**** BEGIN LOGGING AT Wed Feb 22 10:59:57 2006 Feb 22 11:23:15 for example? Feb 22 12:04:15 morgen ... ( oder so ) Feb 22 12:04:23 morning Feb 22 12:04:36 or morning, indeed ... Feb 22 12:05:02 * Angua has a question, concerning compiling additional modules for the NSLU2 debianslug image Feb 22 12:05:28 shoot Feb 22 12:05:47 actually about the Dallas one-wire modules. practically, ( i know how to compile a kernel ) i just enable the modules in the .config file Feb 22 12:06:14 now, i have already compiled the debianslug image completely, and it works like a charm Feb 22 12:06:37 the .config is in openembedded/packages/linux/ixp4xx-kernel/2.6.15/defconfig Feb 22 12:06:58 ah, which is, i think, exactly what i need Feb 22 12:07:18 so changing the .config won't do a thing, ( as i found out ;) Feb 22 12:07:43 if everything can be built as modules, please send the diff on the mailinglist :)' Feb 22 12:07:55 happily Feb 22 12:07:58 or, send the diff anyways Feb 22 12:08:24 as it is, it will be the perfect box for some home-automation related things Feb 22 12:08:44 since it draws no mentionable power ... Feb 22 12:08:49 It's rather versatile :) Feb 22 12:09:13 ( and i like these arm boxes, it is essentially my third, after the peplink manga, which does a lovely job as firewall/router and the nokia 770 ) Feb 22 12:09:22 ooh, you got a 770 Feb 22 12:09:25 * NAiL is envious Feb 22 12:09:45 it's nice Feb 22 12:10:26 I can't afford one right now :( Feb 22 12:10:48 (and even if I could, I'd have to order it from some other country. Nokia doesn't sell 'em here) Feb 22 12:10:52 (.no) Feb 22 12:15:14 they don't sell it in norway ? Feb 22 12:15:16 that's strange Feb 22 12:16:00 well, you could probably get it in sweden Feb 22 12:16:33 yeah Feb 22 12:16:42 ( with the help of a friend ;) Feb 22 12:17:22 i am looking forward to the new firmware for it Feb 22 12:17:45 but i am pleased, after getting familiar with scratchbox, it was not a problem getting vpnc to run on it Feb 22 12:17:54 I know a few people in .fi and .se, so I can probably get them to send me one when I can afford it Feb 22 12:17:58 yep Feb 22 12:19:15 well, it is almost a replacement for my iPAQ, the only thing i still use it for, since i have the nokia, is for GPS ... unfortunately that is an area where you can't really get maps etc yet Feb 22 12:21:22 but hopefully that will change as well, eventually anyway. until then, i will use it Feb 22 12:23:18 anyway, thanks, if it works well, i will get in touch again for the diff Feb 22 12:23:24 * Angua smiles and waves Feb 22 17:56:42 03azummo * 10kernel/2.6.16/ (94-loft-setup.patch 93-loft-maclist.patch 96-loft-leds.patch): fixed loft support Feb 22 22:19:38 03bzhou * 10unslung/make/ (py-ruledispatch.mk py-protocols.mk py-pastescript.mk): adding newer turbogears dependencies Feb 22 22:57:37 03azummo * 10kernel/2.6.16/series: loft leds bugfix Feb 22 23:01:31 03rwhitby * 10kernel/2.6.16/series: Removed non-existent patches. Feb 22 23:01:35 03bzhou * 10unslung/make/py-pastescript.mk: corrected svn site and setup.py install param Feb 23 01:08:07 HI all, I have just been able to build and install a V2.3R63-uNSLUng-6.5-alpha Feb 23 01:11:02 congratz Feb 23 01:12:01 * NAiL senses a "but" Feb 23 01:23:53 .oO(as long as NAiL does not smell like a "butt") Feb 23 01:24:55 eyh Feb 23 01:28:21 jk Feb 23 01:29:23 has anyone posted any 6.5 binaries up anywhere yet Feb 23 01:29:25 oeg:don't forget to sign the "guestbook" at http://groups.yahoo.com/group/nslu2-linux/database?method=reportRows&tbl=12 ---- once we get enough confirmed successful builds and installs we can promote 6.5 to beta status and make a binary available for download! Thanks! Feb 23 01:30:16 Yer I was going to do that, but I wanted to make sure there was nothing else you wanted done Feb 23 01:31:03 ouchy Feb 23 01:31:08 Feedback, if you encountered any problems or if it behaved in a manner you didnt' expect -- other than that, it's all yours! Feb 23 01:31:48 Note that 6.5 kernel modules are not in the feeds for the alpha release, so you'll have to get those off the machine you built on; other than that, most ipkgs should work. Feb 23 01:32:56 ok, where will they be in the build I suppose unslung/tmp/? Feb 23 01:40:02 unslung/tmp/deploy/ipkg if memory serves... Feb 23 01:56:35 oeg: sending an email to nslu2-linux@ with an installation report would be good. It gives the other users confidence in the release if it's a smooth installation for the alpha testers. Feb 23 01:57:02 mwester: the 6.5 kernel modules should be in the unstable feed. Feb 23 02:26:38 rwhitby: the report is in my words ? or is there a log I can get ? Feb 23 02:28:04 in your words. Just your impressions and observations. Feb 23 02:45:36 dwery: I changed 0x1fff to 0x1ffe too Feb 23 02:45:48 ack Feb 23 02:45:49 I assume I have to now, since I'm using readw Feb 23 02:45:59 yes Feb 23 02:47:57 compiling is so sloooow :( Feb 23 02:50:22 uploading kernel Feb 23 02:50:27 * NAiL crosses fingers Feb 23 02:54:28 oops, wrong probe addr Feb 23 02:57:46 INFTL: inftlcore.c $Revision: 1.19 $, inftlmount.c $Revision: 1.18 $ Feb 23 02:57:46 Using configured DiskOnChip probe address 0x50000000 Feb 23 02:57:53 doesn't look like it detected it Feb 23 02:59:00 are you LE or BE now?/ Feb 23 02:59:21 LE Feb 23 02:59:37 I'd try BE with printks in the detection routine Feb 23 03:02:55 hmm Feb 23 03:02:56 well Feb 23 03:03:09 it does get to the probing-for-mplus part atleast Feb 23 03:11:41 ok, uploading BE kernel Feb 23 03:14:02 well Feb 23 03:14:07 the BE kernel gets to the same point Feb 23 03:15:00 ie, it sees the ChipID as 0, and then tries to probe for an MPlus Feb 23 03:15:30 (which is what I expect it to do) Feb 23 03:15:31 mm.. retry your detection routine? Feb 23 03:16:00 I'm gonna see what it returns, adding a printk Feb 23 03:16:18 s/ a /another/ Feb 23 03:16:20 NAiL meant: I'm gonna see what it returns, addinganotherprintk Feb 23 03:16:26 heh Feb 23 03:18:46 hmm, the second probing returns 0 too :( Feb 23 03:20:09 trying with the other routines Feb 23 03:20:33 NAiL: you can confirm that on the ds101 the default latency is 0 ? Feb 23 03:20:44 on the artop? Feb 23 03:20:45 yes Feb 23 03:21:05 Doesn't hurt setting it to 0x90 though Feb 23 03:21:23 (ie, I apparently get ~2mb extra throughput) Feb 23 03:21:41 but that might be hdparm being crazy Feb 23 03:22:59 this is odd Feb 23 03:23:29 with the #defines that worked with my detection routine, I just get ChipID 0 Feb 23 03:24:59 hmm Feb 23 03:25:06 ChipID is u_char though Feb 23 03:27:31 mmm Feb 23 03:27:45 LE or BE now? Feb 23 03:28:20 still BE Feb 23 03:29:11 i'd try using swab16 around ReadDOC Feb 23 03:29:29 DoC is LE iirc Feb 23 03:30:06 #define ReadDOC_(adr, reg) swab16(readw((unsigned long)(adr + reg) ^ 0x02)) Feb 23 03:30:10 #define WriteDOC_(d, adr, reg) writew(swab16(d), (unsigned long)(adr + reg) ^ 0x02) Feb 23 03:30:13 Those then? Feb 23 03:30:21 DoC is LE, yes Feb 23 03:30:51 no, if you are compiling BE Feb 23 03:31:00 you should use the same routines that worked in your code Feb 23 03:31:05 yeah, I am Feb 23 03:31:10 but place a swab16 in ReadDOC only Feb 23 03:31:14 #define ReadDOC_(adr, reg) readw((void __iomem *)(adr) + (reg)) Feb 23 03:31:14 #define WriteDOC_(d, adr, reg) writew(d, (void __iomem *)(adr) + (reg)) Feb 23 03:31:16 oh Feb 23 03:31:27 but no ^2? Feb 23 03:31:33 exactly Feb 23 03:31:45 maybe you will need a swab16 in WriteDOc too, but we will check that later Feb 23 03:32:08 yup, compiling Feb 23 03:36:00 progress :) Feb 23 03:36:03 Using configured DiskOnChip probe address 0x50000000 Feb 23 03:36:03 ChipID: 41: Feb 23 03:36:03 Probing for MPlus Feb 23 03:36:03 ChipID: 41 Feb 23 03:36:05 DiskOnChip found at 0x50000000 Feb 23 03:36:08 No NAND device found!!! Feb 23 03:36:10 No valid DiskOnChip devices found Feb 23 03:36:17 it found the bugger! :-D Feb 23 03:36:35 so, now to figure out why it can't find a NAND device Feb 23 03:37:48 that's in nand_base.c Feb 23 03:38:07 now try a swab16 as in writew(swab16(d), .. to see if anything chnaged Feb 23 03:38:58 hmm, I don't think it will, but I can check Feb 23 03:39:54 It does a few writes in the detection routine, so with swab16 in the write, I think it'll fail Feb 23 03:40:01 right... Feb 23 03:40:06 now Feb 23 03:40:19 it appears that nand_base reads a byte... Feb 23 03:42:09 still "No NAND device found" Feb 23 03:42:19 i think it is time to go to bed... :) Feb 23 03:42:56 heh, yeah Feb 23 03:43:01 talk to you tomorrow ;) Feb 23 03:43:22 * mwester is wondering what NAiL is trying to connect to a slug... Feb 23 03:43:22 'night all Feb 23 03:43:48 great night anyway... yaaaawn.... zzzzzz Feb 23 03:44:06 mwester: I'm not trying to connect anything to a slug. I'm trying to get DiskOnChip working on the DS-101 :) Feb 23 03:44:11 nite, dwery-zzzz Feb 23 03:44:23 sleep well :) Feb 23 03:44:36 (I know I won't, too much progress) Feb 23 03:44:45 I'll have to check that out, I've not really looked at the DS-101 too much. Good luck! Feb 23 03:45:20 thanks. It's starting to look good though :-D Feb 23 03:48:05 ok, dwery-zzzz... when you get back. When I corrected the swab16 stuff, it apparently found the NAND chip Feb 23 03:49:08 now, if I could only get this to work in LE, I could access the logs Feb 23 03:49:41 remove the swabs and do the ^ 2 trick :) Feb 23 03:50:18 oh Feb 23 03:50:21 thanks Feb 23 03:50:39 getting loads of debug output tho Feb 23 03:50:40 nand_read_ecc: from = 0x00d8c000, len = 512 Feb 23 03:50:40 nand_read_ecc: from = 0x00d90000, len = 512 Feb 23 03:50:48 too much, actually ;) Feb 23 04:21:07 yay Feb 23 04:21:23 ok, the DoC is detected in LE, and I get a *heap* of debugging info Feb 23 04:48:26 dwery-zzzz: Ok, for when you get up: http://pastebin.com/568082 Feb 23 06:43:11 03bzhou * 10unslung/make/tethereal.mk: tethereal cross built Feb 23 06:58:10 03oleon * 10unslung/ (sources/torrent/torrent_watchdog make/torrent.mk): fix empty MAILTO string Feb 23 07:10:43 03bzhou * 10unslung/ (Makefile make/py-scgi.mk): py-scgi ready for testing Feb 23 07:12:59 jp30-work, please test and promote when you get a chance ... more to come Feb 23 07:38:25 03orepvik * 10kernel/2.6.16/97-ds101-buttons.patch: Fix powerled blinking on powerbutton press Feb 23 09:22:00 NAiL: I saw the debug Feb 23 09:22:09 yeah Feb 23 09:22:19 the two partitions there..? Feb 23 09:22:38 the last partition is ~804kb, and is filled with 0xFF Feb 23 09:22:47 (iirc, blank) Feb 23 09:23:14 there's actually three mtdblocks Feb 23 09:23:23 mtd0 is the entire flash (16777216 bytes) Feb 23 09:23:42 mtd1 is some subset, slightly smaller Feb 23 09:23:47 and mtd2 is 804kb Feb 23 09:24:28 two secs and I'll give you a bit more info Feb 23 09:24:57 ty Feb 23 09:25:27 take a look at http://pastebin.com/568254 for the entire bootlog Feb 23 09:27:51 NAiL: did you tried BE? Feb 23 09:28:05 same dmesg, IIRC Feb 23 09:28:19 I can't boot BE to a proper fs, so I can't see what happens in the end Feb 23 09:28:40 ok.. we need to find how the nand subsys communicates with the chips Feb 23 09:28:52 ? Feb 23 09:29:38 ah Feb 23 09:29:39 yes Feb 23 09:29:41 :) Feb 23 09:29:47 * NAiL slowly gets it Feb 23 09:30:24 now, there's gotta be *some* kind of partition info, since the tacked-on part of redboot reads two partitions (kernel and rd.gz) before booting Feb 23 09:35:03 NAiL: in doc2000_readbuf_dword there's a readl Feb 23 09:35:11 i don't think it gets triggered Feb 23 09:35:24 but you should stick a printk there as if it is, it will not work Feb 23 09:37:39 there'salso a readl in doc200x_ident_chip, don't think it is triggered but... Feb 23 09:38:04 hmm Feb 23 09:38:14 wouldn't nand_base.c be the place to look? Feb 23 09:39:24 yes.. in theory the nand subsystem should use pointer to functions provided by the doc driver to read them Feb 23 09:40:34 ah Feb 23 09:41:36 heh Feb 23 09:41:42 my kernel doesn't compile right now ;) Feb 23 09:41:53 ah, there we are Feb 23 09:42:21 I had mucked around to make the defines set properly (I hope) in doc2000.h Feb 23 10:13:59 [g2]: SCSI device sda: 1000944 512-byte hdwr sectors (512 MB) Feb 23 10:33:17 <[g2]> dwery CONGRATS! Feb 23 10:36:19 ty **** ENDING LOGGING AT Thu Feb 23 10:59:56 2006