**** BEGIN LOGGING AT Sun Jan 24 02:59:57 2010
Jan 24 04:17:47 <swalker> updated openwrt/upstream, http://pastehtml.com/view/5t3cr9l.html
Jan 24 04:24:51 <swalker> nbd: can you remove the two now empty wpa_supplicant subdirs?
Jan 24 04:27:30 <nbd> swalker: will do
Jan 24 08:08:09 <builder> build #28 of s3c24xx-trunk is complete: Failure [failed compile_11]  Build details are at http://tksite.gotdns.org:8010/builders/s3c24xx-trunk/builds/28
Jan 24 09:14:42 <CIA-88> lars * r19309 /trunk/target/linux/xburst/config-2.6.32: [xburst] Disable LOCALVERSION_AUTO
Jan 24 10:01:24 <CIA-88> nico * r19310 /trunk/target/linux/ubicom32/config-2.6.30: [target] ubicom32: add 2 missing config symbols
Jan 24 10:02:24 <CIA-88> nico * r19311 /trunk/target/linux/generic-2.6/config-2.6.32: [target] generic-2.6: add missing 2.6.32 config symbol
Jan 24 10:41:08 <stintel> miniupnpd no longer compiles: http://openwrt.pastebin.ca/1763800
Jan 24 14:04:47 <swalker> stintel: #6561 has a patch
Jan 24 14:07:36 <stintel> swalker: cheers, I'll try
Jan 24 14:25:34 <CIA-88> nbd * r19312 /trunk/package/hostap-driver/Makefile: fix another kernel version dependency (thx, maddes)
Jan 24 14:28:48 <CIA-88> nbd * r19313 /trunk/package/base-files/files/sbin/wifi: fix a typo in the hwmode handling in /sbin/wifi (thx, maddes)
Jan 24 15:07:14 <CIA-88> kaloz * r19314 /trunk/toolchain/gcc/ (5 files in 2 dirs): add support for gcc 4.4.3
Jan 24 17:16:32 <dolphin> blogic: heard you are working on rt288x/rt305x port
Jan 24 17:16:45 <blogic> i am indeed
Jan 24 17:16:47 <dolphin> I've got Buffaly WZR-AGL300NH cracked open here
Jan 24 17:16:58 <blogic> is it 305x or 288x ?
Jan 24 17:16:59 <dolphin> with serial connection, and I started adding support for it
Jan 24 17:17:01 <dolphin> 288x
Jan 24 17:17:04 <blogic> ok
Jan 24 17:17:13 <blogic> does it boot yet ?
Jan 24 17:17:21 <dolphin> yep, it boots openwrt
Jan 24 17:17:26 <dolphin> but no devices
Jan 24 17:17:36 <dolphin> seems you weren't very careful writing ralink_eth stuff, it won't compile without rt305x
Jan 24 17:18:04 <dolphin> are you working against trunk too?
Jan 24 17:19:12 <dolphin> I'm loading ramdisk images to it using the stock u-boot
Jan 24 17:19:58 <blogic> ok, good luck
Jan 24 17:20:12 <blogic> and try to be careful
Jan 24 17:20:41 <dolphin> well, I'm pretty green to all this openwrt stuff
Jan 24 17:20:45 <blogic> and yes, the ethernt driver is 305x only, due to lack of time
Jan 24 17:20:58 <blogic> the devices might not appear if the mach is not registered
Jan 24 17:21:03 <dolphin> I could try to patch it, if you're working against trunk too
Jan 24 17:21:24 <dolphin> I just added wzr_agl300nh based on the rt_n15
Jan 24 17:21:31 <blogic> good
Jan 24 17:21:34 <dolphin> and seems that thing now jams at boot, great job by me :P
Jan 24 17:21:35 <blogic> so, where is the problem ?
Jan 24 17:22:01 <dolphin> well, I'm pretty confused how everything is scattered around the openwrt tree
Jan 24 17:22:28 <blogic> its not scattered
Jan 24 17:23:04 <dolphin> well, it still seems to me bit like it, but I'm getting the hang of it
Jan 24 17:23:16 <blogic> hehe
Jan 24 17:23:26 <blogic> it is very structured once you get the hang of it
Jan 24 17:23:38 <blogic> what switch does the board you use have ?
Jan 24 17:24:21 <nbd> and how much flash does it have?
Jan 24 17:24:29 <nbd> if it has 8 mb, you need to be careful with mtd support
Jan 24 17:24:50 <rtz2> nbd: why?
Jan 24 17:25:04 <blogic> because ralink is stuppid and boots from the cnter of the mtd window
Jan 24 17:25:15 <rtz2> ohh
Jan 24 17:25:18 <dolphin> its exactly the same as rt-n15, hardware wise, IIRC
Jan 24 17:25:25 <blogic> the window is 8MB and they boot from 4MB-64k
Jan 24 17:25:35 <dolphin> it's 4MB flash
Jan 24 17:25:39 <blogic> ralink made an ugly patch
Jan 24 17:25:43 <blogic> or brainslayer did
Jan 24 17:25:45 <blogic> cant remeber
Jan 24 17:26:36 <rtz2> blogic: by the way, you were also the guy working on danube support, right?
Jan 24 17:27:18 <dolphin> from boot:
Jan 24 17:27:22 <dolphin> "U-Boot 1.1.3 for BUFFALO AIRSTATION (Feb 26 2008 - 16:01:14)
Jan 24 17:27:22 <dolphin> Board: RT2880 DRAM:  32 MB"
Jan 24 17:27:35 <blogic> rtz2: yes
Jan 24 17:28:12 <dolphin> RTL8366RS driver SetVLan initialized
Jan 24 17:28:17 <dolphin> that'd be the switch I think
Jan 24 17:28:54 <dolphin> the wiki would have quite some to correct
Jan 24 17:29:14 <rtz2> blogic: what's the state of this port?
Jan 24 17:29:16 <dolphin> memory is 32 instead of 16, this thing has serial, and jtag too
Jan 24 17:29:30 <dolphin> http://wiki.openwrt.org/oldwiki/completetableofhardware <- this page I mean
Jan 24 17:30:26 <dolphin> going to fix it a little
Jan 24 17:30:30 <blogic> rtz2: working
Jan 24 17:30:32 <blogic> dsl works
Jan 24 17:30:37 <blogic> voip is being worked on
Jan 24 17:30:46 <blogic> we have 750 boards that were donated
Jan 24 17:30:55 <blogic> we just fixed switch support and plan to flash them all soon
Jan 24 17:31:16 <blogic> rtz2: main problem is, that not many boards are supported yet, as there are not many
Jan 24 17:31:28 <blogic> but we are slowly progressing on the arcaydian ones
Jan 24 17:31:44 <rtz2> blogic: the standard vodafone router in germany uses it
Jan 24 17:31:53 <nbd> dolphin: there is a driver for the switch in the ar71xx target. that driver could be moved to the generic section, then the rt2880 code could take advantage of it as well
Jan 24 17:31:59 <blogic> rtz2: yes
Jan 24 17:32:03 <blogic> rtz2: donate one :)
Jan 24 17:32:12 <rtz2> blogic: I have one here, but I would have to open the case by force so I didn't play with it yet
Jan 24 17:32:12 <blogic> rtz2: i live in hamburg so postage is cheap
Jan 24 17:32:23 <dolphin> looks like I can't edit the wiki then...
Jan 24 17:32:53 <xMff> dolphin: not oldwiki, add the entry here; http://wiki.openwrt.org/toh/start
Jan 24 17:32:55 <rtz2> blogic: it's some stupid clip system
Jan 24 17:33:04 <rtz2> blogic: I will take a look myself next week or so, if I have some time :)
Jan 24 17:33:09 <blogic> sure
Jan 24 17:33:22 <blogic> rtz2: make sure to work out if the pci has an internal or external clock
Jan 24 17:33:38 <blogic> its a common curlpit when starting to use ifxmips
Jan 24 17:33:39 <rtz2> blogic: I will come back to you once it's open
Jan 24 17:33:42 <blogic> sure
Jan 24 17:34:53 <blogic> rtz2: can you show me a picture on the web of the box ?
Jan 24 17:35:01 <blogic> use gogole picture search or summin
Jan 24 17:35:08 <blogic> so i can see if it is a ODM
Jan 24 17:36:27 <Ascendion>  anyone running vbox debian with bridged networking for development ?? I'm trying to get set up to where it looks like my vbox has a physical eth port and can get a dhcp assignment from the router
Jan 24 17:37:10 <rtz2> blogic: there were some very good pictures of it somewhere
Jan 24 17:37:13 <rtz2> give me a minute
Jan 24 17:37:31 <xMff> Ascendion: not for development but I use the bridged feature
Jan 24 17:38:20 <rtz2> blogic: http://www.arcor-user-forum.de/dsl-easybox-802-gehaeuse-geoeffnet-hardware-photos-2758/
Jan 24 17:38:44 <Ascendion> anything special I should know ?? I'm running vista and it has the dhcp broadcast bug, and neither vista nor my vox are getting dhcp on that physical port
Jan 24 17:39:14 <Ascendion> not that I think the vista bug should effect the vbox -- but you never know :)
Jan 24 17:39:55 <xMff> Ascendion: uhm... I once had the problem that the bridge thing did not work but a kernel update helped
Jan 24 17:40:03 <xMff> Linux host + windoes guest here
Jan 24 17:40:06 <blogic> rtz2: thx
Jan 24 17:40:40 <Ascendion> oops :) win host linux guest :)
Jan 24 17:40:58 <Ascendion> k.. I keep trying other things :)
Jan 24 17:41:12 <xMff> I have no idea how to even start debugging on vista
Jan 24 17:41:29 <dolphin> blogic: it seems that I can't choose the machine model for ramips from menuconfig yet?
Jan 24 17:41:49 <blogic> indeed
Jan 24 17:41:50 <xMff> dolphin: you need to enable the "show broken targets and packages" feature
Jan 24 17:41:55 <blogic> CMD_LINE
Jan 24 17:42:04 <xMff> erm nvm
Jan 24 17:42:08 <blogic> xMff: model != board
Jan 24 17:42:15 <dolphin> did even that, but I still can't tell it that my device is specificly WZR-AGL300NH
Jan 24 17:42:16 <xMff> yea
Jan 24 17:42:26 <blogic> dolphin: you nees to put the string in the kernel command line
Jan 24 17:42:45 <Ascendion> ahhh forget it -- got it figured :)
Jan 24 17:42:48 <dolphin> right, how would I do that
Jan 24 17:43:02 <dolphin> some u-boot magic?
Jan 24 17:44:32 <blogic> hmmm
Jan 24 17:44:40 <blogic> ideally in the .config in the kernel folder :D
Jan 24 17:44:44 <blogic> that is how i do it
Jan 24 17:46:43 <dolphin> wouldn't that be ditched every time I rebuild
Jan 24 17:47:00 <blogic> yes
Jan 24 17:47:10 <blogic> gabor made that part
Jan 24 17:47:17 <blogic> juhosg
Jan 24 17:47:22 <blogic> or however it is spelled
Jan 24 17:47:50 <dolphin> rite
Jan 24 17:48:34 <blogic> ermm
Jan 24 17:48:41 <blogic> ideally ou can hand it over from uboot
Jan 24 17:48:44 <blogic> but dont ask me how
Jan 24 17:49:10 <xMff> dolphin: you can try to set it using kernel_menuconfig
Jan 24 17:49:20 <xMff> * make k...
Jan 24 17:50:01 <dolphin> I just tried, but only the asus board appears
Jan 24 17:50:16 <dolphin> obviously I still need to hack some places (this is why I said it seems scattered to me)
Jan 24 17:50:51 <blogic> hehe
Jan 24 17:50:59 <blogic> i will update my tree in a sec and have a look
Jan 24 17:51:21 <dolphin> I don't have commit rights to trunk
Jan 24 17:51:34 <nbd> you can send patches to the list
Jan 24 17:51:34 <dolphin> so it's on my HD at the moment
Jan 24 17:51:45 <dolphin> well, what use is the patch if it doesn't work yet :P
Jan 24 17:51:46 <nbd> preferably split up by subsystem/feature
Jan 24 17:52:09 <nbd> what use is it? it gets reviewed
Jan 24 17:52:10 <nbd> ;)
Jan 24 17:52:17 <xMff> dolphin: I think nobody expects you to deliver 100% on the first attempt
Jan 24 17:52:36 <dolphin> but c'mon guys, the device doesn't even boot after adding this stuff
Jan 24 17:52:55 <blogic> well
Jan 24 17:53:02 <blogic> send it to the list, pastebing it ....
Jan 24 17:53:06 <dolphin> if you just want to see, that's ok, it's basically just the rt-n15 copied and modified for MTD setup
Jan 24 17:53:06 <blogic> and peope can review
Jan 24 17:53:36 <dolphin> would I run some command to make the things from target/linux/ramips/files/arch/mips/ralink/rt288x/Kconfig to go to the build tree?
Jan 24 17:54:47 <blogic> nope
Jan 24 17:54:51 <nbd> i usually edit stuff in the build tree
Jan 24 17:54:55 <nbd> and copy it back when i'm done
Jan 24 17:56:14 <dolphin> well, just editing that file, and doing kernel_menuconfig again, doesn't add a new target under machine selection
Jan 24 17:56:37 <nbd> that's why i edit it in the build tree
Jan 24 17:56:41 <dolphin> ;)
Jan 24 17:57:01 <dolphin> but I don't want to accidentally ditch ite
Jan 24 17:57:40 <nbd> the only ways it can get ditched is if you change the kernel version (or it gets changed on an update) or your run make clean or similar
Jan 24 17:58:13 <nbd> another way to keep track of all the changes that you do is to use quilt in the build tree
Jan 24 17:58:31 <nbd> and save the patch somewhere after making changes
Jan 24 18:01:37 <dolphin> hmm, my patch is going to be very bad
Jan 24 18:01:45 <dolphin> I haven't even figured out the gpio's
Jan 24 18:02:23 <dolphin> tried to look for them in the Buffalos 2.4.x (stripped down 2.4.30) source
Jan 24 18:03:58 <dolphin> how would the kernel decide which board initialization routines to use if all of them are built in?
Jan 24 18:07:26 <nbd> the board name is supposed to be passed on the command line
Jan 24 18:07:41 <nbd> we typically generate a separate image for each board with a patched command line
Jan 24 18:07:46 <nbd> built out of the same kernel binary
Jan 24 18:07:57 <nbd> the ar71xx platform uses this extensively
Jan 24 18:08:38 <dolphin> but the kernel stays the same for less build overhead?
Jan 24 18:08:44 <nbd> yes
Jan 24 18:10:01 <dolphin> now I went and edited the build_dir and it seems to work ok
Jan 24 18:10:29 <dolphin> #wonder if creating the patch is now harder, but that's for some other dayto worry
Jan 24 18:12:12 <dolphin> nbd: which scripts would I stab to add two separate images?
Jan 24 18:12:42 <nbd> the image building stuff is in target/linux/ramips/image/Makefile
Jan 24 18:12:48 <dolphin> rt-n15 was the only one existing, so I guess it defaulted to that or something
Jan 24 18:12:50 <nbd> you can take some hints from target/linux/ar71xx/image/Makefile
Jan 24 18:17:55 <dolphin> might take a while :P
Jan 24 18:18:18 <stintel> swalker: the patch from #6561 seems to fix the compile problem
Jan 24 18:18:39 <xMff> stintel: miniupnpd ?
Jan 24 18:18:44 <stintel> xMff: yep
Jan 24 18:18:53 <xMff> I hate that thing, it breaks down on every occasion
Jan 24 18:18:56 <stintel> also, xtables-addons also fails to compile
Jan 24 18:19:03 <xMff> that too :P
Jan 24 18:19:11 <dolphin> nbd: but those would result in flashable images, right
Jan 24 18:19:20 <dolphin> according to what I can understand the file
Jan 24 18:19:37 <dolphin> I'm nowhere near ready to overwrite the existing firmware yet
Jan 24 18:20:04 <nbd> it's also used to build the final kernel images
Jan 24 18:20:09 <nbd> but there's an easier way of testing this
Jan 24 18:20:14 <nbd> put the board name in the configured kernel cmdline
Jan 24 18:20:41 <nbd> with board=WZR-...
Jan 24 18:20:51 <nbd> then you won't have to patch it for the testing images
Jan 24 18:22:35 <dolphin> which string would that have to be equal with
Jan 24 18:22:44 <dolphin> from MIPS_MACHINE line?
Jan 24 18:22:44 <nbd> you have to modify prom.c
Jan 24 18:23:01 <nbd> i mean arch/mips/ramips/common/prom.c
Jan 24 18:23:06 <nbd> it contains a list of detected boards
Jan 24 18:25:48 <dolphin> where would I find the cmdline?
Jan 24 18:27:48 <dolphin> to be done with patch-cmdline I guess
Jan 24 18:27:50 <nbd> CONFIG_CMDLINE
Jan 24 18:29:26 <dolphin> now found it, great
Jan 24 18:29:43 <dolphin> would my ramdisk build option override that?
Jan 24 18:30:02 <dolphin> because I do see rootfstype=squashfs,jffs2 in the same line
Jan 24 18:30:20 <nbd> initramfs does not have to be specified in the command line
Jan 24 18:30:30 <nbd> since the kernel will boot it in either case if it's built in
Jan 24 18:31:49 <dolphin> ok, now building
Jan 24 18:32:07 <dolphin> btw, Buffalo devs must be lazy: system type             : Palmchip Surfboard processor               : 0 cpu model               : unknown V6.12 BogoMIPS                : 265.42
Jan 24 18:32:46 <cshore> nbd: so you like the preinit?
Jan 24 18:32:50 <nbd> cshore: yes
Jan 24 18:32:59 <dolphin> wouldn't any developer with self-respect change such identifiers even if they thought nobody else would ever see it
Jan 24 18:33:19 <nbd> dolphin: in the embedded world the common attitude is 'nobody is ever going to see that'
Jan 24 18:34:56 <cshore> nbd: awesome: I'll be gone for a bit this afternoon, but when I'm here if there's any problem I'll be around to see what went wrong .... though I'm pretty sure the parts I couldn't test (the changes to preinit.arch) are right too.
Jan 24 18:36:05 <nbd> they look right to me
Jan 24 18:36:41 <blogic> dolphin: they are chinese
Jan 24 18:36:44 <blogic> and use a SDK
Jan 24 18:37:00 <blogic> chinese people dont aim to be good, but aim to be not the worste
Jan 24 18:37:07 <blogic> not loosing is like winning to the chinese
Jan 24 18:37:17 <blogic> in addition tey seem to lack the word for accuracy
Jan 24 18:37:25 <blogic> that is how that kind of stuff happens
Jan 24 18:37:47 <dolphin> heh
Jan 24 18:38:28 <dolphin> now I would like to locate asm-ralink, which doesn't even exist
Jan 24 18:38:57 <dolphin> from prom.c I have #include <asm/mach-ralink/common.h>\n#include <asm/mach-ralink/machine.h>
Jan 24 18:39:14 <dolphin> but such directory name doesn't even exist according to find
Jan 24 18:39:29 <nbd> should be somewhere in arch/mips/include/asm/
Jan 24 18:39:58 <dolphin> thanks, I wonder why find wouldn't locate it
Jan 24 18:39:58 <blogic> find build_dir/ -name common.h | grep ralink
Jan 24 18:47:29 <dolphin> btw, this seems to be first helpful *-devel channel, I have found in freenode, and I've been here way before this was even called freenode (good ol' openproject) :P
Jan 24 18:48:24 <nbd> ;)
Jan 24 18:48:44 <dolphin> I'll soon order WBMR-G300N and will contribute to its port too, for sure
Jan 24 18:48:52 <nbd> sounds good
Jan 24 18:49:04 <dolphin> though, I would imagine the ADSL part makes it little no-go
Jan 24 18:49:14 <nbd> what kind of platform is this?
Jan 24 18:49:25 <dolphin> http://oldwiki.openwrt.org/OpenWrtDocs%282f%29Hardware%282f%29Buffalo%282f%29WBMR%282d%29G300N.html
Jan 24 18:49:41 <dolphin> so, should be easier as it's broadcom
Jan 24 18:52:51 <dolphin> ok, so now when I added support for mtd, it seems to work
Jan 24 18:53:09 <dolphin> but unfortunately, it also tries to use mtd as rootfs, which I didn't want at all
Jan 24 18:53:57 <dolphin> MIPS: machine is Buffalo WZR-AGL300NH
Jan 24 18:54:04 <dolphin> but at least I achieved something
Jan 24 18:57:49 <dolphin> http://pastebin.com/m4b7b01b5
Jan 24 18:57:54 <dolphin> that's what the boot says for now
Jan 24 18:58:02 <dolphin> the MTD table appears similar to the original boot
Jan 24 18:58:55 <dolphin> except, I don't get all those warnings, hopefully it didn't actually modify the flash :P
Jan 24 19:00:03 <dolphin> wonder if changing rootfstype=ramdisk would help
Jan 24 19:00:45 <nbd> did you select ramdisk in toplevel menuconfig in openwrt?
Jan 24 19:00:48 <nbd> under 'target images'?
Jan 24 19:01:09 <dolphin> yes
Jan 24 19:01:28 <dolphin> all other options get removed after doing that
Jan 24 19:01:40 <nbd> ok
Jan 24 19:02:01 <dolphin> but I think that rootfstype=squashfs,jffs2 might screw it up now, when I added the MTD
Jan 24 19:05:19 <dolphin> if I would get the ethernet working enough to be able to boot ramdisk image with tools to backup the mtd files, I'd dare to experiment even more :P
Jan 24 19:09:02 <dolphin> blogic: seems that you made some load address trick for rt305x stuff?
Jan 24 19:09:50 <blogic> ?
Jan 24 19:10:10 <blogic> well, there is a patch in openwrt, that allows load adrr == entrypoint for mips based units
Jan 24 19:10:19 <blogic> i sthat what you mean ?
Jan 24 19:11:42 <dolphin> in target/linux/ramips/image/Makefile
Jan 24 19:11:50 <dolphin> ifeq ($(CONFIG_RALINK_RT305X),y)
Jan 24 19:11:50 <dolphin> define kernel_entry
Jan 24 19:11:50 <dolphin> -a 0x80000000 -e 0x80000000
Jan 24 19:11:56 <dolphin> that stuff
Jan 24 19:12:34 <dolphin> could that have something to do why my setup isn't booting anymore
Jan 24 19:14:57 <dolphin> or would it be the rootfs as mtd block name, that causes some automagic to happen
Jan 24 19:16:50 <rtz2> is this union overlay stuff used in openwrt?
Jan 24 19:16:59 <rtz2> because I don't see any kernel patches for it
Jan 24 19:27:26 <dolphin> bah
Jan 24 19:27:37 <dolphin> changing to rootfstype=ramsf changed nothing
Jan 24 19:27:51 <dolphin> it automatically wants to use partition named rootfs as root
Jan 24 19:28:15 <CIA-88> nbd * r19315 /trunk/package/mac80211/patches/610-fix_warn_on.patch: mac80211: fix a WARN_ON in client mode
Jan 24 19:30:12 <dolphin> nbd: am I right? that when it finds mtdblock named rootfs, it gives some special treatment
Jan 24 19:30:23 <nbd> yes
Jan 24 19:30:33 <nbd> but normally that shouldn't interfere with initramfs
Jan 24 19:31:29 <dolphin> wouldn't booting from ramdisk have to include some parameters too
Jan 24 19:31:53 <nbd> no
Jan 24 19:34:25 <dolphin> "Initrd not found or empty - disabling initrd"
Jan 24 19:34:30 <dolphin> that'd be a bad thing, right
Jan 24 19:34:49 <nbd> yup
Jan 24 19:35:21 <dolphin> somehow I guess that my CONFIG_CMDLINE fiddling killed ramdisk thing
Jan 24 19:37:48 <nbd> weird
Jan 24 19:38:39 <dolphin> or the fact that I did make kernel_menuconfig, wonder if it could be that
Jan 24 19:45:06 <nbd> does the kernel image have the right size for an initramfs based image?
Jan 24 19:48:34 <dolphin> 2.2M bin/ramips/openwrt-ramips-uImage-initramfs.bin
Jan 24 19:50:22 <nbd> ok, looks reasonable
Jan 24 19:54:30 <dolphin> well, I made a clean edit against a clean tree
Jan 24 19:54:35 <dolphin> lets see if it works
Jan 24 19:59:15 <CIA-88> jow * r19316 /trunk/package/iptables/Makefile: [package] iptables: install libip[46]tc into staging dir
Jan 24 19:59:53 <dolphin> nbd, blogic: http://pastebin.com/m5aa3c9dd
Jan 24 19:59:59 <dolphin> the changes so far
Jan 24 20:00:22 <nbd> ok
Jan 24 20:03:45 <dolphin> um, I think I could avoid downloading all the stuff and building the toolchain from scratch each time
Jan 24 20:09:00 <CIA-88> jow * r19317 /packages/net/miniupnpd/Makefile: [packages] miniupnpd: fix compile against iptables 1.4.6 (#6561)
Jan 24 20:10:40 <xMff> hm, do we even have xtables-addons anymore?
Jan 24 20:11:01 <xMff> somebody said it breaks but I can't find it :)
Jan 24 20:12:07 <xMff> ah there it is...
Jan 24 20:12:10 <dolphin> blogic: going to work more on the ralink stuff soon?
Jan 24 20:20:26 <cshore> rtz: I don't know....I left that stuff alone on the guess that some platform I don't know uses it
Jan 24 20:20:45 <cshore> rtz: or that it's for future changes
Jan 24 20:21:46 <dolphin> ok, already figured that I can tell it to use common toolchain and dl library
Jan 24 20:21:55 <dolphin> no luck for to avoid building packages?
Jan 24 20:22:11 <dolphin> s/packages/host tools/
Jan 24 20:29:22 <dolphin> also, when doing build with V=99, why does it ask for the kernel configurations two times, instead of one
Jan 24 20:30:46 <xMff> kernel compile happens in two phases
Jan 24 20:30:50 <xMff> first all modules
Jan 24 20:30:51 <xMff> later the image
Jan 24 20:31:02 <xMff> each time make oldconfig (defconfig?) is called
Jan 24 20:31:25 <xMff> if you per-arch kernel config is missing symbols, the oldconfig will prompt you on those
Jan 24 20:31:48 <dolphin> right
Jan 24 20:32:13 <dolphin> so I should add those I think
Jan 24 20:32:19 <xMff> you should just add the symbols it asks for to target/linux/ramips/rtxyz/config-<kernel_version>
Jan 24 20:32:23 <xMff> yes
Jan 24 20:33:01 <dolphin> is order meaningful?
Jan 24 20:33:24 <xMff> don't think so. I think you can even use "make kernel_menuconfig" to populate that file
Jan 24 20:37:33 <blogic> dolphin: yes, but not too soonish
Jan 24 20:38:15 <dolphin> xMff: which is preferred, patching one line only, or updating the file with kernel_menuconfig
Jan 24 20:38:35 <dolphin> the order of certain elements change, so the patch just rearranges them
Jan 24 20:38:45 <dolphin> and adds the missing one
Jan 24 20:39:21 <xMff> well I think make kernel_menuconfig writes its "native" order while the file was updated manually in the past
Jan 24 20:39:46 <dolphin> soo, kernel_menuconfig it is
Jan 24 20:39:52 <blogic> indeed
Jan 24 20:41:45 <CIA-88> jow * r19318 /packages/net/xtables-addons/ (3 files in 2 dirs): [packages] xtables-addons: update to v1.22 and fix compile against iptables 1.4.6
Jan 24 20:42:17 <xMff> swalker, stintel: miniupnd and xtables-addons should compile again
Jan 24 20:42:21 <cshore> xMff: if I make changes in the build (to be copied back later), will make target/linux/clean kill them?
Jan 24 20:42:38 <xMff> it will only kill kernel stuff
Jan 24 20:42:44 <xMff> the linux source tree
Jan 24 20:42:57 <dolphin> forgot to set the board option
Jan 24 20:43:01 <cshore> xMff: ok that's what I'll be working on, so for that purpose, the answer is yes
Jan 24 20:43:01 <dolphin> Kernel panic - not syncing: No init found.  Try passing init= option to kernel.
Jan 24 20:43:25 <dolphin> but last time I tried, without adding the patches or passing any special options, it booted all good
Jan 24 20:43:35 <dolphin> using initramfs
Jan 24 20:43:48 <dolphin> I think the ramfs is now broken
Jan 24 20:44:09 <cshore> is /init present?
Jan 24 20:44:23 <cshore> initramfs uses /init as init
Jan 24 20:44:32 <dolphin> well, it doesn't know anything about MTD, nor have initrd
Jan 24 20:44:42 <dolphin> so there is no root filesystem to begin with
Jan 24 20:45:04 <cshore> dolphin: what's root?
Jan 24 20:45:33 <dolphin> it's supposed to be ramdisk
Jan 24 20:45:40 <cshore> dolphin: initramfs?
Jan 24 20:45:44 <blogic> dolphin: i never tried ramdisk images
Jan 24 20:45:51 <blogic> no idea if they work on ramips
Jan 24 20:45:58 <dolphin> blogic: well, I tried it last time
Jan 24 20:46:16 <blogic> ok
Jan 24 20:46:19 <dolphin> worked all good from trunk, when I didnt apply any patches
Jan 24 20:46:22 <cshore> dolphin: even ramdisk has a root initial image
Jan 24 20:46:27 <dolphin> sure
Jan 24 20:46:28 <stintel> xMff: cheers ;-)
Jan 24 20:46:37 <dolphin> but obviously, it's not finding it for some reason
Jan 24 20:46:50 <cshore> dolphin: or the /init is missing
Jan 24 20:47:08 <cshore> dolphin: but if you didn't move it or delete then that's not it
Jan 24 20:47:24 <cshore> cshore: the ramdisk is supposed to add it from generic-2.6 IIRC
Jan 24 20:47:40 <dolphin> talking to self?
Jan 24 20:47:41 <CIA-88> jow * r19319 /packages/net/tinyproxy/ (3 files in 2 dirs): [packages] tinyproxy: update to v1.8.0 (#6563)
Jan 24 20:48:19 <dolphin> well, anyway, I'm bit clueless now what to do then :P
Jan 24 20:48:22 <cshore> dolphin: heh, I do that to, but I meant you :)
Jan 24 20:48:38 <dolphin> blogic: did you just cold-heartedly overwrite the MTD stuff?
Jan 24 20:49:03 <cshore> dolphin: is it possible that the generic stuff isn't getting added?
Jan 24 20:49:15 <dolphin> I dunno, you tell me?
Jan 24 20:49:32 <dolphin> like I said, I just started with openwrt very very lately
Jan 24 20:49:45 <dolphin> and I'm very confused of the structure
Jan 24 20:49:58 <cshore> dolphin: I'm not quite sure how the kernel tree stuff works...I just poke it occasionally....I'm more of a userspace guy
Jan 24 20:50:00 <dolphin> "Initrd not found or empty - disabling initrd"
Jan 24 20:50:19 <dolphin> I think that would strongly indicate it isn't finding anything, right?
Jan 24 20:50:27 <cshore> dolphin: yes
Jan 24 20:50:44 <dolphin> soo, why wouldn't ramdisk all of sudden not work
Jan 24 20:51:41 <cshore> xMff: does ramdisk on x86 work with qemu?
Jan 24 20:52:47 <blogic> dolphin: dont get the question and the addons like "cold hearted" "uncareful"
Jan 24 20:53:37 <dolphin> heh, I meant that if you didn't use ramdisk, did you just overwrite the existing flash contents
Jan 24 20:53:42 <cshore> blogic: perhaps he means bravely, fearlessly
Jan 24 20:53:47 <dolphin> yes
Jan 24 20:53:56 <blogic> cshore: i doubt it :)
Jan 24 20:54:06 * blogic does not like adjectives :)
Jan 24 20:54:23 <dolphin> I can try to communicate without adjectives too :P
Jan 24 20:55:47 <dolphin> I'm trying to pertain the original functioncality until I get wifi + switch working, so I would rather not touch the flash unless I'm sure I get it working
Jan 24 20:55:55 <cshore> dolphin: I understand your trepidation about flashing before ramdisk works....I was the same way when I was experimenting with my only router
Jan 24 20:56:08 <dolphin> indeed
Jan 24 20:57:58 <dolphin> but nevertheless, the problem seems to be that ramdisk worked, and doesn't work anymore
Jan 24 20:58:23 <dolphin> and because my patches do not directly touch anything related, I think it might be something broken in the bigger scale
Jan 24 21:00:50 <blogic> dolphin: pot your patches somewhere
Jan 24 21:01:05 <blogic> or post them
Jan 24 21:01:07 <blogic> :)
Jan 24 21:02:26 <dolphin> http://pastebin.com/m51d4df49
Jan 24 21:02:28 <dolphin> there it is
Jan 24 21:03:54 <dolphin> if you really want to use it, just edit the target/linux/ramips/rt288x/config-2.6.30 and append board=WZR-AGL300NH to CMDLINE
Jan 24 21:04:45 <dolphin> but I doubt you want, if you don't have the hardware :P
Jan 24 21:05:07 <dolphin> and I haven't even got my ass up to post the disassembly pictures, nor the serial port pinout
Jan 24 21:05:52 <dolphin> and being quite new/expensive device, I'm afraid others haven't cracked theirs open
Jan 24 21:06:41 <dolphin> I'd be more happy to post it all at once, if I got it as far as the wired network would work
Jan 24 21:07:30 <blogic> dolphin: imho the patch should work even with flash
Jan 24 21:07:33 <blogic> ok, i am off to bed
Jan 24 21:07:38 <blogic> i am too tired to type
Jan 24 21:07:48 <blogic> cya in +8
Jan 24 21:07:53 <dolphin> do you mean without flash?
Jan 24 21:08:02 <blogic> with flash
Jan 24 21:08:06 <blogic> i mean try the flash image
Jan 24 21:08:10 <blogic> nothing bad can happen
Jan 24 21:08:21 <blogic> it wont burn
Jan 24 21:08:28 <dolphin> other than I lose the original, right
Jan 24 21:08:28 <blogic> and i assume the board has jtag
Jan 24 21:08:40 <blogic> the original will be on buffallos website
Jan 24 21:08:49 <dolphin> they provide theirs encrypted
Jan 24 21:11:10 <dolphin> I have second 50h+ work week coming, so I'll be back in a week or so :)
Jan 24 21:19:02 <dolphin> aww, someone killed the trunk so that it doesn't compile at all
Jan 24 21:20:13 <xMff> ?
Jan 24 21:20:46 <dolphin> either that, or -j4 support is bad
Jan 24 21:20:58 <dolphin> starting the build again and it went on
Jan 24 21:21:00 <xMff> the former
Jan 24 21:21:32 <dolphin> so basically, better not use that at all?
Jan 24 21:21:36 <dolphin> -jN that is
Jan 24 21:22:06 <xMff> parallel build sometimes dies randomly due to race conditions introduced by missing inter-package dependencies, e.g. foo needs bar to compile but bar is not finished yet
Jan 24 21:22:55 <xMff> just repeating make often helps
Jan 24 21:23:25 <xMff> I usually build with -j2 and didn't see problems recently
Jan 24 21:27:57 <Ascendion> 1st time builder asks: what did I goof up this time ?? while building miniupnpd-1.4/netfilter/iptcrdr.c: linux/netfilter/x_tables.h: no such file or directory
Jan 24 21:28:09 <Ascendion> but find shows the file
Jan 24 21:32:29 <cshore> there is a today or so fix for that been applied...how old is your tree?
Jan 24 21:32:43 <Ascendion> about 1.5 hours
Jan 24 21:32:48 <xMff> too old :)
Jan 24 21:34:04 <Ascendion> snv updated, packages updated, make V=99
Jan 24 21:35:06 <xMff> stop
Jan 24 21:35:12 <xMff> make target/linux/clean
Jan 24 21:35:15 <xMff> then make
Jan 24 21:35:49 <Ascendion> you sure ?? it seems to be much happier now :)
Jan 24 21:35:55 <dolphin> I'm 5 minutes away from giving up
Jan 24 21:36:13 <Ascendion> or not - making clean
Jan 24 21:36:24 <dolphin> how could my patch cause such side effects
Jan 24 21:37:11 <xMff> dolphin: I don't think it is your patch
Jan 24 21:37:31 <xMff> ramdisk images _never_ worked for me, they booted but the rootfs was never found
Jan 24 21:37:42 <xMff> maybe you're supposed to append root=/dev/ram0 or similar
Jan 24 21:37:50 <dolphin> how the hell it worked initially then? ;)
Jan 24 21:37:54 <dolphin> could be
Jan 24 21:38:00 <xMff> luck? solar storm?
Jan 24 21:38:38 <xMff> but your hacked the cmdline, maybe some important ramfs related bits got lost
Jan 24 21:38:58 <xMff> I suppose you don't have the dmesg of the working state anymore?
Jan 24 21:39:00 <dolphin> well, it removed nothing, just appended
Jan 24 21:39:22 <dolphin> nope
Jan 24 21:39:27 <dolphin> just trying unmodified trunk
Jan 24 21:40:04 <dolphin> but I don't remember it asking about the ralink ethernet module back then either
Jan 24 21:40:11 <dolphin> actually, I'm 100% positive it didn't
Jan 24 21:40:21 <dolphin> so I think last patch broke something
Jan 24 21:41:21 <dolphin> but it's only few weeks ago
Jan 24 21:41:42 <dolphin> and according to the web interface, it has not been changed for a while
Jan 24 21:42:07 <Ascendion> would anyone like to recommend a good usb jtag (with win and linux support) that will get along with my 54gl ??
Jan 24 21:43:24 <mirko> xMff: ping
Jan 24 21:43:41 <xMff> mirko: pong
Jan 24 21:45:10 <mirko> xMff: sent you a pm
Jan 24 21:45:54 <dolphin> right, unmodified trunk isn't working anymore
Jan 24 21:48:08 <cshore> dolphin: what platform again?
Jan 24 21:48:16 <dolphin> ramips/rt288x
Jan 24 21:48:40 <dolphin> my bet is that somebody broke the ramdisk after last weekend, so going to check an earlier state
Jan 24 21:53:08 <dolphin> right, so 18th it still worked
Jan 24 21:55:46 <Ascendion> I guess I didnt update the packages right -- same error
Jan 24 21:58:00 <Ascendion> nope -- at 19319
Jan 24 21:59:42 <xMff> strange it works here
Jan 24 22:00:52 <Ascendion> ya -- realy strange off a clean pull
Jan 24 22:01:57 <dolphin> I'm already almost 100% positive the problem is that ramdisk support got broken
Jan 24 22:02:07 <cshore> xMff: is there a reason the openwrt-light theme as a lowercase 'a' in the upper left corner
Jan 24 22:02:35 <xMff> cshore: most likely not
Jan 24 22:02:55 <dolphin> someone was drunk while writing it?
Jan 24 22:03:04 <xMff> probably
Jan 24 22:03:07 <xMff> or stoned
Jan 24 22:04:20 <cshore> I just checked the theme file...it was an an extraneous a at the beginning of the file
Jan 24 22:04:56 <xMff> sounds like a reasonable explanation
Jan 24 22:06:15 <cshore> I'm updating the image, but I've finally got the tiny fix openwrt-light mediabaseurl one-liner patch done
Jan 24 22:06:30 <cshore> I keep getting sidetracked
Jan 24 22:11:16 <cshore> hey, luci's fast in x86 (on dual-core KVM VM that is)
Jan 24 22:13:40 <xMff> anything is fast there
Jan 24 22:13:41 <xMff> :P
Jan 24 22:13:52 <cshore> xMff:  except compiling
Jan 24 22:14:51 <cshore> xMff:  I'd like to be able to compile toolchain + image in a minute or two
Jan 24 22:15:13 <xMff> ask NASA maybe ;)
Jan 24 22:15:33 <cshore> xMff:  yeah, it'd take some power
Jan 24 22:17:10 <cshore> xMff: I hate waiting for image builds
Jan 24 22:17:23 <xMff> :D
Jan 24 22:17:34 <xMff> for me it is the perfect excuse to get some coffee
Jan 24 22:18:13 <cshore> xMff: my coffee's only a few moments from my desk...I work from home
Jan 24 22:20:55 <cshore> xMff: I do dishes sometimes when I'm building
Jan 24 22:22:13 <dolphin> you guys get paid to work on openwrt?
Jan 24 22:22:27 <xMff> nope
Jan 24 22:22:56 <xMff> well I don't get payment for it, can't talk about others
Jan 24 22:23:07 <cshore> dolphin: I do other work too, but some of what I do on OpenWRT is paid
Jan 24 22:23:31 <cshore> dolphin: but I didn't plan it that way...I just was working on it anyway
Jan 24 22:25:18 <cshore> dolphin: and sometimes stuff I work on for my own reason becomes something work wants for it too (in OpenWRT)
Jan 24 22:25:57 <dolphin> sounds cool
Jan 24 22:26:21 <cshore> cshore:  I'm mainlining computers these days (work and play)
Jan 24 22:26:27 <cshore> dolphin I mean
Jan 24 22:26:44 <cshore> dolphin: mostly OpenWRT, but some proprietary server stuff too
Jan 24 22:26:45 <dolphin> I get paid (at the moment) to hit my head to the qemu concrete (some might call it source too :P)
Jan 24 22:26:57 <cshore> dolphin: cool
Jan 24 22:27:07 <dolphin> it's far from cool, I can tell you
Jan 24 22:27:25 <dolphin> fgrep is used more than text editor :P
Jan 24 22:27:34 <cshore> dolphin: painful: painful?
Jan 24 22:28:13 <dolphin> they've lately been cleaning it up, but it's a mess
Jan 24 22:28:47 <dolphin> if openwrt felt bit scattered in the beginning, it's like scattered _and_ mixed up :)
Jan 24 22:30:33 <cshore> xMff: okay the patch is commited to LuCI
Jan 24 22:33:40 <cshore> xMff: is there any guide to the more arcane callback mechanisms and stuff...when I last tried to figure out how LuCI worked I got kind of lost in the dispatchers and things of that nature
Jan 24 22:34:17 <xMff> hehe
Jan 24 22:34:52 <xMff> no, would need to wirte some architectural docs first
Jan 24 22:36:32 <dolphin> soon the old build is finished
Jan 24 22:36:41 <dolphin> if it works, I'm happy, if not, I'll give up :P
Jan 24 22:38:03 <cshore> dolphin: that's what I keep saying about trying to get JTAG to work on the router I was given to work on
Jan 24 22:38:10 <xMff> cshore: the overall concept is quite simple but the implementation details are sometimes not easy to gasp, a lot of stuff also happens implicitly
Jan 24 22:38:53 <dolphin> heh
Jan 24 22:39:07 <dolphin> well, I'm not sure if I'm willing to spend on MIPS JTAG yet
Jan 24 22:39:34 <dolphin> I think it was you xMff who last time told me that the wifi on my router will be tricky to get to work
Jan 24 22:39:42 <xMff> yes
Jan 24 22:39:52 <cshore> dolphin: connectors, wire, and 100 Ohm resistor and a computer with a parallel port
Jan 24 22:39:53 <dolphin> and how useful would the router be without wifi :P
Jan 24 22:40:02 <dolphin> no parallel port
Jan 24 22:40:25 <dolphin> could get PCI express one, but those are damn expensive (parallel ports that are)
Jan 24 22:40:51 <cshore> dolphin: hmmm...ironically my dual core came with a parallel port
Jan 24 22:41:04 <cshore> dolphin: and I don't need there
Jan 24 22:41:52 <dolphin> I'd be even happy to have the binary drivers working
Jan 24 22:53:10 <dolphin> right
Jan 24 22:53:21 <cshore> xMff: what do you think of making a usbroot not use an overlay (my doing so as part of the new package that is)
Jan 24 22:53:23 <dolphin> so someone fscked ramfs up between 18th and today
Jan 24 22:53:46 <dolphin> though, it still reports no initrd etc, but works
Jan 24 22:54:15 <cshore> xMff: speaking of ramdisk, does ramdisk work on x86?
Jan 24 22:54:41 <cshore> dolphin: it was working on brcm63xx yesterday
Jan 24 22:55:01 <dolphin> well, it wasn't working for rt288x today :P
Jan 24 22:55:13 <cshore> dolphin: but I couldn't get it going on x86
Jan 24 22:55:56 <xMff> cshore: never tried ramdisk on x86
Jan 24 22:56:24 <cshore> xMff: what kernel version is x86 at?
Jan 24 22:56:52 <xMff> 2.6.31.12
Jan 24 22:56:54 <cshore> dolphin: I'm wondering if it's kernel version thing...brcm63xx is at .30
Jan 24 22:57:19 <dolphin> so is rt288x
Jan 24 22:57:34 <dolphin> 30.10 to be specific
Jan 24 22:57:58 <cshore> dolphin: I believe brcm63xx is the same
Jan 24 22:58:10 <cshore> dolphin: so that shoots that theory
Jan 24 22:59:21 <dolphin> I tried patching the old version and going to run it
Jan 24 23:01:03 <CIA-88> jow * r19320 /packages/net/miniupnpd/Makefile: [packages] miniupnpd: add the kernel headers shipped by iptables to CFLAGS
Jan 24 23:01:28 <xMff> Ascendion: this commit should fix the miniupnpd issue
Jan 24 23:02:22 <xMff> cshore: why no overlay?
Jan 24 23:03:36 <cshore> xMff: I was thinking it would make things easier, but I realized that's not so...but with USB there'd be enough room that it would likely not be needed for space
Jan 24 23:04:03 <xMff> cshore: why not
Jan 24 23:04:34 <cshore> xMff: er, as in why not make it easier, or why not needed for space?
Jan 24 23:04:50 <xMff> why not as in let's try it
Jan 24 23:04:57 <cshore> xMff: okay
Jan 24 23:05:11 <xMff> one advantage of overlay is that for example the config stays in sync
Jan 24 23:06:02 <cshore> xMff: that's true...but with USB root you're probably applying additional packages anyway
Jan 24 23:06:10 <xMff> yea
Jan 24 23:07:10 <cshore> xMff: keeping the packages with the right version for the firmware is going to interesting
Jan 24 23:07:56 <cshore> xMff: I was thinking of have a separate binary vs. user data partition scheme, so that the binary was treated like jffs2 is now
Jan 24 23:08:07 <cshore> xMff: which is to say wiped on flash
Jan 24 23:08:30 <cshore> xMff: and to have an auto-download of packages specified and build time, as part of the firstboot
Jan 24 23:08:40 <cshore> and=at
Jan 24 23:12:16 <cshore> xMff: but I'm not sure how best to specify the packages to add....it'd be nice if I could select the the same way we select packages for inclusion with the image, only with a different letter (e.g. y-*, m-M, n-' ', u-U = USB)
Jan 24 23:12:34 <cshore> in menuconfig
Jan 24 23:13:45 <cshore> xMff:  but I think what's actually possible, is to specify file with each package on a separate line
Jan 24 23:14:04 <cshore> xMff: and just opkg install $(cat file)
Jan 24 23:15:11 <xMff> yes
Jan 24 23:16:02 <dolphin> I can't wait for all the money and glory I shall get by adding support for this
Jan 24 23:17:08 <cshore> dolphin: and girls >-)
Jan 24 23:19:24 <dolphin> someone going to push the patch in if it works?
Jan 24 23:20:10 <cshore> dolphin's excited now :-P
Jan 24 23:20:31 <dolphin> I sure am
Jan 24 23:20:45 <dolphin> it's 1:20 am, and I'm supposed to be working on qemu by 6:00 am
Jan 24 23:20:52 <cshore> ouch
Jan 24 23:20:58 <dolphin> and sleep in the between too :P
Jan 24 23:21:42 <dolphin> AHHH
Jan 24 23:21:58 <dolphin> MIPS: machine is Buffalo WZR-AGL300NH
Jan 24 23:22:24 <dolphin> and the mtd blocks exist
Jan 24 23:24:05 <dolphin> ohmygod /proc/mtd is identical to what's with the stock firmware
Jan 24 23:24:17 <dolphin> mtd5: 00000000 00000000 "rootfs_data" <- except this odd thing
Jan 24 23:24:43 <dolphin> cshore: go go, push the patch so I get to sleep
Jan 24 23:25:17 <dolphin> http://pastebin.com/m51d4df49
Jan 24 23:25:18 <cshore> dolphin: I don't have that commit access
Jan 24 23:25:36 <dolphin> someone must have
Jan 24 23:25:37 <cshore> dolphin, just brcm63xx and packages
Jan 24 23:25:41 <cshore> maybe xMff
Jan 24 23:26:02 <dolphin> xMff, blogic?
Jan 24 23:26:12 <dolphin> or nbd
Jan 24 23:27:15 <xMff> hmm
Jan 24 23:27:49 <xMff> state is: boots and has flashmap?
Jan 24 23:28:35 <dolphin> yes
Jan 24 23:28:54 <dolphin> but someone fscked up the trunks ramdisk support
Jan 24 23:29:21 <dolphin> that patch to add the code, and the board=WZR-AGL300NH modification to take the flash to use
Jan 24 23:29:56 <dolphin> like mentioned, it now produces an identical mtd map to the stock one
Jan 24 23:31:06 <dolphin> will try to stab the wired network code to work with this board too, and then with wifi and after that should be all set, but may take while...
Jan 24 23:31:39 <dolphin> there is an extra g-class wifi too, but I don't really need it myself, could of course try to add support for it if I have extra time
Jan 24 23:31:52 <CIA-88> jow * r19321 /trunk/target/linux/ramips/ (6 files in 4 dirs): [ramips] add initial support for the WZR-AGL300NH - patch by Joonas Lahtinen <joonas.lahtinen@gmail.com>
Jan 24 23:32:38 <dolphin> thanks, I wet my pants
Jan 24 23:33:07 <xMff> don't get used to it ;)
Jan 24 23:33:37 <dolphin> that working patches get accepted? I know it's bit rare in Open Source world, yes ;)
Jan 24 23:33:53 <xMff> no, I talk about the wet pants
Jan 24 23:35:14 <dolphin> "root@OpenWrt:/# mount /dev/mtdblock4 /tmp/original
Jan 24 23:35:14 <dolphin> SQUASHFS error: Major/Minor mismatch, older Squashfs 2.1 filesystems are unsupported"
Jan 24 23:35:21 <dolphin> great stuff
Jan 24 23:35:44 <xMff> probably a good time to back the stuff up
Jan 24 23:36:00 <dolphin> yes, now that I have netcat
Jan 24 23:36:01 <xMff> I think you can define an additional mtd partition that covers the whole flash area
Jan 24 23:36:06 <xMff> should make backup easier
Jan 24 23:36:45 <dolphin> one quick question
Jan 24 23:36:52 <dolphin> mtd4: 002f0000 00010000 "rootfs"
Jan 24 23:36:52 <dolphin> mtd5: 00010000 00010000 "user_property"
Jan 24 23:37:07 <dolphin> wouldn't the last one be rootfs_data by openwrt terms
Jan 24 23:37:29 <cshore> dolphin: would user_property be nvram?
Jan 24 23:37:31 <xMff> hm rrotfs_data is usually allocated in the free space left after rootfs
Jan 24 23:37:34 <dolphin> mtd4: 002f0000 00010000 "rootfs"
Jan 24 23:37:34 <dolphin> mtd5: 00000000 00000000 "rootfs_data"
Jan 24 23:37:34 <dolphin> mtd6: 00010000 00010000 "user_property"
Jan 24 23:37:43 <dolphin> yeah, that'd be it
Jan 24 23:38:05 <dolphin> there already are uboot and uboot_environ
Jan 24 23:38:18 <dolphin> + the factory partition which is of same size as uboot_environ
Jan 24 23:38:34 <dolphin> I think those are the nvram stuff
Jan 24 23:38:51 * cshore complains that his jtag cable tests fine but his telsey board doesn't give anything except 1's
Jan 24 23:39:01 <CIA-88> nbd * r19322 /trunk/package/madwifi/patches/ (7 files): madwifi: fix some potential null pointer derefs with wds
Jan 24 23:39:30 <cshore> dolphin: actually looking at the start address it's rather odd
Jan 24 23:39:46 <dolphin> those are size and erasize
Jan 24 23:39:56 <cshore> dolphin: oh, my bad
Jan 24 23:40:33 <cshore> dolphin: I think uboot_environ is only for uboot
Jan 24 23:40:40 <dolphin> http://pastebin.com/m6205a746
Jan 24 23:40:42 <cshore> dolphin: and factory could be factory defaults
Jan 24 23:40:50 <dolphin> indeed
Jan 24 23:40:59 <dolphin> I don't think there is any other place for user settings
Jan 24 23:41:14 <dolphin> that's from the original boot
Jan 24 23:42:11 <xMff> dolphin: what wireless chip is used on that board?
Jan 24 23:42:26 <dolphin> there's the integrated RT2880F
Jan 24 23:42:28 <Ascendion> hmmmmm the only locations that even have x_tables.h are in th build_dir tree, and the compile command is looking for all includes some place off of staging_dir
Jan 24 23:42:36 <dolphin> ah, wireless
Jan 24 23:42:46 <xMff> Ascendion: https://dev.openwrt.org/changeset/19320
Jan 24 23:42:51 <dolphin> I think that is integrated too, btw
Jan 24 23:43:27 <xMff> dolphin: is there a /proc/pci listing or something?
Jan 24 23:43:38 <dolphin> yep, just getting it
Jan 24 23:43:58 <dolphin> http://pastebin.com/m33fafa28
Jan 24 23:44:33 <dolphin> I should really get my ass up and document the disassembly / serial header and all the basic info of the stock firmware to a wiki page
Jan 24 23:44:39 <Ascendion> hmmm this time the update got the changes -- wonder why it didnt the last time around :)
Jan 24 23:44:57 <dolphin> RT2525 2.4GHz transceiver + RT2560 MAC/BBP
Jan 24 23:44:58 <xMff> Ascendion: I actually committed that change just a few minutes ago
Jan 24 23:45:16 <dolphin> and the first one, isn't listed in PCI lists, which makes me think it's integrated one
Jan 24 23:45:17 <Ascendion> NOW he commits the change :)
Jan 24 23:45:22 <dolphin> as would 2880 mean too
Jan 24 23:45:36 <xMff> dolphin: hmm rt2525... then you might even have luck with mac80211
Jan 24 23:45:52 <xMff> dolphin: there seems support for it already
Jan 24 23:46:15 <cshore> xMff: does hostapd support rt2661t?
Jan 24 23:46:24 * Ascendion dozes off while the compilers grind
Jan 24 23:46:35 <dolphin> xMff: both of those?
Jan 24 23:47:09 <xMff> dolphin: seems so
Jan 24 23:47:14 <dolphin> I think that would be the extra wifi I don't really care about, but great if it has support! ;)
Jan 24 23:47:22 <dolphin> because it reports a/g
Jan 24 23:47:28 <cshore> xMff: it'd be the nl80211 I think
Jan 24 23:47:31 <xMff> I mean it is maybe not fast or instable, but it would be something
Jan 24 23:47:37 <dolphin> sure
Jan 24 23:47:50 <dolphin> but I definitely need the n device working too
Jan 24 23:50:12 <xMff> cshore: I think you should check out rt2x00.serialmonkey.com
Jan 24 23:50:29 <xMff> http://rt2x00.serialmonkey.com/wiki/index.php/AP-mode_Howto
Jan 24 23:50:48 <dolphin> http://trac.fonosfera.org/fon-ng/browser/trunk/openwrt/target/linux/fonera2n/patches-2.6.21/100-ralink.patch
Jan 24 23:51:03 <dolphin> xMff: didn't you mention the support would be there?
Jan 24 23:51:16 <xMff> dolphin: using binary drivers
Jan 24 23:51:32 <dolphin> rt2860v2_ap           629680   1
Jan 24 23:51:51 <dolphin> is there open source one in the works?
Jan 24 23:52:10 <xMff> yes, the rt2x00 project, they develop a driver family using the mac80211 stack
Jan 24 23:53:08 <dolphin> wouldn't that like, limit the speed big time?
Jan 24 23:53:27 <xMff> maybe
Jan 24 23:54:15 <dolphin> my point #1 when buying this was, that it's capable of GB Ethernet and N-WLAN :P
Jan 24 23:54:35 <xMff> then you need to hack on 2.6.21
Jan 24 23:55:16 <dolphin> uff, so it isn't even blob + open source component?
Jan 24 23:55:44 <xMff> no, blob that wants to talk to a specific kernel version
Jan 24 23:55:53 <dolphin> argh
Jan 24 23:56:10 <dolphin> and it's 12MB anyway, stripping better help quite some :P
Jan 24 23:58:50 <dolphin> it's not like, it couldn't be reverse engineered I think, wonder how much sued it would get me
Jan 24 23:58:57 <cshore> xMff: does OpenWRT support the rt2x00 driver?
Jan 24 23:59:29 <nbd> yes
Jan 24 23:59:57 <cshore> nbd: great, now to see if I can get  serial terminal on that blasted thing....it doesn't like my jtag
Jan 25 00:00:06 <nbd> dolphin: the 11n wifi driver isn't quite there yet, but at some point i will resume my work on it
Jan 25 00:00:29 <dolphin> where is the beginning of it?
Jan 25 00:00:54 <nbd> the patches that i did were merged upstream
Jan 25 00:01:07 <nbd> it's just that the driver is still incomplete
Jan 25 00:01:23 <dolphin> any more specific pointer to look at
Jan 25 00:01:33 <dolphin> than it's "upstream" ? ;)
Jan 25 00:02:08 <xMff> http://git.kernel.org/?p=linux/kernel/git/linville/wireless-testing.git
Jan 25 00:02:12 <nbd> it's the rt2800pci/rt2800soc driver
Jan 25 00:02:25 <nbd> in openwrt it's compiled along with the mac80211 package
Jan 25 00:02:28 <nbd> which uses compat-wireless
Jan 25 00:03:37 <dolphin> so it'd bring the integrated wifi alive
Jan 25 00:04:08 <nbd> at some point i got as far as being able to scan
Jan 25 00:04:11 <dolphin> with device id 0x2880
Jan 25 00:04:25 <dolphin> sounds like a start
Jan 25 00:04:35 <dolphin> did you reverse engineer or what?
Jan 25 00:04:51 <nbd> no, i have access to sources of hw info
Jan 25 00:05:23 <nbd> and there's a reference driver for sta mode
Jan 25 00:05:24 <dolphin> under NDA?
Jan 25 00:05:27 <nbd> which is in staging
Jan 25 00:05:37 <nbd> and there are leaked copies of the internal driver
Jan 25 00:05:42 <nbd> in various gpl sources
Jan 25 00:05:46 <nbd> of routers
Jan 25 00:14:28 <cshore> how does one verify what pins correspond to what JTAG line?  My jtag cable checks out (with it connected to the board, I got from the back-of-the-board pin to the DB25 connect and it checks out)
Jan 25 00:19:05 <cshore> with it connect with what I think is pin1 and and the semi-standard 12-pin connector pinout it acts like there's nothing there (same results as having the jtag cable not connected).  reversed is all 0's so that doesn't help either
Jan 25 00:21:07 <cshore> (6-pins, no idea which is nRST, if it's not standard layout)
Jan 25 00:50:41 <CIA-88> cshore * r19323 /packages/libs/gconv-modules/Makefile:
Jan 25 00:50:41 <CIA-88> gconv-modules depends on glibc. Set gconv-modules version from glibc
Jan 25 00:50:41 <CIA-88> version.
Jan 25 00:50:41 <CIA-88> Signed-off-by: Zintis Petersons <zintis.petersons@e-mail.lv>
Jan 25 02:15:52 <builder> build #26 of x86-trunk is complete: Failure [failed compile_12]  Build details are at http://tksite.gotdns.org:8010/builders/x86-trunk/builds/26
Jan 25 02:42:52 <nbd> cshore: still awake?
Jan 25 02:43:02 <cshore> nbd: yep
Jan 25 02:43:51 <nbd> oh, nevermind. i just tested your preinit stuff, but the errors that i'm getting may be because i forgot to run make oldconfig ;)
Jan 25 02:44:31 <cshore> nbd: ok...Only about about 9:45 so I'm up for a while yet
Jan 25 02:45:30 <nbd> i think i'll do one more test and then go to bed
Jan 25 02:45:34 <nbd> 3:45 am here ;)
Jan 25 02:46:14 <cshore> nbd: I'm trying to get away from that kind of schedule myself...it's hard
Jan 25 02:47:56 <rtz2> nbd: is lua part of the base system or did I enable something that needs it?
Jan 25 02:48:54 <nbd> rtz2: ucitrigger is enabled by default, it requires lua
Jan 25 02:48:56 <nbd> can be disabled
Jan 25 02:49:08 <rtz2> ok, thanks
**** ENDING LOGGING AT Mon Jan 25 02:59:56 2010