**** BEGIN LOGGING AT Sat Sep 17 02:59:56 2005 Sep 17 03:01:03 [g2]: i'm going to post the patches... the first one converts nslu2-setup.c and nslu2-io.c to a version compatible with the latest kernel Sep 17 03:02:43 <[g2]> dwery great! I was talking to dyoung last night and he's interested in 2.6.13 too Sep 17 03:02:57 I was about to just do it. Sep 17 03:03:11 but I'm distracted making this happuage PVR USB thing work Sep 17 03:03:35 <[g2]> with 4/5 people interested we should at least coordinate a tiny bit so we're all not wasting time doing the same thing Sep 17 03:03:39 in theory the nslu2-setup.c from 2.6.12.2 should work straight away. Sep 17 03:03:56 dyoung: no, I had to change it a little bit Sep 17 03:04:17 [g2]: if it's ok for you, I'm going to send the first patch Sep 17 03:04:41 <[g2]> dwery it's open source Sep 17 03:04:49 :) I mean post it to the ml Sep 17 03:04:57 <[g2]> oh absolutely Sep 17 03:05:22 <[g2]> this is a meritocracy Sep 17 03:05:43 <[g2]> nslu2-linux that is Sep 17 03:06:08 :) Sep 17 03:10:17 ok, post sent. Sep 17 03:11:23 Now I would like to generate a patch which converts a vanilla,mainstream 2.6.13.1 to an NSLU2 enabled kernel. Sep 17 03:11:53 or, better, a set of patches Sep 17 03:17:38 Thats the usual strategy. :-) Sep 17 03:21:15 regarding the ixp4xx-regs.patch .. I suppose this affects also other ixp4xx based architectures, isn't it? Sep 17 03:22:59 I mean.. that base regoster the file changes, is 0xFFBFE000 on the NSLU2 and 0xFFBFD000 on the others, or is just plain wrong and the patches fix it? Sep 17 03:24:25 Yes, thats incorrect and it was patched to fix it to correct. It was correct for a little while, but the bug got re-introduced. Sep 17 03:24:42 (its incorrect for *all* IXP4xx) Sep 17 03:25:45 ok, that makes it easy to have it mainstream :) :) Sep 17 03:25:56 dividi et impera :D Sep 17 03:27:00 I spoke with Deepak about it a while ago, but I dunno what happen after that. Sep 17 03:29:35 I'm checking .14-rc1 to see if the bug is still there Sep 17 03:30:13 yes,it is. a lot of ixp4xx changes anyway. So, if that piece shoudl go mainstream, who is the right person to contact? Sep 17 03:31:28 Deepak I suppose :) Sep 17 03:31:41 Or should we just post in the linux arm ml? Sep 17 03:32:11 Thats probably a good place. Sep 17 03:32:24 iirc, the person that noticed it in the first place was mark rakes. Sep 17 03:32:26 Ok. Can you do it? Sep 17 03:32:54 something simple.. like.. while working on... to the best of our knowledge this problem affect all ixp.... please apply. Sep 17 03:32:55 :D Sep 17 03:33:00 someone whos already got a presense there should do it; like [g2] or lennert or beewoolie. Sep 17 03:34:12 I would also CC Deepak and maybe a couple of other guys Sep 17 03:34:14 I dont subscribe to that list; but read it regulalry. Sep 17 03:34:47 [g2]: ? Sep 17 03:34:49 we ought to ask lennert in the nslu2-developer list. Sep 17 03:35:03 Hes kind of the de-facto interface to L-A-K folk. Sep 17 03:35:33 ok Sep 17 03:36:07 missing-exports now.. Sep 17 03:36:20 I might be speaking out of turn there; but thats my current understanding. Sep 17 03:36:28 I'm blissfully ignorant. :-) Sep 17 03:37:08 :) Sep 17 03:37:35 missing-exports.patch exports a couple of symbols.. who's going to use them? Sep 17 03:46:57 3dyoung, I was backreading (was away for lunch) and saw you are working on the hauppauge usb pvr Sep 17 03:47:19 3 is this the one with the embedded mpeg2 encoder ? Sep 17 03:48:45 (it has usb2 in its name) Sep 17 03:49:18 I am looking into turning the slug in a pvr and actually have been looking at what hw to use Sep 17 03:49:26 Yes Sep 17 03:49:29 my avermedia box is definitely not going to help Sep 17 03:49:34 eFfeM: Thats the plan. Sep 17 03:49:45 so I considered either hauppauge or the pinnacle one Sep 17 03:49:47 once I get it going; then mythtv needs exploration Sep 17 03:49:49 what is the status? Sep 17 03:49:59 ok Sep 17 03:50:03 status... Sep 17 03:50:16 Its still being integrated Sep 17 03:50:25 mythtv might be tricky; I think it is an X app Sep 17 03:50:32 I expect to have it in openslug in the next 18 hours Sep 17 03:50:43 I once heard from someone who had it running on buffalo Sep 17 03:50:46 I'm about to pass out now; so it'll probably need to wait until I wake up Sep 17 03:50:57 Ho was it on the buffal Sep 17 03:50:58 o Sep 17 03:52:00 I got a demo, it looked quite good. Don't know if he wants to share source (he did this at work so I don't know if his company will allow him giving it to me Sep 17 03:52:21 he could do recordings and timeshifting Sep 17 03:52:32 exposed through upnp Sep 17 03:53:46 okay it should be workable on a nslu2 thats been de-underclocked then eh? Sep 17 03:54:14 yeah, without problem. when I asked about performance he said load was less than 5% Sep 17 03:54:38 basically recording is just reading the mpeg2 stream from usb and scribbling it to HD Sep 17 03:54:54 okay, justu got it to compile right now. Now for the packing.... Sep 17 03:54:59 with 4Mbit/sec that would not be too much of a problem Sep 17 04:05:26 mythtv backend is what you want Sep 17 04:08:16 it's already in oe BTW Sep 17 04:08:22 jacques, thanks Sep 17 04:08:27 pb put it there IIRC Sep 17 04:08:52 actually the discussion went on in both #openslug and here; it was continued in openslug Sep 17 04:10:07 ah Sep 17 04:15:54 <[g2]> dwery can you e-mail me that patch ? It's not stored in the message :( Sep 17 04:19:14 sure Sep 17 04:19:17 <[g2]> THX Sep 17 04:19:23 give me a sec Sep 17 04:19:29 <[g2]> p Sep 17 04:19:31 <[g2]> np Sep 17 04:20:00 <[g2]> dwery btw I'm booting directly to Debian on hda1 on the Loft Sep 17 04:20:32 wow Sep 17 04:20:35 LE? Sep 17 04:20:39 <[g2]> BE Sep 17 04:21:01 <[g2]> but Reboot is built with the 1.5 IAL Sep 17 04:21:21 <[g2]> I'd imagine I could build Reboot in LE mode Sep 17 04:21:27 is redboot free or do you have to pay a licence? Sep 17 04:21:43 <[g2]> I've got to check the licensing Sep 17 04:22:06 anyway you don;t need it to be LE, as long as you swap the kernel Sep 17 04:22:10 <[g2]> since I'm selling hw I an distribute IAL source with Intel's license Sep 17 04:23:11 i;m giving a lot to the x1205 driver Sep 17 04:23:11 <[g2]> On a previous board I had an Intel GigE mini PCI card running Sep 17 04:23:17 I see conflicting patches Sep 17 04:24:40 [g2], would directly booting the kernel from disk be doable with the slug as well? Sep 17 04:24:58 what I would like to have is boot from disk if the disk is there and otherwise from flash Sep 17 04:25:17 that way no need to reflash if a new kernel is no good Sep 17 04:25:40 <[g2]> eFfeM it's possible with bootloader updates Sep 17 04:25:58 sounds interesting Sep 17 04:26:22 but I'm not really sure if I am up to that challenge yet Sep 17 04:26:48 I don't have easy access to jtag Sep 17 04:26:48 <[g2]> it's very low-level (close to the hw) Sep 17 04:27:00 <[g2]> one would definetly need JTAG Sep 17 04:27:01 yeah, i figured that out already Sep 17 04:27:27 <[g2]> that's why the Loft has JTAG and serial built in Sep 17 04:27:43 <[g2]> "Freedom to tinker" Sep 17 04:27:55 yup Sep 17 04:27:56 [g2]: what do you use to control jtag under linux? Sep 17 04:28:27 <[g2]> dwery I haven't used it that much Sep 17 04:28:31 OpenWince JTAG Tools. Sep 17 04:28:44 <[g2]> I've got a utility that can reflash my OEM boards Sep 17 04:29:19 <[g2]> the one slug I've got modified for JTAG wasn't fully debugged to use jtag from OpenWinCE Sep 17 04:29:20 i';ve contributed to jtag tools a while ago Sep 17 04:30:15 <[g2]> I've got a Raven that I've yet to test Sep 17 04:30:36 <[g2]> I want to make a very cheap JTAG hookup for the Loft Sep 17 04:30:54 [g2]: i used the parallel port and a bunch o pins Sep 17 04:31:03 <[g2]> yeah something like that Sep 17 04:31:33 <[g2]> there'll be multiple products but something like that will be the cheapest Sep 17 04:32:03 <[g2]> you'll probably want to join us in #openjtag Sep 17 04:32:33 [g2]: eh eh :) i'm not using jtag a lot nowadays.. so it's best I keep focusing on the nslu2 Sep 17 04:32:35 <[g2]> one of the guys was able to directly load the mini-I cachce and boot to it via JTAG Sep 17 04:32:40 <[g2]> on the NSLU2 Sep 17 04:32:47 wow Sep 17 04:33:09 hope i adds the info to the wiki Sep 17 04:33:34 <[g2]> I think it's been there for a couple months Sep 17 04:33:38 ah ok :) Sep 17 04:34:01 <[g2]> I think JTAG debugging would be very helpful too Sep 17 04:34:16 <[g2]> whenever there's a lockup, just go grab a hold of the processor Sep 17 04:34:47 <[g2]> there's probably a bunch of performance uses also Sep 17 04:37:00 got the patch? Sep 17 04:38:04 <[g2]> yeah thx Sep 17 04:41:11 03dyoung 07org.openembedded.dev * r240f9dcb... 10/packages/pvrusb2-mci/ (pvrusb2-mci-20050911/Makefile.patch pvrusb2-mci_20050911.bb): pvrusb2: Add support for the Happauge USB2 PVR Device Sep 17 04:52:47 ok, I made some fixes here and there and rebuilding 2.6.13.2 Sep 17 04:52:49 .1 Sep 17 05:34:01 ok, rtc seems working fine now. Sep 17 05:41:09 03koen 07org.openembedded.dev * r597b1f05... 10/conf/machine/ipaq-pxa270.conf: conf/machine/ipaq-pxa270.conf: hardcode preferrence for 2.6.12-hh3 Sep 17 05:42:40 anyone tried the FastFPE floating point emulation? Sep 17 05:45:16 given that the nslu has <64mb of ram, why indirect PCI addressing is used? Sep 17 05:45:29 (i'm throwing out questions today...) Sep 17 06:16:13 03dyoung 07org.openembedded.dev * r18e90def... 10/packages/pvrusb2-mci/ (pvrusb2-mci-20050911/Makefile.patch pvrusb2-mci_20050911.bb): pvrusb2-mci: Add some ugly hackery to define LINUX_ARM_ARCH for nslu2, this should be properly addressed in the future Sep 17 06:52:19 dwery: If you want to see some noticeable speed improvements you should not use kernel level fpu emulation and use a softfloat toolchain, but you'll need to recompile debian packages for that Sep 17 06:57:50 yvasilev: thanks, but i prefer using madplay instead of recompiling. I've just tried the FastFPE emulation but it seems it hangs the kernel. Sep 17 06:59:07 now i'm going to find a suitable and stable .config and release some patches Sep 17 07:00:00 dwery: the problem is also with ALU code, hardfloat gcc "uses" coprocessor registers to "speed up" things, so it generates lots of exceptions and the kernel needs to emulates the registers. Sep 17 07:01:01 dwery: the speed difference for pure ALU code is about 10-20% Sep 17 07:01:45 yvasilev: understood. what kernel are you running now? Sep 17 07:02:06 2.6.11.12 Sep 17 07:02:40 would you like to test 2.6.13.1? Sep 17 07:02:51 sure Sep 17 07:02:58 are you le or be? Sep 17 07:03:17 * yvasilev is waiting for dwery to release patches Sep 17 07:03:21 le Sep 17 07:03:44 did you used peter's page as a reference? Sep 17 07:03:50 yes Sep 17 07:04:01 great. Sep 17 07:04:22 i need to recompile two or tree more times and I should be able to have a good set of patches Sep 17 07:04:32 :-) Sep 17 07:05:35 I have now a working kernel. Will take a vanilla one, and start assembling the patches. Sep 17 07:05:47 if md5sum is ok, the new kernel should be good as well :) Sep 17 07:08:54 if you have some spare time, you can start downloading a vanilla 2.6.13.1 Sep 17 07:10:41 good point, since I intend to upload it over zmodem to the nslu2 and compile it there Sep 17 07:10:54 at you risk :D Sep 17 07:11:07 how log does it take for the nslu2 to compile? Sep 17 07:12:04 have not tested it with just the kernel, but with the slug compiling other packages on the same time it takes some time and swapping Sep 17 07:12:17 * dwery suggest using another machine :) Sep 17 07:12:27 then how do you load the kernel in? Sep 17 07:12:50 for now, tftp Sep 17 07:21:14 do you have root on sda2? Sep 17 07:22:01 sda3, so yes Sep 17 07:23:25 you'll have to change the CMDLINE in .config then Sep 17 07:24:08 yes, that's I know how to Sep 17 07:32:14 do you know what the export symbols patch is for? Sep 17 07:32:28 no Sep 17 07:32:34 'cos I can't find nothing that uses those symbols.. Sep 17 07:35:09 ok, the kernel is working. where can I send you the patches? Sep 17 07:35:23 yvasilev@duke.math.cinvestav.mx Sep 17 07:35:57 ok, just a moment Sep 17 07:38:31 <[g2]> yvasilev you're building the kernel native ? Sep 17 07:38:45 [g2]: yes Sep 17 07:38:47 ok, mail sent Sep 17 07:38:51 <[g2]> on sarge ? Sep 17 07:38:51 patching is easy Sep 17 07:39:01 take vanilla, apply patches in order Sep 17 07:39:13 do a make nslu2_le_defconfig Sep 17 07:39:14 dwery: fot it, thanks Sep 17 07:39:19 change .config to suit you needs Sep 17 07:39:25 [g2]: on gentoo/uclibc Sep 17 07:40:10 <[g2]> I've got Debian Sarge running on my Loft proto which is a 533 IXP425 with 64MB mem Sep 17 07:40:44 [g2]: LE? Sep 17 07:40:50 <[g2]> it's BE Sep 17 07:40:55 <[g2]> it could be LE Sep 17 07:41:09 <[g2]> but I'm BE oriented Sep 17 07:41:12 [g2]: can you try my 2.6.13.1 in BE? Sep 17 07:42:19 <[g2]> I've got to port to the IXP425 first and there's one driver a CF IDE driver that'll need forward porting from 2.4 Sep 17 07:42:28 [g2]: i mean on the nslu2 Sep 17 07:42:45 <[g2]> dwery it's all pretty much the same thing Sep 17 07:42:47 [g2]: i need someone to test be Sep 17 07:43:15 <[g2]> Ah the patches you sent ? Sep 17 07:43:27 [g2]: no, the new set of patches Sep 17 07:43:34 [g2]: which allows compilation from a vanilla treew Sep 17 07:43:43 <[g2]> I haven't seen that set yet Sep 17 07:44:03 [g2]: I hope so, unless you lurk into yvasilev mailbox :) Sep 17 07:44:15 [g2]: i'm sending them Sep 17 07:44:22 <[g2]> thx Sep 17 07:45:12 they're numbered Sep 17 07:45:27 you'll have to skip over the little endian patch Sep 17 07:45:34 and modify the defconfig to suit BE Sep 17 07:46:07 i plaan to have only a BE defconfig in the next set, with a patch to convert it to LE. Sep 17 07:52:47 when i'll have BE confirmed ok, i'll poste the patches to the ml Sep 17 07:53:35 <[g2]> Ok.. the default ixp425 config for 2.6.12 is building natively right now Sep 17 07:53:57 obviously the CMDLINE in y patches suit my syste, Sep 17 07:54:23 <[g2]> right yours won't have root=/dev/hda1 :) Sep 17 07:54:28 no :) Sep 17 07:54:36 <[g2]> but mine will :) Sep 17 07:54:48 would it be better to keep using the original openslug commandline in the patches? Sep 17 07:55:08 <[g2]> definitely keep the noirqdebug Sep 17 07:55:35 <[g2]> I've got a full Redboot so I can just exec -c "cmdline ....." Sep 17 07:55:56 nice :) Sep 17 07:56:26 <[g2]> and I can reflash config scripts etc... Sep 17 07:56:40 <[g2]> it's quite a sane environment Sep 17 07:56:57 <[g2]> the kind I was hoping for a year ago Sep 17 07:57:12 <[g2]> better late than never :) Sep 17 07:57:18 yeah. Sep 17 07:57:33 <[g2]> though APEX was/is a godsend Sep 17 07:57:50 I've been told that. Sep 17 07:58:31 <[g2]> that passes a cmd line too Sep 17 07:58:48 once BE is confirmed and patches release, I think i'll start looking into an LE intel library. Sep 17 07:59:48 and into pushing x1205 and a few other little bits into mainstream. Sep 17 08:01:23 i'm sure you're compiling BE right now, isn't it? :D Sep 17 08:01:35 <[g2]> nod. Sep 17 08:01:39 :D Sep 17 08:01:39 <[g2]> it's actually 2.6.12 Sep 17 08:01:55 yvasilev: ? Sep 17 08:02:05 <[g2]> I'll need to pull .13.1 and do it again Sep 17 08:02:13 thanks :) Sep 17 08:03:06 <[g2]> dwery .... Sep 17 08:03:10 <[g2]> this is COOL Sep 17 08:03:31 <[g2]> I'll have to get my fatslug all fired up and doing the same thing Sep 17 08:03:33 :-D Sep 17 08:03:50 you've a fatslug? Sep 17 08:03:51 <[g2]> so the fatslug can build the 13.1 BE kernel on Debian Sep 17 08:03:58 dwery: yes? Sep 17 08:04:07 <[g2]> and it runs APEX and boot the kernel from jffs2 Sep 17 08:04:07 yvasilev: how's your build going? Sep 17 08:04:23 uploading the kernel Sep 17 08:04:26 [g2]: ok, I dont wanna know more :) Sep 17 08:04:49 <[g2]> so I'll be able to build from debian then load the kernel from jffs2 Sep 17 08:05:15 <[g2]> just cp bzImage /initrd/boot Sep 17 08:05:22 where did you found the chips to fatten it? Sep 17 08:05:34 <[g2]> I've got a bunch Sep 17 08:06:02 <[g2]> I ordered 25 and had a local professional rework create the fatslug with serial and JTAG connector Sep 17 08:06:07 they're too small for my equipment Sep 17 08:06:24 <[g2]> I was planning on selling them but it's too expensive due to the rework cost Sep 17 08:06:39 <[g2]> so I just built units from scratch Sep 17 08:06:50 you can send the chips Sep 17 08:06:53 sell Sep 17 08:07:12 <[g2]> sure Sep 17 08:07:36 <[g2]> most are still in the static bags Sep 17 08:07:41 32 mb should be enough.. my amiga only had 512k :) Sep 17 08:07:55 <[g2]> there the 16M16s Sep 17 08:08:01 <[g2]> instead of the 8M16s Sep 17 08:09:34 is the BE driver of the ob-board ethernet already in the kernel? Sep 17 08:09:51 <[g2]> ob-board ? Sep 17 08:10:14 on-board :) Sep 17 08:10:51 <[g2]> I'll have to check about that, it may be or the chip may be initialized via Redboot Sep 17 08:11:07 <[g2]> they're certainly built into the Redboot Sep 17 08:11:10 it should be there, at least in the openslug kernel Sep 17 08:11:13 03hrw 07org.openembedded.dev * r9d57b6c2... 10/packages/ipkg/ (ipkg-native_0.99.154.bb ipkg_0.99.154.bb): Sep 17 08:11:13 added ipkg 0.99.154 Sep 17 08:11:13 - this version has fixed Provides: problem from 0.99.152 and Depends: from .153 Sep 17 08:11:13 - changset 5b224b4c91dec0b22ff45bec26fb18a01889e139 can be reverted now Sep 17 08:11:35 <[g2]> you can't include the IAL modules in the kernel Sep 17 08:11:55 <[g2]> they have to be loaded as modules the licensing prohibits it Sep 17 08:11:56 oh Sep 17 08:12:01 understood Sep 17 08:12:12 <[g2]> and they are built separately by Openslug Sep 17 08:12:53 <[g2]> I'm up to fs/ext2 Sep 17 08:13:12 <[g2]> make that fs/ext3 Sep 17 08:15:54 you mean in the CMDLINE? Sep 17 08:21:23 it's curious that the usb audio adapter works when connected directly and not when on the hub Sep 17 08:21:51 Self powered hub? Sep 17 08:21:53 yes Sep 17 08:22:04 Very odd then Sep 17 08:22:04 other peripherals work fine. i've just tried an usb key Sep 17 08:22:27 I can see it not working if the hub was bus powered but if it's self powered then there shouldn't be any problem with that Sep 17 08:22:40 <[cc]smart> do we have a nice tool for logrotation for openslug ? Sep 17 08:22:56 <[g2]> there's a great one in ODS Sep 17 08:22:57 it keeps saying "cannot submit datapipe for urb 0, err = -38" Sep 17 08:23:11 <[g2]> logrotate Sep 17 08:23:26 <[g2]> it think that might be a package Sep 17 08:23:52 dwery: http://article.gmane.org/gmane.linux.alsa.devel/25841 Sep 17 08:24:05 Thats why it's not working Sep 17 08:24:13 As to how to fix *shrug* Sep 17 08:24:39 <[cc]smart> i saw the logrotate script but i felt the actual program is missing. i'll check DB. always forget that... Sep 17 08:25:25 Tiersten: whoa.. what a nast bug Sep 17 08:25:47 well, without it I wouldn't have probably ever built 2.6.13 Sep 17 08:25:58 More of a missing feature than a bug. Very poorly documented tho Sep 17 08:26:06 yeah. Sep 17 08:26:18 I guess i'll have to add another two ports on the slug... Sep 17 08:26:27 or use an usb 1.1 hub Sep 17 08:26:29 Yeah. How big is the audio dongle? Sep 17 08:27:08 <[cc]smart> aw Sep 17 08:27:09 about 6x4cm Sep 17 08:27:16 <[cc]smart> busybox should deliver it Sep 17 08:27:30 <[cc]smart> but it seems the setup we have for busybox doesn't includ eit Sep 17 08:27:56 You could probably wedge it inside the NSLU2 casing if you took it apart Sep 17 08:28:25 Tiersten: i think so. i'll see. Since this thing will go in my car, i prefer to keep things apart. Sep 17 08:28:37 ah okay Sep 17 08:28:44 Just remember not to overload the PSU Sep 17 08:29:04 I'll have to get supply from the car Sep 17 08:29:15 and use an hub for most of the devices Sep 17 08:29:23 Yeah. A 5V PSU at > 2A would be fine Sep 17 08:29:32 <[cc]smart> dwery: your dongle is supported ? Sep 17 08:29:42 ah okay. Hub would make it easier Sep 17 08:29:56 [cc]smart: yes, it sound fine :) Sep 17 08:30:16 Just remember to use madplay :) Sep 17 08:30:21 Or tremor if you want Ogg Sep 17 08:30:26 <[cc]smart> what brand ? was trying to make me slug a network player, but the on i tried wasn't doing well... Sep 17 08:30:52 [cc]smart: brand? :D I bought it on ebay. It says C-media when plugged in Sep 17 08:31:01 <[cc]smart> :) Sep 17 08:31:04 <[cc]smart> rright Sep 17 08:31:10 [cc]smart: it's of poor quality, i need to resolder the jack Sep 17 08:31:29 C-media is the chip inside Sep 17 08:31:40 <[cc]smart> ya, just as many :) Sep 17 08:31:40 Tiersten: yes Sep 17 08:31:49 CM108 Sep 17 08:33:29 if you want, I can search the dealr on ebay.. it has a lot of them Sep 17 08:34:17 <[cc]smart> shipping such a thingy to here would probably cost more than 5 of them brand new in a store :) Sep 17 08:34:42 :D I didn't find a lot of them here in Italy.. so I had to use ebay Sep 17 08:34:43 You could get an iMic. I think that's been made to work :) Sep 17 08:35:00 <[cc]smart> iMac ? Sep 17 08:35:02 <[cc]smart> :) Sep 17 08:35:03 <[cc]smart> heck Sep 17 08:35:06 <[cc]smart> what names that Sep 17 08:35:10 iMic Sep 17 08:35:14 It's a Griffin thing Sep 17 08:35:22 <[cc]smart> ye, got you, but still its funny Sep 17 08:35:33 <[cc]smart> griffin Sep 17 08:35:36 <[cc]smart> ...taking note Sep 17 08:35:39 http://www.griffintechnology.com/products/imic/ Sep 17 08:35:48 $39.99 Sep 17 08:35:54 Bit expensive :) Sep 17 08:36:23 <[cc]smart> well, name and design point to the same directio... so why should the prrice tag point diffrent ? Sep 17 08:36:40 ? Sep 17 08:36:51 <[cc]smart> iMic ~ iMac Sep 17 08:37:06 It's a bit expensive for what it is physically Sep 17 08:37:07 <[cc]smart> just as design with that clear plastics and look Sep 17 08:37:23 What you're getting is guaranteed operation with a Mac Sep 17 08:39:14 <[cc]smart> i've got acces to a IIe... shall try :) Sep 17 08:41:22 03rpurdie 07org.openembedded.dev * r154208ed... 10/packages/linux/linux-openzaurus_2.6.13-mm3.bb: linux-oz-2.6: Fix the pcmcia card resume problem (introduced by the pcmcia maintainer). Add tosa to the list of devices to enforce a kernel size limit on. Sep 17 08:41:25 03rpurdie 07org.openembedded.dev * r84f9b4d8... 10/packages/orinoco/ (3 files in 2 dirs): Add firmware generation package for spectrum wireless LAN cards. Sep 17 08:50:15 i'm inspecting my adapter. It seems they've wired it for mono... Sep 17 08:52:33 useful Sep 17 08:53:06 Tiersten: yeah. cheap. I'm downloading the datasheet to see if I can fix it Sep 17 09:00:11 AAARGH the outputs were short-circuited! Sep 17 09:00:19 I must admit they were cheap... Sep 17 09:00:55 The stick the PCB to the wall and throw solder at it until it sticks method? Sep 17 09:01:03 :D Sep 17 09:01:16 I managed to separate them Sep 17 09:01:19 seems ok now Sep 17 09:01:24 Cool Sep 17 09:01:38 Now I should found a way to control the mixer vi aHID Sep 17 09:01:54 It's not part of the usual ALSA mixer? Sep 17 09:02:14 Tiersten: no Sep 17 09:02:25 Tiersten: thos usb audio devices are usually controlled thru hid Sep 17 09:02:50 oh okay Sep 17 09:04:29 this thing has also some spare i/o and volume up/down pins left unconnected Sep 17 09:12:22 Does the driver for it support reading those pins? Sep 17 09:13:23 Tiersten: no , the audio usb driver is generic. Those pin can be used via HID.. according to the datasheet, it shouldn't be difficult. Sep 17 09:14:06 Tiersten: I think i'll write something for that. My priority, at the moment, is to get 2.6.13.1 tested and released to the ml. Sep 17 09:14:43 cool Sep 17 09:15:17 Tiersten: you're volunteering, isn't it? :) Sep 17 09:15:55 LOL. No :) Sep 17 09:16:03 Tiersten: ach :) Sep 17 09:18:51 no beta testers today here.. :) Sep 17 09:20:23 btw usbnet doesn't work at all in 2.6.13 Sep 17 09:20:56 I'm curious if it works on 2.6.13 BE Sep 17 09:25:14 dwery: not good, kernels >= 2.6.12 need gettext to work, which I don't have on uclibc, so need some extra patching before testing your kernel Sep 17 09:26:26 yvasilev: can you try to build non-natively? Sep 17 09:27:14 dwery: that's what i'll probably end doing for this quick test Sep 17 09:27:20 yvasilev: thanks Sep 17 09:28:42 speaking about IAL.. I've heard someone tried it LE. Is it right? Sep 17 09:29:55 beewoolie I think Sep 17 09:30:07 dwery: this is what [g2] told me yesterday about it http://rafb.net/paste/results/huNvUx14.html Sep 17 09:31:38 ok, i'll dig around. Sep 17 09:31:53 it's now a priority for me, due to usbnet not working anymore. Sep 17 09:32:59 dwery: 2.0 or 1.5? Sep 17 09:34:22 yvasilev: dunno, the easiest one. I just need ethernet working.. I gues 2.0 could be the best option, since it should be quite newer Sep 17 09:37:38 dwery: yesterday I was digging with 2.0, and it looks like the patch form http://www.wnk.at/ixp400_accesslib_kernel26/ will need some changes, and the GPL_ixp400LinuxEthernetDriverPatch-1_4.zip and GPL_ixp400LinuxIntegrationPatch-1_3.zip patches for the kernel should be better unpacked to another dir and a module be made directly a la 2.6 modules Sep 17 09:37:51 so there should be no need to touch the kernel Sep 17 09:38:01 yvasilev: yes, it should be all external Sep 17 09:38:17 but had not enough time to correct the patches Sep 17 09:46:16 03rpurdie 07org.openembedded.dev * r6b658da7... 10/.mt-attrs: Add firmware generation package for spectrum wireless LAN cards. Sep 17 09:46:19 03rpurdie 07org.openembedded.dev * r39f8b722... 10/packages/hostap/hostap-modules_0.4.4.bb: hostap: Add fixes so it builds correctly for kernels > 2.6.13 Sep 17 09:46:21 03rpurdie 07org.openembedded.dev * rc9c47277... 10/conf/machine/zaurus-clamshell-2.6.conf: zaurus-clamshell-2.6.conf: Switch to the in kernel verison of orinoco/spectrum wireless LAN driver (tested and works). Sep 17 09:51:00 I noticed Karen Spearel is the author of the x1205 driver.. anybody know his nick? Sep 17 09:51:04 03koen 07org.openembedded.dev * r0a6a4345... 10/ (2 files in 2 dirs): Sep 17 09:51:04 packages/matchbox-wm/matchbox-wm_svn.bb: make PV conform to OE standards Sep 17 09:51:04 conf/distro/preferred-gpe-versions-2.7.inc: use mb-wm svn till mallum releases a new version Sep 17 09:52:42 * eFfeM did it! Made the slug battery powered! Sep 17 09:53:03 Yahooo Wiki! Wiki! Wiki! Sep 17 09:53:11 lol Sep 17 09:53:26 i thought someone would say that ... Sep 17 09:53:33 eh eh :) Sep 17 09:53:49 Were you the one who asked about batter power on the ml? Sep 17 09:53:52 yep Sep 17 09:54:04 just thought, what the heck, lets try it Sep 17 09:54:09 I've just read the thread Sep 17 09:54:25 USB power is not connected directly to the power input jack Sep 17 09:54:28 the slug takes 0.5 amps Sep 17 09:54:48 and can run on as little as 3.5 V Sep 17 09:54:56 that was without any peripheral Sep 17 09:54:57 amazing. Sep 17 09:55:14 Is that with ethernet running though? Sep 17 09:55:23 The ethernet phy takes a fair bit of power Sep 17 09:55:29 yes Sep 17 09:55:53 actually i'm not really well equiped to do power measurements Sep 17 09:55:56 hmm... somebody measured it before and came up with about 1A Sep 17 09:56:00 just one measly power connector Sep 17 09:56:15 eFfeM: The highest Vin you used? Sep 17 09:56:38 Hmm. Guess not. It varied from 0.5 to 0.6A Sep 17 09:56:43 power supply gives 5.3 Sep 17 09:57:09 with batteries I started with about 6V unconnected, but that would drop to 4 once the slug was running Sep 17 09:57:25 the batteries were not fully fresh Sep 17 09:57:32 eFfeM: I was planning to power it from the car, maybe filtering the supply line a bit. Sep 17 09:58:13 Any 5V regulator should be okay Sep 17 09:58:25 If you can get one then a 12V -> 5V DC-DC converter would be best Sep 17 09:58:36 Tiersten: If the onboard regulator is switching, I would throw in 12V directly. Sep 17 09:58:55 I'm not really a hw guy, but if I understood the messages in the group correctly this should be possible Sep 17 09:59:05 I was not *that* daring thoug Sep 17 09:59:11 I don't believe it's 12V capable Sep 17 09:59:14 eFfeM: :) i'll have to see the datasheet Sep 17 09:59:31 (already fried my power supply (made a nice source for the connector though) Sep 17 09:59:33 Running the NSLU2 at less than 5V isn't recommended either if you want to attach any USB devices Sep 17 09:59:47 There is a ref to the datasheet in the wiki, but I could not access it yesterday. Sep 17 10:02:55 it is at http://www.feeling-tech.com.tw/EN/Product/ProductList.asp?iWebId=1&sDownloadType=2&#A but I only can d/l a few KB Sep 17 10:04:04 get 29.1 kb in 1 sec ... Sep 17 10:04:09 do you have the chip name ? Sep 17 10:05:39 FP5452 Sep 17 10:06:09 see http://www.nslu2-linux.org/wiki/Info/LinksToHardwareComponentDocumentation Sep 17 10:06:56 I tried googling for it, no success. Sep 17 10:07:06 usualy google can find almost all datasheets Sep 17 10:08:06 well the feeling-tech url gives you the chip site Sep 17 10:08:12 chip supplier Sep 17 10:08:24 i get a connection refused... Sep 17 10:10:22 no problem here, Tiersten, could you give the url a try? Sep 17 10:11:02 I don't know why you're trying anyway. The chip if I remember correctly can support to around 30V but you'll fry any USB device you plug in Sep 17 10:11:25 The FP5452 is to provide 3.3V Sep 17 10:11:42 ok and the 5 v is going to the FP and directly to the USB Sep 17 10:11:51 no Sep 17 10:11:52 Pretty much yes Sep 17 10:12:00 the USB connector is connected to the FP Sep 17 10:12:04 It's not directly connected but it doesn't have another regulator in it Sep 17 10:12:06 so 12v could feed the slug but no peripherals ? Sep 17 10:12:18 You sure? Somebody mapped it out and it went to an internal layer in the PCB Sep 17 10:12:38 Well, I used a continuity tester Sep 17 10:12:57 and I'm quite sure it's connected to the FP. Sep 17 10:13:10 I will recheck once i have the FP datasheet Sep 17 10:15:18 talked with the i2c maintainer. seems we may not need a driver id for the x1205 driver after all... i'll chack that, Sep 17 10:18:30 I'll check that with the original author. Sep 17 10:19:13 Of the x1205 driver? Sep 17 10:19:22 optware-question: I'm looking for the makefile that builds php-mysql. Does it exist? Sep 17 10:19:23 yes Sep 17 10:20:43 just checked, without ethernet the power consumption is 400 mA Sep 17 10:21:43 I've to go now.. see ya Sep 17 10:21:49 Bye Sep 17 10:21:51 cu Sep 17 10:42:53 made the wiki page: http://www.nslu2-linux.org/wiki/HowTo/MakeABatteryPoweredSlug Sep 17 10:43:19 Cool Sep 17 10:43:40 Can you measure the voltage of your batteries and the voltage of the USB port? Sep 17 10:43:58 If the USB port is 5V still then it'll maybe be okay running the NSLU2 off 12V.... Sep 17 10:44:05 Although you might need to add a heatsink to it Sep 17 10:44:10 battery voltage yes, see the wiki for details Sep 17 10:44:20 ya but I want to know both at the same time :) Sep 17 10:44:41 usb, no, I have only a simple multimeter here and I don't even know where to measure. Sep 17 10:44:43 If battery voltage is definitely < 5V and the USB port is 5V then it's not a directish connection between the two Sep 17 10:45:00 and together is impossible at all (I only have 1 multimeter) Sep 17 10:45:20 We,ll leave it to dwery to test that :-) Sep 17 10:45:33 It's easy. Measure the two outer pins Sep 17 10:45:43 I don't need to at exactly the same time. Just the same run would do Sep 17 10:46:16 I'll see if I can do that. not now Sep 17 10:46:22 k. no hurry Sep 17 10:46:40 want to grab some coffee first and finish the message to the group Sep 17 10:47:05 k Sep 17 10:59:51 % uname -a Sep 17 10:59:56 Linux slug 2.6.13.1 #1 Sat Sep 17 19:51:18 CEST 2005 armv5tel GNU/Linux <- wee Sep 17 11:26:22 03rpurdie 07org.openembedded.dev * r912340c2... 10/packages/linux/ (6 files in 2 dirs): linux-oz-2.6: Update to 2.6.14-rc1-mm1 which the external hostap will work better against. Update defconfigs to include orinoco, spectrum and hostap modules (although still use the external hostap for now). Sep 17 11:29:00 Tiersten, just had a quick look at the usb port. I don't think I want to measure it. Too much risk that I touch the outer shielding (which I suppose is grounded) Sep 17 11:29:12 No desire to fry a usb port ... Sep 17 11:39:06 eFfeM-away: You should be able to measure it from the back of the PCB Sep 17 15:55:22 hello.. anyone is awake? :) Sep 17 15:56:04 I'm pretending to be asleep Sep 17 15:56:07 whoops Sep 17 15:56:56 eheh :) I'm still searching for volunteers to test 2.6.13.1 :) Sep 17 16:12:01 dwery: if I had serial console, I'd do it :) Sep 17 16:12:09 but I dont Sep 17 16:12:48 :) Sep 17 16:12:53 You don't need it Sep 17 16:13:07 IF usbnet works BE :D Sep 17 16:14:25 usbnet? Sep 17 16:15:10 just kidding.. the usb net drivers does not work in 2.6.13 . I don't know if it's because i'm trying it on a little endian nslu2 or if it's a general problem. Sep 17 16:17:25 ah.. Sep 17 16:25:53 well.. it seems it's working now... Sep 17 16:44:07 dwery: Karen Spearel is kas11, and she is best contacted by email. Sep 17 16:44:52 rwhitby: thanks Sep 17 16:45:11 rwhitby: do you know if the driver is still being worked on? Sep 17 16:55:38 dwery: dunno, kas11 had a sf project (OpenN2 I think) Sep 17 16:57:02 rwhitby: i'll ask. the current driver is quite ugly and mixes and rtc dev and an i2c driver in the same file. Sep 17 16:59:06 dwery: check PM Sep 17 16:59:58 dwery: did you get my /msg ? Sep 17 17:00:51 rwhitby: yep Sep 17 17:50:01 i'm going to remove the rtc parts from the x1205 driver in order to split it. Sep 17 17:50:15 that means the nslu will boot without rtc for a while Sep 17 17:50:19 what could happen? Sep 17 17:50:38 nothing. it'll work without it but you'll be living in 1970 Sep 17 17:51:09 that's fine. I always thought I should have born a little bit earlier :) Sep 17 17:52:12 heh Sep 17 17:57:59 is there a doc somewhere telling what you have to change to enable distcc builds of unslung/openslug? Sep 17 18:34:21 <[g2]> dwery just run ntp after you boot **** ENDING LOGGING AT Sun Sep 18 02:59:56 2005