**** BEGIN LOGGING AT Wed Jul 31 02:59:58 2013
Jul 31 03:16:25 <russell--> wgt634u build seems to be broken, boots to no shell on serial console
Jul 31 03:55:07 <builder> build #304 of rb532 is complete: Failure [failed shell_12]  Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/304
Jul 31 03:56:50 <builder> build #304 of ppc44x is complete: Failure [failed shell_12]  Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/304
Jul 31 04:08:26 * russell-- recalls seeing someone else complain about this, it's answering on the network, just not on serial
Jul 31 04:32:38 <russell--> not finding it now
Jul 31 04:33:39 <russell-->  /etc/inittab has ::askconsole:/bin/ash --login
Jul 31 04:33:51 <russell--> yet, no serial console shell, wtf?
Jul 31 04:39:33 <MartijnVdS> russell--: and everthing else *is* starting?
Jul 31 04:54:09 <russell--> MartijnVdS: yes
Jul 31 04:54:21 <russell--> afaict
Jul 31 04:57:04 <russell--> trying again with a vanilla config
Jul 31 05:10:45 <builder> build #318 of uml is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/uml/builds/318
Jul 31 05:27:44 <russell--> vanilla wgt634u build is dying with crypto/ocf/cryptosoft.ko errors
Jul 31 05:28:07 <russell--> ERROR: "crypto_alloc_ahash" [crypto/ocf/cryptosoft.ko] undefined!
Jul 31 05:50:33 <russell--> ERROR: "crypto_ahash_digest" [crypto/ocf/cryptosoft.ko] undefined!
Jul 31 05:50:33 <russell--> ERROR: "crypto_ahash_setkey" [crypto/ocf/cryptosoft.ko] undefined!
Jul 31 06:03:07 <russell--> disabling ocf lets it build
Jul 31 06:03:14 <russell--> not sure what that's giving up
Jul 31 06:03:29 <russell--> linux-ocf does not seem to be actively maintained anymore
Jul 31 06:15:38 <russell--> oh, interesting
Jul 31 06:16:48 <russell--> when i disabled ocf, serial console returns
Jul 31 07:06:12 <builder> build #323 of ppc40x is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/ppc40x/builds/323
Jul 31 07:10:29 <builder> build #288 of sibyte is complete: Failure [failed shell_12]  Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/288
Jul 31 08:55:07 <OpenWrt-Commits> nbd r37621 trunk/toolchain/ gcc/patches/4.6-linaro/040-gcc_bug_49696.patch gcc/patches/4.6.3/040-gcc_bug_49696.patch * gcc: merge a bugfix for a MIPS specific internal compiler error
Jul 31 09:02:22 <builder> build #337 of orion is complete: Failure [failed compile_8]  Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/337
Jul 31 09:52:52 <OpenWrt-Commits> wigyori r37622 trunk/target/linux/ imx23/patches/101-soc-audio-dts.patch imx23/modules.mk imx23/profiles/01-olinuxino.mk imx23/patches/100-soc-audio-support.patch * imx23: add SoC sound support
Jul 31 10:37:29 <Devastator> I'm open to suggestions on this: https://lists.openwrt.org/pipermail/openwrt-devel/2013-July/020862.html
Jul 31 10:44:56 <OpenWrt-Commits> jow r37623 trunk/package/system/ubox/Makefile * ubox: update to git head
Jul 31 13:57:53 <mazilo> Is it possible to port pofcontroller (http://www.poforwarding.org) to OpenWRT? It is a cross-platform, Java-based OpenFlow Controller based on the floodlight controller project.
Jul 31 14:04:37 <cyrusff> mazilo: java on openwrt?
Jul 31 14:05:16 <mazilo> I reckon so.
Jul 31 14:10:40 <OpenWrt-Commits> jow r37624 trunk/package/base-files/ files/sbin/sysupgrade Makefile
Jul 31 14:10:40 <OpenWrt-Commits> base-files: abort sysupgrade if file specified with --restore-config is not found
Jul 31 14:28:56 <zinx> wheee building for BR-6475nD again, going to apply wlan patches once I make sure I got all the changes right
Jul 31 14:29:37 <zinx> Devastator: just add an extra param for size :/
Jul 31 14:30:21 <zinx> you should see the madness the BR-6475nD requires :P
Jul 31 14:56:18 <builder> build #257 of iop32x is complete: Failure [failed shell_12]  Build details are at http://buildbot.openwrt.org:8010/builders/iop32x/builds/257
Jul 31 15:23:49 <Devastator> zinx how?
Jul 31 15:44:57 <aport> is compiling against an external kernel tree supported? I see that you can tell the buildroot to do that, but nothing seems to work
Jul 31 15:45:11 <aport> make kernel_menuconfig fails with a perl syntax error
Jul 31 16:01:18 <builder> build #260 of ep93xx is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/ep93xx/builds/260
Jul 31 16:53:33 <zinx> uhhhhh
Jul 31 16:53:43 <zinx> how the hell am i supposed to specify where the rootfs is now
Jul 31 16:53:54 <jow_laptop> context?
Jul 31 16:54:10 <zinx> ralink, rt3883, moved to dts stuff
Jul 31 16:54:33 <zinx> openwrt wants a "factory" partition, so i can't specify a separate rootfs partition (it would overlap)
Jul 31 16:54:46 <zinx> erh, "firmware", sorry
Jul 31 16:54:52 <jow_laptop> huh?
Jul 31 16:54:57 <jow_laptop> mtd partitions can overlap just fine
Jul 31 16:55:02 <zinx> but dts won't allow it
Jul 31 16:55:35 <zinx> none of the dts files in ramips/dts do it, either - i have no idea how they're getting the location of the rootfs
Jul 31 16:55:47 <zinx> perhaps their image formats do not have the ridiculous restrictions mine does
Jul 31 16:56:06 <zinx> although, now that i look, it's saying there are no partitions at all >_<
Jul 31 16:56:20 <zinx> [    2.100000] of-flash 1c000000.nor-flash: do_map_probe() failed
Jul 31 16:56:22 <zinx> oops
Jul 31 16:56:51 <zinx> now if i can figure out why :/
Jul 31 17:11:57 <zinx> yyeah no idea why it's failing :(
Jul 31 17:20:56 <builder> build #290 of xburst is complete: Failure [failed compile_9]  Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/290
Jul 31 17:35:51 <Devastator> zinx can you help me figure this? you said to add an extra param for size, but how Makefile detect router flash size?
Jul 31 17:40:23 <zinx> Devastator: it doesn't
Jul 31 17:41:59 <OpenWrt-Commits> hauke r37625 trunk/target/linux/brcm47xx/base-files/etc/uci-defaults/09_fix_crc * brcm47xx: run mtd fixtrx linux every time
Jul 31 17:42:16 <Devastator> can you describe your idea?
Jul 31 17:42:24 <zinx> Devastator: add a parameter, provide the correct value.
Jul 31 17:43:01 <OpenWrt-Commits> hauke r37626 trunk/package/kernel/ broadcom-wl/patches/012-compat-3.10.patch broadcom-wl/Makefile * broadcom-wl: fix compile error with kernel 3.10
Jul 31 17:43:55 <zinx> so cfi_probe_chip is failing to find anything, but the pre-existing debug isn't printing anything either
Jul 31 17:44:33 <Devastator> there must be something wrong with spi config
Jul 31 17:44:37 <zinx> looks like it's only probing 0x0, not 0x1c000000 :x
Jul 31 17:45:11 <zinx> hmm maybe not
Jul 31 17:45:49 <Devastator> anyway, I fail to understand how to fix my own problems, better not try to help somebody else's problems heheh
Jul 31 17:46:03 <KanjiMonster> zinx: dts doesn't stop you from defining overlapping partitions; and it works fine for me
Jul 31 17:46:53 <zinx> KanjiMonster: errors with overlapping partitions here.
Jul 31 17:47:24 <Devastator> if on the same node, it will give you error
Jul 31 17:47:58 <zinx> so leave out kernel :/
Jul 31 17:48:06 <KanjiMonster> zinx: the problem is likely you try to use two nodes with the same name (e.g. two partition@X { }; nodes)
Jul 31 17:48:15 <zinx> anyway, that's not going to help much if flash isn't being detected
Jul 31 17:48:52 <Devastator> I've hacked m25p80.c for mine hehe
Jul 31 17:48:57 <KanjiMonster> but works fine if you give them different names (e.g. kernel@X { reg = <X Y>; }; linux@X { reg = <X Z>; }
Jul 31 17:49:50 <KanjiMonster> zinx: btw, are you sure this is parallel flash and not e.g. spi flash?
Jul 31 17:49:53 <zinx> KanjiMonster: still doesn't help if flash isn't being detected at all :)
Jul 31 17:50:00 <zinx> KanjiMonster: absolutely certain
Jul 31 17:50:18 <zinx> (i had all this working before with the non-dts stuff)
Jul 31 17:50:23 <KanjiMonster> maybe wrong bus size?
Jul 31 17:50:30 <zinx> nope, bus size is correct, verified that as well
Jul 31 17:51:25 <Devastator> hum.. check cfi_cmdset_002 I think
Jul 31 17:51:44 <Devastator> or cfi_util.h
Jul 31 17:51:47 <zinx> cfi_cmdset_002 is bugged on this router
Jul 31 17:52:01 <Devastator> I can't remember which one has the flashchips definitions
Jul 31 17:52:08 <zinx> or at least it was with ralink's kernel
Jul 31 17:52:31 <zinx> or i'm misreading something
Jul 31 17:52:51 <Devastator> KanjiMonster can you help with my logic problem? :)
Jul 31 17:53:37 <KanjiMonster> Devastator: what logic problem?
Jul 31 17:53:58 <zinx> yeah misreading
Jul 31 17:54:26 <zinx> don't know that it's getting that far :x
Jul 31 17:55:07 <Devastator> KanjiMonster how to make this code suitable for both 4M flash and 8M flash: https://lists.openwrt.org/pipermail/openwrt-devel/2013-July/020862.html
Jul 31 17:55:57 <KanjiMonster> Devastator: relplace the hardcoded 8meg with $(4), then add it to all call sites?
Jul 31 17:56:06 <KanjiMonster> ah wait
Jul 31 17:56:13 <KanjiMonster> you also need to change the default layout
Jul 31 17:56:37 <Devastator> and the offset to the maximum image size
Jul 31 17:59:15 <KanjiMonster> so maybe then replace the layout with $(4), size with $(5) then define BuildFirmware/Seama/squashfs8M as $(Call BuildFirmware/Seama/squashfs,$(1),$(2),$(3),ralink_default_fw_size_8M,8060864) and do the appropriate one for 4M?
Jul 31 18:00:31 <Devastator> it's an option
Jul 31 18:02:12 <Devastator> or just another define with ralink_default_fw_size_4M?
Jul 31 18:02:41 <Devastator> to prevent 8M routers breakage :P
Jul 31 18:03:03 <Devastator> ugly solution though
Jul 31 18:12:02 <OpenWrt-Commits> jow r37627 trunk/package/base-files/ files/sbin/sysupgrade Makefile
Jul 31 18:12:02 <OpenWrt-Commits> base-files: implement -l (--list-backup) sysupgrade parameter to list the files that would be backed up
Jul 31 18:13:38 <Devastator> at least is a start, I'll post your suggestion to the mailing list, if it's ok by you
Jul 31 18:16:35 <builder> build #257 of mcs814x is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/mcs814x/builds/257
Jul 31 18:52:40 <zinx> so i've verified that it's not detecting the QRY with cfi->interleave, cfi->device_type, base, and map->bankwidth all correct :X
Jul 31 19:01:47 <zinx> it's reading zeros :/
Jul 31 19:04:35 <KanjiMonster> zinx: maybe you are missing some pinmux settings or so?
Jul 31 19:05:13 <zinx> maybe
Jul 31 19:05:41 <zinx> i'm not clear on how those are separated now, they used to all be set at the same time
Jul 31 19:08:31 <zinx> no documentation either, apparently
Jul 31 19:24:06 <zinx> KanjiMonster: nope, not that.
Jul 31 19:27:20 <Devastator> chip id? chip vendor also correct?
Jul 31 19:27:32 <Devastator> what was your kernel when it worked?
Jul 31 19:27:34 <zinx> no idea, that stuff is supposed to be detected for cfi
Jul 31 19:27:38 <Devastator> what is your kernel now?
Jul 31 19:27:42 <zinx> 3.7.x, 3.10.4
Jul 31 19:28:04 <zinx> also, boot log from the stock kernel:
Jul 31 19:28:14 <zinx> ralink flash device: 0x1000000 at 0x1c000000
Jul 31 19:28:14 <zinx> Ralink SoC physically mapped flash: Found 1 x16 devices at 0x0 in 16-bit bank
Jul 31 19:28:14 <zinx>  Amd/Fujitsu Extended Query Table at 0x0040
Jul 31 19:28:14 <zinx> number of CFI chips: 1
Jul 31 19:29:13 <Devastator> you will have to check that query table I guess
Jul 31 19:29:33 <Devastator> try to dig into kernel changelog to see the changes on mtd
Jul 31 19:29:37 <zinx> in 3.7.x i just did rt3883_register_pflash(0)
Jul 31 19:30:03 <zinx> as far as i can tell, it's all checking the same stuff
Jul 31 19:30:33 <zinx> Devastator: it's not even getting that far, something's pretty strange :/
Jul 31 19:31:18 <Devastator> try to compile the kernel with mtd debug
Jul 31 19:31:29 <zinx> i already did, and i've added a good bit of my own
Jul 31 19:34:27 <Devastator> I'm gonna post Kanji's suggestion to the mailing list, I don't know what else I can do to help you zinx
Jul 31 19:34:51 <zinx> no worries
Jul 31 19:44:38 <Devastator> sent, I hope I didn't top post this time :)
Jul 31 20:25:53 <builder> build #243 of mpc52xx is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/mpc52xx/builds/243
Jul 31 20:30:17 <builder> build #228 of octeon is complete: Failure [failed shell_12]  Build details are at http://buildbot.openwrt.org:8010/builders/octeon/builds/228
Jul 31 21:16:29 <russell--> how to debug procd?  i am seeing lack of serial console shell on some builds for brcm47xx, seeing an 'open' failure approximately where i normally see press-button-for-shell thing, but no clue what failed to open.
Jul 31 21:49:54 <russell--> procd: - init -
Jul 31 21:49:55 <russell--> open: No such file or directory
Jul 31 21:50:31 <russell--> can still ssh/telnet in (depending on files built into image)
Jul 31 22:00:42 <KanjiMonster> russell--: happened to me once, too, but I wasn't able to reproduce it again
Jul 31 22:07:06 * russell-- is able to reproduce at well :(
Jul 31 22:07:11 <russell--> will*
Jul 31 22:07:16 <OpenWrt-Commits> nbd r37628 trunk/package/network/services/uhttpd/Makefile
Jul 31 22:07:16 <OpenWrt-Commits> uhttpd: update to the latest version, fixes a crash on early cgi/lua process termination
Jul 31 22:07:20 <OpenWrt-Commits> nbd r37629 trunk/package/libs/libubox/Makefile
Jul 31 22:07:20 <OpenWrt-Commits> libubox: update to the latest version, fixes a fd deletion race condition in uloop
Jul 31 22:07:49 <OpenWrt-Commits> nbd r37630 branches/attitude_adjustment/package/libubox/Makefile
Jul 31 22:07:50 <OpenWrt-Commits> libubox: update to the latest version, fixes a fd deletion race condition in uloop
Jul 31 22:12:45 <russell--> why tf doesn't the open error message say what it's trying to open?
Jul 31 22:17:54 <zinx> so i'm beginning to think ioremap() is somehow not working :x
Jul 31 22:22:47 <KanjiMonster> russell--: it probably fails to open the tty
Jul 31 22:35:04 <russell--> why might that be?
Jul 31 22:35:28 * russell-- finds the CONFIG_CMDLINE in kernel_menuconfig, rebuilding
Jul 31 22:35:57 <russell--> added init_debug=2, also disabled some message silencing options
Jul 31 22:39:30 * russell-- assumes some subtle timing issue where /dev isn't populated when it goes to fire off the shell?  maybe the same thing that's borking my remote syslogs?
Jul 31 22:54:14 <russell--> init_debug=2 is no particular help
Jul 31 22:54:40 <russell--> no more verbose at the "open: No such file or directory"
Jul 31 22:59:16 <russell--> aha, running askfirst /dev/ttyS0 (from the ssh session) gives me a "Please press Enter to activate this console." on the serial console
Jul 31 22:59:48 <russell--> but it segfaults
Jul 31 23:01:25 <russell--> ah, i needed to give it something to run
Jul 31 23:31:19 <zinx> this is maddening :P
Jul 31 23:41:07 <zinx> i've verified that older kernel still boots and finds the flash just fine, via physmap + CFI probe
Jul 31 23:50:13 <russell--> printf debugging ftw
Jul 31 23:50:30 <zinx> sure
Jul 31 23:50:50 <zinx> but there's only so much you can do when reads from exactly the same location result in 0 instead of what they're supposed to result in
Jul 31 23:53:48 <russell--> zinx: sorry, /me was talking to myself
Jul 31 23:54:05 <zinx> oic :P
Jul 31 23:54:17 <zinx> anyone else here deal with rt3883 devices?
Jul 31 23:54:22 <zinx> specificly ones without spi flash
Jul 31 23:54:34 <russell--> "open: No such file or directory" appears to be coming from procd_inittab()
Aug 01 02:39:22 <builder> build #298 of avr32 is complete: Failure [failed compile_9]  Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/298
Aug 01 02:39:38 <builder> build #352 of at91 is complete: Failure [failed compile_5]  Build details are at http://buildbot.openwrt.org:8010/builders/at91/builds/352
**** ENDING LOGGING AT Thu Aug 01 02:59:58 2013