**** BEGIN LOGGING AT Tue May 17 02:59:58 2011
May 17 03:03:46 <gzanan> What's 004-ar71xx_spi_controller.patch exactly do?
May 17 03:21:50 <gzanan> jow,are you around?
May 17 05:17:28 <CIA-89> obsy * r26914 /packages/net/transmission/ (Makefile files/transmission.config files/transmission.init): [packages] transmission: update to 2.30
May 17 05:34:59 <builder_> build #27 of au1000 is complete: Exception [exception interrupted]  Build details are at http://buildbot.openwrt.org:8010/builders/au1000/builds/27
May 17 05:34:59 <builder_> build #35 of atheros is complete: Exception [exception interrupted]  Build details are at http://buildbot.openwrt.org:8010/builders/atheros/builds/35
May 17 05:35:01 <builder_> build #35 of cobalt is complete: Exception [exception interrupted]  Build details are at http://buildbot.openwrt.org:8010/builders/cobalt/builds/35
May 17 05:35:03 <builder_> build #34 of orion is complete: Exception [exception interrupted]  Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/34
May 17 06:02:48 <builder_> build #34 of pxcab is complete: Failure [failed compile_3]  Build details are at http://buildbot.openwrt.org:8010/builders/pxcab/builds/34
May 17 06:05:18 <builder_> build #34 of ps3 is complete: Failure [failed compile_3]  Build details are at http://buildbot.openwrt.org:8010/builders/ps3/builds/34
May 17 06:32:49 <builder_> build #34 of lantiq is complete: Failure [failed compile_4]  Build details are at http://buildbot.openwrt.org:8010/builders/lantiq/builds/34
May 17 07:50:29 <russell--> weird, i am having some trouble with ipv6 autoconf
May 17 07:50:58 <russell--> tcpdump sees the router advertisement message with the right prefix, but the node doesn't pick up the address
May 17 07:52:35 <xMff> russell--: accept_ra set to 2 on this interface in sysctl?
May 17 07:54:03 * russell-- looks
May 17 07:54:47 <russell--> accept_ra set to 1
May 17 07:56:33 <xMff> try 2
May 17 07:56:43 <xMff> that wll make it accept them even if forwarding is on
May 17 07:57:29 <xMff> if that works you can set "option accept_ra 1" (will be translated to 2) in the corresponding interface section
May 17 07:57:40 <russell--> i think forwarding is off though
May 17 07:58:12 <russell--> yeah, ipv6 forwarding is zero
May 17 07:58:28 <russell--> trying accept_ra = 2
May 17 08:03:10 <russell--> echo 2 > /proc/sys/net/ipv6/conf/br-lan/accept_ra seems to have done the trick
May 17 08:04:02 <russell--> thanks for the pointer
May 17 08:04:17 <russell--> it would have taken a while to find that one :-)
May 17 08:04:44 <xMff> right now openwrt will set it to 2 if the interface is set to proto dhcp
May 17 08:05:03 <xMff> else not
May 17 08:05:49 <russell--> seems like a few weeks ago (or maybe months now) all i had to do was turn off ipv6 forwarding
May 17 08:06:49 <xMff> I found that the "all or nothing" approach is unrealisitical
May 17 08:07:02 <xMff> therfore I backported some changes to introduce the 2 value
May 17 08:07:52 <russell--> i'm noticing that the br-lan macaddrs aren't as stable as i might like either, actually, maybe i should just bite-the-bullet and set the address statically
May 17 08:08:16 <xMff> what are the members of br-lan ?
May 17 08:08:42 <russell--> eth0.x, wlanN, bat0, typically
May 17 08:09:09 <russell--> bat0 gets a randomized macaddr
May 17 08:09:29 <russell--> i suspect that's the source of the instability
May 17 08:10:02 <russell--> (bat0 comes from batman-adv, btw)
May 17 08:12:15 <xMff> I thought the bridge sticks to the mac of its first added port
May 17 08:12:35 <xMff> hm but maybe bat0 exists indeed before eth0.1 as its in kernel space
May 17 08:12:57 <xMff> and it will probably appear before eth0 in /proc/net/dev so its brought up first
May 17 08:14:18 <russell--> yeah, br-lan is getting bat0's macaddr
May 17 08:18:50 <KanjiMonster> xMff: for some reason bridges use the "lowest" mac from their members
May 17 08:20:22 <russell--> these are the lowest ;-)
May 17 08:20:26 <KanjiMonster> philipp64|laptop: pong
May 17 08:20:37 <russell--> aren't* i mean
May 17 08:21:03 <russell--> bat0 has some crazy high-numbered macaddr
May 17 08:21:20 <KanjiMonster> ah, hm
May 17 08:21:22 <russell--> eg da:c1:41:7e:a5:2c
May 17 08:21:40 <KanjiMonster> xMff: ah, hm that seems to be fixed
May 17 08:22:11 <xMff> then its just timing I suspect
May 17 08:22:24 <xMff> russell--: bat0 comes before eth0 in /proc/net/dev ?
May 17 08:23:17 <russell--> nope
May 17 08:23:21 <russell--> next to last
May 17 08:23:27 <xMff> k, hm
May 17 08:23:48 <xMff> well I think you really have to fix the mac for now
May 17 08:23:56 <russell--> yeah
May 17 08:24:21 <russell--> ip link set br-lan address 00:ba:dc:0f:fe:e0 works
May 17 08:24:39 <russell--> and i understand there's a uci option too?
May 17 08:27:20 <KanjiMonster> macaddr
May 17 08:27:56 <russell--> yeah
May 17 08:39:55 <russell--> oh, that's interesting.  i change the br-lan and the bat0 macaddr changes along with it!
May 17 08:40:29 <CIA-89> nbd * r26915 /trunk/package/mac80211/patches/571-ath9k_fix_adhoc_nexttbtt.patch: ath9k: fix ad-hoc nexttbtt calculation, which broke beacon transmission in some instances
May 17 08:44:54 <CIA-89> nbd * r26916 /branches/backfire/package/mac80211/patches/ (2 files): mac80211: refresh patches
May 17 08:44:55 <CIA-89> nbd * r26917 /branches/backfire/package/mac80211/patches/ (2 files): ath9k: backport beacon related fixes from r26913, r26915
May 17 08:53:43 <russell--> xMff: /me's previous observations were on a ubiquity bullet M5HP, wgt634u br-lan macaddr does not track bat0 afaict
May 17 08:53:53 <russell--> ubiquiti*
May 17 09:51:33 <builder_> build #38 of s3c24xx is complete: Failure [failed compile_10]  Build details are at http://buildbot.openwrt.org:8010/builders/s3c24xx/builds/38
May 17 10:36:37 <CIA-89> blogic * r26918 /trunk/tools/ (5 files in 3 dirs):
May 17 10:36:37 <CIA-89> [tools]
May 17 10:36:37 <CIA-89> * adds mkelfimage
May 17 10:37:07 <builder_> build #25 of etrax is complete: Failure [failed compile_3]  Build details are at http://buildbot.openwrt.org:8010/builders/etrax/builds/25
May 17 10:38:05 <CIA-89> blogic * r26919 /trunk/target/linux/x86/ (12 files in 7 dirs):
May 17 10:38:05 <CIA-89> [x86]
May 17 10:38:05 <CIA-89> * adds support for Artec Thincan
May 17 10:38:05 <CIA-89> Signed-of-By: Lauri Võsandi <lauri.vosandi@gmail.com>
May 17 10:44:11 <blogic> loswillios: ping
May 17 10:44:14 <blogic> http://patchwork.midlink.org/patch/1000/
**** ENDING LOGGING AT Tue May 17 10:44:26 2011

**** BEGIN LOGGING AT Tue May 17 10:44:26 2011
**** ENDING LOGGING AT Tue May 17 10:45:25 2011

**** BEGIN LOGGING AT Tue May 17 10:45:53 2011
May 17 10:51:28 <CIA-89> blogic * r26920 /trunk/package/kernel/modules/crypto.mk:
May 17 10:51:28 <CIA-89> [modules]
May 17 10:51:28 <CIA-89> * add missing aes dependency to mv_cesa
May 17 10:51:28 <CIA-89> Signed-off-by: Jan Willies <jan@willies.info>
May 17 10:52:35 <blogic> loswillios: please verify
May 17 10:53:06 <blogic> i only did compile time testing, so try to flash the image to verify runtime please
May 17 11:11:44 <loswillios> blogic: will do
May 17 11:12:42 <CIA-89> swalker * r26921 /packages/net/miniupnpc/ (3 files in 2 dirs):
May 17 11:12:42 <CIA-89> [packages] miniupnpc: update to 1.5
May 17 11:12:42 <CIA-89> * add libminiupnpc package
May 17 11:12:42 <CIA-89> * use PKG_INSTALL & MAKE_FLAGS
May 17 11:12:42 <CIA-89> * add copyright info
May 17 11:12:42 <CIA-89> * refresh patches
May 17 11:13:03 <blogic> loswillios: thx :)
May 17 11:14:52 <loswillios> blogic: ah, misread. I did verify it already.
May 17 11:15:24 <loswillios> hunting down a crc32c loading dependency issue now
May 17 11:15:47 <CIA-89> matteo * r26922 /trunk/target/linux/ar71xx/files/drivers/net/ag71xx/ (ag71xx.h ag71xx_ar7240.c ag71xx_phy.c): ar71xx: detect link on LAN ports
May 17 11:17:06 <CIA-89> nbd * r26923 /trunk/package/mac80211/patches/572-ath9k_tx_last_beacon.patch: ath9k: implement tx_last_beacon() to allow mac80211 to respond to probe requests in ad-hoc mode without creating too much spam
May 17 11:18:15 <CIA-89> matteo * r26924 /trunk/target/linux/ar71xx/files/arch/mips/ar71xx/mach-ubnt.c: ar71xx: don't probe the LAN PHY
May 17 11:27:49 <builder_> build #24 of ep93xx is complete: Failure [failed compile_4]  Build details are at http://buildbot.openwrt.org:8010/builders/ep93xx/builds/24
May 17 11:40:08 <swalker> nbd: the lantiq build is broken with r26911
May 17 11:41:09 <nbd> i'll look into it later, thanks
May 17 12:05:53 <jow_laptop> KanjiMonster: can you comment on https://dev.openwrt.org/ticket/9403 ?
May 17 12:07:00 <KanjiMonster> jow_laptop: unfortunately no, I don't know much about ath9k
May 17 12:07:13 <KanjiMonster> well, I could comment that ;)
May 17 12:16:16 <jow_laptop> KanjiMonster: oh maybe I confused it
May 17 12:16:54 <jow_laptop> KanjiMonster: erm no, I meant the actual issue with the eraseblock size
May 17 12:17:08 <jow_laptop> users tend to throw totally unrelated issues into tickets
May 17 12:17:41 <jow_laptop> so apart from the ath9k comments, is the underlying issue fixed by the proposed patch and is the proposed patch in trunk now?
May 17 12:17:43 <KanjiMonster> jow_laptop: see my patch sent to -devel?
May 17 12:18:08 <KanjiMonster> jow_laptop: http://patchwork.midlink.org/patch/991/
May 17 12:18:24 <jow_laptop> ah ok
May 17 12:19:08 <voyo> hello
May 17 12:19:26 <KanjiMonster> for backporting to backfire the actual flash driver/definition needs also to be backported; it currently doesn't even get recognized there
May 17 12:19:59 <jow_laptop> good
May 17 12:21:10 <voyo> I have stupid question to developers ;)    why kernel modules are not included all (like with make modules_install), but you need to create package for each one ? All kernel options are separate anyway, with kernel_menuconfig, so what it does matter ?
May 17 12:23:36 <jow_laptop> because they would never fit on an ordinary openwrt device?
May 17 12:24:09 <jow_laptop> also merely enabling modules causes side effects in the core kernel iamge reaching from simple size increases to lowered network throughput
May 17 12:25:12 <jow_laptop> and kernel_menuconfig is useful for special use cases or porting, it is not intended for normal operation
May 17 12:25:17 <voyo> you dont have to enable ALL modules. what I mean - you know which modules you enabled in kernel_menuconfig, so deal with this.
May 17 12:25:34 <jow_laptop> there is no way to programamtically map enabled symbols to modules
May 17 12:26:01 <jow_laptop> and your proposal lacks ways to support binary distributions and repositories
May 17 12:27:33 <voyo> its not a proposal, just a question.  Im just bored creating another one module package for another missing module I need ;)
May 17 12:28:06 <voyo> but talking about prosal - what about some option in menu_config, like "install all configured kernel modules" ?
May 17 12:28:20 <jow_laptop> make a patch which can be discussed
May 17 12:30:34 <jow_laptop> I can't imagine how it would work right now
May 17 12:30:42 <voyo> probably I'll do this anyway , but Im not a owrt developer , I still lack of important knowedledge , and my patch probably will be not well welcomed
May 17 12:31:57 <jow_laptop> another issue unaddressed with the "install all" approach is the module autoloading
May 17 12:32:06 <voyo> maybe its all because my strange needs, Im using owrt as "distribution" builder, for my own  router.  Im not working on ordinary (small) arm boxes.
May 17 12:32:08 <jow_laptop> openwrt has no modprobe and depmod (way too big)
May 17 12:32:31 <voyo> yes it has, ie. from busybox
May 17 12:32:35 <jow_laptop> so without appropriate kmod definitions, stuff is not autoloaded
May 17 12:33:21 <voyo> but indeed, some time ago I added my own modutils package.. ;)
May 17 12:39:08 <voyo> who has made x86 targets ? (uml, generic, etc)
May 17 12:43:01 <CIA-89> loswillios * r26925 /packages/utils/cryptsetup/Makefile: [packages] cryptsetup: update to 1.3.0
May 17 12:43:35 <jow_laptop> I made xen
May 17 12:44:36 <jow_laptop> I mean the xen_domu subtarget
May 17 12:46:55 <voyo> ok, those targets are generaly for testing purposes , or its used for something 'production'  ?
May 17 12:46:55 <voyo> what I mean - having big image for x86 should be not a big problem, and sometimes its good , speedup a lot testings process.
May 17 12:57:37 <CIA-89> nbd * r26926 /trunk/scripts/metadata.pl: metadata: do not emit feature flag selects for targets which have subtargets - fixes a buildbot error cause by ssb not being blacklisted for lantiq falcon
May 17 13:10:23 <gzanan> I am building a 2.6.39-rc7 image but get some irq.c errors
May 17 13:16:16 <blogic> gzanan: let me get the vodoo priest
May 17 13:16:23 <blogic> he is good at reading remote minds
May 17 13:16:43 <blogic> maybe he can enlighten us about the arch you use or provide the pastebin link to the error you get
May 17 13:16:46 <blogic> :)
May 17 13:17:01 <gzanan> Here's are some errors I got
May 17 13:17:02 <gzanan> http://pastebin.com/uHwNJSfp
May 17 13:17:39 <jow_laptop> voyo: yes, we use kvm and xen_domu in some high availability balancer setups
May 17 13:19:59 <blogic> gzanan: ah
May 17 13:20:01 <blogic> .39
May 17 13:20:02 <blogic> :)
May 17 13:20:07 <blogic> ok, lmo changed the irq api
May 17 13:20:15 <blogic> well, not just kernel
May 17 13:20:40 <gzanan> blogic:get a solution?
May 17 13:21:03 <KanjiMonster> gzanan: use ath79 instead of ar71xx ;)
May 17 13:21:06 <blogic> yes
May 17 13:21:09 <blogic> use ath79
May 17 13:21:17 <blogic> its the upstream version of ar71xx
May 17 13:25:59 <gzanan> There's no signs of ath79 in makemenu.
May 17 13:26:33 <jow_laptop> right because its not integrated yet
May 17 13:34:07 <gzanan> It looks like 2.6.38 has integrated ath79?I copy 38's config and patches and do a slightly adaption.All the kernel modules and appz has compiled to ipk.
May 17 13:35:15 <gzanan> When come to the very last,may be linking all the stuff to an image,I got the above mentioned errors.
May 17 13:35:51 <jow_laptop> well at the very last the actual kernel is built
May 17 13:36:59 <gzanan> kernel_2.6.39-rc7-1_ar71xx.ipk and other kmod moudules has shown up in bin/packages.
May 17 13:37:21 <jow_laptop> kernel_* is jsut an empty meta package
May 17 13:37:48 <builder_> build #23 of mpc52xx is complete: Failure [failed compile_3]  Build details are at http://buildbot.openwrt.org:8010/builders/mpc52xx/builds/23
May 17 13:50:21 <gzanan> Under ath79,there are only two options: [*] Atheros AP81 reference board
May 17 13:50:23 <gzanan>   │ │    [*] Atheros PB44 reference board
May 17 13:50:50 <gzanan> No tp-link device...
May 17 13:50:57 <jow_laptop> yes
May 17 13:51:04 <jow_laptop> because ath79 was only partly merged
May 17 13:57:06 <gzanan> We can still add definitions of ar71xx to 2.6.39 like svn did?https://dev.openwrt.org/changeset/26746
May 17 14:04:03 <jow_laptop> probably
May 17 14:11:28 <KanjiMonster> IMHO porting the missing parts from ar71xx to ath79 for 2.6.38+ would make more sense, since we have to do it anyway sooner or later
May 17 14:12:25 <KanjiMonster> but it's a lot of work ;)
May 17 14:16:51 <gzanan> If I just pass through these irq errors,what can I do?
May 17 14:29:57 <acoul> @gzanan: http://git.ozo.com/?p=openwrt.git;a=commitdiff;h=b3264e6ad6462980c3e0c8c6e841c939171cfe05
May 17 14:30:08 <acoul> http://pastebin.ca/2062515
May 17 14:44:16 <gzanan> acoul:great!
May 17 15:03:14 <gzanan> Ozonet is a branch of openwrt or just a git source?
May 17 15:05:35 <acoul> it's openwrt with some tested bleeding edge addons
May 17 15:08:59 <gzanan> Fantastic!
May 17 15:20:57 <gzanan> acoul:Do you have time to port aria2 ? I can't get it work properly after r25706.
May 17 15:24:41 <gzanan> The one built by openwrt buildroot fail to download larger file.
May 17 15:27:42 <gzanan> gzanan@axishero-qq-810091899:~/ozonet$ git clone git://git.ozo.com/openwrt.git
May 17 15:27:43 <gzanan> git.ozo.com[0: 212.70.193.243]: errno=Connection timed out
May 17 15:27:52 <gzanan> :(
May 17 15:32:13 <acoul> try now ;-)
May 17 15:33:54 <gzanan> gzanan@axishero:~/ozonet$ git clone git://git.ozo.com/openwrt.git/target/linux/ar71xx
May 17 15:33:57 <gzanan> Cloning into ar71xx...
May 17 15:33:58 <gzanan> fatal: The remote end hung up unexpectedly
May 17 15:59:23 <RealOp> xMff, jow_laptop, Couldnt get my device to boot when i tested out the patches last night. I couldnt find my serial cable so idk what happened other than it was my patches that kept it from booting.
May 17 15:59:57 <RealOp> I suspect that the issue is that rootfs_data is used by that NULL pointer in the flash mappping
May 17 16:41:33 <loswillios> hm, is 2.6.39 using a different lzma compression (than 2.6.37)? I always get Uncompressing Linux...LZMA data is corrupt in uboot
May 17 17:32:19 <russell--> what's the right way to enable ATH5K_DEBUG_FOO?  There isn't a knob in the regular menuconfig.
May 17 17:41:39 <KanjiMonster> russell--: enable debug for ath-common - it enables debug for both ath5k and ath9k
May 17 17:49:53 <KanjiMonster> nbd: ping
May 17 18:07:48 <russell--> KanjiMonster: ah yes, indeed.  the debug thingies i was looking for are toggleable in /sys/kernel/debug/ieee80211/phy0/ath5k/debug
May 17 18:15:07 <KanjiMonster> russell--: ah, so enable debugfs for mac80211 ;(
May 17 18:15:11 <KanjiMonster> ;)
May 17 18:20:25 <nbd> KanjiMonster: pong
May 17 18:22:35 <KanjiMonster> nbd: in package/mac80211/Makefile - CONFIG_ATH5K_DEBUG gets set to =y for the make_opts if either CONFIG_PACKAGE_ATH_DEBUG or CONFIG_PACKAGE_MAC80211_DEBUGFS is set, but for BUILDFLAGS only defined when the latter is enabled
May 17 18:26:45 <CIA-89> nbd * r26927 /trunk/package/mac80211/Makefile: mac80211: sync the CONFIG_ATH5K_DEBUG make override with the buildflags override (thx, KanjiMonster)
May 17 18:27:04 <KanjiMonster> yw
May 17 18:48:35 <philipp64|laptop> nbd: ping
May 17 18:49:54 <nbd> philipp64|laptop: pong
May 17 19:07:39 <builder_> build #39 of at91 is complete: Failure [failed compile_4]  Build details are at http://buildbot.openwrt.org:8010/builders/at91/builds/39
May 17 19:20:13 <philipp64|laptop> nbd: so please walk me throw what I need to change again to add Geos LED support?
May 17 19:20:25 <nbd> no time
May 17 19:20:38 <nbd> look at the parts, understand them, then complete your driver
May 17 19:20:53 <philipp64|laptop> that's a truism.
May 17 19:24:41 <builder_> build #38 of ubicom32 is complete: Failure [failed compile_3]  Build details are at http://buildbot.openwrt.org:8010/builders/ubicom32/builds/38
May 17 21:32:04 <Syburg> whois noz
May 17 21:34:16 <RealOp> xMff, i have the kernel oops log now. you around ?
May 17 21:37:19 <RealOp> KanjiMonster, hey
May 17 21:38:23 <KanjiMonster> hey RealOp
May 17 21:38:53 <voyo> hey, I have trouble creating some package for myself, Im getting this while compilation:
May 17 21:38:53 <voyo> checking libintl.h usability... no
May 17 21:38:53 <voyo> checking libintl.h presence... yes
May 17 21:38:53 <voyo> - can someone look at this for me pls ? :(
May 17 21:40:25 <RealOp> KanjiMonster, so im trying to code up a fix for a device.  do you have any experience with brcm947xx?
May 17 21:40:34 <KanjiMonster> RealOp: limited
May 17 21:40:59 <voyo> http://pastebin.com/4GCEURH2
May 17 21:41:07 <KanjiMonster> voyo: look at the configure.log; there should be some output on why it fails to use it
May 17 21:41:49 <RealOp> KanjiMonster, ok well im trying to make another partition in the flash mapping.
May 17 21:42:45 <voyo> KanjiMonster: ha actually thats good idea..
May 17 21:42:59 <RealOp> lol?
May 17 21:43:17 <RealOp> KanjiMonster, http://paste.ubuntu.com/609194/ <- take a peek to see if you notice any bad coding
May 17 21:43:29 <RealOp> plz^
May 17 21:45:20 <RealOp> KanjiMonster, im thinking that the line 17 is causing a kernel panic. idk im lost.
May 17 21:46:04 <KanjiMonster> RealOp: you could try allocating the array dynamically
May 17 21:46:23 <KanjiMonster> RealOp: where exactly does it panic?
May 17 21:46:45 <RealOp> number of CFI chips: 1
May 17 21:46:46 <RealOp> Flash device: 0x800000 at 0x1fc00000
May 17 21:47:30 <KanjiMonster> RealOp: don't you have some stack trace from the panic?
May 17 21:47:39 <RealOp> yes, ill pastebin
May 17 21:48:51 <RealOp> KanjiMonster, http://paste.ubuntu.com/609198/
May 17 21:49:20 <KanjiMonster> RealOp: okay, enable debug symbols for the kernel ;)
May 17 21:49:39 <KanjiMonster> then you'll get actual function names instead of addresses
May 17 21:49:39 <RealOp> lol i know thats why i didnt mention it before
May 17 21:50:19 <KanjiMonster> well, then it's time for a rebuild ;)
May 17 21:50:33 <xMff> actually there might be a way
May 17 21:50:45 <xMff> RealOp: did you clean yet?
May 17 21:51:31 <RealOp> xMff, kernel clean?
May 17 21:51:42 <xMff> ah nvm, you'll need a cross gdb so you can also just rebuild as well
May 17 21:51:44 <RealOp> or clean up whitespace from the code? i did that yes
May 17 21:54:10 <RealOp> xMff, enable kernel debugging  under kernel_makemenu -> kernel hacking. correct?
May 17 21:56:46 <voyo> KanjiMonster: I see in config.log :  /home/voyo/owrt-x86/owrt/build_dir/linux-x86_xrio/linux-2.6.31.6/include/linux/stddef.h:4:28: error: linux/compiler.h: No such file or directory
May 17 21:58:09 <xMff> RealOp: yea
May 17 21:58:18 <xMff> RealOp: no
May 17 21:58:27 <xMff> RealOp: global build settings in normal openwrt menuconfig
May 17 21:58:32 <RealOp> o
May 17 21:58:33 <voyo> not sure if that makes a sense.. its already file included from linux/ , and I have set CFLAGS=-I/home/voyo/owrt-x86/owrt/build_dir/linux-x86_xrio/linux-2.6.31.6/include/linux
May 17 21:58:33 <RealOp> kk
May 17 21:59:40 <RealOp> xMff, need to do a make clean or will it detect that change?
May 17 21:59:42 <voyo> pls help, I'v spend all day on this stupid package and Im still nowhere :(
May 17 21:59:51 <xMff> target/linux/clean
May 17 22:00:07 <xMff> voyo: seems the kernel was never compiled/prepared
May 17 22:00:07 <KanjiMonster> voyo: 2.6.31.6? sounds rather old, what openwrt version are you using?
May 17 22:01:06 <voyo> xMff: all packages (including kernel) are compiled. only this one. (its open-vmware-tools). I think its *really* badly writen for cross compiling..
May 17 22:01:21 <xMff> sure it is
May 17 22:01:40 <xMff> did you include kernel.mk in your openwrt makefile?
May 17 22:01:51 <voyo> xMff: Im messing hard with Makefile, I'll paste it all . 1sec
May 17 22:02:40 <voyo> http://pastebin.com/qzJLZdx0
May 17 22:03:19 <voyo> lots of comments.. from all day..
May 17 22:11:50 <builder_> build #35 of ps3 is complete: Failure [failed compile_3]  Build details are at http://buildbot.openwrt.org:8010/builders/ps3/builds/35
May 17 22:14:25 <voyo> xMff: any tip ?
May 17 22:14:39 <xMff> voyo: you shouldn't pass .../include/linux  but just .../include
May 17 22:15:16 <xMff> and get rid of that crazy overrides
May 17 22:16:18 <xMff> remove the Build/Configure section entirely and add your --without-foo and --disable-bar flags as  CONFIGURE_ARGS += --disable-foo ...
May 17 22:16:35 <xMff> then attempt a clean,compile cycle and paste the resulting config.log
May 17 22:17:33 <RealOp> flash0.nvramcopy     New CFI flash at 1C000000 offset 007E0000 size 64KB
May 17 22:17:34 <RealOp> flash0.nvram         New CFI flash at 1C000000 offset 007F0000 size 64KB
May 17 22:17:52 <xMff> RealOp: that looks ok so far
May 17 22:18:07 <voyo> xMff: ok, right,I'll made this change.   currently Im still getting:  configure: error: libintl.h not found. Make sure you have the gettext headers installed
May 17 22:18:18 <RealOp> playing with the CFE while it compiles, found that info
May 17 22:18:28 <xMff> voyo: that one can be solved by include $(INCLUDE_DIR)/nls.mk
May 17 22:18:41 <xMff> voyo: it will setup cflags accordingly
May 17 22:19:06 <xMff> include it after package.mk
May 17 22:19:24 <RealOp> on every reboot the cfe will replace nvram if it dont match nvramcopy, so thats why nvram commit dont stick after reboot
May 17 22:23:31 <voyo> xMff: hm.. seems Im missing this file (nls.mk),  I'll try to get it (or should I rather update my whole git repo ?)
May 17 22:23:53 <xMff> voyo: uh
May 17 22:25:04 <xMff> voyo: ok then just do  TARGET_CFLAGS += -I$(STAGING_DIR)/usr/lib/libintl/include
May 17 22:25:34 <voyo> I get  https://dev.openwrt.org/browser/trunk/include/nls.mk , trying it now
May 17 22:25:47 <xMff> won't help
May 17 22:26:03 <RealOp> xMff, http://paste.ubuntu.com/609210/
May 17 22:26:13 <xMff> as libintl and libiconv got splitted into libintl-full and stub + libiconv-full and -stub
May 17 22:26:13 <RealOp> looks as useless as the last one? wtf
May 17 22:26:17 <xMff> they also got relocated
May 17 22:26:47 <xMff> RealOp: make menuconfig, global build settings, build kernel with symbol table information   make target/linux/clean world
May 17 22:26:52 <voyo> xMff: oki, not helped. trying with this TARGET_CFLAGS now..
May 17 22:27:47 <RealOp> <3
May 17 22:44:27 <voyo> xMff: much better :) , but still not 100%, now I have yet  /home/voyo/owrt-x86/owrt/staging_dir/toolchain-i386_gcc-4.3.3+cs_eglibc-2.8/usr/bin/../lib/gcc/i486-openwrt-linux-gnu/4.3.3/../../../../i486-openwrt-linux-gnu/bin/ld: cannot find -lintl
May 17 22:44:27 <voyo> collect2: ld returned 1 exit status
May 17 22:44:54 <xMff> TARGET_LDFLAGS += -L$(STAGING_DIR)/usr/lib/libintl/lib
May 17 22:45:12 <voyo> I also cleaned up those overrides (left only LDFLAGS when saw this error)
May 17 22:46:46 <xMff> the other ldflag overrides are not needed
May 17 22:48:16 <voyo> xMff:  ok :)  small step forward
May 17 22:48:24 <RealOp> xMff, http://paste.ubuntu.com/609220/
May 17 22:48:26 <voyo> cannot find -liconv now, but I think I can manage now
May 17 22:48:35 <xMff> same principle
May 17 22:48:38 <voyo> yup
May 17 22:48:49 <RealOp> cant use that get_router function twice or error in that function?
May 17 22:49:02 <xMff> -I$(STAGING_DIR)/usr/lib/libiconv/include to cflags and  -L$(STAGING_DIR)/usr/lib/libiconv/lib to ldflahs
May 17 22:49:25 <xMff> exactly that is solved by the nls.mk btw in recent backfire and trunk
May 17 22:49:52 <xMff> so you can remove those TARGET_{C,LD}FLAGS += ... lines once you update
May 17 22:49:57 <voyo> yup ok , trying that now, I hope thats last errors..
May 17 22:51:48 <voyo> xMff: probably I will not upgrade soon yet, have my own git repo, with many own packages,  targets, and other changes. Cannot let myself for more issues rightnow...
May 17 22:51:59 <xMff> ok
May 17 22:53:33 <voyo> ok, that passed. now yet:  vmxnet.c:45:22: error: asm/page.h: No such file or directory
May 17 22:53:33 <voyo> should I add this to TARGET_CFFLAGS ?
May 17 22:53:44 <xMff> no
May 17 22:54:08 <xMff> this seems to be a kernel driver component
May 17 22:54:15 <voyo> yes
May 17 22:54:22 <xMff> you have to build it differently
May 17 22:54:40 <voyo> this whole package, is mix of tools + kernel drivers
May 17 22:56:51 <xMff> did you add the --with-linuxdir to CONFIGURE_ARGS ?
May 17 22:57:09 <voyo>         --with-linuxdir=$(LINUX_DIR) \
May 17 22:58:26 <xMff> paste your current makefile
May 17 22:58:53 <xMff> and also the full commandline producing the error above
May 17 23:00:11 <voyo> http://pastebin.com/e56aXN3v
May 17 23:00:24 <voyo> make V=99 package/open-vm-tools/{clean,compile}
May 17 23:00:40 <xMff> I don't mean that commandline
May 17 23:00:45 <xMff> I mean the gcc invocation
May 17 23:01:18 <xMff> also remove --target --host and --build
May 17 23:01:23 <voyo> http://pastebin.com/0PTipipV
May 17 23:01:23 <xMff> those are supplied automatically
May 17 23:01:41 <voyo> ok
May 17 23:02:33 <xMff> andd add  -I$(LINUX_DIR)/include  to TARGET_CFLAGS
May 17 23:03:23 <voyo> oki.. checking
May 17 23:07:39 <voyo> not helped , still missing files in asm/*
May 17 23:08:57 <xMff> well only asm/page.h ?
May 17 23:09:07 <voyo> I\ll try yet to add  -I$(LINUX_DIR)/arch/$(LINUX_KARCH)/include
May 17 23:09:18 <xMff> the error you posted failed on entirely different files
May 17 23:09:22 <voyo> vmxnet.c:45:22: error: asm/page.h: No such file or directory
May 17 23:09:22 <voyo> vmxnet.c:46:25: error: asm/uaccess.h: No such file or directory
May 17 23:09:22 <voyo> vmxnet.c:47:23: error: asm/delay.h: No such file or directory
May 17 23:09:54 <voyo> hm there is more,..
May 17 23:09:57 <voyo> /home/voyo/owrt-x86/owrt/build_dir/linux-x86_xrio/open-vm-tools-8.4.2-261024/modules/linux/shared/compat_autoconf.h:36:31: error: linux/autoconf.h: No such file or directory
May 17 23:10:48 <xMff> on modern kernels its  generated/autoconf.h
May 17 23:10:50 <RealOp> xMff, does 	if (!strncmp(boardnum, "04FN", 4) && boardtype == 0x042f && boardrev == 0x10 && boardflags == 0 && sdram_init == 0x0009) {  need another set of ( ) ??
May 17 23:11:17 <xMff> on older ones it was linux/config.h  iirc
May 17 23:12:06 <voyo> what do you suggest ?
May 17 23:12:59 <xMff> RealOp: no
May 17 23:13:28 <xMff> RealOp: however is boardnum properly initialized?
May 17 23:13:41 <RealOp> yes
May 17 23:13:47 <xMff> voyo: the kernel has been built correctly already=
May 17 23:13:48 <xMff> ?
May 17 23:14:13 <RealOp> u32 boardnum = 0;
May 17 23:14:29 <xMff> RealOp: wait
May 17 23:14:31 <RealOp> if (nvram_getenv("boardnum", buf, sizeof(buf)) >= 0)
May 17 23:14:31 <RealOp> 		boardnum = simple_strtoul(buf, NULL, 0);
May 17 23:14:40 <xMff> RealOp: u32 != string
May 17 23:14:49 <RealOp> oooo
May 17 23:15:01 <xMff> RealOp: calling strncmp on a u32 variable will try to dereference a random invalid memory address
May 17 23:15:21 <xMff> since the u32 value is assumed to be a memory address of the start of a char array
May 17 23:15:26 <voyo> xMff:  yes. under owrt. also this open-vm-tool is building against this kernel (in owrt directory) outside owrt  (not cross compiling).
May 17 23:15:49 <xMff> voyo: don't understand that sentence
May 17 23:15:52 <voyo> xMff:  thats whole build output http://pastebin.com/DEBa7BMb
May 17 23:16:15 <xMff> you compile on the target?!
May 17 23:16:56 <voyo> xMff:  I can do "make V=99"  without this package enabled. is that answer your question ?
May 17 23:17:03 <voyo> and all is build correctly
May 17 23:17:12 <xMff> RealOp: so boardnum is interpreted as interger, it can never be "0x4FN"
May 17 23:17:26 <xMff> erm *"04FN"
May 17 23:17:37 <RealOp> i understand
May 17 23:18:32 <xMff> RealOp: what is  nvram get boardnum ?
May 17 23:18:46 <xMff> voyo: I assume the module just does not work with your kernel version
May 17 23:19:17 <xMff> voyo: at this point its an incompatibility between the vmware source and your kernel verison, not a cross build issue
May 17 23:19:19 <RealOp> varies on each simpleshare as far as i can tell, it always begings with 04FN though
May 17 23:19:39 <xMff> are you sure its board*num* ?
May 17 23:19:45 <RealOp> yes
May 17 23:19:52 <xMff> even the broadcom vendor code reads that as integer
May 17 23:20:02 <xMff> 04FN makes no sense
May 17 23:20:24 <voyo> xMff:  I was ABLE to build this vmware package outside owrt (under ubuntu, and some old fedora). for this same kernel.
May 17 23:21:00 <RealOp> xMff, boardnum=04FN51804579_SD_C-04
May 17 23:21:09 <RealOp> https://dev.openwrt.org/ticket/9323
May 17 23:21:43 <xMff> voyo: well find out why "vmmemctl" says "Using 2.6.x kernel build system." and "vmxnet" does "Using standalone build system."
May 17 23:22:15 <voyo> xMff: I think something is wrong with some paths set... see line 38,  #error "No Module support in this kernel.   but surely modules are enabled for this kerne #error "No Module support in this kernel.  I can see that in "make kernel_menuconfig".
May 17 23:22:57 <xMff> voyo: no its not
May 17 23:23:11 <xMff> voyo: it built at least one *.ko sucessfully
May 17 23:23:43 <xMff> and them issing module support error is due to "linux/autoconf.h: No such file or directory"
May 17 23:23:43 <voyo> xMff:  yes indeed.. strange :-/
May 17 23:23:50 <xMff> the first error matters, not the last one
May 17 23:25:26 <RealOp> xMff, so change u32 boardnum
May 17 23:25:28 <RealOp> to
May 17 23:25:32 <RealOp> char *boardnum ??
May 17 23:26:12 <voyo> xMff: feck... got it..  thats my own patch... http://www.mail-archive.com/open-vm-tools-devel@lists.sourceforge.net/msg00679.html
May 17 23:27:05 <xMff> well
May 17 23:27:28 <xMff> the error you run into only happens if VM_KBUILD=no
May 17 23:28:01 <xMff> you can try this:
May 17 23:28:01 <xMff> MAKE_VARS += VM_KBUILD=26
May 17 23:28:49 <voyo> voyo@build:~/owrt-x86/owrt/build_dir/linux-x86_xrio$ grep 2.6.35 open-vm-tools-8.4.2-261024/* -r | wc -l
May 17 23:28:49 <voyo> 91
May 17 23:28:49 <voyo> $ uname -r
May 17 23:28:49 <voyo> 2.6.35-28-generic-pae
May 17 23:29:18 <voyo> while kernel under owrt is 2.6.31 ...
May 17 23:29:55 <xMff> ok
May 17 23:30:05 <xMff> this would explain some issues, yes
May 17 23:34:48 <voyo> xMff:  is any variable defining kernel version ? to use in Makefile
May 17 23:35:06 <xMff> LINUX_VERSION
May 17 23:36:40 <voyo> right, I think I tried this before already..
May 17 23:36:53 <voyo> #       --with-kernel-release=$(LINUX_VERSION) \
May 17 23:37:02 <voyo> seems Im chasing my tail..
May 17 23:42:23 <xMff> well just patch it
May 17 23:42:50 <xMff> vanilla sources do:  VM_UNAME = $(shell uname -r)  and  HEADER_DIR = /lib/modules/$(VM_UNAME)/build/include
May 17 23:43:12 <xMff> change it to:  VM_UNAME ?= $(shell uname -r)  and  HEADER_DIR ?= /lib/modules/$(VM_UNAME)/build/include
May 17 23:43:19 <xMff> then add in your openwrt makefile:
May 17 23:43:45 <xMff> MAKE_VARS += VM_UNAME=$(LINUX_VERSION) HEADER_DIR=$(LINUX_DIR)/include
May 17 23:45:37 <voyo> I applied my patch , to   package/open-vm-tools/patches/
May 17 23:45:37 <voyo> fix-kernel-include-path.patch
May 17 23:45:51 <xMff> your patch is incomplete
May 17 23:46:03 <voyo> hm.. ?
May 17 23:46:30 <xMff> nvm
May 17 23:46:37 <xMff> I don't think you need this patch at all
May 17 23:46:51 <xMff> just add  MAKE_VARS += HEADER_DIR=$(LINUX_DIR)/include
May 17 23:46:59 <voyo> hm...?
May 17 23:47:04 <voyo> ok I'll try
May 17 23:47:12 <voyo> without this patch ? oki..
May 17 23:48:43 <xMff> and VM_UNAME=$(LINUX_VERSION) too
May 17 23:48:55 <voyo> ok
May 17 23:48:55 <xMff> its used for the driver version string as it seems
May 17 23:49:07 <xMff> DRIVERNAME = $(DRIVER)-$(VM_UNAME)
May 17 23:49:23 <voyo> MAKE_VARS += VM_UNAME=$(LINUX_VERSION)
May 17 23:49:41 <xMff> yea
May 17 23:49:47 <xMff> and HEADER_DIR
May 17 23:49:56 <voyo> yup
May 17 23:53:53 <voyo> http://pastebin.com/eMsNVrdi
May 17 23:53:56 <voyo> no joy :-/
May 17 23:54:43 <voyo> thats what is set...
May 17 23:54:43 <voyo> TARGET_CFLAGS += -I$(STAGING_DIR)/usr/lib/libintl/include  -I$(LINUX_DIR)/include
May 17 23:54:43 <voyo>         --with-kernel-release=$(LINUX_VERSION) \
May 17 23:54:43 <voyo>         --with-linuxdir=$(LINUX_DIR)
May 17 23:54:57 <voyo> MAKE_VARS += HEADER_DIR=$(LINUX_DIR)/include
May 17 23:54:57 <voyo> MAKE_VARS += VM_UNAME=$(LINUX_VERSION)
May 17 23:56:15 <voyo> before at least one .ko file was compiled ;)
May 18 00:01:57 <xMff> moment
May 18 00:02:56 <xMff> well
May 18 00:03:30 <xMff> see modules/Makefile.am in the vmware source tarball
May 18 00:03:38 <xMff> it does not pass through HEADER_DIR
May 18 00:03:43 <xMff> only the VM_UNAME
May 18 00:04:12 <RealOpty> xMff, change u32 boardnum to char boardnum ??
May 18 00:04:19 <xMff> if you apply your patch you also have to patch the Makefile template a directory above to passthrough the LINUXDIR you introduced
May 18 00:04:26 <xMff> RealOpty: its not that easy
May 18 00:04:42 <xMff> RealOpty: read the string into another char * varaible and operate on that
May 18 00:05:12 <voyo> xMff: I removed my patch now, its without it
May 18 00:05:32 <xMff> voyo: yes and I now looked at the source tarball and noticed that its braindead
May 18 00:06:26 <voyo> xMff:  open-vm ?  yes agree. totaly unfriendly to crosscompilation
May 18 00:09:37 <xMff> the relevant part is:
May 18 00:09:37 <xMff> $(MODULES):
May 18 00:09:37 <xMff>     $(MAKE) VM_UNAME=$(KERNEL_RELEASE) MV=mv RM=rm    \
May 18 00:09:37 <xMff>        OVT_SOURCE_DIR=$(abs_top_srcdir) $(EXTRA_ARGS) \
May 18 00:09:37 <xMff>        -C "$(modulesrc)/$(MODULES_OS)/$@"
May 18 00:09:41 <xMff> in modules/Makefile.am
May 18 00:09:47 <xMff> your patch does not touch that
May 18 00:10:24 <xMff> if you add a LINUXDIR=$(LINUXDIR) to the command above it will problaby work
May 18 00:11:23 <voyo> hm..
May 18 00:14:08 <voyo> adding LINUXDIR=$(LINUXDIR) to modules/Makefile.am not helped
May 18 00:14:31 <xMff> well debug it a bit
May 18 00:14:42 <xMff> its clearly some issue in the variable passing of the vmware source
May 18 00:14:47 <xMff> the openwrt part is settled
May 18 00:15:00 <xMff> I am out for now
May 18 00:15:31 <xMff> as long as it whines about autoconf.h its not using the right linux dir
May 18 00:15:43 <voyo> yes agree
May 18 00:15:50 <voyo> tnx a lot for your help
May 18 00:16:19 <xMff> ideally find out how configure stores the linuxdir param
May 18 00:16:33 <xMff> and why its not propagated down to modules/*/Makefile
May 18 00:16:43 <xMff> I suspect its not apssed on in modules/Makefile
May 18 00:17:02 <voyo> hm.. possible.. oki, I'll check that
May 18 00:17:22 <voyo> Makefile:237: *** Recursive variable `LINUXDIR' references itself (eventually).  Stop.
May 18 00:17:33 <voyo> lol. we both should go sleep now ;)
May 18 00:23:21 <builder_> build #35 of pxcab is complete: Failure [failed compile_3]  Build details are at http://buildbot.openwrt.org:8010/builders/pxcab/builds/35
May 18 00:52:36 <RealOpty> xMff,
May 18 00:52:37 <RealOpty> mtd: partition "rootfs_data" created automatically, ofs=3E0000, len=410000
May 18 00:52:37 <RealOpty> 0x0000003e0000-0x0000007f0000 : "rootfs_data"
May 18 00:52:37 <RealOpty> 0x0000007e0000-0x0000007f0000 : "nvram"
May 18 00:52:37 <RealOpty> 0x0000007f0000-0x000000800000 : "nvrambackup"
May 18 00:52:42 <RealOpty> :)
May 18 00:54:48 <RealOpty> bleh
May 18 00:57:31 <RealOpty> rootfs_data is overlaping 'nvram'
May 18 01:02:34 <RealOpty> xMff, wheres the source that does the creation of rootfs_data?
**** ENDING LOGGING AT Wed May 18 02:59:57 2011