**** BEGIN LOGGING AT Thu Oct 19 02:59:57 2006 Oct 19 06:41:37 03blaster8 * r495 10kernel/trunk/patches/2.6.18/ (5 files): Move Redboot Patches to submit upstream Oct 19 06:55:32 03blaster8 * r496 10kernel/trunk/patches/2.6.18/ (7 files): Major cleanup and refresh of Loft patches Oct 19 07:13:24 03blaster8 * r497 10kernel/trunk/patches/2.6.19/ (3 files): Add 2.6.19 redboot patches Oct 19 08:23:15 03blaster8 * r498 10kernel/trunk/ (README STATUS patches/2.6.18/defconfig): Update defconfig, docs Oct 19 08:26:17 [g2]: Oct 19 08:26:28 libtorrent/rtorrent finally finished building on my slug Oct 19 08:27:02 so, i'm resigning my recent foray into x-compilers into history. Oct 19 08:27:16 long live the slug! Oct 19 08:45:22 <[g2]> barnseenio congrats! Oct 19 08:45:59 <[g2]> barnseenio I'm up early today Oct 19 08:47:04 [g2]: I need you to sign off the 01-mtdpart patch and add a description and sign off 02-redboot patch Oct 19 08:47:16 they are in the 2.6.19 directory (not 2.6.18) Oct 19 08:47:23 Thanks for your work on the SVN repo Oct 19 08:49:12 <[g2]> blaster8 Ok I'll take a look at that Oct 19 08:54:23 dwery & [g2]: http://pastebin.ca/209699 Oct 19 08:54:30 One last thing from me Oct 19 08:54:40 I'm not moving the loft config patch to 2.6.19 Oct 19 08:55:28 We need to integrate the libata CF patch, then as the ixdp425-setup.c suggests, produce a separate avila setup file (with Loft support) Oct 19 08:55:56 That should go into 2.6.19, then go upstream Oct 19 08:56:10 ok :) Oct 19 09:01:02 <[g2]> blaster8 re: the CF patch, sounds good Oct 19 09:01:54 blaster8: I need to convert the libata cf to the driver model.. I thought I did did.. but maybe was the ep93xx libata pata driver Oct 19 09:02:17 I saw a libata ixp4xx driver somewhere Oct 19 09:02:45 http://trac.nslu2-linux.org/kernel/browser/archived/intel_eth/patches/2.6.18/96-pata-ixp4xx.patch Oct 19 09:03:51 yep, that's mine :) Oct 19 09:04:10 so, less work? - just need to move that file to drivers/ata and check it works ?? Oct 19 09:04:44 yes. The last time I used that drivers was months ago on the Loft Oct 19 09:04:49 I need to check if it still works Oct 19 09:05:36 ok, we should be able to get loft support in piggy-backed on the CF card updates :) Oct 19 09:06:46 :) Oct 19 09:07:37 i don't know how is the cf connected on the Avila but I suspect the same way as the loft Oct 19 09:07:50 ask [g2] - I think the 2 boards are very similar Oct 19 09:08:49 <[g2]> the Loft is a custom Avila board Oct 19 09:09:13 so the cf should be connected the same way? Oct 19 09:20:45 OK, Christian got back to me Oct 19 09:20:58 He says we should switch to data coherent mode for increased performance Oct 19 09:23:28 doh! :) Oct 19 10:10:25 03blaster8 * r499 10kernel/trunk/ (4 files in 2 dirs): Add ad741x hwmon chip support to 2.6.18 Oct 19 10:19:21 03blaster8 * r500 10kernel/trunk/Makefile: Revert Makefile Oct 19 13:05:51 03azummo * r501 10kernel/trunk/patches/2.6.18/13-hwmon-ad7418.patch: AD7417/8 driver Oct 19 14:57:06 Where can I find the most recent version of the new network driver? The version with the MAC address patch still hasn't hit the main repository. Oct 19 14:58:16 the latest we are using is the one in the svn repo Oct 19 14:58:18 .19 dir Oct 19 15:00:36 .19 or .0.2.0? Oct 19 15:01:32 Unforturnately I have 2 slugs and the MAC Address seemes fixed to 00:0C:BE:00:01:01. My network does not seem to approve of the twins. Oct 19 15:02:14 I was told that 0.2.1 fixed that. Oct 19 15:15:26 joshin: directory 2.6.19, driver 0.2.1 . it has a mac randomizer. Oct 19 15:15:34 anyway you can change the mac via ifconfig Oct 19 15:16:10 dwery, I swear I have tried with .0.2.0 and it hasn't worked. It is driving me batty. Oct 19 15:21:46 Aha, I have to down the interface before I can change it. So "ifconfig eth0 down && ifconfig eth0 hw ether 00:0C:BE:00:01:0a && ifconfig eth0 up" works Oct 19 15:22:25 yep Oct 19 15:30:07 hi fooks Oct 19 15:33:27 howdy likewise Oct 19 15:36:47 joshin: hi Oct 19 16:06:42 03blaster8 * r502 10kernel/trunk/ (9 files in 3 dirs): Add AD7418 support to 2.6.18/2.6.19 Oct 19 16:08:30 03blaster8 * r503 10kernel/trunk/Makefile: Revert Makefile again Oct 19 20:13:58 morning Oct 19 20:14:44 i'd like to know if some1 allready found an alternative to apex for enabling memory mod on the slug Oct 19 20:14:45 ?? Oct 19 20:18:50 <[g2]> Szafran what's the issue ? Oct 19 20:19:03 <[g2]> The bootloader needs to configure memory properly Oct 19 20:19:15 <[g2]> which means you'll need an updated bootloader Oct 19 20:19:32 <[g2]> APEX probably has the most air time on the nslu2 Oct 19 20:19:57 <[g2]> and you should definitely be JTAG ready if you are adding memory Oct 19 20:20:11 i allready have the mem mod Oct 19 20:20:12 <[g2]> you could even use the JTAG for testing for shorts Oct 19 20:20:25 i don't have the ability to configure apex Oct 19 20:20:42 <[g2]> Szafran I don't understand what you mean Oct 19 20:20:53 i've tried 3 different perspectives for a serial mod and none worked Oct 19 20:21:10 <[g2]> so you don't have serial ? Oct 19 20:21:14 nope Oct 19 20:21:31 <[g2]> phone call Oct 19 20:38:23 the thing with the serial mod is that each time there was only one way connection - i couldn't write anything to the slug - i've tried a module from a siemens phone cable, a module with max3232 compatible and finally i bought a module with the original max3232 - all the same - only one way connection - wasted over a month for test and stuff - checked the connections betwen cable and ixp processor and everything looked ok. i don't have a Oct 19 20:44:47 did you get the level converter? Oct 19 20:47:25 Szafran, do you get a plus sign right at power on? i do, fwiw. Oct 19 20:47:35 yep Oct 19 20:47:54 ah, but it's input that's not working. Oct 19 20:48:02 yes Oct 19 20:48:05 wierd. Oct 19 20:48:11 i know Oct 19 20:48:18 3x weird Oct 19 20:48:18 you are using a level converter and all that i assume.... Oct 19 20:48:26 yes Oct 19 20:49:15 also assume you've tried reversing the input pair of leads. Oct 19 20:49:29 you mean rx/tx ?? Oct 19 20:49:37 if so ... then yes Oct 19 20:49:50 no, the tx+/tx- order Oct 19 20:49:59 also... yes Oct 19 20:50:09 we know the rx works (-; Oct 19 20:50:16 ;) Oct 19 20:50:25 what kind of DTE are you using, minicom or something? Oct 19 20:50:42 DTE ? Oct 19 20:51:05 oh, old timer expression : data terminal equipment. in serial communications such terms are important Oct 19 20:51:17 just think terminal. Oct 19 20:51:32 or a x86 serial port Oct 19 20:51:48 i'm using soft on my pc Oct 19 20:51:56 if that's what u mean Oct 19 20:52:11 setting the parity and word size correct to 8/n/1? Oct 19 20:52:32 tried a few proggys on windows and linu systems with different settings... nothing helped Oct 19 20:52:39 *linux Oct 19 20:52:41 okay. Oct 19 20:53:26 well - it looks like a hardware problem. you could try another slug... or another level converter. Oct 19 20:53:27 sorry Oct 19 20:54:16 i've allready delpeted the money resources for this, so i need to find an alternative now Oct 19 20:54:25 *depleted Oct 19 20:55:28 that's a bummer. tried a different serial cable? Oct 19 20:55:35 is there a way to make the kernel to use the command line in the /boot/config* file instead of using the command line passed to it by the boot loader Oct 19 20:55:49 yep... different cables, different chips Oct 19 20:56:04 bad luck. Oct 19 20:58:55 any1, anything ? i'm desperate - 32mb is not enough, sure could use the power of the 128mb that is allready on the board :/ Oct 19 21:01:11 Szafran: its APEX on the board? Oct 19 21:01:33 Szafran: or RedBoot? Oct 19 21:01:57 redboot Oct 19 21:02:15 flashing apex without serial makes no sense Oct 19 21:02:40 you can manually set the SDRAM controller register if you get into the command prompt Oct 19 21:03:04 if the default Redboot is installed with has memory modify function Oct 19 21:03:06 s Oct 19 21:03:38 how ? and would it be done permanently or do i have to do it every time i boot the slug ? Oct 19 21:04:35 i have debianslug beta 2 from 2006.03.11 flashed Oct 19 21:08:48 likewise: can you tell me how can i do it ? or give me a link for site with description on how to do it ? Oct 19 21:09:01 03bzhou * r4256 10optware/trunk/ (4 files in 2 dirs): python24: upstream upgrade to 2.4.4 Oct 19 21:10:44 szafran: It's low level, you would use the redboot to modify IXP4xx registers. You need to read the IXP4xx docs on the SDRAM memory controller. "mm 0xcc000000 0x00000018" for example in RedBoot. Oct 19 21:11:14 szafran: but forget it, this is low-level stuff Oct 19 21:12:17 so if sthing goes wrong there's no way of coming back ? Oct 19 21:13:00 Szafran: of course, just reset and everything is fine Oct 19 21:13:33 oh... so the changes aren't permanent Oct 19 21:14:16 but if i'd find that it's working fine is there a way to make it permanent after doing some test ?? Oct 19 21:25:25 Szafran: yes, build a new bootloader Oct 19 21:26:06 szafran: or modify the Redboot fconfig script, that way you do not need to build the bootloader Oct 19 21:31:57 i don't know if that version of redboot is able to edit fconfig - as i remember i've tried that once a long time ago and it said that it does not recognize the command Oct 19 21:39:05 03bzhou * r4257 10optware/trunk/ (Makefile make/py-ply.mk): added and promoted py-ply Oct 19 22:14:46 <[g2]> bummer Szafran left Oct 19 22:15:05 <[g2]> Szafran double check the flow-control on the serial connection Oct 19 22:15:10 <[g2]> it needs to be off Oct 19 22:17:06 i'm not sure that will make any diff. as there are no hw flow lines connecting the devices. Oct 19 22:17:15 but you never know :-) Oct 20 02:08:07 dumfrac: hi Oct 20 02:08:24 rod: hi Oct 20 02:11:55 dumfrac: thanks for sending the slugimage microcode support bug report to the developers list. Oct 20 02:13:47 rod: I hope that it helps - I reviewed 02-redboot-partition.patch and I think that it will help Oct 20 02:14:17 i'm going to try an image without the NPE-B firmware, and then I'll be brave and try one with the NPE-B firmware included Oct 20 02:22:57 ok - a Debian kernel with 01-mtdpart-redboot-config-byteswap.patch and 02-redboot-partition.patch works - the MTD partitions are parsed correctly Oct 20 02:23:15 now to try with the NPE-B firmware included (which didn't work for me last night) Oct 20 02:25:07 I'm not sure that 02-redboot-partition.patch is a valid patch. Oct 20 02:25:20 (I need to check the redboot spec/code to see what it does) Oct 20 02:25:56 It's possible that the definition of the FIS dir allows you to disable an entry in the middle by writing FF (just like you erase a file in DOS) Oct 20 02:27:29 so it may work, but not for the right reasons... Oct 20 02:27:49 right Oct 20 02:28:07 And if that's the case, then it shouldn't be accepted upstream Oct 20 02:29:16 it does work - the MTD partitions are parsed correctly - I want to verify that the microcode got written to the right place Oct 20 02:29:24 http://ecos.sourceware.org/docs-latest/redboot/fis-delete-command.html Oct 20 02:29:37 I can imagine the fis delete command just writes FF Oct 20 02:31:11 can you check the source for redboot ? Oct 20 02:33:52 looking for that now Oct 20 02:45:16 rod: is the start of the microcode location in the flash image supposed to have a SERCOMM header ? if so, slugimage has written the microcode to the image correctly **** ENDING LOGGING AT Fri Oct 20 02:59:58 2006