**** BEGIN LOGGING AT Tue Jul 23 02:59:59 2013 Jul 23 03:00:05 anyone have an rt5350 device? Jul 23 03:00:25 me Jul 23 03:04:21 does wifi work on it? Jul 23 03:04:34 yes Jul 23 03:05:01 but I couldn't configure proper mac address to the interface Jul 23 03:05:09 hmm Jul 23 03:05:16 which device? Jul 23 03:05:36 D-Link DIR-610 A1 Jul 23 03:06:43 i got this thing today: http://www.rakuten.com/pr/product.aspx?sku=245965249 Jul 23 03:07:24 i thought it would be largely the same as the tp-link device (ar71xx) Jul 23 03:31:49 build #310 of uml is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/uml/builds/310 Jul 23 05:30:03 is /etc/init.d/log stop not supposed to stop logread? Jul 23 05:30:14 because it doesn't seem to Jul 23 05:32:43 maybe the designers of procd should fill-in the wiki page that describes it? http://wiki.openwrt.org/doc/techref/procd Jul 23 07:40:39 blogic: pinging as requested Jul 23 07:41:21 yes Jul 23 07:41:26 i am not at work yet ;) Jul 23 07:41:33 okay, np Jul 23 07:41:35 lets chat in 2-3 hours Jul 23 07:41:36 ok ? Jul 23 07:41:47 build #280 of sibyte is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/280 Jul 23 07:41:58 * russell_ may be asleep then but we'll see Jul 23 07:42:14 * russell_ is on US west coast time Jul 23 07:42:33 plus hacker-offset Jul 23 08:11:55 russell_: otherwise wehn you wake up Jul 23 08:12:09 i'll be awake for at least another 12-14 hours Jul 23 08:15:42 blogic: okay, thanks! Jul 23 09:28:48 looking for a way to tune a 16M RAM openwrt router to survive a high network traffic Jul 23 09:55:40 build #282 of xburst is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/282 Jul 23 10:23:15 juhosg r37514 trunk/include/image.mk * image: fix a typo in JFFS2OPTS to fix jffs2 rootfs generation Jul 23 10:30:39 russell_: ing Jul 23 10:30:41 *ping Jul 23 10:31:51 tripolar r37515 packages/net/ git/Makefile git/patches/100-convert_builtin.patch git/patches/200-disable_fasthash.patch * [packages] update git to 1.8.3.4 and refresh patches Jul 23 10:36:07 ping blogic Jul 23 10:38:17 acoul: hello Jul 23 10:39:49 about the 16M RAM atheros routers dying on high network bandwidth, have you been beaten by this at all? Jul 23 10:40:02 yes Jul 23 10:40:06 its the kernel bloat Jul 23 10:40:12 do oyu have a log of it dying ? Jul 23 10:40:16 its most likely a oom Jul 23 10:41:32 no clue on how to deal with it, though I have access to two of those routers with the issue right now Jul 23 10:42:17 can you get a log of the crash ? Jul 23 10:42:46 it's not a crash it's memory starvation Jul 23 10:43:20 yes Jul 23 10:43:23 its an oom Jul 23 10:45:16 I can recreate the problem with a high ping "ping -f -s 31000 ip-on-wireless-end" and watch the free mem going down instantly Jul 23 11:01:21 russell_: i just mailed you a test patch Jul 23 11:01:37 please add it to the procd folder and then rebuild and show me the log of the unit where it fails Jul 23 11:03:57 build #329 of orion is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/329 Jul 23 12:24:38 blogic r37516 trunk/package/ system/ubox/Makefile system/procd/Makefile * update ubox and procd to latest git revision Jul 23 13:08:52 blogic r37517 trunk/target/linux/ramips/base-files/etc/uci-defaults/02_network * ramips: fix HG255D VLAN initial detect issue Jul 23 13:13:32 nbd r37518 trunk/include/kernel-defaults.mk * kernel: fix remaining initramfs related breakage Jul 23 13:34:25 acinonyx r37519 packages/utils/rtl-sdr/Makefile * [packages] rtl-sdr: Update to latest revision Jul 23 14:27:53 KanjiMonster: i'm testing a Comtrend VR-3025un (http://wiki.openwrt.org/toh/comtrend/vr3025un) with latest trunk and almost all works fine. The problem is that i'm experiencing some data corruption in rootfs_data when i reboot the router once it is configured. Jul 23 14:28:16 If you are busy don't answer. Jul 23 14:30:41 Pteridium: try adding your board to etc/uci-defaults/09_fix_crc Jul 23 14:31:11 assuming you mean CFE claims crc errors and refuses to boot the image Jul 23 14:31:35 i've done it without success Jul 23 14:31:40 :( Jul 23 14:31:59 tried mtd fixtrx linux too Jul 23 14:32:31 what the actual error you get? Jul 23 14:32:55 http://pastebin.com/RpXSGig4 Jul 23 14:33:53 the files changed with UCI or luci are corrupted after reboot Jul 23 14:34:18 the name and the content Jul 23 14:34:51 can you try to manually mount the overlay ? Jul 23 14:34:53 or Jul 23 14:35:06 are the files inside /overlay/etc/config/ also broken ? Jul 23 14:36:07 ok, i go to see /overlay/etc/config/. one moment, please. Jul 23 14:36:28 blogic: take in mind that surely is my fault Jul 23 14:37:18 i would say a bu but if you claim the fault fine with me ;) Jul 23 14:38:04 blogic: in /overlay/etc/config/ the files are fine. Jul 23 14:38:18 with all the procd-switch related breakage, it's currently the best time for blaming blogic ;) Jul 23 14:38:52 http://pastebin.com/dpmMPAaP Jul 23 14:39:06 :) Jul 23 14:39:12 how i can help? Jul 23 14:41:20 KanjiMonster: blogic had a lot of patience with me so i want to help him, not blame. Jul 23 14:42:37 and you too, of course Jul 23 14:42:57 I just wanted to troll a bit, too ;) Jul 23 14:43:14 xD Jul 23 14:43:43 KanjiMonster: http://linux-lantiq.org/troll.jpeg Jul 23 14:43:48 :-) Jul 23 14:44:03 Pteridium: call "mount" and paste the output Jul 23 14:44:10 and what kernel version is this ? Jul 23 14:44:20 and can oyu show me a full bootlog please ? Jul 23 14:46:35 blogig: the info you need. http://pastebin.com/f4Qbe0UR Jul 23 14:47:18 blogic: the info you need. http://pastebin.com/f4Qbe0UR* Jul 23 14:51:29 Do you think is a procd problem? i'm almost sure that the fault is mine. :-/ Jul 23 14:56:03 one moment, i go to test r37519 Jul 23 14:56:07 i dont see how it could be procd Jul 23 15:26:22 blogic: ping ... i am awake again ;-) Jul 23 15:28:35 i am just trying what i should have tried previously, which is to replicate the problem with a vanilla build, just to be sure it isn't one of my local files/* tweaks (seems unlikely, but ...) Jul 23 15:32:45 blogic, KanjiMonster: here is a complete bootlog: http://pastebin.com/Zn6A9Ube, and this called my attention: http://pastebin.com/QN8YmjVV "xref (0 dead, 2 orphan)" Jul 23 15:33:21 http://pastebin.com/Zn6A9Ube * Jul 23 15:34:26 blogic: same thing with none of my files/*, fails to start remote syslog on boot, works after restart Jul 23 15:37:57 mount_root: No jffs2 marker was found Jul 23 15:38:03 what image are you flashing ? Jul 23 15:38:10 there is no deadc0de marker Jul 23 15:38:53 sorry, don't understand, deadcode? Jul 23 15:39:19 still newbie :( Jul 23 15:40:34 normally an image has kernel + squashfs + 0xdeadc0de Jul 23 15:40:53 this tells mount_root -> "there is an uninitialized jffs2 here" Jul 23 15:40:58 and that deadcode is missing Jul 23 15:41:06 did you add this as a new board ? Jul 23 15:41:11 or is it in owrt already Jul 23 15:41:22 no, new board Jul 23 15:41:29 ok Jul 23 15:41:43 then i missed something Jul 23 15:41:43 i think maybe you generate the image incorrectly Jul 23 15:42:00 show me you patch to target/linux/bcm63sxx/image/Makefile please Jul 23 15:42:32 ok, one moment... Jul 23 15:43:19 thx ! Jul 23 15:45:42 blogic: https://github.com/Pteridium/openwrt/commit/e11596088fc53f397314877c7e71187772cc12f4 still i haven't the patch for the latest trunk rev Jul 23 15:47:28 maybe CFEFIXUP instead CFE? Jul 23 15:47:50 blogic: replicated with a generic .config as well Jul 23 15:52:16 KanjiMonster: ping Jul 23 15:52:31 KanjiMonster: Pteridium's image is missing the 0xdeadc0de tag Jul 23 15:52:35 oh, just spotted the patch in the scrollback Jul 23 16:04:08 blogic: could be used this comment as workaround? https://dev.openwrt.org/ticket/9143#comment:8 Jul 23 16:07:01 ( echo -n -e "# Airstation FirmWare\nrun u_fw\nreset\n\n" | dd bs=512 count=1 conv=sync; dd if=/home/danielg4/openwrt-rt5350/bin/ramips/openwrt-ramips-rt305x-whr-g300n-squashfs-sysupgrade.bin; ) > /home/danielg4/openwrt-rt5350/build_dir/target-mipsel_dsp_uClibc-0.9.33.2/linux-ramips_rt305x/whr-g300n-tftp.tmp Jul 23 16:07:01 0+1 records in Jul 23 16:07:02 1+0 records out Jul 23 16:07:02 512 bytes (512 B) copied, 3.101e-05 s, 16.5 MB/s Jul 23 16:07:02 dd: opening `/home/danielg4/openwrt-rt5350/bin/ramips/openwrt-ramips-rt305x-whr-g300n-squashfs-sysupgrade.bin': No such file or directory Jul 23 16:07:47 i'm sure i'm missing something obvious Jul 23 16:08:23 you can use jffs2mark for that Jul 23 16:08:30 its a tool on the rootfs Jul 23 16:10:07 ok: jffs2mark: /dev/mtdblock3 - marking with deadc0de Jul 23 16:11:03 then reboot Jul 23 16:11:06 try again Jul 23 16:11:19 at least the error should be gone now Jul 23 16:11:58 before: mount_root: No jffs2 marker was found, after: mount_root: jffs2 is not ready - marker found Jul 23 16:12:19 A lot of thanks blogic! Jul 23 16:14:39 Pteridium: well Jul 23 16:14:52 now edit some files and see if they are intact after a reboot Jul 23 16:15:28 yes, i go to make tests Jul 23 16:15:54 later i'll tell you the results Jul 23 16:16:25 thanks to KanjiMonster an you for your time. Jul 23 16:16:34 and* Jul 23 16:24:27 blogic: your debugging printf's never fire ... at least they don't show up on the serial console during boot Jul 23 16:28:19 blogic: your debugging printf's never fire ... at least they don't show up on the serial console during boot Jul 23 16:31:37 russell__: hmmm Jul 23 16:31:53 russell__: you copied it to package/system/procd/patches/ Jul 23 16:31:55 ? Jul 23 16:32:03 russell__: and .... Jul 23 16:32:10 1) boot unit Jul 23 16:32:21 2) call /etc/init.d/log restart Jul 23 16:32:26 3) logger test 1 Jul 23 16:32:33 4) /etc/network/restart Jul 23 16:32:38 5) logger test 2 Jul 23 16:32:51 and see if test 2 also shows up in the remote syslog please Jul 23 16:32:52 i had to rework your patch a little, since it was in git format Jul 23 16:32:58 but it applied Jul 23 16:33:12 * russell__ trying restart (recall that works) Jul 23 16:33:31 i want to know if it works afer a network restart Jul 23 16:33:52 and if that works then move /etc/init.d/log to a later time Jul 23 16:34:01 as in edit START= to 80 or so Jul 23 16:34:05 see if that works Jul 23 16:34:17 does procd have a stderr in the context it is running? Jul 23 16:35:40 oh, wait, /me looking in the right place now Jul 23 16:35:48 hang on Jul 23 16:37:46 i'm seeing your debugging messages on the remote syslog Jul 23 16:38:05 but split across lines in a highly unreadable way ;-) Jul 23 16:39:10 lol Jul 23 16:39:11 ok Jul 23 16:40:26 this is just from bootup: http://pastebin.com/bC5UMqjw Jul 23 16:40:33 no "logger" testing Jul 23 16:42:22 "logger test 1" results in nothing in remote syslog Jul 23 16:47:00 /etc/init.d/log restart gives this on remote syslog: http://pastebin.com/Cr0CZWn8 Jul 23 16:48:11 ok Jul 23 16:48:22 try moving it to after the firewall start please Jul 23 16:48:43 hand on let me boot a unit to test it here and tell you how the patch looks Jul 23 16:49:41 hmmm Jul 23 16:49:49 its behind firewall already Jul 23 16:49:57 change START=21 Jul 23 16:50:08 and use START=99 please Jul 23 16:50:11 and see if that works Jul 23 16:50:31 build #252 of iop32x is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/iop32x/builds/252 Jul 23 16:50:55 * russell__ used 95 Jul 23 16:51:53 and that also breaks ? Jul 23 16:52:18 yes Jul 23 16:52:37 logger test 3 doesn't show up or generate debugging messages Jul 23 16:52:53 ok Jul 23 16:52:56 ermmm Jul 23 16:53:09 no i am puzzled Jul 23 16:53:14 * russell__ too Jul 23 16:53:15 *now Jul 23 16:53:28 can you install tcpdump Jul 23 16:53:38 boot the image without any changes applied Jul 23 16:53:40 i can run it upstream, and nothing shows up Jul 23 16:53:48 on the device i mean Jul 23 16:54:05 build an unmdified image and boot it on the unit that fails Jul 23 16:54:26 you mean without your patch? Jul 23 16:54:27 and then run tcpdump in it and run "logger foo" and send me the pcap file ? Jul 23 16:54:30 yes Jul 23 16:54:31 without Jul 23 16:54:39 tcpdump on what interface? Jul 23 16:54:42 wan? Jul 23 16:54:46 the one you log to Jul 23 16:54:51 okay Jul 23 16:54:52 if you log over wan then on wan Jul 23 16:54:58 and you use dhcp on wan ? Jul 23 16:55:01 yes Jul 23 16:55:08 try also with a static ip Jul 23 16:55:13 see if that works Jul 23 16:55:29 maybe we can narrow the cause down this way Jul 23 16:56:47 this device has lots of flash so reflashing+rebuilding jffs takes about 5 minutes ;-) Jul 23 16:56:56 great ! :) Jul 23 16:57:06 i am here for a few more hours Jul 23 16:57:07 ;) Jul 23 16:57:15 * russell__ has an image with tcpdump in it already built Jul 23 17:01:43 flashing Jul 23 17:01:57 btw, do you see me ping when my nick is russell-- ? Jul 23 17:03:07 ping where ? Jul 23 17:03:15 here in this channel Jul 23 17:03:31 yes Jul 23 17:03:41 i think so Jul 23 17:03:47 i was offline for a few days Jul 23 17:03:57 okay, /me switches back Jul 23 17:04:00 my first day back online today Jul 23 17:04:15 ah, that explains ... vacation++ Jul 23 17:04:42 russell--: well Jul 23 17:04:46 works does it not Jul 23 17:05:44 i cant see you say anything Jul 23 17:05:49 say something :) Jul 23 17:10:46 sorry testing Jul 23 17:11:00 you see me? Jul 23 17:11:14 blogic: hello Jul 23 17:11:33 10:11 russell--: blogic: hell Jul 23 17:11:35 o Jul 23 17:11:37 lol Jul 23 17:12:31 did you see me saying "sorry testing"? Jul 23 17:12:55 19:10 < russell--> sorry testing Jul 23 17:12:55 19:11 < russell--> you see me? Jul 23 17:12:55 19:11 < russell--> blogic: hello Jul 23 17:12:59 i can see you Jul 23 17:13:02 with the other link Jul 23 17:13:05 * blogic is confused Jul 23 17:13:07 anyhow Jul 23 17:14:12 okay, so i rebooted, ran tcpdump -i eth1 -vv udp port 514, then ran logger test 1, nothing Jul 23 17:14:20 oh Jul 23 17:14:23 ok Jul 23 17:14:27 i never tested on wan Jul 23 17:14:28 only lan Jul 23 17:14:33 i will test wan sid enow ;) Jul 23 17:14:41 thanks for helping narrow it doen Jul 23 17:14:44 thanks for helping narrow it down Jul 23 17:14:54 that's with dhcp Jul 23 17:14:55 lets see what i find ;) Jul 23 17:14:57 ok Jul 23 17:15:06 try static then before i do anything Jul 23 17:15:08 * russell__ has firewall turned off (as per norm) Jul 23 17:15:12 okay Jul 23 17:17:49 also, i notice that /etc/init.d/log stop doesn't actually stop logread Jul 23 17:18:10 ok Jul 23 17:19:44 static network config results in no change Jul 23 17:20:00 == not working on boot Jul 23 17:20:11 ok Jul 23 17:20:18 * blogic sets up a testbed Jul 23 17:20:58 works on restart, like dhcp Jul 23 17:22:37 the one difference i saw in the strace was in rt_N's in a rt_sigaction, but i don't know how to decode those Jul 23 17:23:04 it looked suspicious, anyway Jul 23 17:23:59 also note that it works on ubnt-airrouter, normally Jul 23 17:24:24 but not when i wrap logread in an strace Jul 23 17:26:23 that is, i editted /etc/init.d/log to insert "/usr/bin/strace -ff -s128 -o /tmp/strace-logread" in front of the "$PROG" in start_service_remote Jul 23 17:34:25 that is, ubnt-airrouter works on boot without the strace, but doesn't work on boot with the strace. both work on /etc/init.d/log restart Jul 23 17:47:09 The_Lizard: ping Jul 23 17:47:19 i think i know what might be your problem Jul 23 17:47:24 show me a full bootlog please Jul 23 18:12:20 build #255 of ep93xx is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/ep93xx/builds/255 Jul 23 19:03:50 blogic: hm, I'll take a look at it Jul 23 19:06:12 KanjiMonster: thanks Jul 23 19:06:25 if this is the cause for the breakage we still have a bug ;) Jul 23 19:06:39 as it should still work even without the deadc0de marker Jul 23 19:07:02 there's no special deadc0de code in bcm63xx Jul 23 19:07:11 its in the generic patches Jul 23 19:07:16 yes Jul 23 19:07:28 the prepare_generic_squashfs function does it for you Jul 23 19:07:37 i saw is called somewhere unrelated in the code though Jul 23 19:07:39 not sure Jul 23 19:08:00 Pteridium never reported if the bug went away after that Jul 23 19:12:42 blogic: what exactly is the issue? Jul 23 19:14:01 his overlay is corrupted Jul 23 19:14:09 /overlay has the files intact Jul 23 19:14:16 /etc/config/ has broken files Jul 23 19:15:08 you said something about deadc0de Jul 23 19:15:55 well, while looking at it we realized that the deadcode marker is misisng Jul 23 19:18:26 the deadc0de marker gets overwritten when jffs2 formats the flash Jul 23 19:22:53 yes Jul 23 19:23:01 but its not there on firstboot for him somehow Jul 23 19:23:29 all the images I checked have it Jul 23 19:23:35 weird Jul 23 19:23:41 then something else is wrong Jul 23 19:31:00 build #253 of mcs814x is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/mcs814x/builds/253 Jul 23 19:55:01 jogo r37520 trunk/package/kernel/linux/modules/netfilter.mk * kernel: ebtables depends on bridge Jul 23 20:11:28 blogic: no luck. Also tried setting CFEFIXUP instead CFE but still the same problem. Could be the flash chip damaged and how to test it? Jul 23 20:11:55 I'm feeling useless... :( Jul 23 20:12:54 Pteridium: CFEFIXUP just writes a slightly different cfe header that will overwrite CFE's boardid with a different one, and that's all Jul 23 20:13:08 ah Jul 23 20:14:45 tomorrow i will try more things; today i'm upset of this router. Jul 23 20:52:38 build #239 of mpc52xx is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/mpc52xx/builds/239 Jul 23 21:13:41 blogic: im here now... Jul 23 21:13:46 hi Jul 23 21:13:49 bootlog ? Jul 23 21:13:50 :) Jul 23 21:14:02 i am going ot bed in 15 minutes Jul 23 21:14:07 so hurry or ping me tomorrow Jul 23 21:23:37 There is a bootlog at the wiki Jul 23 21:23:42 but it isn't the latest anymore Jul 23 21:23:45 will update it now Jul 23 21:23:55 link ? Jul 23 21:24:37 http://wiki.openwrt.org/toh/arcadyan/vgv7519 Jul 23 21:25:15 what do you think is wrong? Jul 23 21:28:47 hmmm Jul 23 21:28:49 not Jul 23 21:28:50 [ 0.192000] phy-xrx200 gphy-xrx200.5: requesting lantiq/vr9_phy11g_a1x.bin Jul 23 21:28:57 this is for vr9 v1.1 Jul 23 21:29:02 and the board has 1.1 Jul 23 21:29:12 i thought maybe oyu loaded the wrong firmware Jul 23 21:29:20 blogic: just opdated Jul 23 21:29:58 It was incorrect earlier (as the one from the easy80920, but I took the one as specified in daniels u-boot tree Jul 23 21:34:30 Should I do some basic configuration before it should work? Jul 23 21:34:49 as in setting the mac addresses etc? they are now around 00:11:22:33:44:55 Jul 23 21:35:45 or could it be some default firewall behaviour? **** ENDING LOGGING AT Wed Jul 24 02:59:59 2013