**** BEGIN LOGGING AT Wed Mar 09 23:59:56 2005 Mar 10 01:22:21 got a problem bitbaking openslug-image: Nothing provides m4-native Mar 10 01:27:16 can someone add m4 to yhe symlink package? Mar 10 01:27:24 yhe -> the Mar 10 02:08:17 hey Mar 10 02:08:27 <[g2]> VoodooZ_work, hi Mar 10 02:08:46 <[g2]> beewoolie-away, thx for the email to arm linux Mar 10 02:08:58 I haven't processed it. Mar 10 02:09:14 David's a good guy (TM) Mar 10 02:09:28 <[g2]> He seems real good Mar 10 02:10:10 <[g2]> It's a bummer about the performance Mar 10 02:10:36 <[g2]> I'd like to run that issue down and possibly verify with NEC Mar 10 02:18:45 what you guys talking about? usb chip has some problem? Mar 10 02:19:39 <[g2]> jacques, see usr/src/linux/Documentation/usb/ehci.txt Mar 10 02:19:40 jacques: we're getting spurrious interrupts and we don't know why. Mar 10 02:20:24 <[g2]> I think David wrote of the issue 2+ years ago Mar 10 02:20:27 [g2], what performance are you referring to? usb? Mar 10 02:21:07 <[g2]> The NEC when working with the PCI bus can only have 6-7 urb outstanding instead of the 13 that fit in the microframe Mar 10 02:21:25 <[g2]> hence max through put 20ish MB Mar 10 02:21:42 ugh Mar 10 02:21:56 <[g2]> I'm a *TOTAL* idiot, that's just my understanding of the issue Mar 10 02:22:37 David said that there seems to be another issue. It can only perform IO on one bust at a time. Mar 10 02:22:43 Either PCI or USB. Mar 10 02:22:50 <[g2]> Also, I don't know if this is fixed it later spins of the chip, which rev silicon we are running etc ... Mar 10 02:22:52 Cuts the throughput wa down. Mar 10 02:23:51 <[g2]> beewoolie-away, I think that issue leads to the only 6 or 7 USB transactions out of the 13 (about half) Mar 10 02:24:24 I see what you mean. It's a synchronous bus, right? Mar 10 02:24:45 I recall the "original" NEC chip ended in like 100 and there was a better one ending in like 101 Mar 10 02:25:10 <[g2]> dunno, I think it's simply half-duplex (chip talks to PCI or USB) but not both at the same time Mar 10 02:25:11 I read about this stuff when I bought a usb2.0 card for my desktop Mar 10 02:25:48 <[g2]> the half-duplex being a *great* over simplification of the problem Mar 10 02:30:26 It's pretty significant, though. Makes the throughput no more than 240Mib/s. Mar 10 02:49:39 <[g2]> FatSlug-GW has arrived Mar 10 03:36:52 * beewoolie-away celebrates Mar 10 03:38:00 <[g2]> Now I think I just need to remap some pins on the headers Mar 10 03:39:23 <[g2]> 1st serial, 2nd JTAG Mar 10 05:14:12 . Mar 10 05:42:35 . Mar 10 05:55:05 Can anyone here advise me on a bricked slug? Mar 10 05:55:26 <[g2]> define bricked Mar 10 05:55:27 If its really bricked, get JTAG. Mar 10 05:56:10 I built and flashed openslug thinking that if there was any problems with the image I could always telnet into redboot Mar 10 05:56:22 I don't seem to be able to get the network up anymore Mar 10 05:56:37 I think I should have tested it in memory first ;-) Mar 10 05:57:02 did you verify telnet to redboot access before flashing? Mar 10 05:57:34 A week or two ago yes - network doesn't seem to come up anymore though Mar 10 05:57:53 how did you flash? Mar 10 05:58:12 Using the unslung web page in maintenence mode Mar 10 05:58:46 so your network light doesnt come on anymore? Mar 10 05:58:53 then it's highly unlikely that redboot was overwritten Mar 10 05:59:33 dyoung - Correct. Not even on boot. Mar 10 05:59:45 Thats Bad. Mar 10 06:00:03 You re sure the network cabling is good, etc? Mar 10 06:00:29 I've tried two cables directly to the PC - both with similar results. Mar 10 06:01:06 Are they crossover cables? Mar 10 06:01:31 One is, one isn't and I've tried both with a cross-over adapter Mar 10 06:01:48 The usb light comes on when I plug in a flash disk. Mar 10 06:02:08 Oh? in that case, something is stuffed up with your network. Mar 10 06:02:28 Okay - I'll keep playing. Mar 10 06:02:37 Work out your network issues first. Mar 10 06:02:49 The network light should come on almost immediately on powerup Mar 10 06:03:10 yeah, if usb light is coming on, it's booting, which means redboot is OK Mar 10 06:03:21 Thanks guys - great job on unslung and looking forward to openslug. Mar 10 07:10:59 [g2]: how's the GW board? Mar 10 07:19:21 <[g2]> I'm just about done with my serial patch Mar 10 07:19:28 <[g2]> then I'll try to boot it Mar 10 07:19:38 <[g2]> and hopefully see Redboot :) Mar 10 07:19:50 <[g2]> It's a sweet little board Mar 10 07:19:59 <[g2]> I'm quite excited about it Mar 10 07:20:11 <[g2]> there's even a 4K bit eeprom :) Mar 10 07:20:28 Are they available yet? Mar 10 07:20:36 <[g2]> the boards ? Mar 10 07:20:40 ooh, MAC Address storage place. Mar 10 07:20:41 Yes. Mar 10 07:20:57 <[g2]> they are really geared to sell to OEMs Mar 10 07:21:11 Howed you get your paws on one? Mar 10 07:21:24 yeah Mar 10 07:21:37 <[g2]> the guys was impressed with our l337 skillz Mar 10 07:22:07 <[g2]> :) Mar 10 07:22:35 I dont suppose you scored one for me too. ;-) Mar 10 07:25:26 <[g2]> I thought about it Mar 10 07:25:35 Drat. ;-) Mar 10 07:26:07 <[g2]> Have you guys looked at the user manual ? Mar 10 07:26:43 hi all Mar 10 07:26:52 hey Mar 10 07:28:28 hanjo: do you need any assistance in getting OpenSlug built? Mar 10 07:28:33 [g2] I did see the user manual. Mar 10 07:29:15 the experience I had with the EHCI issue can be summed like this: I've seen the same problems on debian 2.4.27, while they don't exist on my laptop that is FC2 2.6.10 Mar 10 07:29:26 [g2] OpenSlug should easily be portable to this board. Mar 10 07:30:12 rwhitby-web: so in the PC world, I was suspecting a chipset issue more then a kernel one Mar 10 07:30:56 rwhitby-web: A pointer to some nice docs about building openslug would be great Mar 10 07:31:36 hanjo: http://www.nslu2-linux.org/wiki/OpenEmbedded/HomePage Mar 10 07:31:58 anyone else having build probl;ems with latest pull? Mar 10 07:32:07 looks like an automake issue in findutils Mar 10 07:36:27 jacques: haven't done a clean build in a while. kergoth was doing some automake stuff lately. Mar 10 07:36:45 rwhitby-web: I am having dependency ERRORS for the *-native packages Mar 10 07:37:19 hanjo: which packages? Mar 10 07:37:41 rwhitby-web: m4, autoconf ... etc Mar 10 07:37:54 are you using nslu2-package-symlinks? Mar 10 07:38:13 (I think there is a symlink missing since kergoth made some changes - I'm checking now) Mar 10 07:38:17 rwhitby-web: yes Mar 10 07:39:28 can someone add a symlink for m4 and push please? Mar 10 07:40:58 rwhitby-web, I'm not using symlinks Mar 10 07:41:03 doing it now Mar 10 07:41:13 jacques: separate issues Mar 10 07:41:20 dyoung-zzzz: thx Mar 10 07:41:42 pushed. Mar 10 07:41:48 jacques: which packages do I need to rebuild to see the problem you are referring to? Mar 10 07:42:18 hanjo: bk pull in nslu2-package-symlinks and it should build (unless you get the same problem as jacques too) Mar 10 07:44:20 rwhitby-web, hard to say since so many things are interrelated - for example I dunno if the issue's with findutils or automake or autoconf Mar 10 07:44:29 rwhitby-web: how long does it take for the changes to propagate? I get "nothing to pull" Mar 10 07:45:43 oops my fault Mar 10 07:45:47 just a moment Mar 10 07:46:19 try again now Mar 10 07:46:56 rwhitby-web: that worked Mar 10 07:47:17 dyoung-zzzz: tnx Mar 10 07:50:58 rwhitby-web: still no go with the image Mar 10 07:51:09 now it fails on: Mar 10 07:51:10 Importing patch /home/hanjo/src/slug/nslu2-package-symlinks/packages/gnu-config/gnu-config/config-guess-uclibc.patch (stored as config-guess-uclibc) Mar 10 07:51:11 chmod: cannot access `patches/config-guess-uclibc': No such file or directory Mar 10 07:51:57 NOTE: package gnu-config-native-0.1cvs20050221-r2: task do_patch: failed Mar 10 07:52:22 do I need more recent version of patch from the default in FC2? Mar 10 07:52:39 I have 2.5.4 Mar 10 07:54:02 BTW the patch is there (in symlinks) Mar 10 08:10:07 anyone else experiencing problems with quilt? Mar 10 08:15:31 Nope. quilt works fine...except with symlinked sources. Mar 10 08:19:42 beewoolie-away: like it never worked? or it is a temporal issue Mar 10 08:28:54 hanjo: did you do bk -r co -q ? Mar 10 08:34:38 rwhitby-web: I've switched to patcher instead of quilt. Seems to be working for now... Mar 10 08:35:28 weird. quilt works fine for me here and also on our official build machine. they are both deian sarge (testing). Mar 10 08:36:04 Don't know what the problem is. I am on a FC2 at the moment Mar 10 08:39:47 rwhitby-web: Tnx for all the help. I am heading for bed now. The build will take some time now on my slow box. I'll get back tomorrow with the results. My slug is at work, so I can not test the image tonight anyhow. Mar 10 08:40:19 by all Mar 10 08:40:22 np - thanks for looking into the ehci issue for us. Mar 10 08:40:40 well. it is for me :) Mar 10 08:41:59 If the EHCI stuff works, we will be buying 50 slugs to provide backbone connection for our 150 wireless sensor node testbed at TU Berlin Mar 10 08:42:34 cool - someone with a real incentive to get the ehci bug fixed :-) Mar 10 08:43:56 :) Mar 10 09:12:17 would anyone object to changning CMDLINE_ROOT = to CMDLINE_ROOT ?= in openslug-kernel_2.6.11.bb so iti can be overridden in local.conf ? Mar 10 09:12:59 CMDLINE_CONSOLE is already ?= Mar 10 09:14:47 good idea. Mar 10 09:20:41 dyoung-web: know anything about the Texas Instruments %u201CTNETD7301GDU%u201D ? It's in the new version of the WAG54G, one of which might my dropping into my hands soon gratis ... Mar 10 09:20:59 whoa - what happened to that number. Mar 10 09:21:15 Texas Instruments TNETD7301GDU Mar 10 09:26:03 <[g2]> # cat /proc/cpuinfo Mar 10 09:26:04 <[g2]> Processor : XScale-IXP4xx/IXC11xx rev 1 (v5b) Mar 10 09:26:04 <[g2]> BogoMIPS : 266.24 Mar 10 09:26:04 <[g2]> Features : swp half thumb fastmult edsp Mar 10 09:26:04 <[g2]> Hardware : Intel IXDP425 Development Platform Mar 10 09:26:06 <[g2]> Revision : 0000 Mar 10 09:26:08 <[g2]> Serial : 0000000000000000 Mar 10 09:26:53 <[g2]> # cat /proc/meminfo Mar 10 09:26:53 <[g2]> total: used: free: shared: buffers: cached: Mar 10 09:26:53 <[g2]> Mem: 65007616 13881344 51126272 0 126976 10821632 Mar 10 09:26:53 <[g2]> Swap: 0 0 0 Mar 10 09:26:53 <[g2]> MemTotal: 63484 kB Mar 10 09:26:54 <[g2]> MemFree: 49928 kB Mar 10 09:35:39 whoa! what kind of slug does 266 bogomips? Mar 10 09:36:21 not to mention 64MB of RAM. Mar 10 09:36:35 have you got a doppelganger slug twins? Mar 10 09:38:55 <[g2]> http://www.gateworks.com/avila_gw2347.htm Mar 10 09:39:15 <[g2]> it'x actually a 422 Mar 10 09:39:23 <[g2]> it's actually a ixp422 Mar 10 09:43:06 [g2]: is it booting already? Nice work. Mar 10 09:43:22 <[g2]> well it came with firmware :) Mar 10 09:44:49 <[g2]> beewoolie-away, But now there a couple things to do. Mar 10 09:44:58 <[g2]> 1/ Build a 2.6.11.2 kernel for it Mar 10 09:45:03 <[g2]> 2/ Try out the JTAG Mar 10 09:45:19 <[g2]> An you know what goes with JTAG right ? Mar 10 09:56:54 well that was fun Mar 10 09:57:28 [g2], what proc is on that board? Mar 10 09:58:56 <[g2]> ixp422 Mar 10 09:59:09 <[g2]> they also sell 420s and 425. Mar 10 09:59:15 <[g2]> The 425s run to 533 Mar 10 10:01:27 [g2], is it running at 266MHz? if so it would really give credence to the "slug only running at 133MHz" argument Mar 10 10:02:36 website seems to indicate a 500+ clock on that board Mar 10 10:03:20 [g2] when you hook up the JTAG, report back the ident codes and ought to be able to see what stepping is set at. Mar 10 10:03:27 ok, if it's 533MHz then slug is 266MHz as we thought Mar 10 10:04:26 Or in fact, what exact model do you have? Mar 10 10:05:12 the spec sheet says one of the 422 models is the 266mhz version. Thats the GW2347-C4R2F1E2 Mar 10 10:05:48 <[g2]> peteru-home, the 425 clocks at 533 Mar 10 10:05:56 <[g2]> this is just the 422 Mar 10 10:06:07 <[g2]> I think it's clocked a 266 Mar 10 10:06:34 hmmmmmmmmmm Mar 10 10:06:44 then that indicates the slug is at 133MHz :-\ Mar 10 10:07:36 * [g2] looks at chip Mar 10 10:07:47 <[g2]> 422BB Mar 10 10:09:08 Aha, the 422 is only availible in 266Mhz. Mar 10 10:09:54 so... Mar 10 10:10:06 I can't believe we've been running the slug at 133MHz all this time Mar 10 10:10:13 <[g2]> It's rumored that the 420, 422, and 425 are all the same die Mar 10 10:11:11 * peteru-home is looking forward to the TurboSlug hack. Mar 10 10:12:05 <[g2]> peteru-home, that may not be worth it Mar 10 10:12:14 I guess if the NSLU2 really is 1/2 clocked and can be software hacked to go full speed, then it really deserved the slug nickname it got in the first place :-) Mar 10 10:12:29 [g2], why not? Mar 10 10:12:45 <[g2]> jacques, I'd guess thermal and power Mar 10 10:13:03 water cooled slug anyone? Mar 10 10:13:20 thermal *maybe* Mar 10 10:13:42 how much better cooled is your board running at 266MHz? Mar 10 10:13:46 I thought so. Mar 10 10:13:55 <[g2]> jacques, I'm talking about the power IC and what happens under the extra amps Mar 10 10:14:10 jacques, the WRV54G has a 266Mhz cpu in it and it claims 266 BogoMips too. Mar 10 10:14:22 <[g2]> that's a 425 Mar 10 10:14:25 make that a 266Mhz IXP425 CPU in it. Mar 10 10:14:56 but either way, its still another xscale running at 266Mhz saying its 266Bogomips. Mar 10 10:15:00 I'd be more concerned about memory than either thermal or power Mar 10 10:15:14 <[g2]> well these go up to 128MB Mar 10 10:15:18 so how do we soup it up? Mar 10 10:15:23 * dyoung-web looking Mar 10 10:15:25 <[g2]> this has got 64MB and 16MB Flash Mar 10 10:15:51 <[g2]> plus the JTAG connector and the usb 1.1 brought out Mar 10 10:16:01 I meant Mar 10 10:16:07 <[g2]> and the two mini pci's of course Mar 10 10:16:23 I would be more concerned about the ram on the slug not handling 266MHz than thermal or power Mar 10 10:16:47 <[g2]> jacques, well this unit can also run POE Mar 10 10:17:13 <[g2]> I had a little discussion with the vendor about power ranges Mar 10 10:17:22 * peteru-home had a good thumb through the "Hacking the XBox" book at the bookstore today. hot acid baths for your components together with a close examination with a microscope to find the firmware ROM! Now, that's extreme hacking :-) Mar 10 10:18:05 <[g2]> how about hacking the 800Mhz Hyper-Transport bus.... That's magic Mar 10 10:18:20 * [g2] loves bunnie's book Mar 10 10:19:29 I'd really like to try running the slug at 266MHz if we can figure out how Mar 10 10:19:50 yes, the bus tap with "epoxy connector" is neat Mar 10 10:20:03 <[g2]> and the FPGA Mar 10 10:20:14 if it can't handle it, at least we'll know Mar 10 10:20:50 is jacques volunteering his slug for overclocking test duties? Mar 10 10:21:15 Is the master oscillator 33Mhz on the slug? Mar 10 10:21:16 yes I am - what's the worst that can happen? Mar 10 10:21:39 I keep getting power glitches here Mar 10 10:21:48 I'm guessing thats what Y5 is. Mar 10 10:21:56 all my stuff is on UPS but CAM-POS-CAST's crap apparently isnt Mar 10 10:21:56 <[g2]> oh this looks sweet :) Mar 10 10:22:15 <[g2]> RedBoot(tm) bootstrap and debug environment [ROM] Mar 10 10:22:15 <[g2]> Red Hat certified release, version 1.92p1 - built 16:53:03, Mar 26 2004 Mar 10 10:22:15 whenever the power glitches, my modem loses sync Mar 10 10:23:52 Jacques,. the ram wont be a problem, the sdram clock is 4x the master osc; so its gonna be 133Mhz either way. Mar 10 10:24:20 dyoung-web, interesting Mar 10 10:24:41 so it's running 1:1 right now Mar 10 10:30:31 [g2] When you have some hires images, I'm interested. Mar 10 10:31:15 <[g2]> dyoung-web, ok.. it's pretty uninteresting Mar 10 10:31:33 <[g2]> it's a pretty clean laid out board Mar 10 10:31:49 <[g2]> it's also suprisingly small Mar 10 10:32:10 <[g2]> it think it'd fit in the slug case Mar 10 10:33:34 <[g2]> this thing is way faster though Mar 10 10:33:49 <[g2]> it boots in 20 seconds to a prompt Mar 10 10:34:20 <[g2]> and about 13 is + to end of Redboot wait Mar 10 10:34:46 <[g2]> so it booting from after Redboot starts to # in about 7-8 seconds Mar 10 10:35:23 <[g2]> 8.65 Mar 10 10:46:42 is there a 66Mhz Clock or PLL onboard? Mar 10 10:47:37 <[g2]> don't see one on a quick check Mar 10 11:03:54 <[g2]> LOL Mar 10 11:07:20 <[g2]> Anyone know the price difference between NAND and NOR ? Mar 10 11:07:32 2 transistors? Mar 10 11:07:32 2 gates, isn't it? Mar 10 11:07:42 yeah, what he said Mar 10 11:07:53 <[g2]> like what 16MB NOR cost versus 32MB NAND ? Mar 10 11:09:07 peteru-home: need one of these? http://www.wireless-italia.com/modules.php?name=Sections&op=viewarticle&artid=57&page=1#articlestart Mar 10 11:14:52 rwhitby-away: need usb2, so I got one of these: http://www.netgear.com.au/products/prod_details.asp?prodID=242 Mar 10 11:15:27 Once I have more time, I'll get puppy onto it. Mar 10 11:16:01 I still would have prefered a USB 802.11g dongle, but it seems there just aren't any supported under Linux. Mar 10 11:16:44 the 108Mbps Netgear retails for about the same price as a NSLU2 and you get a lot more hardware. Mar 10 11:17:44 2 ethernet ports (one with 4 port 10/100 switch), auto MDI-X, one USB2.0 port and 802.11g in a mini-PCI slot. Mar 10 11:17:52 [g2] a 128Mbit Strataflash NOR is around $25, and a 128Mbit Toshiba NAND flash is around $8. Mar 10 11:18:09 two serial ports and one of them already has a 4 pin header soldered on :-) Mar 10 11:18:10 the 256Mbit Toshiba NAND is $9. Mar 10 11:18:11 <[g2]> dyoung-web, thx Mar 10 11:18:38 <[g2]> so it's like 1/3 or 1/4 the price Mar 10 11:19:16 the 512Mbit is $11, so its not exactly scaling. Mar 10 11:19:34 but I guess you could generalize its 3-4 tmes as much. Mar 10 11:19:35 <[g2]> 1/3, 1/4, 1/6 ..... :) Mar 10 11:20:08 <[g2]> and except for the driver complexity, are there other drawbacks ? Mar 10 11:20:19 <[g2]> power, number of write cycles ? Mar 10 11:56:20 <[g2]> dyoung-web, ping Mar 10 11:56:30 * dyoung-web grunt Mar 10 11:56:41 <[g2]> Do you know that this means ? Mar 10 11:56:43 <[g2]> -#define IXP4XX_EXP_CFG_BASE_VIRT (0xFFBFD000) Mar 10 11:56:43 <[g2]> +#define IXP4XX_EXP_CFG_BASE_VIRT (0xFFBFE000) Mar 10 11:56:53 <[g2]> why we changed it for the nslu2 ? Mar 10 11:56:59 Yeah, its a bug. Mar 10 11:57:04 it was fixed in 2.6.10 Mar 10 11:57:09 but they broke it again in 2.6.11 Mar 10 11:57:34 <[g2]> so FE is the proper addr Mar 10 11:57:43 I dont remember the exact details, but itwas mentioned in the list, etc Mar 10 11:57:53 yes. Mar 10 11:58:00 <[g2]> "in the list" ? Mar 10 11:58:09 look at te 2.6.10 codebase. Mar 10 11:58:12 The yahoo list Mar 10 11:58:34 A Real Arm Guy found it and informed the yahoo nslu2 list. Mar 10 11:58:36 <[g2]> the yahoo mailing list ? Mar 10 11:58:54 <[g2]> ah Mar 10 11:58:59 which is how the patch got into the 2.6.9 patch in the first place. Mar 10 12:00:57 <[g2]> hmmmm Mar 10 12:01:06 <[g2]> I wonder if that's causing the EHCI ? Mar 10 12:01:15 <[g2]> by changing the timing on the PCI bus Mar 10 12:02:11 <[g2]> I think it won't take very many changes and I'll be able to run the 2.6.11.2 kernel on the gateworks board Mar 10 12:04:54 I seriously doubt it. Mar 10 12:05:56 <[g2]> you doubt I'll be able to make the changes to make 2.6.11.2 run ? Mar 10 12:06:07 <[g2]> :) Mar 10 12:06:23 No, I seriously doubt the configuration addresse base would cause a EHCI timing problem. Mar 10 12:07:19 <[g2]> I'm just teasing you :) Mar 10 12:07:25 <[g2]> nod. Mar 10 12:25:12 peteru-home: so puppy probably won't run on openslug due to a problem you've found in the usb in 2.6 ? Mar 10 12:26:02 (in the puppy logs for those who are interested) Mar 10 12:28:46 * [g2] checking Mar 10 12:30:42 back later Mar 10 13:27:39 wouldnt you know it. Mar 10 13:28:14 2 days after buying my big ram the price goes down by $12; just the cost of shipping. Mar 10 13:29:23 <[g2]> doh! Mar 10 13:32:06 any reason why I couldn't do OpenSlug with the Linksys 2.4 kernel? Mar 10 13:32:28 <[g2]> just 1 or 2 Mar 10 13:32:56 <[g2]> currently mod-utils and stuff is 2.6 based Mar 10 13:33:34 <[g2]> and I don't know if stuff like hotplug-ng, sysfs, etc... Mar 10 13:33:40 <[g2]> dyoung-web, ping Mar 10 13:34:25 <[g2]> RedBoot> load -r -v -b 0x800000 /var/www/localhost/switchbox.ext2.gz Mar 10 13:34:25 <[g2]> Using default protocol (TFTP) Mar 10 13:34:25 <[g2]> - Mar 10 13:34:25 <[g2]> Raw file loaded 0x00800000-0x00833f2b, assumed entry at 0x00800000 Mar 10 13:34:25 <[g2]> RedBoot> fis load zimage Mar 10 13:34:25 <[g2]> RedBoot> exec 0x1600000 Mar 10 13:34:27 <[g2]> Using base address 0x01600000 and length 0x000a4018 Mar 10 13:34:30 <[g2]> Uncompressing Linux............................................... done, booting the kernel. Mar 10 13:34:31 <[g2]> Linux version 2.4.24-uc0 (root@debian) (gcc version 3.3.3) #1 Fri Jun 25 01:13:19 PDT 2004 Mar 10 13:34:33 <[g2]> CPU: XScale-IXP4xx/IXC11xx revision 1 Mar 10 13:34:35 <[g2]> Machine: Intel IXDP425 Development Platform Mar 10 13:34:38 <[g2]> alloc_bootmem_low Mar 10 13:35:02 <[g2]> switchbox ran fine :) Mar 10 13:35:17 <[g2]> except it was missing some drivers Mar 10 13:36:56 ack Mar 10 13:37:02 Hi. Mar 10 13:37:11 <[g2]> did you see the paste ? Mar 10 13:37:23 Yes Mar 10 13:37:25 Cool! Mar 10 13:37:29 youre openslugged! Mar 10 13:37:35 <[g2]> not yet Mar 10 13:37:53 oh, thats the factory one Mar 10 13:38:03 <[g2]> factory one ? Mar 10 13:38:07 with switchbox. Mar 10 13:38:15 the kernel it came with I'm assuming. Mar 10 13:38:21 I didnt throughly read what you did. Mar 10 13:38:38 <[g2]> I just ran switchbox-4.10 with their 2.4 kernel Mar 10 13:38:46 yeah, I see that now. ;-) Mar 10 13:38:48 Cool. Mar 10 13:39:02 <[g2]> Hmmmmm Mar 10 13:39:14 <[g2]> Hmmmmm Mar 10 13:39:27 <[g2]> Do you have a copy of the ixp drivers ? Mar 10 13:39:30 I'm suprised you havent just taken a openslug image and exploded it and flashed the bits in the right place. Mar 10 13:39:55 I have a 2.6.11.2 ixp driver set. Mar 10 13:40:00 <[g2]> well actually that's kinda what I'm in the middle of :) Mar 10 13:40:15 <[g2]> but actually I looked at porting the kernel to the platform for a little while Mar 10 13:40:34 It shouldnt take more than 20minutes to do. Mar 10 13:40:42 <[g2]> exactly Mar 10 13:40:46 its pretty normal. Mar 10 13:40:54 <[g2]> I'm just going through all the changes Mar 10 13:41:11 okay maybe 30 minutes because you'll need to do something about the PCI interrup config. Mar 10 13:41:28 <[g2]> that's all the same Mar 10 13:41:36 <[g2]> except they use all 4 GPIOs Mar 10 13:41:42 You'll probably want to return INTD Mar 10 13:41:46 <[g2]> and the come from two adapters Mar 10 13:42:01 <[g2]> which don't have anything plugged in so I'm not too worried about that Mar 10 13:42:08 nod! Mar 10 13:43:31 Back later, Food Now. Mar 10 14:08:33 [g2]: there's ixp drivers on nudi ... Mar 10 14:11:51 <[g2]> rwhitby-treo, for the 2.4 kernel ? Mar 10 14:12:18 <[g2]> rwhitby-treo, for the 2.4 kernel ? Mar 10 14:13:14 um, no for 2.6 - but you could build them for 2.4 in unslung ... Mar 10 14:13:39 <[g2]> I'm wondering if I'll get a kernel version mismatch Mar 10 14:13:58 <[g2]> I've decided to just port the full kernel Mar 10 14:14:22 nod Mar 10 14:14:50 <[g2]> it'll probably only take a couple hours to be 98% done the proper way Mar 10 14:15:16 * rwhitby-treo thinks that OpenSlug development is going to suffer a slowdown :-) Mar 10 14:16:22 but I guess it will still be OE ... Mar 10 14:20:28 [g2]: What I am seeing is that in 2.6 a process can do usb_bulk transfers via the usbfs device, then when done, release the interface, close the device and exit. When a new process tries to use the same device and interface, the usb_bulk transfers die with -ECONNRESET Mar 10 14:21:59 so this should manifest on the desktop too? Mar 10 14:22:01 USBDEVFS_RESET restores the usb device operation back to normal, but the doco clearly says not to use that call - plus the syslog ends up with warnings Mar 10 14:22:06 yes it does Mar 10 14:22:30 2.6.11-gentoo-r2 has the problem here Mar 10 14:23:04 <[g2]> peteru-home, are you on the arm linux ml ? Mar 10 14:23:48 as far as I can tell, the kernel should only use -ECONNRESET for scatter gather I/O, which I am not touching, but god only knows what happens in the UHCI hcd Mar 10 14:24:13 <[g2]> Is this OHCI ? Mar 10 14:24:16 peteru-home: do any previous 2.6 versions work on the desktop? Mar 10 14:24:24 [g2]: nope, not on any kernel mailing lists - don't have time to read them. does the arm list have a good SNR ? Mar 10 14:24:38 <[g2]> for kernel issues Mar 10 14:25:00 <[g2]> but it depends on you focus Mar 10 14:25:19 <[g2]> however, David B. the guy that wrote the code is there Mar 10 14:25:41 iManufacturer 3 Linux 2.6.11-gentoo-r2 ehci_hcd Mar 10 14:25:41 iProduct 2 Intel Corp. 82801EB/ER (ICH5/ICH5R) USB2 EHCI Controller Mar 10 14:25:49 <[g2]> well he write the ehci-hcd.c code Mar 10 14:26:04 <[g2]> it's late :)... wrote Mar 10 14:27:53 rwhitby-treo: haven't had a chance to test on anything else than 2.6.11 and probably won't be able to in the future Mar 10 14:28:44 just wondering how such a basic thing could be so broken ... Mar 10 14:29:05 basic as in fundamental, not easy. Mar 10 14:29:37 <[g2]> FreeBSD slug! Mar 10 14:29:49 <[g2]> OpenBSD slug! Mar 10 14:31:09 I'm not implying that the kernel is broken. It could be the toppy or my code that has bugs, it's just that something changed between 2.4 and 2.6 that results in failures. Mar 10 14:32:17 we're seeing 2.6 ehci failures with pl2303 serial cables ... Mar 10 14:32:31 ... on a hub. Mar 10 14:33:46 I wonder if a usb 2.0 non-mass-storage peripheral would work plugged in directly with ehci? I don't have any except for disk and hub. Mar 10 14:34:29 that would isolate whether it is a basic ehci problem or a ehci hub problem Mar 10 14:36:27 need USB2 to Ethernet Mar 10 14:37:28 yeah, that would be a good test Mar 10 14:39:21 <[g2]> USB2 to Ethernet worked before Mar 10 14:39:50 on 2.6? Mar 10 14:40:00 <[g2]> sure Mar 10 14:40:26 hmm - how about on a 2.0 hub? Mar 10 14:40:28 <[g2]> I even did host-to-host USB 2.0 Mar 10 14:42:32 wonder if the multiple streams of stuff going through the one interface to the hub is like what peteru is seeing ... Mar 10 14:42:56 * [g2] sets up Mar 10 14:46:10 <[g2]> jffs2 full, cleaned up and rebooting Mar 10 14:46:43 Ding! Mar 10 14:46:56 speaking of multiple things going to usb Mar 10 14:48:08 what I saw in 2.4 was that if puppy was doing bulk transfers via usb and another process did cat /proc/bus/usb/device, the bulk_transfer would lose data Mar 10 14:49:43 but, surely that shouldn't be an issue, because I also had puppy copy tens of gigabytes of data from usb port one to a disk on usb port two Mar 10 14:49:50 so we could do the same thing with a large usb ethernet transfer ... Mar 10 14:50:03 different port? Mar 10 14:50:16 <[g2]> DOH! Mar 10 14:50:27 <[g2]> the module isn't there ! Mar 10 14:51:06 which module? Mar 10 14:51:22 <[g2]> root@LKG0F9D2C:/lib/modules/2.6.11.2# find . -name "*.ko" Mar 10 14:51:23 <[g2]> ./drivers/ixp400/ixp400.ko Mar 10 14:51:23 <[g2]> ./kernel/drivers/net/ixp425_eth.ko Mar 10 14:51:23 <[g2]> ./kernel/drivers/usb/host/ehci-hcd.ko Mar 10 14:51:23 <[g2]> root@LKG0F9D2C:/lib/modules/2.6.11.2# Mar 10 14:52:21 I've also had a report from a debian desktop user where he complained that his puppy transfers would die when he was browsing the internet. as it turns out, he has a usb mouse and it was interfering on the usb bus Mar 10 14:52:22 <[g2]> there may be something busted with OE/Openslug and the extra rev on the kernel the 11.2 Mar 10 14:53:32 <[g2]> or maybe my upgrade just failed Mar 10 14:53:33 peteru: was his toppy and mouse on a hub? Mar 10 14:53:56 which module is missing? Mar 10 14:54:01 <[g2]> usbnet Mar 10 14:54:02 remember that csr is not in the feed ... Mar 10 14:54:04 <[g2]> usbcore Mar 10 14:54:20 ah - but you have your own local feed Mar 10 14:54:21 <[g2]> I was pulling from my own seet Mar 10 14:54:26 <[g2]> set Mar 10 14:55:19 do you set OPENSLUG_EXTRA_* in your local.conf? Mar 10 14:56:04 if so, you need to put those modules in there too, cause they are only in by default if you don't set those vars Mar 10 14:56:12 <[g2]> OPENSLUG_EXTRA_DEPENDS += "strace thttpd " Mar 10 14:56:12 <[g2]> OPENSLUG_EXTRA_RDEPENDS += "strace thttpd " Mar 10 14:56:39 yeah - that won't work cause local.conf is read first Mar 10 14:56:58 that order thing again ... Mar 10 14:57:26 <[g2]> # Move due to the ?= Mar 10 14:57:26 <[g2]> #OPENSLUG_EXTRA_DEPENDS += "strace thttpd " Mar 10 14:57:26 <[g2]> #OPENSLUG_EXTRA_RDEPENDS += "strace thttpd " Mar 10 14:58:16 usbnet is no longer in base image, it is just a default extra. if you set extra, you need to explicitly ask for usbnet too Mar 10 14:58:43 <[g2]> so commenting those out won't do it ? Mar 10 14:59:06 yeah it will, but you'll loOse strace and thhtpd Mar 10 14:59:24 <[g2]> yeah I know that silly :) Mar 10 14:59:26 (if that matters) Mar 10 14:59:39 nod Mar 10 14:59:46 <[g2]> I do like having strace around Mar 10 14:59:54 <[g2]> thttpd was for playing Mar 10 15:01:08 rwhitby-treo: no idea of the usb topology. he said he has usb mouse, printer and toppy, but the printer was not in use. Mar 10 15:07:43 g2: ethernet work on hub? Mar 10 15:08:47 <[g2]> flashing Mar 10 15:14:54 <[g2]> Odd. Mar 10 15:15:11 <[g2]> Have the drivers are in 2.6.11 and the other ones are in 2.6.11.2 Mar 10 15:15:14 <[g2]> that's not good Mar 10 15:16:44 <[g2]> usb 3-1: new high speed USB device using ehci_hcd and address 2 Mar 10 15:16:44 <[g2]> usbnet: version magic '2.6.11 ARMv5 gcc-3.4' should be '2.6.11.2 ARMv5 gcc-3.4' Mar 10 15:18:07 <[g2]> I'll have to check tomorrow sometime Mar 10 15:18:17 <[g2]> time for sleep Mar 10 15:18:22 <[g2]> sorry Mar 10 15:18:32 <[g2]> nite Mar 11 00:45:02 hi all Mar 11 00:45:39 g'day! Mar 11 00:45:53 It looks like I have a bad luck. My nightly build failed at the patch phase of ixp425-eth-1.1-r1 Mar 11 00:46:09 hmm. you're really having problems with patch. Mar 11 00:46:40 it is strange Mar 11 00:47:10 bothe the gcc and the kernel are heavely patched and I had no problems on those packages Mar 11 00:48:10 the error log says: patch did not apply, try again with force (-f -i ...) Mar 11 00:49:09 do you have the version of patch on the RequiredSoftware page? Mar 11 00:50:52 No. I have 2.5.4, stock FC2. But someone told me that it should be fine Mar 11 00:52:41 I'll install 2.5.9 and see Mar 11 00:53:01 morning Mar 11 00:56:48 hanjo: I've never tried OpenEmbedded on FC, and most people here use Debian or Gentoo I think. Mar 11 00:58:37 are you all using 2.5.9? Mar 11 01:01:26 2.5.9 here Mar 11 01:03:46 I've started the build again with 2.5.9. Still no go.... I think I am going to clean up tmp and start from scratch Mar 11 02:40:22 hey, we should really meet at OLS or something Mar 11 02:40:24 who's going? Mar 11 02:40:55 who's paying for my plane ticket :-) Mar 11 02:41:22 you should've written a paper :) Mar 11 02:41:32 What month is that in? Mar 11 02:41:42 do they pay for your plane ticket if you do? Mar 11 02:41:54 if you ask them pretty please with sugar on top Mar 11 02:42:05 there are known cases of them paying Mar 11 02:42:10 at least they let you in for free :) Mar 11 02:42:28 dyoung: 20-23 july iirc Mar 11 02:43:05 Hmm. Mar 11 02:43:15 I wonder if I can pretend its work related. Mar 11 02:43:33 you don't happen to work for a CELF member, do you? Mar 11 02:44:41 Nope. Mar 11 02:44:46 drat Mar 11 02:48:01 Cool - my Asus router can load rootfs from an external usb storage :-) Mar 11 02:48:45 And it also has an NFS server in the custom firmware. I think I've found a new home for the cornice drive ... Mar 11 02:48:52 http://wl500g.dyndns.org/ Mar 11 02:49:38 cool Mar 11 02:49:58 oh, do you have that asus router/printserver/fileserver thingy? Mar 11 02:50:09 wl500g, yeah Mar 11 02:50:26 http://nslu2-linux.thestuffguy.com/gallery/slug-central/IMGP3299?full=1 Mar 11 02:50:35 It's second from the left. Mar 11 02:51:51 Hmm - I wonder if it can handle things attached a usb 2.0 hub with ehci .... Mar 11 02:58:52 ah, it's only usb 1.1, so it doesn't have ehci at all ... Mar 11 03:08:31 Build machine is much more comfortable with the extra 512MB I put in. Mar 11 03:10:01 Hmm - time to create a wl500g distro for openembedded .... Mar 11 03:32:53 ooh yeah...*way* more comfortable with 1G in there. Mar 11 03:33:23 and so the "openasus" distro for the "wl500g" machine is born ... Mar 11 03:33:35 heh really? Mar 11 03:33:51 yeah, I need some way to build binaries for my wl500g Mar 11 03:34:08 it'll be packges only, not firmware - I'll leave that up to Oleg :-) Mar 11 03:34:11 Nice. Mar 11 03:34:40 Reverse of unslung. Mar 11 03:34:59 should be a nice easy clone on the wrt54oe distro Mar 11 03:35:09 (as far as the OE toolchain goes) Mar 11 04:17:11 bitbake failes on findutils :( Mar 11 04:22:54 hi [g2] Mar 11 04:23:39 <[g2]> Morning perlguru Mar 11 04:25:03 hey [g2] Mar 11 04:25:16 <[g2]> Morning rwhitby-away Mar 11 04:25:59 <[g2]> rwhitby-away, I tried to check the logs from dyoung but the seem hosed the last couple days Mar 11 04:55:59 what ARCH is the wl500g ? Mar 11 04:56:10 mipsel Mar 11 04:56:20 same as wrt54g Mar 11 04:56:44 both use the broadcom chipset Mar 11 04:56:53 the wgt634u is also a mips/broadcom Mar 11 04:57:14 does that mean that oe already has the cross toolchain I'll need? ;-) Mar 11 04:57:16 what version kernel? Mar 11 04:57:26 2.4 Mar 11 04:57:30 very likely so Mar 11 04:57:50 is it uclibc based? Mar 11 04:58:08 yes - it uses LEAF Mar 11 04:58:19 which is a variant of shorewall Mar 11 04:58:34 you may well be able to ipkg install openasus packages when I'm done then .... Mar 11 04:59:30 I was hoping I can just add puppy and (to be written) puppy-ui packages to it's base config :-) Mar 11 05:04:24 <[g2]> LEAF include shorewall Mar 11 05:04:30 <[g2]> LEAF includes shorewall Mar 11 05:05:29 <[g2]> that's something kergoth should help ppl do Mar 11 05:05:38 <[g2]> build LEAF distros Mar 11 05:05:46 <[g2]> in OE Mar 11 05:05:50 <[g2]> of course Mar 11 05:17:49 kergoth seems to have a very large list of things to do Mar 11 05:43:32 morning Mar 11 06:00:14 peteru-home: puppy -c dir works in OpenAsus :-) Mar 11 06:10:25 <[g2]> rwhitby-away, Cool! Mar 11 06:27:27 excellent: I think I might have a solution for the 2.6 problems with hung USB state. Mar 11 06:27:42 <[g2]> peteru-home, THAT'd be excellent Mar 11 06:28:02 peteru-home: just working on an ipkg exe for OpenAsus now ... Mar 11 06:28:16 can you get an /opt directory set up on the WGT634U ? Mar 11 06:28:35 [g2]: I'm affraid it's a puppy specific fix, not the ehci issue :( Mar 11 06:28:54 we're looking at using /opt for OpenAsus packages Mar 11 06:30:10 rwhitby-away: dunno about the wgt634u just yet. it uses the .lrp packaging system, which I think decompresses each package to the RAM disk. Mar 11 06:30:26 I'll have to study that after I get some sleep Mar 11 06:30:42 if you can mount something as /opt, then that should be enough ... Mar 11 06:31:08 maybe I'll need to invent something like ungeared :-) Mar 11 06:31:19 the unslung of netgear world Mar 11 06:32:02 nogear Mar 11 06:32:15 <[g2]> rwhitby-away, before you head off for sleep too... Mar 11 06:32:33 belt - and you can give your router a good belting ;) Mar 11 06:32:46 <[g2]> openslug is still borked with regard to the 2.6.11 / 2.6.11.2 modules thingy Mar 11 06:33:32 I'm going to check in the puppy changes to CVS - see if you can make the new puppy work with openslug hub/hubless Mar 11 06:35:50 <[g2]> did you see linux-colinux in there ? Mar 11 06:36:19 [g2]: you'll need to talk to dyoung about the modules thing. I have no idea why it would do that. Are you sure that all your stuff was rebuilt? Mar 11 06:36:51 * rwhitby-away scp's the ipkg executable to the asus .... Mar 11 06:36:51 <[g2]> yes. Just did a fresh pull and rebuild, kernel got rebuilt Mar 11 06:37:16 <[g2]> I'm guessing the name on the bb file should be openslug-kernel_2.6.11.2 Mar 11 06:38:11 yeah, probably Mar 11 06:39:31 rwhitby-away: puppy changes are in CVS, but I won't tag until I had a chance to test more. give it a shot tomorrow morning if you have the time. Mar 11 06:40:36 OK, I'm off to bed. Mar 11 06:40:39 ok, maybe. Mar 11 06:40:43 ~emulate rwhitby Mar 11 06:41:01 if puppy works on OpenAsus, then I might just use that for the moment :-) Mar 11 06:41:07 DOH! no jbot here :\ Mar 11 06:41:26 is it not only USB1.1 ? Mar 11 06:41:41 yeah, but I only use puppy for epg uploads Mar 11 06:41:51 I suppose that's sufficient for EPG, not so great for 40GB transfers Mar 11 06:41:59 ok. night all Mar 11 06:42:21 it's good night from me and it's good night from him. Mar 11 06:42:36 <[g2]> night guys Mar 11 06:45:21 * peteru-home is off to bed and a quick read of the new kernel book by Robert Love. Mar 11 06:45:35 probably get through 2-3 pages before falling asleep :-) Mar 11 06:45:48 <[g2]> that's a good book Mar 11 06:45:57 <[g2]> where are you @ ? Mar 11 06:46:08 * [g2] was reading the interrupt chapter the other day Mar 11 06:46:33 bought the book this morning, so about 2 pages into chapter 1 :) Mar 11 06:46:35 <[g2]> trying to get a handle on the EHCI issue Mar 11 06:46:42 <[g2]> enjoy Mar 11 06:46:55 <[g2]> did you pickup the latest linux device drivers book too ? Mar 11 06:47:23 <[g2]> it's all based on the 2.6 kernel and 2.6.10 code :) Mar 11 06:47:27 not avail in Australia yet - waiting for the boat to deliver them... Mar 11 06:47:27 <[g2]> 3rd ed. Mar 11 06:47:44 <[g2]> :( Mar 11 06:47:55 the usb chapter of that book is online at oreily Mar 11 06:48:03 <[g2]> nod Mar 11 06:48:04 been reading that today too. Mar 11 06:48:30 <[g2]> funny 1/2 the world apart, but on the same page(s) Mar 11 06:48:53 <[g2]> If that's not group think I don't know what is :) Mar 11 06:49:19 but that book has been getting mixed reviews - people either love it or hate it, no middle ground. Mar 11 06:49:59 <[g2]> well since OpenSlug is 2.6.11 it's timely for me :) Mar 11 06:50:15 anyway, off to bed now. Mar 11 06:50:23 <[g2]> sweet dreams Mar 11 06:50:37 <[g2]> of kernel modules dancing in your head Mar 11 06:50:45 * peteru-home will dream of struct alignment no doubt Mar 11 08:51:10 <[g2]> jacques, ping Mar 11 09:06:48 [g2], hi Mar 11 09:30:32 <[g2]> jacques, I was wonder if you could help me identify some stuff for the native build environ on openslug ? Mar 11 09:30:48 <[g2]> bb autoconf automake bash bison bzip2 coreutils diffutils findutils flex gawk grep m4 make ncurses openssh perl quiltperl sed tar Mar 11 09:31:10 <[g2]> and bb gcc gdb Mar 11 09:31:13 <[g2]> all build Mar 11 09:31:31 <[g2]> I was thinking about adding them to openslug-packages Mar 11 09:32:12 ok Mar 11 09:32:17 <[g2]> is there something else ? there were goff stdc++ liddb and wget-ssl in your list on the wiki Mar 11 09:32:35 <[g2]> s/goff/groff/ Mar 11 10:12:14 <[g2]> jacques, back Mar 11 11:37:07 has anyone else done a clean build of openslug-image lately? Mar 11 11:37:18 I'm still getting configure failure on findutils Mar 11 11:37:29 it's been days Mar 11 11:39:36 same here. Mar 11 11:40:05 I always do full builds. That would explain why I'm always the one to complain about build failures. :( Mar 11 11:40:41 VoodooZ_work, yeah sometimes I think you and I are the only ones that do clean builds Mar 11 11:41:27 since it only takes about 63 minutes, I almost always do a clean build Mar 11 11:41:59 I mentioned the findutils thing and the actual error (looks like in automake) on #oe but got no response ( a couple of days ago) Mar 11 11:42:25 yeah. I know how it feels. Mar 11 11:43:03 I'm stuck doing builds on my dual xeon which is a bit slower with all the extra junk. Mar 11 11:43:51 the quad opteron builds in less than 30minutes but because nobody has a clue why the AMD64 binaries are bigger I can't really use it as the kernel overflows. :( Mar 11 11:44:13 wow, that's impressive Mar 11 11:44:22 strange about the 64bit thing tho Mar 11 11:44:39 maybe because the optimizations work differently in the current 64bit gcc ? Mar 11 11:45:29 I was reading an article where ppl were complaining it wasn't a good test because the guy used differnt optimizations for 64 and 32 bit and he said certain opts arent currently supported in 64bit mode Mar 11 11:45:59 tho that might not make sense in this case, since you are cross-compiling.... Mar 11 11:46:09 tho, is your cross-gcc 64bit ? Mar 11 11:47:22 yeah, I wish somebody could find the problem as It's way over my head. I asked the #oe and #gcc guys but nobody had a clue. Mar 11 11:47:58 Somebody said it might be because the cross compiler is generating 64bit ARM binaries. The .text word size is bigger in every binaries produced. Mar 11 11:48:16 it should not be generating 64bit binaries Mar 11 11:48:30 that would be a gcc bug Mar 11 11:48:39 this could also be binutils-related Mar 11 11:48:44 yeah. Mar 11 11:49:19 but I'm afraid I can't do much as gcc is a huge beast and little do I know about it. Mar 11 11:57:00 do we use the latest binutils? Mar 11 11:57:21 probably not. Mar 11 11:57:47 you might try it. Mar 11 11:57:49 I know some of the stuff is fixed but I'm not sure where. Mar 11 11:58:28 That's what my latest build uses: binutils-cross-2.15.94.0.1-r0 Mar 11 12:01:34 saw that too Mar 11 12:01:43 http://sources.redhat.com/ml/binutils Mar 11 12:04:32 thanks. I'll have a quick look Mar 11 12:05:45 quick :) Mar 11 12:07:34 binutils-2.15.96.tar.bz2 Mar 11 12:07:44 is latest Mar 11 12:10:00 ok. Mar 11 12:10:09 Lots of stuff in that mailing list. ouch. Mar 11 12:50:15 [g2] ping Mar 11 12:50:35 [g2] we talked about whether or not it needs to be openslug-nerle_2.6.11.2.bb and decided it didnt. Mar 11 12:50:53 [g2] and when I make builds it works for me. Mar 11 12:51:03 I have /lib/modules/2.6.11.2 Mar 11 12:51:19 <[g2]> and 2.6.11 Mar 11 12:51:43 No, 2.6.11.2 Mar 11 12:51:51 I dont have 2.6.11 dir Mar 11 12:52:04 <[g2]> you haven't done a bb -clean on your openslug-kernel Mar 11 12:52:07 If you have a 2.6.11, you didnt clean something. Mar 11 12:52:31 whats in the 2.6.11 dir? Mar 11 12:52:34 <[g2]> no the perferred version locks that kernel at 2.6.11 Mar 11 12:52:59 Oh, youre right. Mar 11 12:53:06 I set my preffered version to 2.6.11.2 Mar 11 12:53:12 (i think lemme check) Mar 11 12:53:37 No, I dont set the preferred version. Mar 11 12:53:54 so whats in lib/modules/2.6.11 ? Mar 11 12:54:09 <[g2]> all the kernel build modules Mar 11 12:54:21 weird. Mar 11 12:54:26 I wonder why it works for me Mar 11 12:54:41 <[g2]> ixp425 ipx4xx and ehci for some reason were winding up in 11.2 Mar 11 12:54:50 ixp is explainable. Mar 11 12:55:12 echi is weird. Mar 11 12:55:46 <[g2]> yeah it's a little weird Mar 11 12:56:05 welklk, I initially had a 2.6.11.2 bb file, but it worked for me both ways so we talked about it and decided to just stuff it into the existing 2.6.11 bb file Mar 11 12:56:16 Does it work for anyone else? Mar 11 12:56:36 <[g2]> the system would boot and run ixp and ehci, but no other drivers Mar 11 12:56:50 BTW before doing this I do rm tmp/{work,stamps}/{openslug-*,ixp*} . Mar 11 12:56:51 <[g2]> it wasn't a real bad failuer Mar 11 12:57:21 the thing is though is it puts everything in the right place for me, and thats just weird. Mar 11 12:57:48 <[g2]> well you had 11.2 as the perferred kernel right ? Mar 11 12:58:04 <[g2]> it worked as told Mar 11 13:00:00 nono , said I dont set prefferred kernel anymore. Mar 11 13:00:09 No, I dont set the preferred version. Mar 11 13:02:01 <[g2]> did you pull ? Mar 11 13:02:15 <[g2]> what changes are out standing Mar 11 13:03:01 Yes, I just did this last night Mar 11 13:03:07 I dont have outstanding changes Mar 11 13:03:30 <[g2]> well I checked in the perferred version today Mar 11 13:03:30 not related to this at least, I have a half working nail package outstanding, but that does not count. ;-) Mar 11 13:03:55 Where did you prefer it? Mar 11 13:04:49 <[g2]> http://nslu2-linux.bkbits.net:8080/openembedded/patch@1.3046.1.1?nav=index.html|ChangeSet@-1d|cset@1.3046.1.1 Mar 11 13:06:46 Oh okay you made a new set, thats cool Mar 11 13:07:05 in that case, we should probably repair 2.6.11.bb back to really use 2.6.11 Mar 11 13:07:31 <[g2]> I thought about that, but I was too lazy :( Mar 11 13:07:56 <[g2]> it's in the repo right ? Mar 11 13:12:04 <[g2]> dyoung-zzzz, so except for some errors which I think are PCI bus errors relating to the ECHI issue Mar 11 13:12:28 <[g2]> my disks and usbnet device come up find on my powered 2.0 hub Mar 11 13:12:47 <[g2]> I think I just need to ifconfig eth2 Mar 11 13:15:34 strange Mar 11 13:15:48 <[g2]> I know I'm wierd Mar 11 13:16:00 What brand hub is that? Mar 11 13:16:22 <[g2]> belkin F5U124-OE Mar 11 13:16:32 <[g2]> funny about the OE part eh ? **** ENDING LOGGING AT Fri Mar 11 23:59:57 2005