**** BEGIN LOGGING AT Mon Jul 01 02:59:58 2013 Jul 01 08:23:52 ping blogic Jul 01 08:24:36 tripolar: hi Jul 01 08:31:38 blogic: hi did you see the patch i sent to the ml? Jul 01 08:31:56 yes Jul 01 08:32:01 it has whitespace errors in it Jul 01 08:32:12 and it calls fsck without checking if the file exists Jul 01 08:32:19 i will reply on the ML in a few hours Jul 01 08:32:32 thats true Jul 01 08:32:42 should i check for e2fsck? Jul 01 08:32:54 i will mail some comments later on Jul 01 08:33:16 okay Jul 01 09:15:34 blogic r37113 branches/attitude_adjustment/ include/version.mk feeds.conf.default package/base-files/image-config.in Jul 01 09:15:35 AA: 12.09.01 Jul 01 09:15:35 Jul 01 09:15:35 set repo and version to 12.09.1 Jul 01 09:15:35 Signed-off-by: John Crispin Jul 01 09:15:37 Jul 01 10:05:42 build #308 of orion is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/308 Jul 01 10:27:10 where can I find valid values for ralink,led-polarity? router ports blinks fine, but it's lit even without cable Jul 01 10:29:56 another thing is, is wireless mac supposed to be on soc_wmac.eeprom? Jul 01 10:54:11 jow r37114 trunk/package/ network/services/hostapd/Makefile network/services/hostapd/files/hostapd.sh * hostapd: truncate default mac file before adding entries to it (#13797) Jul 01 11:04:38 jow r37115 trunk/package/libs/gettext-full/Makefile Jul 01 11:04:38 gettext-full: always use shipped libcroco, unbreaks build if a different libcroco is installed on the host (#12539) Jul 01 13:22:43 jogo r37116 trunk/ (22 files in 4 dirs) * kernel: update linux 3.10 to 3.10 final Jul 01 13:43:16 jow r37117 packages/ipv6/ndppd * Drop ndppd, it moved to the openwrt-oruting feed Jul 01 14:40:49 blogic: the feeds.conf.default points to svn://svn.openwrt.org/openwrt/branches/packages_12.09.1 but that doesn't exist (yet?) Jul 01 14:41:08 is it going to exist, or should I leave it at 12.09 for now? Jul 01 14:43:35 true Jul 01 14:43:39 i will fix that for 12.09 Jul 01 14:52:58 sorry, "fix that for 12.09" ==> feeds.conf.default will go back to pointing to 12.09? Jul 01 15:00:27 blogic r37118 branches/attitude_adjustment/feeds.conf.default * AA: fix feed url Jul 01 15:14:08 urh, why did the libjson version get changed from 0.9 to 0.11 _within_ the AA branch? Jul 01 15:14:23 it just says, "backported from trunk" Jul 01 15:14:36 but that changes the .so names even! Jul 01 15:17:55 karlp: because netifd updates depended on the newer version Jul 01 15:19:24 did that reallllly need to be done _within_ the "stable" aa branch? Jul 01 15:19:45 0.9 to 0.11 is a massive change in that project. Jul 01 15:20:42 not really much I can do about it anyway, track or die I guess :) Jul 01 15:24:18 I think it was a requirement for the new ipv6 stuff Jul 01 15:24:29 0.11 was meant to include a compatibility layer for this sort of thing, is that not being installed? Jul 01 15:24:58 the diff looks like it is only putting in the new names f0cdf7214aad65cc739dd7f7fff468eba4e047f0 Jul 01 15:25:11 the compat layer is not enabled Jul 01 15:25:16 or at least not packaged Jul 01 15:25:31 what do you need it for? Jul 01 15:26:09 my own apps that built on earlier parts of the AA branch, and also built againt backfire too Jul 01 15:27:31 i guess one way to deal with this would be to rename libjson to libjson-c Jul 01 15:27:39 and add a libjson package that selects libjson-c Jul 01 15:27:42 and contains the compat libs Jul 01 15:27:48 would that work for you? Jul 01 15:28:42 sounds like it should work, yeah. Jul 01 15:29:01 ok, then i'll implement it this way Jul 01 15:29:11 I want to get rid of backfire compat myself anyway, so sooner or later it can all update, I just wasn't really expecting this to change within the AA branch. Jul 01 15:37:49 nbd r37119 trunk/package/ libs/libubox/Makefile libs/libjson-c/Makefile Jul 01 15:37:49 libjson-c: rename the libjson binary package to libjson-c, add a new libjson package with the compatibility library (to deal with the rename) Jul 01 15:39:02 florian r37120 trunk/target/linux/ar71xx/image/Makefile * ar71xx: fix initramfs images generation after r37049 Jul 01 15:39:28 nbd r37121 branches/attitude_adjustment/ package/libubox/Makefile package/libjson-c/Makefile Jul 01 15:39:28 libjson-c: rename the libjson binary package to libjson-c, add a new libjson package with the compatibility library (to deal with the rename) Jul 01 15:39:32 karlp: fixed Jul 01 15:40:58 awesome, thanks. Jul 01 16:12:33 nbd: http://pastebay.net/1244648 is that how the compat's meant to work? Jul 01 16:12:51 I got a libjson and libjson-c .a/.so built ok, Jul 01 16:13:28 but the libjson.so is basically empty Jul 01 16:13:28 hrm, seems to be rpath crap Jul 01 16:13:40 libjson.so is supposed to be empty Jul 01 16:13:47 it should normally pull in libjson-c.so Jul 01 16:16:43 objdump -x on libjson.so says that libjson-c.so.0 is NEEDED, which is correct I think. Jul 01 16:23:20 is that rpath stuff that I'm meant to do in my own app? Jul 01 16:35:32 yes Jul 01 16:35:41 if you add it at at build time it should be fine Jul 01 16:36:07 try adding -Wl,-rpath,$(STAGING_DIR)/usr/lib Jul 01 16:36:13 sorry Jul 01 16:36:15 -rpath-link Jul 01 16:36:17 instead of -rpath Jul 01 16:46:12 nbd: no, same behaviour Jul 01 16:49:09 did you verify that it was passed to the linker command properly? Jul 01 16:50:51 looks good: http://pastebay.net/1244651 Jul 01 16:52:19 try getting it added before -ljson Jul 01 16:55:36 nope, no help. Jul 01 16:56:01 I tried --as-needed and --copy-dt-needed-entries too, but no luck yet Jul 01 16:58:32 hm Jul 01 16:58:42 i guess you will have to use -ljson-c then Jul 01 16:59:01 but with the compat stuff at least old builds of the package are binary-compatible with the new release Jul 01 17:11:03 yeah, I can make this work, it's just a lot more of a hassle than it shoudl have been. my fedora18 laptop only has 0.10 packaged, which still has the old names only, so now I need yet more customizations in the makefiles for no particularly good reason :| Jul 01 17:12:58 I fixed another package using libjson like this: $(CC) -o binary -ljson foo.o || $(CC) -o binary -ljson-c foo.o Jul 01 17:13:15 not very beautiful but extremely simple and allows you to avoid pkg-config etc. Jul 01 17:31:22 with cmake it's also very easy Jul 01 17:32:24 build #314 of brcm47xx is complete: Failure [failed shell_10 compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/brcm47xx/builds/314 Jul 01 17:32:37 build #310 of orion is complete: Failure [failed shell_10 compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/310 Jul 01 17:35:13 build #278 of rb532 is complete: Failure [failed shell_10 compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/278 Jul 01 17:35:30 build #278 of ppc44x is complete: Failure [failed shell_10 compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/278 Jul 01 17:37:40 cmake might be nice, but it's kinda beside the point. Jul 01 17:38:12 but, I'm just going to have to deal with it :) Jul 01 17:38:14 build #282 of x86 is complete: Failure [failed shell_10 compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/x86/builds/282 Jul 01 17:38:15 build #292 of uml is complete: Failure [failed shell_10 compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/uml/builds/292 Jul 01 17:38:21 it's far too late now for could have should have would have Jul 01 17:41:01 build #281 of ar7 is complete: Failure [failed shell_10 compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/ar7/builds/281 Jul 01 17:41:15 build #264 of sibyte is complete: Failure [failed shell_10 compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/264 Jul 01 17:43:31 build #274 of avr32 is complete: Failure [failed shell_10 compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/274 Jul 01 17:44:24 build #266 of xburst is complete: Failure [failed shell_10 compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/266 Jul 01 17:46:48 build #265 of ixp4xx is complete: Failure [failed shell_10 compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/ixp4xx/builds/265 Jul 01 17:46:57 build #264 of kirkwood is complete: Failure [failed shell_10 compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/kirkwood/builds/264 Jul 01 17:49:35 build #258 of ar71xx is complete: Failure [failed shell_10 compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/ar71xx/builds/258 Jul 01 17:49:43 build #259 of au1000 is complete: Failure [failed shell_10 compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/au1000/builds/259 Jul 01 17:57:23 build #240 of ep93xx is complete: Failure [failed shell_10 compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/ep93xx/builds/240 Jul 01 17:58:31 build #237 of iop32x is complete: Failure [failed shell_10 compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/iop32x/builds/237 Jul 01 17:58:45 florian r37122 trunk/include/prereq-build.mk * Revert "prereq-build: check for "bc" (#13767)" Jul 01 18:01:17 build #51 of mpc85xx is complete: Failure [failed shell_10 compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/mpc85xx/builds/51 Jul 01 18:06:18 build #231 of gemini is complete: Failure [failed shell_10 compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/gemini/builds/231 Jul 01 18:43:51 heffer: hi Jul 01 18:43:55 hi :) Jul 01 18:44:15 jogo r37123 trunk/target/linux/generic/config-3.10 * kernel: add missing config symbol for 3.10 Jul 01 18:44:26 jogo r37124 trunk/target/ (41 files in 2 dirs) * ixp4xx: add support for linux 3.10 Jul 01 18:52:35 heyho Jul 01 18:52:53 anybody an idea what hardware uses the i2s links of ar71xx chipsets? Jul 01 18:52:54 roh: aloha Jul 01 18:53:06 roh: not seen any apart of eval kits Jul 01 18:53:14 Hi all, I am trying to build a lantiq target, but I seem to be missing drv_mei_cpe-1.2.0.tar.gz, anyone knows how to obtain it? Jul 01 18:53:21 * roh wants to build a audio-link (proprietary/own standard/can be free) with ultra-low-latency spec Jul 01 18:53:27 The_Lizard: its on mirrro2 ? Jul 01 18:53:36 blogic: too bad.. any idea if there is code which uses them? Jul 01 18:53:43 blogic: roh: IIRC mikrotik uses it for the beeper present on some routerboards Jul 01 18:53:51 KanjiMonster: do they ? Jul 01 18:53:58 I think so, yes Jul 01 18:53:59 what i want to do is basically: building a wireless microphone for stage use out of it. Jul 01 18:54:29 The_Lizard: let me check Jul 01 18:54:42 use 5mhz or less wide carriers when possible, add some fec, and add some codec for differential mic in and differential out on the other side. latency should be <1-5ms Jul 01 18:54:57 KanjiMonster: a beeper on an i2s link? whut? Jul 01 18:55:09 blogic: ok Jul 01 18:55:13 using it as gpio and bitbang it for a buzzer? Jul 01 18:55:26 The_Lizard: http://downloads.openwrt.org/sources/drv_mei_cpe-1.2.0.tar.gz ? Jul 01 18:55:47 the morons of regtp take away all the wireless mic frequencies one by one.. the 700mhz band will be the next to go Jul 01 18:56:03 I guess that will do, should I place it in the /dl/ folder? Jul 01 18:56:08 yes Jul 01 18:56:16 but it should download it automagicallly really Jul 01 18:56:17 so people who use these for events are searching for alternatives.. and analog ones are expensive to replace /modify for different freq. Jul 01 18:56:57 I thought that, but apparently that didn't happen.. Jul 01 18:57:15 hmmm Jul 01 18:57:19 The_Lizard: copy it to dl/ Jul 01 18:57:25 which board do you have ? Jul 01 18:57:28 3370 ? Jul 01 18:57:51 VGV7519, which is in base pretty equal to the EASY80920 Jul 01 18:57:58 if somebody has another bright idea for other socs which could work for such a project, i'm all ears... currently i am in the investigations phase...see whats possible with what devboards. Jul 01 18:58:00 An arcadyan board Jul 01 18:58:16 dsl_cpe_control_vrx-4.11.4.tar.gz is missing as well Jul 01 18:58:17 roh: blogic: oh, misremembered, it's the SLIC they are using Jul 01 18:58:29 hrr Jul 01 18:58:31 hm, well should be i2c Jul 01 18:58:41 at least they enable i2c Jul 01 18:59:01 The_Lizard: http://downloads.openwrt.org/sources/ Jul 01 18:59:09 they are all there Jul 01 18:59:10 i see defines for the i2s (digital audio link) on the ar71xx patches.. but i fear there are no drivers yet Jul 01 18:59:13 yes, it was just for info.. Jul 01 18:59:46 weird Jul 01 18:59:51 i will retry that later on Jul 01 19:05:20 It were the only 2 files that were missing... it is building my tree now Jul 01 19:05:34 weird Jul 01 19:06:11 I saw it downloading a lot of other stuff later on.. Jul 01 19:11:07 The tree is built... time to try it on the device * starts praying * Jul 01 19:23:02 Cool, my first kernel panic... Jul 01 19:23:33 yay Jul 01 19:23:40 where does it panic ? Jul 01 19:23:47 Something with my fs Jul 01 19:23:59 possibly useful to start explaining some things about the flashes on board Jul 01 19:24:08 http://wiki.openwrt.org/toh/arcadyan/vgv7519 Jul 01 19:24:47 This is the page I have made for it Jul 01 19:25:19 There are 2 8MB nor flashes on the board Jul 01 19:25:32 one at 0xB0000000, one at 0xB4000000 Jul 01 19:25:45 Currently everything is in the first flash... Jul 01 19:26:06 it all starts with u-boot (from 0x0 to 0x40000) Jul 01 19:26:22 then the u-boot env (0x60000-0x70000) Jul 01 19:26:45 I decided to place my linux kernel at 0x80000 Jul 01 19:27:42 at the end i have placed a block of nvram (0x7f0000-0x80000) Jul 01 19:28:56 nvram ends at 0x800000 ;) Jul 01 19:29:09 In the linux block, while booting, it has decided that the linux partition should contain the kernel from 0x80000 - 0x270000 Jul 01 19:29:29 and the rootfs should be from 0x270000-0x7f0000 Jul 01 19:30:18 the error shown: SQUASHFS error: unable to read id index table Jul 01 19:30:40 with a list of partitions (8 in total) ending with driver? Jul 01 19:30:55 No filesystem could mount root, tried: squashfs Jul 01 19:33:08 blogic: I have flashed the root.squashfs file to my nor flash, at the correct location, so any idea? Jul 01 19:33:33 build #299 of ramips is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/ramips/builds/299 Jul 01 19:36:09 The_Lizard: I think you need to implement this http://lxr.linux.no/#linux+v3.10/drivers/mtd/maps/physmap.c#L96 in lantiq-flash.c, where it takes more than one resource and then concats them together to one big mtd device Jul 01 19:36:56 It will then use both flashes as one? Jul 01 19:39:49 I will give it a try... Jul 01 19:40:38 The_Lizard: the the trick then is to have two lines to reg, one for the first flash and one for the second one; i.e. reg = <0 0x0 0x800000>, <0 0x800000 0x800000>; Jul 01 19:44:02 KanjiMonster: what would be a good flash layout then? Just give all space away to the linux partition? Jul 01 19:44:56 The_Lizard: the linux partition should cover everthing that's usable, to ensure rootfs_data will be completely erased when flashing a new image Jul 01 19:45:04 usable to the os ;) Jul 01 19:45:10 so kernel + rootfs + rootfs_data Jul 01 19:45:47 ok, so adjustment to above: linux from 0x80000 size 0xf70000, nvram at 0xff0000 size 0x10000 Jul 01 19:46:36 why nvram ? Jul 01 19:46:48 just so you ca go back to the original fw ? Jul 01 19:47:39 build #320 of at91 is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/at91/builds/320 Jul 01 19:47:40 hi blogic Jul 01 19:47:43 what should i change to get the patch accepted? Jul 01 19:47:55 the stuff i said Jul 01 19:48:02 stat e2fck before i call it Jul 01 19:48:04 add a check for the fsck binary Jul 01 19:48:11 and the whitespace errors Jul 01 19:48:38 +}else if (pid > 0) { Jul 01 19:48:44 there is a missing space Jul 01 19:49:12 ahh okay Jul 01 19:49:34 blogic: I saw it in another flash layout, but as you say it, I guess it is unnecessary... Jul 01 19:55:33 The_Lizard: well its optional Jul 01 19:55:36 just wondered Jul 01 19:55:37 build #211 of octeon is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/octeon/builds/211 Jul 01 19:56:27 There was an other bootloader, but all of that is gone now, so no need for it to remain (I do have a flash dump) Jul 01 19:56:39 blogic: should i print an error if it doesn't exist? Jul 01 19:56:54 nope Jul 01 20:13:18 KanjiMonster: Should I basically just throw about everything over to the lantiq file? Jul 01 20:13:43 I see there are some fixes around reading and writing... I guess they should remain ;) Jul 01 20:20:05 blogic: the problem i just found now - busybox install e2fsck at /sbin/fsck and e2fsprogs at /usr/sbin/e2fsck - should i check for both? Jul 01 20:20:40 as fe Jul 01 20:21:11 fsck isn't enabled in the default image most users will use e2fsprogs Jul 01 20:52:15 KanjiMonster: I am unsure about what to do with the module stuff in the bottom of the file... Jul 01 20:53:37 florian r37125 trunk/include/kernel-defaults.mk * include: force a kernel-generated cpio removal and rebuild Jul 01 20:55:08 The_Lizard: the only parts relevant is that it iterates over the resources and then at the end if there are more than one found device, calls mtd_concat_create to make them seem like one device Jul 01 20:56:08 Ok, I have moved over almost all of the code... so I hope it works... most of it seems somewhat the same... Jul 01 21:06:11 I guess the part defined as CONFIG_MTD_PHYSMAP_COMPAT is required??? Jul 01 21:10:13 The_Lizard: no Jul 01 21:13:11 KanjiMonster: Not good: VFS: Cannot open root device "(null)" or unknown-block(0,0): error -6 Jul 01 21:13:42 The_Lizard: what did the lantiq probe say? Jul 01 21:14:05 I don't see a probe in my log :S Jul 01 21:14:18 I have seen it earlier (in another tree) Jul 01 21:14:21 also you should stay with booting initramfs kernels over network until you got the flash fixed Jul 01 21:14:39 so you don't accidentially kill anything important ;) Jul 01 21:14:40 I am using u-boot to load the uImage Jul 01 21:15:09 but appearently not a initramfs uImage Jul 01 21:15:27 shoud i take the vmlinux-initramfs? Jul 01 21:15:28 those come with an embedded rootfs in the kernel Jul 01 21:16:18 yes Jul 01 21:16:23 well, if it is an uImage Jul 01 21:17:10 how can I tell if it is an uImage? Jul 01 21:17:38 uboot will complain if it isn't Jul 01 21:17:51 well, lets try then ;) Jul 01 21:18:36 blogic: i sent a new version to the ml Jul 01 21:21:08 guess it isn't Jul 01 21:21:47 I do have some uImage-board-jffs2-size files... are these the ones? Jul 01 21:22:38 blogic: sent another version to fix another whitespace Jul 01 21:26:31 KanjiMonster: How do I make sure the driver will output it's debugging data? Jul 02 01:00:44 build #297 of ppc40x is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ppc40x/builds/297 Jul 02 01:30:39 build #308 of atheros is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/atheros/builds/308 Jul 02 01:42:19 build #298 of cobalt is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/cobalt/builds/298 **** ENDING LOGGING AT Tue Jul 02 02:59:58 2013