**** BEGIN LOGGING AT Fri Aug 12 02:59:59 2011
Aug 12 03:37:48 <CIA-89> jow * r27964 /branches/backfire/package/block-extroot/ (Makefile files/extmount.sh): [backfire] fix block-extroot on brcm-2.4 (#9454)
Aug 12 06:42:07 <builder_> build #64 of pxcab is complete: Failure [failed compile_10]  Build details are at http://buildbot.openwrt.org:8010/builders/pxcab/builds/64
Aug 12 07:24:33 <builder_> build #56 of avr32 is complete: Failure [failed compile_4]  Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/56
Aug 12 14:36:35 <CIA-89> jow * r27965 /trunk/package/base-files/ (Makefile files/etc/profile): [package] base-files: fix arp() procedure to properly detect /usr/bin/arp
Aug 12 14:37:47 <CIA-89> jow * r27966 /branches/backfire/package/base-files/ (Makefile files/etc/profile): [backfire] backport r27965
Aug 12 14:49:37 <fullstop> I've made good progress with the WZR-HP-G300NH2
Aug 12 14:49:55 <fullstop> It boots and the lan works.  I'll test the wan in a moment.
Aug 12 14:49:57 <Chocky> \o/
Aug 12 14:50:15 <fullstop> gpios are all different though
Aug 12 14:50:21 <fullstop> so I have no LEDs.. :)
Aug 12 14:50:39 <Chocky> gotta have flashing lights
Aug 12 14:51:10 <fullstop> it is actually very similar to the ag300
Aug 12 14:52:49 <fullstop> boot log: http://pastebin.com/1qwxGbV2
Aug 12 14:52:50 <karlp> is there any software support for periodically blinking leds/gpios? rather than just turning them on/off yourself at various events?
Aug 12 14:53:02 <fullstop> karlp: yes
Aug 12 14:53:23 <fullstop> you can have them blink on timers, based on rx/tx etc..
Aug 12 15:30:11 <karlp> fullstop: got any pointers to some examples that blink on timers?
Aug 12 15:35:26 <builder_> build #60 of x86 is complete: Success [build successful]  Build details are at http://buildbot.openwrt.org:8010/builders/x86/builds/60
Aug 12 15:40:24 <fullstop> karlp: look in /sys/class/leds/
Aug 12 15:40:32 <fullstop> and see what you have in there.  You'll need that first.
Aug 12 15:40:58 <fullstop> There is a trigger "file" there.  echo "timer" into it
Aug 12 15:42:08 <fullstop> and then you have delay_on and delay_off files, which you can echo the delay values, in ms, there.
Aug 12 15:48:43 <karlp> right, might help if I include kmod-leds-gpio first then :)
Aug 12 15:49:00 <fullstop> that would be important, yes.  ;-)
Aug 12 15:52:17 <karlp> ooh kmod-ledtrig-morse :) sexy
Aug 12 15:52:55 <fullstop> there's actually quite a few led triggers.
Aug 12 16:33:15 <fullstop> anyone here familiar with ap94 ?
Aug 12 16:33:26 <fullstop> what is cal_data ?
Aug 12 16:34:13 <Chocky> moo
Aug 12 16:35:44 <fullstop> cal not cow
Aug 12 16:36:28 <karlp> calibration?
Aug 12 16:36:40 <fullstop> possibly
Aug 12 16:39:19 <CIA-89> hauke * r27967 /trunk/target/linux/brcm47xx/patches-3.0/0014-ssb-move-flash-to-chipcommon.patch:
Aug 12 16:39:19 <CIA-89> brcm47xx: fix flash detection on chipcommon less devices.
Aug 12 16:39:19 <CIA-89> This closes #9906
Aug 12 16:39:19 <CIA-89> Thank you for the report Kim
Aug 12 16:40:49 <karlp> so, I rebooted with kmod-dev-leds, but /sys/class/leds is empty, is there something other than just compiling in that module that is needed?
Aug 12 16:46:45 <fullstop> karlp: which board are you using?
Aug 12 17:06:04 <fullstop> I should be able to read the art partition in the board setup, but I must be doing something wrong.
Aug 12 17:07:13 <fullstop> "art" on my board is at 0x0050000.  hexdump /dev/mtd4 | grep 4c -> 0001200 0cb8 1470 e016 0201 0000 001f 4ce6 7640
Aug 12 17:08:15 <fullstop> 4c:e6:76 ... is my mac address, so the offset from 0x0050000 is 0x120C.  But I'm reading all 0x00 at boot time.
Aug 12 17:09:05 <fullstop> KSEG1ADDR is kind of magic to me.  Does it map a memory address to a flash address?
Aug 12 17:11:04 <nbd> using KSEG1 you get a virtual uncached address range for a particular physical address
Aug 12 17:11:31 <fullstop> nbd: Am I doing something incorrectly?
Aug 12 17:11:51 <fullstop> nbd: I'm following what I have read in other board definitions.
Aug 12 17:17:20 <CIA-89> nbd * r27968 /trunk/package/libubox/Makefile: libubox: update to fix a compile error
Aug 12 17:18:37 <nbd> when are you reading it?
Aug 12 17:18:52 <nbd> memory mapped access to the flash chip is only available early in the boot process
Aug 12 17:19:05 <nbd> later on the flash driver disables it and only talks spi with it directly
Aug 12 17:20:57 <nbd> btw. the offset is exactly the same as on wzr-hp-ag300h
Aug 12 17:21:00 <nbd> (for eeprom1
Aug 12 17:21:18 <nbd> the physical offset of the flash chip is 0x1f000000
Aug 12 17:48:58 <karlp> fullstop: it's an ar2317 board, not a standard router.  I can use gpioctl to turn on/off leds by hand though
Aug 12 17:51:18 <karlp> but, friday night here, time to get out of here
Aug 12 18:01:02 <fullstop> nbd: in wrzhpg300nh2_setup, very early on
Aug 12 18:02:00 <fullstop> nbd: are you certain about it being in the same location as the ag300h?  the "art" partition is in a different location on this board.
Aug 12 18:03:05 <fullstop> nbd: http://pastebin.com/JNmTSWiW is what I currently have for the mtd partitions.
Aug 12 18:03:29 <fullstop> and I can find the hardware address inside of that mtd partition.
Aug 12 18:04:41 <fullstop> nbd.. I understand what you wrote now.  I did not know of the physical offset.
Aug 12 18:09:14 <fullstop> nbd: that worked.  Thanks
Aug 12 18:49:46 <WillieNL> nbd: any idea wat is going wrong in https://dev.openwrt.org/browser/trunk/toolchain/gcc/patches/4.4.6/930-avr32_support.patch on x86 it make the build fail.
Aug 12 18:50:41 <WillieNL> regrename.c -> error: 'IFCVT_ALLOW_MODIFY_TEST_IN_INSN' undeclared (first use in this function)
Aug 12 18:55:03 <KanjiMonster> fullstop: sounds like you are making good progress :)
Aug 12 19:03:20 <fullstop> KanjiMonster: So far, yes.
Aug 12 19:03:43 <fullstop> KanjiMonster: I'm confused between AR71XX and AR92XX
Aug 12 19:03:58 <fullstop> I'm not seeing the wlan phy
Aug 12 19:04:17 <fullstop> which, I assume, is because I've not added it to the platform.  ;-)
Aug 12 19:04:35 <fullstop> It is an AR9280 wlan module
Aug 12 19:05:04 <KanjiMonster> fullstop: actually afaik the ath9k chip is connected through pcie, so it should be automatically picked up when scanning the pcie
Aug 12 19:05:25 <fullstop> but, calling ar9xxx_add_device_wmac crashes because the SOC doesn't match anything in the list and the kernel calls BUG()
Aug 12 19:05:52 <fullstop> KanjiMonster: interesting
Aug 12 19:07:03 <KanjiMonster> fullstop: that's expected, ar724x doesn't have integrated wireless, only ar913x and ar933x/ar934x have (afaik)
Aug 12 19:07:36 <fullstop> KanjiMonster: good to know.
Aug 12 19:08:43 <fullstop> Okay, so it should be found when it scans the pcie bus?
Aug 12 19:09:42 <KanjiMonster> fullstop: have you tried "pb9x_pcie_init();"? (-> https://dev.openwrt.org/browser/trunk/target/linux/ar71xx/files/arch/mips/ar71xx/mach-pb92.c#L102 )
Aug 12 19:09:59 <fullstop> no.  Let me give that a shot
Aug 12 19:10:41 <fullstop> I tried ap94_pci_init
Aug 12 19:16:15 <builder_> build #55 of ixp4xx is complete: Exception [exception interrupted]  Build details are at http://buildbot.openwrt.org:8010/builders/ixp4xx/builds/55
Aug 12 19:16:15 <builder_> build #50 of iop32x is complete: Exception [exception interrupted]  Build details are at http://buildbot.openwrt.org:8010/builders/iop32x/builds/50
Aug 12 19:16:17 <builder_> build #49 of etrax is complete: Exception [exception interrupted]  Build details are at http://buildbot.openwrt.org:8010/builders/etrax/builds/49
Aug 12 19:16:19 <builder_> build #48 of ep93xx is complete: Exception [exception interrupted]  Build details are at http://buildbot.openwrt.org:8010/builders/ep93xx/builds/48
Aug 12 19:17:43 <fullstop> KanjiMonster: made no difference
Aug 12 19:18:06 <fullstop> perhaps CONFIG_AR71XX_DEV_PB9X_PCI is not set
Aug 12 19:18:24 <KanjiMonster> fullstop: might be, but it ususally should be
Aug 12 19:18:34 <fullstop> It built the .o file, so it is set.
Aug 12 19:19:20 <KanjiMonster> fullstop: do you have a bootlog?
Aug 12 19:20:00 <fullstop> http://pastebin.com/SXEax6yu
Aug 12 19:20:21 <fullstop> Skipping PCI bus scan due to resource conflict
Aug 12 19:20:30 <fullstop> registering PCI controller with io_map_base unset
Aug 12 19:20:33 <fullstop> probably not good
Aug 12 19:21:48 <KanjiMonster> fullstop: stupid question, lsmod lists ath9k? (because I don't see any output from any ath module)
Aug 12 19:22:07 <fullstop> KanjiMonster: yes
Aug 12 19:22:10 <KanjiMonster> I'm not sure if it does any if doesn't find anything
Aug 12 19:22:12 <KanjiMonster> okay
Aug 12 19:22:39 <fullstop> http://pastebin.com/HzDqrZ6Y
Aug 12 19:25:07 <KanjiMonster> fullstop: did you remove the ap94_pci_init()?
Aug 12 19:26:13 <fullstop> KanjiMonster: I just did.  ;-)
Aug 12 19:26:16 <fullstop> Building again
Aug 12 19:27:07 <fullstop> I felt like I needed to pass along the wmac information to somebody, but perhaps not.
Aug 12 19:31:34 <KanjiMonster> fullstop: since ar9280 is a real pcie device, you only need to properly register the pcie bus, and it should be found automagically ;)
Aug 12 19:34:13 <fullstop> it appears the same, minus the message about the resource conflict.
Aug 12 19:34:34 <fullstop> http://pastebin.com/d9taPPAY
Aug 12 19:35:59 <fullstop> Let me add lspci to the image.
Aug 12 19:36:18 <CIA-89> nbd * r27969 /trunk/include/site/ (powerpc-openwrt-linux-gnuspe powerpc-openwrt-linux-uclibcspe): include/site: add powerpc-openwrt-linux-*spe sitefiles for mpc85xx
Aug 12 19:37:06 <KanjiMonster> fullstop: hm, google hits suggest "pci 0000:00:00.0: [168c:ff1c] type 0 class 0x000200" already is the wifi, so I'm wondering why ath9k doesn't see it
Aug 12 19:37:45 <fullstop> I'm not terribly familiar with how ath9k works, but I assume that there is a table of supported pci ids?
Aug 12 19:38:39 <nbd> :ff1c is a broken id that shows up if the device has no eeprom and the eeprom based fixup hasn't been run yet
Aug 12 19:38:48 <nbd> i mean the flash based fixup
Aug 12 19:39:23 <nbd> fullstop: it would probably be better if you do your porting on trunk and then backport to backfire afterwards
Aug 12 19:39:32 <fullstop> nbd: that is what I am doing
Aug 12 19:39:48 <nbd> ok
Aug 12 19:40:09 <fullstop> what  do you mean by eeprom fixup?
Aug 12 19:40:35 <nbd> so the wifi chip cannot access its eeprom (because it's not there)
Aug 12 19:40:50 <nbd> so the software has to read the eeprom data from the 'art' partition and apply the register writes that are in there
Aug 12 19:40:58 <fullstop> not there because of design or because I inadvertently blew it away?
Aug 12 19:41:03 <nbd> there's some stuff in the ar71xx code to take care of that
Aug 12 19:41:07 <fullstop> gotcha
Aug 12 19:41:14 <nbd> let me check for the right functions to use
Aug 12 19:41:37 <nbd> you can probably use ap91_pci_init()
Aug 12 19:41:38 <fullstop> I assumed that this is what ap_94_pci_init was going
Aug 12 19:41:49 <fullstop> let met look at 91
Aug 12 19:41:54 <nbd> ap94 is the wrong one
Aug 12 19:41:59 <nbd> ap94 is for ar7132
Aug 12 19:42:04 <nbd> whereas ap91 is for ar724x
Aug 12 19:42:25 <fullstop> 94 > 91 and 724x > 7132
Aug 12 19:42:37 <fullstop> so.. you can see how I ended up at ap94.  ;-)
Aug 12 19:42:57 <nbd> yeah, well, the atheros naming scheme for chips vs platforms is not as simple as that
Aug 12 19:44:30 <fullstop> kind of like how nvidia versions video cards, I suppose.
Aug 12 19:44:36 <nbd> :)
Aug 12 19:44:48 <fullstop> 280 > 460
Aug 12 19:45:01 <fullstop> The 2nd digit is far more important
Aug 12 19:45:59 <fullstop> 2 questions.  pb = "public board", so what is "ap" ?  And ART = "Atheros Radio ???"
Aug 12 19:49:07 <nbd> public board? i'm not sure what pb stands for, but that sounds a bit weird
Aug 12 19:49:12 <nbd> AP is access point
Aug 12 19:49:20 <nbd> ART is atheros radio test
Aug 12 19:49:29 <nbd> which is the tool that generates the data which is stored in that partition
Aug 12 19:50:45 <fullstop> nbd: getting much closer... http://pastebin.com/katCCK1r
Aug 12 19:51:50 <fullstop> No pdev->bus->self
Aug 12 19:54:25 <fullstop> despite that warning, the wlan works
Aug 12 20:28:02 <fullstop> is there any calculated way to determine the ethernet phy mask?
Aug 12 20:51:43 <nbd> what kind of switch does it have?
Aug 12 20:55:01 <KanjiMonster> ar8316
Aug 12 20:55:32 <KanjiMonster> nbd: at least that one get's detected
Aug 12 20:57:10 <nbd> then the phy mask should be the same as with any other device using an ar8316
Aug 12 20:57:20 <nbd> if it's using both MACs
Aug 12 21:00:39 <fullstop> nbd: in that case, is it normal for the wan port to behave identically to the lan port unless one configures vlan?
Aug 12 21:01:30 <KanjiMonster> fullstop: that sounds like it doesn't use both MACs
Aug 12 21:02:49 <fullstop> I don't think that there is a second mac in "art", but I could be wrong.
Aug 12 21:03:04 <fullstop> let me check uboot
Aug 12 21:03:55 <fullstop> there is one
Aug 12 21:04:02 <fullstop> let me look in art
Aug 12 21:04:33 <fullstop> haha.. it's at byte 0
Aug 12 21:05:20 <fullstop> also, usb is not quite working.  It finds the host controller, but no device is ever powered up.
Aug 12 21:05:40 <nbd> maybe you need to enable gpio 2 there as well
Aug 12 21:05:42 <fullstop> From the dd-wrt source, Sebastian modified usb-ehci.c and added the following: http://pastebin.com/F4HeV87N
Aug 12 21:05:58 <fullstop> I tried gpio2, as was used in ag300
Aug 12 21:06:06 <nbd> maybe it's a different gpio
Aug 12 21:06:10 <nbd> i don't think you need to add those lines
Aug 12 21:06:19 <fullstop> so what do I do?  try every gpio?
Aug 12 21:06:34 <nbd> yeah, maybe
Aug 12 21:09:47 <fullstop> Well, I suppose that there can not be that many... ;-)
Aug 12 21:13:05 <fullstop> nbd: can I set gpios from the command line or do I have to do many builds?
Aug 12 21:14:36 <nbd> you can test it from the command line
Aug 12 21:14:48 <nbd> you can go to /sys/class/gpio or something lkke that
Aug 12 21:14:51 <nbd> and register gpios
Aug 12 21:15:04 <fullstop> danke
Aug 12 22:13:17 <builder_> build #56 of ar71xx is complete: Success [build successful]  Build details are at http://buildbot.openwrt.org:8010/builders/ar71xx/builds/56
Aug 12 22:30:38 <Chocky> why is libssl so fat?
Aug 13 00:26:39 <WillieNL> hmm builds go almosth flawless here atm :)
Aug 13 02:40:32 <philipp64|laptop> KanjiMonster: I can't reproduce what you're seeing with the patch.
**** ENDING LOGGING AT Sat Aug 13 02:59:57 2011