**** BEGIN LOGGING AT Tue Sep 20 02:59:56 2005 Sep 20 03:36:31 03frederic 07org.openembedded.dev * r344cfe46... 10/packages/gnome/ (gnome-vfs-dbus/gssapi.patch gnome-vfs-dbus_2.8.4.4.bb): Sep 20 03:36:31 gnome-vfs-dbus: do not use GSSAPI_LIBS if gssapi does not compile Sep 20 03:36:31 http://bugs.openembedded.org/show_bug.cgi?id=250 Sep 20 04:10:13 reading the logs, it seems yvasilev got the bug a few hours ago and was searching for a be volunteer... any news on that? Sep 20 04:11:28 nope Sep 20 04:11:40 I'm not set up to compile the stuff, I guess Sep 20 04:11:51 I could test it if someone compiled it Sep 20 04:13:26 ok. let's wait for yvasilev to come back then. Sep 20 04:20:18 morning inge Sep 20 04:25:05 ingeba: are you still running Unslung? Sep 20 04:35:32 conradl: what was that about you leaving your job? Sep 20 04:39:42 yeah had a guts-full ...and left...i do things like that. Sep 20 04:39:58 kinda the same way I left siemens Sep 20 04:40:18 my brother hates them with a passion. Sep 20 04:40:33 Siemens? Sep 20 04:40:37 I'm not surprised :P Sep 20 04:40:41 yeah Sep 20 04:41:04 their office in brisbane had issues. Sep 20 04:42:15 i had setup a replacement job. so 1 weeks holiday and back to the grind. wanna buy a PC? Sep 20 04:44:02 heh, not exactly, why? Sep 20 04:44:41 i will be with a whitebox mob from next tuesday. Sep 20 04:44:50 ah Sep 20 04:45:04 Can't say I've ever bought a pre-built box Sep 20 04:45:52 a nslu2 would be pre-built? Sep 20 04:46:08 ah yes, that's embedded Sep 20 04:46:25 The amd x64 3700+ desktop I have today, is merely an upgraded version of my 386dx/33... Not a single part is left though :P Sep 20 04:48:42 i also have issues like that. new case every 6 months, new bits every month. Sep 20 04:54:36 "upgraded 386" ?? Sep 20 04:54:40 geez dude. Sep 20 04:54:44 BUY A CASE! Sep 20 04:54:58 or is that differnt too? Sep 20 04:55:06 that has been upgraded a bunch of times Sep 20 04:55:20 along with everything else... Sep 20 04:55:25 Okay. hehe I thought you were implying you were using the same case! Sep 20 04:55:30 heh, no Sep 20 04:56:04 I was getting ready to paypal you a case! ;-) Sep 20 04:56:18 hahaha Sep 20 04:56:27 no, I have several cases Sep 20 04:58:27 a nice silent zalman with heatpipes everywhere About 1500 here. Sep 20 04:58:44 zalman what? Sep 20 04:59:33 zalman make a v nice v dear case with heatpipe cooling. for about 1500 AU Sep 20 04:59:55 oooh Sep 20 05:00:06 How much is 1 AU? Sep 20 05:00:15 in eur, usd or nok ;) Sep 20 05:00:22 .77 us or .44 gbp Sep 20 05:00:34 whats a nok? Sep 20 05:00:52 conradl: Norwegian Krone Sep 20 05:01:04 nice. Sep 20 05:02:42 that's an... expensive case Sep 20 05:03:59 i did two last year for a video editing mob with dual xeons 4g ram etc...lots of HD 72K AU all up Sep 20 05:05:11 ah Sep 20 05:05:23 I'm using one 120mm fan to cool my PC Sep 20 05:05:33 that's all I need ;) Sep 20 05:06:07 power supply fan? but its knee deep in ice 6 months of the year? Sep 20 05:06:34 no, fanless PSU Sep 20 05:06:41 watercooled everything Sep 20 05:07:05 not a kit i hope Sep 20 05:07:27 'tis a kit. And it's a cheap one Sep 20 05:07:39 ouch. Sep 20 05:07:44 Tt Bigwater Sep 20 05:08:46 it works like a charm though Sep 20 05:08:53 they look all right, i went stupid and got 1/2" gear with v high flow. Sep 20 05:09:23 ah Sep 20 05:11:08 anyway, the 120mm fan is virtually silent, as I run it on the minimum RPM. Also, my case is padded with asphalt and foam ;) Sep 20 05:11:29 asphalt?! Sep 20 05:11:34 yeah Sep 20 05:11:44 like the stuff you make a road out of? Sep 20 05:11:47 yes Sep 20 05:11:51 without the stones, of course Sep 20 05:11:54 how... ... ?? Sep 20 05:12:28 Hope you didnt fill it with the hot asphalt, or do watercooling stop everything from melting anyway =) Sep 20 05:13:55 dyoung-zzzz: Something like this: http://www.quietpc.com/uk/materials.php Sep 20 05:14:27 the "very high density Acoustic Barrier Mass" is asphalt Sep 20 05:14:47 doubles as double-sided tape Sep 20 05:14:48 Oh, you call that stuff ashfault ? Sep 20 05:15:21 it's the same sticky goo. only without the stones Sep 20 05:15:31 Looks like HDPE to me. Sep 20 05:15:39 ashphalt is tar + rocks Sep 20 05:15:50 usually. Sep 20 05:16:04 hdpe? Sep 20 05:16:18 High Density Poly Ethylene. Sep 20 05:17:23 possible Sep 20 05:17:36 felt like tar. smelt like tar. Sep 20 05:17:46 hmm, weird. Sep 20 05:17:49 the burning question is ...snow to the knees and raindeer at the backdoor? NAiL? Sep 20 05:18:01 reindeer stink Sep 20 05:18:04 conradl: not the last 10 years, no. Sep 20 05:18:12 dyoung-zzzz: yep. Kinda like ferrets. Sep 20 05:18:44 ... But so tasty..... Sep 20 05:19:20 hmm. Sep 20 05:19:22 i get dropbears in the backyard!... Sep 20 05:20:04 Reindeers taste like a cross between Dolphin and Koala ... Sep 20 05:20:26 "Dropbears, cute but deadly!" Sep 20 05:20:27 Whoa, have you really had dolphin? Sep 20 05:20:43 taste like chicken. Sep 20 05:20:55 I've been wondering about dolphin. Sep 20 05:20:55 whale tastes like shit :P Sep 20 05:21:38 because the japanese word for dolphin literally means "ocean pig"; but since pig is domesticated animal for food; I was kinda wondering if dophin is yummy too. Sep 20 05:21:59 dyoung-zzzz: reindeer is ok. deer is way better though ;) Sep 20 05:27:40 bleh Sep 20 05:31:33 bleh Sep 20 05:31:59 my connection is acting up Sep 20 06:01:17 03koen 07org.openembedded.dev * re2a816af... 10/packages/ (2 files in 2 dirs): Sep 20 06:01:17 packages/gpe-contacts/: update hildon version to 0.42 Sep 20 06:01:17 packages/gpe-mini-browser/: update hildon version to 0.16 Sep 20 06:37:17 the irssi package works too Sep 20 06:37:53 in OE? Sep 20 06:38:02 in debonaras Sep 20 06:38:11 aah Sep 20 06:54:29 Belgieeee Sep 20 07:56:47 There is no mt-daapd package in debian, but there is a .deb (for i386) on the mt-daapd homepage. Tips on strategy there? What route would be most useful to most people? Sep 20 07:57:55 Adopting the package into debian from source, would that make for an easy build for debonaras aswell? Sep 20 08:09:34 03bzhou * 10unslung/ (4 files in 2 dirs): upstream upgrade from 0.4.25 to 0.4.26 Sep 20 08:21:06 dwery: ping Sep 20 08:23:22 yvasilev: what's the status on the csr stuff? Sep 20 08:26:04 NAiL: still need to verify that my conjecture is right, and then it'll need some serious patching Sep 20 08:27:45 NAiL: I'll try to build be modules for you to test ;-) Sep 20 08:30:51 That'd be brilliant :) Sep 20 08:32:50 NAiL: what is your uname -a so I know the kernel version I should use Sep 20 08:32:53 yvasilev: pong Sep 20 08:33:47 dwery: hi, I probably found it, and you'll be surprised but it's endiannes issue :-D Sep 20 08:34:15 Linux slugave 2.6.12.2 #1 Tue Sep 20 02:43:48 CEST 2005 armv5teb unknown unknown GNU/Linux Sep 20 08:34:18 I must have known :D Sep 20 08:35:29 yvasilev: tell me :) Sep 20 08:35:30 the problem is that most components are big endian connected, and this includes ram and the nics Sep 20 08:36:08 (that's why you need to byteswap the kernel when you load it to flash) Sep 20 08:36:36 ok Sep 20 08:36:39 yvasilev: I'll be gone for a couple of hours though... can test stuff when I get back Sep 20 08:37:17 so, the binary microcode we are loading needs to be byteswaped (still not sure if all of it or just some parts) Sep 20 08:37:22 NAiL: ok Sep 20 08:37:29 uh Sep 20 08:37:38 NAiL: I'll have the modules for that time Sep 20 08:37:43 cool Sep 20 08:37:59 yvasilev: i thought about it but since LE i supposed to be supported.... Sep 20 08:38:18 dwery: you can see it by the marker of the start of the npe images Sep 20 08:39:04 also we need to find if it's snapgear or intel who are duing it wrong Sep 20 08:39:07 in the microcode, right? Sep 20 08:39:27 s/duing/doing/ Sep 20 08:39:33 yes Sep 20 08:39:54 can you point me to a specific file/location? Sep 20 08:40:09 <[g2]> dev/null :) ? Sep 20 08:40:19 [g2]: :-D Sep 20 08:40:28 <[g2]> heh Sep 20 08:40:40 also we were doing it a little bit wrong, there is a much better way to compile it, if you care to start from scratch, i'll guide you Sep 20 08:41:03 [g2]: that's the only one that works!!! Sep 20 08:41:18 <[g2]> /dev/null Sep 20 08:41:22 yvasilev: ok, but first let me check the microcode issue :D Sep 20 08:42:16 compile ixp400_xscale_sw/src/npeDl/ixNpeDlImageConverter.c Sep 20 08:43:16 and run it Sep 20 08:44:03 (this was going to be done in the process ;-) ) Sep 20 08:44:50 it'll give you a IxNpeMicrocode.dat Sep 20 08:46:43 which target will build that file? Sep 20 08:46:50 naitive Sep 20 08:47:05 x86 Sep 20 08:47:47 oh! Sep 20 08:47:57 do you have a Makefile ? :-D Sep 20 08:48:49 gcc compile ixp400_xscale_sw/src/npeDl/ixNpeDlImageConverter.c -o ixNpeDlImageConverter Sep 20 08:48:57 gcc ixp400_xscale_sw/src/npeDl/ixNpeDlImageConverter.c -o ixNpeDlImageConverter Sep 20 08:49:07 oh.. i thoung it required the other /c Sep 20 08:49:08 .c Sep 20 08:49:18 ok, it'll need an include Sep 20 08:49:51 just add the appropriate -I ;-) Sep 20 08:50:00 ok Sep 20 08:50:02 compiled Sep 20 08:50:22 run it, so you get IxNpeMicrocode.dat Sep 20 08:51:21 got it. I think I saw this fil ein one of the miriad of build I made in th epast nights.. Sep 20 08:52:34 now you can see the #define NPE_IMAGE_MARKER 0xfeedf00d in /src/npeDl/IxNpeDlImageMgr.c Sep 20 08:52:55 just a sec Sep 20 08:53:35 yes Sep 20 08:53:57 and if you hexdump the first bytes you get that exact number swapped Sep 20 08:55:14 oh Sep 20 08:56:30 note that ixNpeDlImageConverter converts the file to big endian, also note that www.intel.com/design/network/applnots/25423701.pdf clearly states that the nics are connected in big endian Sep 20 08:56:42 ok Sep 20 08:56:57 so you think we have to byteswap it? Sep 20 08:57:41 so it's more of a problem that we have mixed endiannes, cpu in bug an mem in little Sep 20 08:58:41 but this thing is supposed to work LE, so they must have addressed it Sep 20 08:58:48 yes, also unless it's byteswapped, the kernel will not accept it if we give it to it with the /dev/ixNpe device Sep 20 08:59:29 why /dev/ixnpe ? isn't the microcode built in the module? Sep 20 09:00:03 I'm nor so sure that this statement is correct, yes it is supposed to work with montavista's kernel and no snapgear patches Sep 20 09:00:28 but montavista's kernel has a diff of about 60MB to vanilla Sep 20 09:01:02 :-O Sep 20 09:01:35 so I'm not sure about where it is byte swapped (if it is) Sep 20 09:01:48 if you want another test Sep 20 09:01:59 load your modules Sep 20 09:02:14 ifconfig eth0 x.x.x.x up Sep 20 09:02:50 watch ifconfig Sep 20 09:03:09 loading.. Sep 20 09:03:22 you'll see that the number of dropped packets changes Sep 20 09:04:26 yes Sep 20 09:05:13 if you transfer this huge number to hex, you'll see that just the first byte changes and the other ones are 0x00, so if you byte swap it, you'll get meaningful results Sep 20 09:06:24 oh Sep 20 09:06:54 wasn't this number generated in _eth.c? Sep 20 09:07:56 so, what I want to do, is to check our patch set on a be slug, and if it works, then all is the same except endiannes, all the components are source compiled, so they are should be ok, but the npe should load and initialize correctly and so generate interrupts to the driver Sep 20 09:07:57 yes Sep 20 09:08:46 (actually maybe in ial) but remember that we are working in mixed endiannes Sep 20 09:09:27 ok. did you already tried to byteswap the microcode? Sep 20 09:09:37 and it will be written by the ial/_eth to big endian memory, and then read by the kernel Sep 20 09:09:40 yes Sep 20 09:10:30 no success i suppose :) Sep 20 09:11:20 the problem is that I need to byteswap it for the kernel to recognize it, and then possibly byteswapp it again all, or just the right parts of it so it stays in be Sep 20 09:11:25 not at all Sep 20 09:11:36 none at all Sep 20 09:12:12 so, what do you think about all this? Sep 20 09:12:33 i'm quite confused. If we trust intel saying that the whole thing is LE compatible, where should they have handled all of this? Sep 20 09:14:36 there is one function, ixNpeDlTestRegWrite, that does the actual writing of the npe and its parameters, a similar one reads from it Sep 20 09:14:54 and it's definition is not in the ial Sep 20 09:15:07 maye we must check if they have defines for BE and LE Sep 20 09:16:39 actually I was unable to find it's definition anyware, but the code still compiles, and uses it, so it most be somewhere Sep 20 09:17:25 so, if you can find it, it'll be of much help Sep 20 09:17:45 i'm going to search Sep 20 09:17:57 (will send an invoice to intel.. :) ) Sep 20 09:18:44 also we van try to get montavista, and do it the intel way just to may sure they are saying reproducible things Sep 20 09:19:31 never worked with montavista... Sep 20 09:21:41 I have a ppc NAS running montavista's 2.4.20, but I never managed to recompile working kernel, but it's diff to vanilla is a little bit more than 60MB Sep 20 09:22:21 yeck! Sep 20 09:22:29 but I had just the kernel sources and no config nor any serial console, so it wasn't an easy task Sep 20 09:23:15 ack Sep 20 09:24:58 yes, it is actually mvl 3.1 (the version intel suggests) Sep 20 09:32:50 mvl 4 is kernel 2.6 which intel probably won't support for a while Sep 20 09:33:01 if at all Sep 20 09:34:12 dwery: also, we should also make the npeDl to make more check (if available) to if it actually runs the npe (there are some check to see if it is stopped) Sep 20 09:34:39 lennert: never imagined montavista will ever move to 2.6! Sep 20 09:35:04 dwery: but it was to late, so I had no time to explore that path Sep 20 09:35:25 yvasilev: i'm not quite sure the firmware gets included in my build... Sep 20 09:36:14 dwery: this is where we unify our building method ;-) Sep 20 09:36:29 :) Sep 20 09:38:16 * yvasilev is making the patches Sep 20 09:40:33 thx Sep 20 09:45:36 yes, the microcode ends in lib/linuxle/npeDl Sep 20 09:45:40 so it gets included Sep 20 09:47:54 actually I still don't know how you managed to compile it into the module :-S Sep 20 09:48:34 I just type make :D Sep 20 09:48:37 typed. Sep 20 09:53:02 yvasilev: deepak wouldn't have ported ixp4xx and ixp2000 to 2.6 otherwise :) Sep 20 09:53:25 yvasilev: it's just that ISVs keep asking for 2.4 Sep 20 09:54:20 lennert: what/who is deepak? Sep 20 09:55:47 deepak saxena, maintainer of the 2.6 ixp4xx/ixp2000 port Sep 20 09:56:28 ah Sep 20 10:03:05 given the way th econverter works, it seems it always write the .dat in order to be able to read it on a big endian processor Sep 20 10:04:35 it always writes it in big andian, yes Sep 20 10:04:52 but check: www.intel.com/design/network/applnots/25423701.pdf Sep 20 10:05:22 you'll see that the nics are always be Sep 20 10:05:35 ok Sep 20 10:05:42 so they must receive their firmware be Sep 20 10:06:03 I would assume so Sep 20 10:08:46 but you already swapped it.. Sep 20 10:10:00 I think the problem is that it reads it from be memory with le cpu Sep 20 10:12:04 pk, patches work, packing and sending them Sep 20 10:12:24 thanks Sep 20 10:17:01 it seems we may need to patch the image but not it's header Sep 20 10:17:04 its. Sep 20 10:17:52 btw when i make, I use libIxp425.a and ixp400.o targets Sep 20 10:18:46 Oh, my first mt-daapd_0.2.3-1_armeb.deb and it was easy even. Ill try get it posted on their project page, with the i386 .debs. Easier to find them all. Sep 20 10:20:53 dwery: sent Sep 20 10:21:04 getting them.. Sep 20 10:21:28 now there is no need for libIxp425.a ;-) Sep 20 10:22:15 :) Sep 20 10:22:28 i made an a little bit to explicit instructions, but this is only so there are no ambiguities left Sep 20 10:22:31 I was looking into IxEthDBFeatures.c Sep 20 10:22:45 after the image is loaded, the code tries to communicate to the npe Sep 20 10:23:59 giving an error it it does not Sep 20 10:24:05 if Sep 20 10:24:06 so far we have not used EthDB, else it crashed (for me), and I attribute it to the fact that npe is el garbage so there is no way to comunicate with ti Sep 20 10:24:22 oh. Sep 20 10:24:36 but ethDB is in my component list Sep 20 10:24:52 that is also probably why pooling mode crashes so badly Sep 20 10:25:41 yes but CONFIG_IXP425_COMPONENT_ETHDB is unset for _eth.c Sep 20 10:26:13 on mine is enabled Sep 20 10:26:48 with no_phy_scan=1? Sep 20 10:27:13 tried both with and w/o Sep 20 10:27:41 ethdb is really not needed anyhow. Sep 20 10:27:50 where does the firmware gets loaded? Sep 20 10:28:38 ok, actually I tried ethdb log before making a working thing, so I don't despond for why it was crashing :-D Sep 20 10:28:48 :) Sep 20 10:28:55 what do you think of this process/patches? Sep 20 10:29:38 i;m looking at it. Sep 20 10:29:42 (ok, working thing := modules that load but not necessarily work (tm) ) Sep 20 10:29:48 does your process also build the firmware? Sep 20 10:29:54 yes Sep 20 10:30:09 along with the ixp400.o Sep 20 10:30:26 ok, i'm going to try it Sep 20 10:31:42 brb Sep 20 10:37:52 in mod1 gcc is used instead of the one specified by LINUX_CROSS_COMPILE.. trying to fix Sep 20 10:52:38 I get an error while cat'ing the microcode: /dev/ixNpe: No such device or address . I've created with mknod before. Sep 20 10:53:04 crw-r--r-- 1 root root 241, 0 Jan 1 01:02 /dev/ixNpe Sep 20 10:53:54 oh well.. ixp400.o is too smal.. I must have a problem somewhere Sep 20 10:58:12 dwery: good point about the gcc part, how big is your ixp400.o? Sep 20 10:59:22 yvasilev: solved. module in ok. no useful result anyhow. brb. Sep 20 11:09:28 ok, here I am Sep 20 11:09:38 rocking like a hurricane Sep 20 11:09:48 :) Sep 20 11:10:00 dwery: yes it gives not breakthrough but it's a little bit easier and we have a common base Sep 20 11:10:16 I swapped the entire microcode, right? Sep 20 11:10:23 did ti worked for you at last as well as the previous modules you have tested? Sep 20 11:10:27 yes Sep 20 11:10:28 yes Sep 20 11:10:50 but the microcode has header and length, compiled and read by little endian code Sep 20 11:11:02 so I should have seen some errors... Sep 20 11:11:37 perfe3ct, then I was able to reproduce up to where we where ;-) Sep 20 11:12:06 bte I always get this, with both mine and yours: ixp400_eth: eth1: ixEthAccPortDisable(0) failed Sep 20 11:12:11 when ifconfig down Sep 20 11:12:24 do you mean the module, or the .dat? Sep 20 11:13:27 i mean codebase Sep 20 11:14:12 yes, have that ixEthAccPortDisable(0) too Sep 20 11:14:53 uh, i guess it's ok then :) Sep 20 11:16:08 actually, it just has the IDs of the microcode it wants to use, and the .dat separates them by that NPE_IMAGE_MARKER word so it's more dynamic Sep 20 11:17:30 did you tried defining IX_NPEDL_TRACE_LEVEL ? Sep 20 11:23:08 no Sep 20 11:24:46 i'm trying.. Sep 20 11:30:27 have to change the tracking code a bit... Sep 20 11:30:34 it uses printf Sep 20 11:32:43 if you manage to get the trace, please send it to me by email, I have to go now Sep 20 11:33:51 ok.. Sep 20 11:34:01 the debugging code is ugly.. Sep 20 11:50:23 it seens the swapped microcode is being loaded succesfuly, since if I send the unswapped one the ethd module does not load. Sep 20 11:55:09 so tracing is not useful, as errors are always reported. Sep 20 11:55:33 well.. have to go now. Sep 20 11:55:34 bye Sep 20 11:55:54 poka Sep 20 11:59:52 <[g2]> DOH! Sep 20 11:59:57 <[g2]> dwery-away, around ? Sep 20 12:00:08 stilll around :) Sep 20 12:00:16 <[g2]> kernel patches Sep 20 12:00:27 who? what? where? Sep 20 12:00:28 :) Sep 20 12:00:35 <[g2]> I'm ready to post on the developers list the Loft changes Sep 20 12:00:44 great Sep 20 12:01:00 <[g2]> I've got two questions for lennert Sep 20 12:01:31 <[g2]> we regarding compress-head and the other map_irqs Sep 20 12:02:10 <[g2]> I can just post to the list and then discuss for 5 minutes ? Sep 20 12:02:29 ok Sep 20 12:02:44 for me. I'll have to go away in about 15 mins Sep 20 12:03:05 <[g2]> Ok... don't worry about it. Sep 20 12:03:25 <[g2]> I'll just post to the list and get comments and incorporate them as appropriate Sep 20 12:03:50 <[g2]> I just wanted to give you a heads up Sep 20 12:03:59 <[g2]> and wanted to bug lennert :) Sep 20 12:04:04 i'm here Sep 20 12:04:15 just mail the diffs and i'll review Sep 20 12:04:23 <[g2]> thx Sep 20 12:18:31 hi Sep 20 12:19:35 I want to build an USB device driver, but when I compiled it I got many errors Sep 20 12:19:45 dont know why Sep 20 12:20:21 sombody here who can help me? Sep 20 12:22:13 on this page I found a template: http://ezs.kr.hsnr.de/TreiberBuch/html/sec.usbsub.html#EX.USBCHECK Sep 20 12:22:54 all include files are on the NSLU2, but I got errors when I compile this template Sep 20 12:25:38 i want to create an device driver for an usb-cd-storage-box Sep 20 12:26:25 I know all string_discriptors and commands for it, but cant get the template to work Sep 20 13:15:27 nobody online? Sep 20 13:16:06 xcodex: i can't help you, sorry, don't know anything about the usb subsystem Sep 20 13:18:23 xcodex: i suspect that nobody here knows the answer to your question, and that's why nobody is responding Sep 20 13:28:48 xcodex: Do you get your module to compile on a x86 box? Sep 20 13:34:20 no directly on the nslu2 Sep 20 13:35:00 yeah Sep 20 13:35:24 but can you get it to compile on an x86 box? Ie... Does the module compile *at* *all*? Sep 20 13:35:51 sorry I have no x86 with linux here Sep 20 13:36:09 the template should work Sep 20 13:36:28 the include files are all in the right place in the NSLU Sep 20 13:37:02 can you pastebin the errors? http://pastebin.ca/ Sep 20 13:37:06 yes Sep 20 13:39:13 here it is: http://pastebin.ca/23467 Sep 20 13:39:46 eventually you can test the template in your box? Sep 20 13:40:17 the template is on this page at the end: http://ezs.kr.hsnr.de/TreiberBuch/html/sec.usbsub.html#EX.USBCHECK Sep 20 13:42:18 nope Sep 20 13:42:26 doesn't compile on my ubuntu box either Sep 20 13:42:33 the template has errors Sep 20 13:43:30 It's not exactly a nslu2-specific problem, so I suggest you contact the author of the article. Sep 20 13:48:01 are you sure? If you put only the include-lines in the teplate without anything else you got errors too Sep 20 13:49:57 I think the include files on the nslu2 have errors Sep 20 13:50:47 in that case, the include files on my ubuntu box also have the same errors Sep 20 13:51:40 and my gentoo box Sep 20 13:52:01 you are using the crosscompile tools for nslu? Sep 20 13:52:17 I tried compiling it natively on the slug Sep 20 13:52:40 and then I tried to get it to compile on my ubuntu and gentoo boxes Sep 20 13:52:45 none of them worked Sep 20 13:53:46 * NAiL notices the template has a goto statement Sep 20 13:54:19 "GOTO statement considered harmful" -- E. Dijkstra Sep 20 13:54:43 is there another way to build usb device for the NSLU2? Sep 20 13:55:03 (usb devide driver) Sep 20 13:55:26 good evening Sep 20 13:55:51 xcodex: tried this one http://www.lrr.in.tum.de/Par/arch/usb/usbdoc/ ? Sep 20 13:56:34 or just google for "linux device driver guide" Sep 20 13:56:43 anyone have any more info about usb-audio on unslung? I understand someone is writing a wiki page (as of february?) Sep 20 13:57:25 yes I was there already but it didnt helped me Sep 20 13:57:49 im not so good in develeoping to start from zero Sep 20 13:59:04 Then I can't help, I'm afraid. Try contacting the author of the first article... Sep 20 14:00:02 its the last chance Sep 20 14:00:08 thanks for your help Sep 20 14:01:05 did sk8joker ever write that wiki? were his changes added to cvs? thanks.. Sep 20 14:03:23 Sound Hardware Sep 20 14:03:23 Any linux-supported usb sound adapter should work (see http://www.qbik.ch/). Unslung users will have to build the "audio.o" kernel module from source. Sep 20 14:03:44 Another (unrelated) question - on unslung 5.5, I'm seeing this in my messages: <86>Sep 20 21:44:29 dropbear[586]: Warning: Reading the random source seems to h Sep 20 14:03:44 ave blocked. If you experience problems, you probably need to find a better entr Sep 20 14:03:44 opy source. Is that a known unslung problem and is there a workaround? didn't find anything from searching nslu2-linux.org (with the Search function) Sep 20 14:04:09 NAiL - thanks. Was there a wiki for building audio.o for unslung? Sep 20 14:04:52 stripwax: not that I know :P Sep 20 14:05:06 http://www.nslu2-linux.org/wiki/HowTo/SlugAsAudioPlayer Sep 20 14:05:22 That's where I found the relevant part, but it doesn't exactly say much... Sep 20 14:05:32 hm, ok, thanks. I'm familiar with that page, was looking for tips on building audio.o .... Sep 20 14:11:20 Anyone on here know how to build the audio.o driver on unslung without building the entire world? Sep 20 14:13:25 The link here in the wiki is broken - it takes me to $Group/$Name http://www.nslu2-linux.org/wiki/FAQ/BuildingKernelModules Sep 20 14:16:19 fixed Sep 20 14:22:16 should I "make build-unslung" or "make unslung"? These two pages seem to differ: http://www.nslu2-linux.org/wiki/Development/HomePage and http://www.nslu2-linux.org/wiki/Development/MasterMakefile Sep 20 14:22:32 iirc, those are exactly the same Sep 20 14:22:37 do make unslung ;) Sep 20 14:24:19 I tried "make build-optware-nslu2" the other day (because make optware-nslu2 didn't seem to be a valid target) Sep 20 14:24:41 [devuser@LKG7DB347 unslung]$ make unslung Sep 20 14:24:42 make: Nothing to be done for `unslung'. Sep 20 14:25:43 NAil: some question: when I compile the template which prog do I need? armv5b...c++,cpp,g++,gcc? Sep 20 14:33:20 I have seen there is usb driver example in the unslung package. The file called dabusb.c in snapgear\linux-2.6.x\drivers\usb\media. Sep 20 14:33:39 But it does gives error, too, at compile Sep 20 14:34:58 stripwax: for building kernel modules, you want the unslung-image target Sep 20 14:35:29 but just unslung or build-unslung should do much the same thing anyway, since all the packages are in optware. Sep 20 14:43:45 rwhitby - I must be missing something here. "make unslung" does nothing (error message as above, same for make build-unslung, and make unslung-image for that matter). I've got as far as downloading the Makefile and it's not clear what target to make *first*. Sep 20 14:44:01 make setup Sep 20 14:44:03 ah. "make update" seems to be doing the biz. Sep 20 14:44:48 NAiL - thanks. I'll try that next, when it's finished downloading Sep 20 14:44:59 make update runs make setup iirc Sep 20 14:45:03 stripwax: strange, I designed the Makefile to build everything required starting from just "make unslung" Sep 20 14:45:13 rwhitby - doesn't work. Sep 20 14:45:24 Hmm - I'll have to take a look at that then Sep 20 14:45:26 NAiL - thanks Sep 20 14:45:35 So setup, update, unslung is what you want Sep 20 14:45:42 and update should run setup Sep 20 14:45:47 rwhitby - thanks Sep 20 14:57:05 hrm - "monotone: connecting to monotone.nslu2-linux.org Sep 20 14:57:06 monotone: network error: failed to connect: Connection refused Sep 20 14:57:06 make: *** [setup-master] Error 1 Sep 20 14:57:06 " Sep 20 14:57:20 monotone still down? Sep 20 14:58:27 looks like it Sep 20 14:58:45 hmm, I'm unable to fix right now Sep 20 14:59:00 rwhitby: ping? Sep 20 15:00:43 pong Sep 20 15:00:57 could you look at monotone? Sep 20 15:01:14 done Sep 20 15:01:22 thx :) Sep 20 15:01:27 stripwax: retry make update ;) Sep 20 15:01:27 thanks Sep 20 15:01:32 am Sep 20 15:02:48 wonder why the cron job didn't restart it ... Sep 20 15:06:08 should it race through this bit or not? last output was "monotone: 16.0k | 8117 | 21 | 4 | 0" and that was several minutes ago now Sep 20 15:06:16 hehe Sep 20 15:06:22 that'll take a little while Sep 20 15:06:41 ok - yeah, monotone sucking cpu so it must be doing something Sep 20 15:07:57 03ph5 07org.openembedded.dev * rd98bddca... 10/packages/linux/ (2 files in 2 dirs): handhelds-pxa-2.6: fix EXTRAVERSION of 2.6.12-hh3 Sep 20 15:07:59 03ph5 07org.openembedded.dev * r8a43dee8... 10/packages/linux/handhelds-pxa-2.6_2.6.12-hh3.bb: handhelds-pxa-2.6_2.6.12-hh3.bb: bump PR to r1 Sep 20 15:08:02 03rpurdie 07org.openembedded.dev * r17725bba... 10/packages/pcmciautils/pcmciautils_010.bb: pcmciautils: Switch to udev instead of hotplug. Sep 20 15:08:04 03rpurdie 07org.openembedded.dev * rb40c2aec... 10/packages/udev/ (20 files in 6 dirs): Sep 20 15:08:04 udev: Rearrange the udev files into a more logical structure. Create copies of Sep 20 15:08:04 the rules we're using rather than the distributed version due to links to Sep 20 15:08:04 external scripts. Add mount.sh which attempts to mount block devices using Sep 20 15:08:04 pmount if available, falling back to mount if not. Fix the --mode switch to Sep 20 15:08:06 mknod in the init script. Sep 20 15:08:08 03rpurdie 07org.openembedded.dev * rf9869b4d... 10/packages/udev/ (udev_058.bb udev_063.bb udev_065.bb udev_070.bb): udev: Bump PRs Sep 20 15:08:11 03rpurdie 07org.openembedded.dev * r8e1ce5fd... 10/packages/pcmciautils/pcmciautils_010.bb: pcmciautils: Bump PR Sep 20 15:08:13 Is it unwise for me to be attempting a native build, by the way? :-) Sep 20 15:08:14 03rpurdie 07org.openembedded.dev * r7c097245... 10/packages/udev/ (6 files in 2 dirs): udev: Add the devfs udev 'comaptiblity' rules to OE as it changes names in different versions of udev. We *really* don't want to be using these. Sep 20 15:08:18 03koen 07org.openembedded.dev * rbd09b615... 10/packages/tslib/ (tslib/h3900/tslib.sh tslib_cvs.bb): Sep 20 15:08:21 packages/tslib/tslib/: remove h2200 linearizer, courtesy Matthew Reimer Sep 20 15:08:23 closes #336 Sep 20 15:08:32 oh, you're doing it natively? Sep 20 15:08:49 that might not work. Dunno if anything unslung compiles natively. Sep 20 15:09:25 packages do, no? i'll see how it goes anyway Sep 20 15:11:04 stripwax: you need to build cross Sep 20 15:11:46 rwhitby - ah, crap. ok. was kinda hoping building audio.o would be quick and easy enough to just do natively Sep 20 15:12:10 unless someone could quickly build audio.o for me? otherwise it'll take me a while to set up a cross-compile system Sep 20 15:12:27 stripwax - is that just the audio kernel module? Sep 20 15:12:36 rwhitby yes Sep 20 15:12:57 like the one in here: http://ipkg.nslu2-linux.org/feeds/unslung/oe/ ? Sep 20 15:14:23 rwhitby - if that is supposed to work on unslung, then the wiki needs updating because it said that that was for openslug users only... and that unslung users need to build Sep 20 15:14:51 that's definitely an unslung kernel modules feed Sep 20 15:15:00 there is a separate feed for openslug Sep 20 15:15:13 Also, unless I'm missing something, after installing kernel-module-audio and doing "insmod audio", I get an error: Sep 20 15:15:19 # insmod audio Sep 20 15:15:19 Using /lib/modules/2.4.22-xfs/kernel/drivers/usb/audio.o Sep 20 15:15:19 insmod: unresolved symbol unregister_sound_dsp Sep 20 15:15:19 insmod: unresolved symbol register_sound_dsp Sep 20 15:15:19 insmod: unresolved symbol register_sound_mixer Sep 20 15:15:20 insmod: unresolved symbol unregister_sound_mixer Sep 20 15:15:43 you need to install some other modules first Sep 20 15:16:21 maybe soundcore? Sep 20 15:16:23 Oh, I just assumed the module was incorrect (as indicated here http://www.nslu2-linux.org/wiki/HowTo/SlugAsAudioPlayer). Surprised the dependent modules aren't installed automatically by ipkg! which other modules pls? Sep 20 15:16:50 yeah, ipkg doesn't handle kernel module dependencies Sep 20 15:17:05 ahh.. the penny drops. thanks rwhitby - I didn't know that! Sep 20 15:17:26 Please correct the wiki, cause it's wrong. Sep 20 15:17:36 once I've got it working, I will, you can be sure of that Sep 20 15:17:42 thanks. Sep 20 15:20:47 [g2]: thanks for posting the loft patches Sep 20 15:21:00 <[g2]> rwhitby, np Sep 20 15:21:26 <[g2]> it turns out there are two or three little files that only one of us need to push upstream Sep 20 15:22:57 ok, more dumb questions. really sorry about this. how do I get hold of ncurses.h? there doesn't seem to be an 'ncurses-dev' package or similar Sep 20 15:34:01 stripwax: no-one has attempted building unslung natively before. Only optware packages and some openslug packages. Sep 20 15:34:30 optware stages ncurses.h so native optware builds can use it Sep 20 15:35:18 rwhitby - I'm no longer building unslung. I'm trying to compile an mpd client app that uses ncurses. was wondering if anyone had ever used the native toolchain to compile any curses app natively Sep 20 15:35:41 dunno - but it should work in optware Sep 20 15:36:16 03rpurdie 07org.openembedded.dev * rb0918c0a... 10/packages/pcmciautils/pcmciautils_010.bb: Sep 20 15:36:16 pcmciautils: Install the binaries into the correct place. Remove coldplug Sep 20 15:36:16 references as they now just create errors. Add dependency on module-init-tools - Sep 20 15:36:16 when busybox's modprobe expands to cover pcmcia alias handling, this can be Sep 20 15:36:16 removed. Sep 20 15:36:18 03rpurdie 07org.openembedded.dev * r0505ee47... 10/packages/module-init-tools/ (5 files in 2 dirs): module-init-tools: Upgrade to 3.2pre9. Change the update-alternatives priority so the genuine tool takes priority over busybox. This gets pcmciatils working properly. Sep 20 15:59:13 where is this on the nslu?: /usr/local/include Sep 20 16:19:30 xcodex: depends whether you are talking optware, unslung, openslug, or opendebianslug .... Sep 20 16:19:37 * rwhitby-away leaves for work Sep 20 16:25:30 unslung Sep 20 16:25:57 I am trying to build a USB driver but it fails Sep 20 16:26:32 I'll get lots of "error: dereferencing pointer to incomplete type" Sep 20 16:26:40 whenn gcc Sep 20 16:27:08 so, there are errors in the code... Sep 20 16:27:57 I must use this command: gcc -o usbdevice_details testlibusb.c -I/usr/local/include -L. -lnsl -lm -lc -L/usr/local/lib -lusb Sep 20 16:28:09 but dont know the right paths Sep 20 16:28:57 you can find testlibusb.c here: http://www.linuxforums.org/tutorials/1/tutorial-21493.html Sep 20 16:32:11 I got this damn error in every line of the code Sep 20 16:32:39 I tested a few codes, and everytime I got this error Sep 20 16:34:22 How can I do this: "You need to add the search path for the usb.h header file aswell" Sep 20 16:34:29 ? Sep 20 16:34:48 has it to do with this? Sep 20 16:37:34 yvasilev: ping Sep 20 16:37:43 I don't know. but hey, it really seems you should read a book about c programming ;) Sep 20 16:46:30 [g2]: ping Sep 20 16:49:32 dont have the time Sep 20 16:49:45 i can c basicly Sep 20 16:50:49 but i have problems with gcc and the include paths Sep 20 16:51:14 xcodex - you want a manual for gcc? Sep 20 16:52:00 on unslung, we don't use /usr/local for *anything*. try /opt/include instead of /usr/local/include and /opt/lib instead of /usr/local/lib Sep 20 16:55:05 i'll try it Sep 20 16:55:30 manual is here, already Sep 20 16:56:52 rwhitby: ping Sep 20 16:57:30 Gosh! I nearly got mpd + kernel-module-audio working on unslung! Sep 20 16:57:58 pity it's nailing at 100% cpu :-( Sep 20 16:58:26 stripwax: dmesg says what? Sep 20 16:58:34 anything in particular? :) Sep 20 17:00:57 what means an empty "#include" in source codes? Sep 20 17:01:04 eg: Sep 20 17:01:06 #include Sep 20 17:01:07 #include Sep 20 17:01:11 dmesg doesn't say anything exciting at this point.. unless you meant a couple of hours ago, when it printed "audio.c: v1.0.0:USB Audio Class driver" which I thought was pretty exciting Sep 20 17:01:44 kangaroo, PM Sep 20 17:02:15 xcodex where are you seeing that? Sep 20 17:02:38 here: http://www.linuxforums.org/tutorials/1/tutorial-21493.html Sep 20 17:03:24 the error makes me crazy Sep 20 17:03:49 everything what has to do with libusb Sep 20 17:04:08 xcodex: it means: the tutorial is b0rken! ;) Sep 20 17:04:32 yep. that C code is wrong Sep 20 17:04:38 but every tutorial Ive tested calls the same error Sep 20 17:04:49 perhaps the <...> part got lost because of html somewhere Sep 20 17:05:06 xcodex - I don't know what you mean by "calls the same error". how do you call an error? Sep 20 17:05:25 could it be that there must a #include ? Sep 20 17:05:35 seems likely Sep 20 17:06:02 but whats the second include? Sep 20 17:06:26 only the author of that code can tell you that. dunno, ?> Sep 20 17:09:36 ok, outta here. i can't build natively for shit. seey'all soon Sep 20 17:22:34 long time, no see geojeff Sep 20 17:31:39 hi ByronT Sep 20 17:31:54 yep, school started and i got sidetracked on various things Sep 20 17:32:04 I have heard of some success with the Griffin iMic and openslug Sep 20 17:32:35 cool. i forgot about the Soundblaster external i bought. probably too late to take it back :) Sep 20 17:32:51 oh, well, it might work too. Sep 20 17:32:57 that it might Sep 20 17:34:43 i still have to get my dev environment set up for openslug, having done so already for unslung Sep 20 17:35:44 someone got an ext soundblaster thingy to work under openslug earlier Sep 20 17:36:33 cool Sep 20 17:46:00 anyone knows who is the maintainer of the ixp4xx arch? Sep 20 17:46:51 depaak Sep 20 17:47:18 NAiLZzZ: thx. can I get him on IRC? Sep 20 17:47:44 no idea Sep 20 17:47:46 Deepak Saxena - dsaxena@plexity.net - http://www.plexity.net Sep 20 17:47:52 ~deepak Sep 20 17:47:57 ~seen deepak Sep 20 17:48:04 deepak was last seen on IRC in channel #debian, 35d 4h 4m 42s ago, saying: 'ok..'. Sep 20 17:48:23 35days... hope it has a different nick :) Sep 20 17:51:06 dwery: he/she posts regularly to LAK Sep 20 17:51:19 ok, i'll use email then Sep 20 17:51:27 dwery: lennert and [g2] have had contact with deepak in the past Sep 20 17:51:49 i would like to discuss about the timer thing.. Sep 20 17:52:20 <[g2]> lennert and beewoolie post often on l-a-k Sep 20 17:52:26 <[g2]> linux arm kernel ml Sep 20 17:52:48 as do jbowler Sep 20 17:52:48 having a FREQ defined for the NSLU2 at config time is not appropriate. do do it at running time we have two choiches: Sep 20 17:52:55 a) duplicate code Sep 20 17:53:07 b) change the arch a little bit Sep 20 17:53:22 this is re the 66.00MHz bit, innit? Sep 20 17:53:26 yes Sep 20 17:54:51 duplicate code if MACH=nslu2? Sep 20 17:55:54 yes Sep 20 17:56:13 the code woul be exactly the same as generix ixp4xx minus a variable. Sep 20 17:57:02 I think i'll propose the arch change. If rejected, i'll duplicate. Sep 20 17:57:25 hmm.. how will you avoid the arch change affecting other ixp-based platforms? Sep 20 17:57:30 (like my ds101) Sep 20 17:58:27 or [g2]'s loft Sep 20 17:58:51 basicaly this variable will be put in a struct inizialized by default with the value common to all archs. Sep 20 17:59:17 so no arch is affected. Sep 20 17:59:40 unless it's MACH nslu2 Sep 20 17:59:50 nslu2 arch will alter this variable in his init code Sep 20 17:59:53 at runtime Sep 20 18:01:40 whaddya think? Sep 20 18:02:04 send a mail to linux arm kernel Sep 20 18:02:08 :) Sep 20 18:02:22 alt. b sounds best to me though Sep 20 18:02:29 <[g2]> dwery btw did you see the loft kernel posts ? Sep 20 18:02:35 <[g2]> kernel patch Sep 20 18:02:44 [g2]: i only gave a quick look Sep 20 18:02:56 [g2]: quite similar to nslu2, as expected Sep 20 18:04:20 <[g2]> nod. Sep 20 18:05:58 I guess I should look at that too, and see if I can make things work on the ds101 Sep 20 18:06:44 the GPIOs are different.. except that, I guess most is the same Sep 20 18:08:10 going to bed.. bye! Sep 20 18:17:55 03byrontodd * 10unslung/make/libogg.mk: libogg.mk - Corrected download site to current one. Sep 20 18:19:45 <[g2]> dwery-zzz, sweet dreams Sep 20 18:32:08 03byrontodd * 10unslung/make/libvorbis.mk: libvorbis.mk: Corrected download site to current one. Sep 20 19:06:02 03byrontodd * 10unslung/make/ntpclient.mk: ntpclient.mk : Correct download site. Sep 20 19:22:55 03byrontodd * 10unslung/make/sudo.mk: sudo.mk : Corrected download site. Sep 20 19:38:59 somebody here? Sep 20 19:39:34 need to find ld.so.conf Sep 20 19:40:13 that's in /etc, normally Sep 20 19:40:47 and a quick google would have told you that... Sep 20 19:41:04 yes but on NSLU it isnt Sep 20 19:41:34 i've searched the whole nslu Sep 20 19:41:49 NSLU2 is using uClibc instead of glibc Sep 20 19:41:52 I need to specifiy the lib path Sep 20 19:42:21 ld.so.conf is only used at runtime, not for compilation Sep 20 19:42:36 yes i know Sep 20 19:42:46 i have compiled something Sep 20 19:42:54 it works without errors Sep 20 19:43:55 but when I start the compiled script it says: Sep 20 19:43:56 ./usbdevice_details: error while loading shared libraries: libusb-0.1.so.4: cannot open shared object file: No such file or directory Sep 20 19:44:19 but the libusb...-file is at /opt/lib Sep 20 19:44:43 dont know why it couldnt find it Sep 20 19:45:06 so, put it in /usr/lib or some other known location - or, set LD_RUN_PATH (or something like that). although i'm not sure it's used by uclibc... Sep 20 19:46:02 i think it's LD_LIBRARY_PATH Sep 20 19:46:02 bullet: uclibc? Sep 20 19:46:23 isnt it? Sep 20 19:46:36 only ucslugc is uclibc, IIRC Sep 20 19:46:45 ah, right. sorry. Sep 20 19:47:17 xcodex: did you ipkg install libusb, or did you install it manually? Sep 20 19:47:21 yes Sep 20 19:47:26 i installed it Sep 20 19:47:36 ipkg install libusb Sep 20 19:47:40 ok Sep 20 19:48:02 the files are installed in /opt/lib/libusb... Sep 20 19:48:15 and ldconfig is run? Sep 20 19:48:25 no idea Sep 20 19:48:31 never heared Sep 20 19:49:04 when should it run? Sep 20 19:49:22 after installing new libs Sep 20 19:50:18 no its not Sep 20 19:50:25 this command was not found Sep 20 19:50:41 i use now ipkg install ldconfig Sep 20 19:50:48 right? Sep 20 19:50:58 ldconfig is part of glibc Sep 20 19:51:14 but its missing on my slug Sep 20 19:51:43 ok now it works Sep 20 19:52:12 after I used ipkg install ldconfig Sep 20 19:52:15 thanks Sep 20 20:50:12 03bzhou * 10unslung/make/abook.mk: upstream upgrade from 0.5.3 to 0.5.4 Sep 20 20:59:37 03bzhou * 10unslung/make/ (bluez-libs.mk bluez-utils.mk): upstream upgrade to 2.21 Sep 20 21:00:41 03bzhou * 10unslung/make/bluez-hcidump.mk: upstream upgrade to 1.25 Sep 20 21:08:39 03bzhou * 10unslung/make/libgc.mk: upstream upgrade from 6.4 to 6.5 Sep 20 21:30:57 03bzhou * 10unslung/make/py-cheetah.mk: upstream upgrade from 0.9.17 to 0.9.18 Sep 20 21:45:14 03bzhou * 10unslung/make/py-serial.mk: upstream upgrade from 2.1 to 2.2 Sep 20 21:48:20 03bzhou * 10unslung/make/py-psycopg.mk: upstream upgrade from 1.1.18 to 1.1.20 Sep 20 22:06:10 03bzhou * 10unslung/make/sqlite.mk: upstream upgrade from 3.2.3 to 3.2.6 Sep 20 22:12:04 03bzhou * 10unslung/make/sqsh.mk: upgrade from 2.1 to 2.1.3 Sep 20 22:17:13 03bzhou * 10unslung/make/libtiff.mk: in case source tarball got moved to old/ Sep 20 22:26:13 03dyoung 07org.openembedded.dev * r8f6c3757... 10/packages/linux/openslug-kernel-2.6.12.2/defconfig: openslug-kernel: Add the missing USB Audio modules to the defconfig Sep 20 22:26:15 03dyoung 07org.openembedded.dev * r9bbf3b49... 10/packages/linux/nslu2-kernel_2.6.12.2.bb: openslug-kernel: adjust the package revision for added usb audio bits Sep 20 22:31:24 03bzhou * 10unslung/make/libtiff.mk: upstream upgrade from 3.7.3 to 3.7.4 Sep 21 02:11:06 03koen 07org.openembedded.dev * r59522ce9... 10/packages/abiword/abiword_2.3.99.bb: packages/abiword/abiword_2.3.99.bb: s/AbiSuite-2.2/AbiSuite-2.4/g to fix packaging Sep 21 02:19:33 a couple of news Sep 21 02:19:41 talked with deepak about the timex question Sep 21 02:19:47 and he agreed on the use of a variable Sep 21 02:19:58 i'm now doing the patch and submitting to him for review Sep 21 02:19:58 yeah Sep 21 02:20:02 nice Sep 21 02:20:09 He also asked about upstream push Sep 21 02:20:18 and I've told him we are working on it Sep 21 02:21:02 dwery: i had to add something similar in http://www.arm.linux.org.uk/developer/patches/viewpatch.php?id=2701/1 and he was fine with that Sep 21 02:21:14 dwery: just put it in the same cache line and you shouldn't really notice Sep 21 02:22:34 ok Sep 21 02:23:15 I need someone else with an NSLU2 to test it.. volunteers? Sep 21 02:26:27 03ph5 07org.openembedded.dev * r9cc938ac... 10/packages/x11/x11_6.2.1.bb: x11: add the XErrorDB/XKeysymDB fix from cvs to 6.2.1, too Sep 21 02:36:10 03rpurdie 07org.openembedded.dev * r1c546e4d... 10/packages/pcmciautils/pcmciautils_010.bb: pcmciautils: Correct DEPENDS Sep 21 02:36:12 03rpurdie 07org.openembedded.dev * r7d60c571... 10/packages/maemo/nokia770-init_1.0.bb: nokia770-init: Correct DEPENDS Sep 21 02:37:22 dwery: did you talk to deepak in a channel, or by email? Sep 21 02:37:25 email Sep 21 02:37:32 quick exchange Sep 21 02:37:39 for those who care, the patch is here: http://rafb.net/paste/results/VSEz3W72.html Sep 21 02:37:46 very small. Sep 21 02:38:51 cool Sep 21 02:39:06 where does ixp4xx_ticks_per_usec get modified? Sep 21 02:41:37 in the nslu2 init code Sep 21 02:41:43 which is not part of the patch Sep 21 02:42:08 that'll be a separate upstream push, right? Sep 21 02:42:12 yep Sep 21 02:46:07 is there a way to change the default ip/gw addresses in redboot, permanently? Sep 21 02:46:33 I guess you would have to patch the bootloader and reflash it. Sep 21 02:46:48 Not recommended unless you have confirmed JTAG access :-) Sep 21 02:48:00 :D I guess I'll keep it the way it is.. it's been years since I last used JTAG Sep 21 02:48:44 I ended up changing my lan addressing to suit the reset state of various devices on it ... Sep 21 02:49:01 :D Sep 21 02:49:13 the problem is the NSLU2 default gw is.. itself Sep 21 02:51:12 i've applied the patch and the kernel is running. Sep 21 02:51:23 How can I be sure it works correctly? :D Sep 21 02:54:51 dwery: you could put the patch in monotone, then lots of people will be testing it :-) Sep 21 02:55:10 :) Sep 21 02:55:27 As Linus says "Testing is for Users" Sep 21 02:56:46 rwhitby: you are an user, isn't it? please do an adjtimex -c on your slug :D Sep 21 02:57:41 <[cc]smart> dwery: is that another patch about the slug timekeeping issue ? Sep 21 02:57:54 yes. Sep 21 02:58:01 I need to be sure it works correctly. Sep 21 02:58:07 <[cc]smart> the "the more right way to do it" i guess Sep 21 02:58:41 <[cc]smart> i mean your patch compared the earlier approach Sep 21 02:59:46 [cc]smart: thanks :) **** ENDING LOGGING AT Wed Sep 21 02:59:56 2005