**** BEGIN LOGGING AT Sat Jun 24 02:59:56 2006
Jun 24 03:51:11 <CIA-5> 03bzhou * r3509 10optware/trunk/ (Makefile make/py-reportlab.mk): added py-reportlab
Jun 24 04:04:26 <rwhitby> Configuring network interfaces... ixp400: module license 'unspecified' taints k.
Jun 24 04:04:26 <rwhitby> ixp400_eth: no version for "ixQMgrInit" found: kernel tainted.
Jun 24 04:04:26 <rwhitby> ixp400_eth: Initializing IXP400 NPE Ethernet driver software v. 1.5
Jun 24 04:04:26 <rwhitby> ixp400_eth: CPU clock speed (approx) = 266 MHz
Jun 24 04:04:26 <rwhitby> ixp400.o:  ERROR, no Microcode found in memory
Jun 24 04:04:27 <rwhitby> ixp400_eth: Error starting NPE for Ethernet port 0!
Jun 24 04:04:42 <rwhitby> That's the 2.1.1 ixp modules.  Now I need to work out how to load the microcode.
Jun 24 04:04:54 <placid_> progress
Jun 24 04:12:02 <rwhitby> Configuring network interfaces... ixp400: module license 'unspecified' taints k.
Jun 24 04:12:02 <rwhitby> ixp400_eth: no version for "ixQMgrInit" found: kernel tainted.
Jun 24 04:12:02 <rwhitby> ixp400_eth: Initializing IXP400 NPE Ethernet driver software v. 1.5
Jun 24 04:12:02 <rwhitby> ixp400_eth: CPU clock speed (approx) = 266 MHz
Jun 24 04:12:02 <rwhitby> [error] ixNpeDlImageMgrImageFind: image library format may be invalid or corrupd
Jun 24 04:12:03 <rwhitby> ixp400_eth: Error starting NPE for Ethernet port 0!
Jun 24 04:12:04 <rwhitby> FATAL: Error inserting ixp400_eth (/lib/modules/2.6.17/kernel/drivers/net/ixp40d
Jun 24 04:12:07 <rwhitby> done.
Jun 24 04:12:51 <rwhitby> lets try again with word-wrap on
Jun 24 04:13:48 <rwhitby> Configuring network interfaces... ixp400: module license 'unspecified' taints ke
Jun 24 04:13:49 <rwhitby> rnel.
Jun 24 04:13:49 <rwhitby> ixp400_eth: no version for "ixQMgrInit" found: kernel tainted.
Jun 24 04:13:49 <rwhitby> ixp400_eth: Initializing IXP400 NPE Ethernet driver software v. 1.5
Jun 24 04:13:49 <rwhitby> ixp400_eth: CPU clock speed (approx) = 266 MHz
Jun 24 04:13:50 <rwhitby> [error] ixNpeDlImageMgrImageFind: image library format may be invalid or corrupt
Jun 24 04:13:53 <rwhitby> ed
Jun 24 04:13:55 <rwhitby> ixp400_eth: Error starting NPE for Ethernet port 0!
Jun 24 04:13:57 <rwhitby> FATAL: Error inserting ixp400_eth (/lib/modules/2.6.17/kernel/drivers/net/ixp400
Jun 24 04:13:58 <rwhitby> _eth.ko): Operation not permitted
Jun 24 04:14:01 <rwhitby> done.
Jun 24 04:14:58 <jacques> not permitted?
Jun 24 04:15:25 <jacques> ah I see, because it's "corrupted"
Jun 24 04:16:07 <rwhitby> at least it hasn't oopsed yet.  Better than I was expecting.
Jun 24 04:17:13 <rwhitby> md5sums match, so the microcode dat file is fine in the flash.
Jun 24 04:17:30 <rwhitby> iface eth0 inet static
Jun 24 04:17:30 <rwhitby>         pre-up modprobe ixp400
Jun 24 04:17:30 <rwhitby>         pre-up mknod /dev/ixNpe c 241 0
Jun 24 04:17:30 <rwhitby>         pre-up cat /lib/modules/2.6.17/kernel/drivers/ixp400/IxNpeMicrocode.dat
Jun 24 04:17:30 <rwhitby> > /dev/ixNpe
Jun 24 04:17:31 <rwhitby>         pre-up modprobe ixp400_eth
Jun 24 04:17:32 <rwhitby>         address 192.168.0.6
Jun 24 04:17:33 <jacques> well, looks like I can stop spending every spare minute working on the linkstation u-boot
Jun 24 04:17:34 <rwhitby>         netmask 255.255.252.0
Jun 24 04:17:36 <rwhitby>         gateway 192.168.0.254
Jun 24 04:17:44 <jacques> some guy in japan just kicked my butt
Jun 24 04:17:48 <rwhitby> why is that?
Jun 24 04:17:53 <rwhitby> heh
Jun 24 04:19:49 <rwhitby> I wonder if ixp400_eth needs to be compiled differently now with external microcode ...
Jun 24 05:14:03 <rwhitby> well, I have no idea where to start debugging this ethernet driver ...
Jun 24 05:26:45 <rwhitby>         /* Initialise and Start NPE */
Jun 24 05:26:45 <rwhitby>         if ((res = ixNpeDlNpeInitAndStart(default_npeImageId[portId])))
Jun 24 05:26:45 <rwhitby>         {
Jun 24 05:26:45 <rwhitby>             P_ERROR("Error starting NPE for Ethernet port %d!\n", portId);
Jun 24 05:26:45 <rwhitby>             return -1;
Jun 24 05:26:46 <rwhitby>  
Jun 24 05:27:28 <rwhitby> Guess I need to trace that call through the source ...
Jun 24 05:39:07 <rwhitby>         /* Locate image */
Jun 24 05:39:07 <rwhitby>         status = ixNpeDlImageMgrImageFind (imageLibrary, imageId,
Jun 24 05:39:07 <rwhitby>                                            &imageCodePtr, &imageSize);
Jun 24 05:39:07 <rwhitby>  
Jun 24 06:00:40 * rwhitby realises he should try debugging this big-endian first, just in case Intel wasn't lying about 2.1.1 not supporting little-endian
Jun 24 06:04:20 <jacques> heh
Jun 24 06:04:56 <rwhitby> I wonder if it's as simple as byte-swapping the microcode data or something ...
Jun 24 06:07:08 <rwhitby> In ixNpeDlImageConverter.c:             /* Convert to format suitable for reading as 32-bit unsigned int by big endian processor */
Jun 24 06:07:08 <rwhitby>  
Jun 24 06:07:29 <rwhitby> that's the program that creates the microcode .dat file
Jun 24 06:08:12 <rwhitby>             IxNpeMicrocode_bigEndianArray[x] = convertedValue;
Jun 24 06:08:12 <rwhitby>  
Jun 24 06:12:29 <rwhitby> Hmm - the microcode.dat file has lots of different bits of microcode in it - it might not even need to be as big as it is.
Jun 24 06:13:14 <gandhijee> yeah
Jun 24 06:13:30 <gandhijee> there is code for the ixp465 in there as well
Jun 24 06:14:09 * rwhitby flashes openslug-4.1-alpha.bin with big endian 2.1.1 ixp modules
Jun 24 06:14:44 <rwhitby> gandhijee: GentooSlugger?
Jun 24 06:15:43 <rwhitby> For those wishing to follow along at home, you need monotone head, and the following in your local.conf file:
Jun 24 06:15:54 <rwhitby> PREFERRED_VERSION_ixp4xx-kernel = "2.6.17"
Jun 24 06:15:54 <rwhitby> PREFERRED_VERSION_ixp4xx-csr = "2.1.1"
Jun 24 06:15:54 <rwhitby> PREFERRED_VERSION_ixp-osal = "2.1.1"
Jun 24 06:15:54 <rwhitby> PREFERRED_VERSION_ixp400-eth = "1.5"
Jun 24 06:16:09 <gandhijee> no, more like kernel.org
Jun 24 06:16:23 <gandhijee> w/ 1.5.1 drivers
Jun 24 06:16:36 <rwhitby> 1.5.1?
Jun 24 06:16:40 <gandhijee> yep
Jun 24 06:16:53 <gandhijee> for the ixp_eth
Jun 24 06:16:59 <gandhijee> on a ixp  465
Jun 24 06:17:18 <gandhijee> i am still hackin my way through it
Jun 24 06:21:44 <rwhitby> ah, good point.  I should be using 1.5.1 too
Jun 24 06:23:06 <rwhitby> gandhijee: is it much different to 1.5?
Jun 24 06:32:11 <rwhitby> Configuring network interfaces... ixp400: module license 'unspecified' taints ke
Jun 24 06:32:11 <rwhitby> rnel.
Jun 24 06:32:11 <rwhitby> ixp400_eth: no version for "ixQMgrInit" found: kernel tainted.
Jun 24 06:32:11 <rwhitby> ixp400_eth: Initializing IXP400 NPE Ethernet driver software v. 1.5
Jun 24 06:32:11 <rwhitby> ixp400_eth: CPU clock speed (approx) = 266 MHz
Jun 24 06:32:12 <rwhitby> [message] ixEthMiiPhyScan, Mii 1: Mii PHY ID 00008201
Jun 24 06:32:14 <rwhitby> ixp400_eth: Found PHY 0 at address 1
Jun 24 06:32:16 <rwhitby> ixp400_eth: ethernet 0 using NPEB and the PHY at address 1
Jun 24 06:32:18 <rwhitby> ixp400_eth: Use default MAC address 00:02:b3:01:01:01 for port 0
Jun 24 06:32:22 <rwhitby> ixp400_eth: MAC address now set to 00:04:5a:0f:b2:c6 for port 0
Jun 24 06:32:24 <rwhitby> ixp400_eth: Use default MAC address 00:02:b3:02:02:02 for port 1
Jun 24 06:32:26 <rwhitby> done.
Jun 24 06:34:23 <rwhitby> So the 2.1.1 csr with microcode loaded from a file works with the 1.5 ethernet
Jun 24 06:36:46 <rwhitby> So it may just be a matter of byteswapping the microcode to send it in ...
Jun 24 06:57:48 <jacques> jbot, test
Jun 24 06:57:49 <jbot> Oh, no! There's a test and I haven't studied!
Jun 24 08:19:40 <CIA-5> 03marceln * r3510 10optware/trunk/make/dovecot.mk: Upstream update to dovecot 1.0.beta9
Jun 24 08:21:37 <rwhitby> ixp400_eth_1.5.1 builds
Jun 24 08:25:43 * rwhitby flashes openslug-4.1.1-alpha.bin with ixp4xx-csr-2.1.1 and ixp400-eth-1.5.1
Jun 24 08:30:05 <rwhitby> pushing changes to monotone ...
Jun 24 08:40:35 <rwhitby> ixp400-eth-1.5.1 works
Jun 24 10:07:53 <rwhitby> Configuring network interfaces... ixp400: module license 'unspecified' taints ke
Jun 24 10:07:53 <rwhitby> rnel.
Jun 24 10:07:53 <rwhitby> ixp400_eth: no version for "ixQMgrInit" found: kernel tainted.
Jun 24 10:07:53 <rwhitby> ixp400_eth: Initializing IXP400 NPE Ethernet driver software v. 1.5.1
Jun 24 10:07:53 <rwhitby> ixp400_eth: CPU clock speed (approx) = 266 MHz
Jun 24 10:07:54 <rwhitby> [message] ixEthMiiPhyScan, Mii 1: Mii PHY ID 00008201
Jun 24 10:07:56 <rwhitby> ixp400_eth: Found PHY 0 at address 1
Jun 24 10:07:58 <rwhitby> ixp400_eth: ethernet 0 is using NPEB and the PHY at address 1
Jun 24 10:08:00 <rwhitby> ixp400_eth: Use default MAC address 00:02:b3:01:01:01 for port 0
Jun 24 10:08:02 <rwhitby> ixp400_eth: MAC address now set to 00:04:5a:0f:b2:c6 for port 0
Jun 24 10:08:04 <rwhitby> done.
Jun 24 10:08:06 <rwhitby> little-endian!
Jun 24 10:08:15 <rwhitby> pre-up devio '<<'/lib/modules/2.6.17/kernel/drivers/ixp400/IxNpeMicrocode.dat 'xp $,4' > /dev/ixNpe
Jun 24 10:09:48 <rwhitby> Just need to work out how to do that in the generation step now
Jun 24 10:12:57 <rwhitby> well, my work here is done for tonight.
Jun 24 12:42:41 <EvilDevil> hmm.. the URL for the GPL_ixp400LinuxEthernetDriverPatch-1_5_1.zip changed to ftp://aiedownload.intel.com/df-support/10159/eng/GPL_ixp400LinuxEthernetDriverPatch-1_5_1.zip (was ftp://aiedownload.intel.com/df-support/9519/eng/GPL_ixp400LinuxEthernetDriverPatch-1_5_1.zip from ixp400-eth_1.5.1.bb) Maybe someone could update this
Jun 24 12:52:23 <malveaux> Hi, has anyone gotten subversion working properly on an openslug?
Jun 24 17:00:03 <CIA-5> 03mwester * r347 10kernel/trunk/patches/2.6.17/76-dsmg600-pwrbtn.patch: DSM-G600 Power Button functionality
Jun 24 17:04:06 <CIA-5> 03mwester * r348 10kernel/trunk/patches/2.6.17/75-dsmg600.patch: DSM-G600 patches updated for .17 kernel, reset button support added and limited LED support
Jun 24 17:05:41 <CIA-5> 03mwester * r349 10kernel/trunk/patches/2.6.17/series: DSM-G600 patch #1 support added
Jun 24 17:45:22 <welbo> NAiL - wanted to say thanks for the work on the debianslug image.  Installation was flawless and system runs great.
Jun 24 17:52:35 * Harvy is back
Jun 24 19:27:47 <gandhijee> rwhitby: you still here?
Jun 24 21:49:01 <rwhitby> gandhijee: back now
Jun 25 00:41:35 * Harvy is away: sleaping but will check tomorrow
Jun 25 01:47:48 <rwhitby> Hmm - looks like the 2.1.1 patch isn't being applied to ixp4xx-csr_2.1.1 (due to OE removing it after applying it).
**** ENDING LOGGING AT Sun Jun 25 02:59:56 2006