my problem is, i've had three dead HDs in my server last week in one night breaking the raid Jan 29 13:50:00 <[cc]smart> so i gotta check by net Jan 29 13:50:02 <[g2]> [cc]smart OUCH! Jan 29 13:50:05 [cc]smart: I think it's just a matter of having a 01_core with a link to the /var/tmp/resolv.conf, but I wasn't sure. Jan 29 13:50:46 I.e. I think openslug(slugos) and nylon need an empty file but everything else needs an empty file in /var/tmp + a link. Jan 29 13:50:56 <[cc]smart> so /var/tmp/resolv.conf is generated during startup and what they want is a link generated in /etc as resolv.conf pointing to /var/tmp/resolv.conf. correct ? Jan 29 13:51:43 jbowler: zImage-loftle or zImage-nslu2le ? Jan 29 13:51:44 I think so - the origiinal (pre-r36) code seemed to do that, then (openslug/nylon) zap the link in do_install_append Jan 29 13:52:05 sinclair73de: nslu2 is for an nslu2 board, loft is for a loft board Jan 29 13:52:14 nas100d is for a nas100d board Jan 29 13:52:28 ok, i have a nslu2 board Jan 29 13:52:44 <[g2]> I think jbowler has 'em all Jan 29 13:53:04 <[g2]> jbowler is _such_ a *stud* Jan 29 13:53:08 sinclair73de: you have to build your own kernel, you need to modify the compiled in command line. Jan 29 13:53:33 hmmm? Jan 29 13:53:44 I.e. you have to change the CMDLINE in defconfig because you aren't booting off /dev/mtdblock4 and there is no way to change the command line on NSLU2 from the boot loader Jan 29 13:53:53 <[cc]smart> if zapping is still in place, then 00_core shows what to do which would be sth. likel root root 644 /var/tmp/resolv.conf /etc/resolv.conf Jan 29 13:54:32 <[cc]smart> with zapping this could be added to core Jan 29 13:54:58 <[cc]smart> removing the zapping and creating a new 01_resolv.conf thing would allow to install this for those that need it only Jan 29 13:56:04 <[cc]smart> ah, sometime i'm not reading precise. will see if pb__ is there Jan 29 13:56:17 the bitbake (.bb) file zapped it - the boot script doesn't need to Jan 29 13:56:29 I think it gets zapped because /var/tmp is on tmpfs Jan 29 14:00:20 this is in defconfig CONFIG_CMDLINE=" noirqdebug console=ttsS0,115200n8" Jan 29 14:01:27 right, so root=/dev/mtdblock4 and rootfstype=jffs2 gets prepended to that in work/ixp4xx-kernel*/*/arch/arm/mach-ixp4xx/nslu2-setup.c Jan 29 14:01:55 So if you add something to the end it will probably work (assuming you remember to byte swap the disk image if you load it into SDRAM). Jan 29 14:03:18 i think thats too much for me :-( Jan 29 14:26:02 jbowler: what is the right cmdline for booting from ram? Jan 29 14:45:11 koen, don't know if this interests you but in #nslu2-linux, snowbike is claiming he has the imic wokring Jan 29 14:45:17 s/wokring/working Jan 29 14:48:45 that's great news Jan 29 14:49:32 * koen is in Amersfoort right now IRCing over gprs Jan 29 14:50:17 koen this is something he mentioned: Jan 29 14:50:18 ok, rebooted, modprobe soundcore; modprobe snd_usb_audio; cat /dev/dsp > test.wav ; cat test.wav > /dev/dsp shows that it is working after reboot Jan 29 14:50:18 snowbike the other change I made was "mv /dev/dsp /dev/dsp0 ; ln -s /dev/dsp1 /dev/dsp" because the imic wants to be /dev/dsp1, and I'm really trying to save that valuable "1" key :) Jan 29 14:50:47 heh Jan 29 14:51:35 I'' try that when I get back home Jan 29 14:51:38 'night all Jan 29 15:06:49 sinclair73de: use initrd=... and either init= or rdinit= depending on what you are doing. You may have to know the exactly length of the initrd (true for a CPIO archive for example), but if you use a regular FS image I think that isn't necessary. You have to know the load point and give a length. Syntax can be found in the kernel source somewhere (search for initrd=) Jan 29 15:07:47 sinclair73de, I think your le boot with apex will just work. see http://groups.yahoo.com/group/nslu2-developers/message/559 Jan 29 16:34:05 Dunno about the insmod command, it should be doing that automatically, but perhaps not on Unslung 5.5 (I run 6.0 alpha) Jan 29 16:35:21 nevermind! **** ENDING LOGGING AT Mon Jan 30 02:59:57 2006