**** BEGIN LOGGING AT Thu Jun 02 02:59:56 2011 **** ENDING LOGGING AT Thu Jun 02 03:41:39 2011 **** BEGIN LOGGING AT Thu Jun 02 03:42:14 2011 Jun 02 06:15:22 hi Jun 02 06:15:39 i need help upgrading 4mb flash to 8 mb flash Jun 02 06:16:15 cab anyone help me with that Jun 02 06:21:12 you just change the flash and map the new size to the environmental variables if the boot loader needs it Jun 02 06:22:55 if you're able to install owrt then you should be able to do that second part too Jun 02 06:38:28 i flashed 8mb chip in flash programmer, and router won't boot Jun 02 06:50:33 of course? Jun 02 06:50:48 the boot loader sees only 4 Jun 02 06:58:40 PsyMan : well it should be enough to boot Jun 02 06:58:56 PsyMan: i got 4mb flash. dumped it and flashed into 8mb flash Jun 02 06:59:08 PsyMan: soldered 8mb flash and it won't boot Jun 02 06:59:12 PsyMan: why? Jun 02 06:59:42 if i solder 4mb flash back it works Jun 02 07:03:06 probably cause the new flash has a different block size or it doesn't work if the environmental variables miss the additional space Jun 02 07:03:26 or the boot loader does some sort of flash size check Jun 02 07:03:59 was the dump you did an openwrt dump or a default firmware dump? Jun 02 07:07:35 cause the default firmware might be checking the size too Jun 02 07:25:13 PsyMan: it was flashed with openwrt but with OEM bootloader AFAIK Jun 02 07:25:28 how to check blocksize btw? Jun 02 07:25:52 check the flash specifications Jun 02 07:26:21 i'd also check if the bootloader works Jun 02 07:26:26 on the new flash Jun 02 07:26:40 (even if th router won't boot) Jun 02 07:27:06 i don't have jtag Jun 02 07:27:08 laters Jun 02 07:27:23 jow_laptop: really bad :( I just tested with a local configure of asterisk (out of openwrt) and the bug occur with gcc-4.6 and 4.5. Working in gcc-4.4. Jun 02 08:25:38 I got kernel panic with kernel 3.0-rc1 on mr3420. Jun 02 08:25:39 http://pastebin.com/kR6DRm4X Jun 02 08:25:50 Any solution? Jun 02 08:26:50 Does it looks like lacking of mtd definition for uboot loading device? Jun 02 08:32:17 @gzanan: can you also post a 2.6.39 bootlog? Jun 02 08:34:20 (06:01:57 PM) acoul: @gzanan: on kernel_menuconfig, Machine selection ---> Atheros AR71xx machine selection ---> make sure your box/device is selected Jun 02 08:34:28 I haven't record last time's 2.6.39 boot log.Need to build one and log,will be later. Jun 02 08:35:16 Yes,I've selected all device including mine on kernel_menuconfig definitely. Jun 02 08:38:42 I'm not familiar with git.Last time I git pull nbd's repo.Then git pull yours.It gives me this notice.How to solve it? Jun 02 08:38:53 http://pastebin.com/B2x12AWd Jun 02 08:42:35 git status Jun 02 08:45:19 you should use openwrt include/netfilter.mk Jun 02 08:45:41 ozonet git tree is synced to it on a daily basis Jun 02 08:47:32 I've deletedithe confilited include/netfilter.mk,what command can pull it back? Jun 02 08:48:46 you can download the latest one and manually add/commit it Jun 02 08:59:40 @gzanan: does 2.6.38 boots ok on your device? Jun 02 09:05:57 Yes Jun 02 09:06:36 Nothing wrong with 2.6.38.7 or before. Jun 02 09:08:19 I see Jun 02 09:11:53 I've git add include/netfilter.mk.then git commit.But include/netfilter.mk still marked as changes to be commited. Jun 02 09:12:43 http://pastebin.com/4jrTQzUV Jun 02 09:25:07 Just gone out to flash my device with K2.6.39 Jun 02 09:25:53 http://pastebin.com/ytQddHxf Jun 02 09:26:30 It's still like last time,only can be accessed by ttl cable. Jun 02 09:27:34 Acoul are you still around? :) Jun 02 09:31:25 always Jun 02 09:32:52 what linux kernel is this? Jun 02 09:52:19 2.6.39 Jun 02 09:53:50 Former question:I've git add include/netfilter.mk.then git commit.But include/netfilter.mk still marked as changes to be commited. Jun 02 09:53:51 2.6.39 boots ok but network is not working? Jun 02 09:55:01 acoul1: did you ever put my 400-rootfs_split.patch into your git? Jun 02 09:55:03 Yes,the switch is not working either. Jun 02 09:56:08 http://www.kernel.org/pub/software/scm/git/docs/git-merge.html Jun 02 09:58:54 @KanjiMonster: yes, but send me the link/patch again to double check Jun 02 09:59:51 acoul1: still at http://inuyasha.ath.cx/~jonas/400-rootfs_split.patch Jun 02 10:04:18 the files are identical Jun 02 10:06:13 hm, okay Jun 02 10:06:40 then I dunno (regarding the not booting problem) Jun 02 10:08:02 my RouterStation Pro & Routerboard RB411 boot fine under linux-3.0-rc1 Jun 02 10:24:55 <^Willie^> GOAL ! Jun 02 10:27:25 <^Willie^> gor the kernel boot output afther enabling serial console next to the ubimailbox.. Jun 02 10:27:29 <^Willie^> got* Jun 02 10:27:54 <^Willie^> weird part.. i het the routers output via gnd pin from my wan led (dlink dir-652) Jun 02 10:27:59 <^Willie^> damn keyboard :S Jun 02 10:28:16 <^Willie^> weird part.. i Get the routers output via gnd pin from the wan led (dlink dir-652) Jun 02 11:19:49 <^Willie^> results first part is the ttyUM0 (ubicom dongle) readable part is from the led pins http://pastebin.com/08bihM4L Jun 02 11:24:44 <^Willie^> hmm and do think about moving the port order.. not shure if that make the readable uart working from boot since i have no idea how u-boot is configured :S Jun 02 11:24:56 <^Willie^> *port order in the kernel command line Jun 02 12:08:37 build #39 of uml is complete: Failure [failed shell_4] Build details are at http://buildbot.openwrt.org:8010/builders/uml/builds/39 Jun 02 12:37:05 When I tried to use 'autoreconf -fiv' to build mp4v2 package, it complains with this message 'autoreconf: configure.ac: not using Gettext'. Does anyone here have any idea how to fix this? Jun 02 12:51:14 How to go back to previous git tree? Jun 02 13:41:01 <^Willie^> {Nico}_: can you notice/pm me when your at your keyboard ? i have some questions about the ubicom32 builds. Jun 02 13:41:19 <^Willie^> before i'm going to redo your work etc etc.. Jun 02 13:42:02 build #42 of ramips is complete: Failure [failed shell_4] Build details are at http://buildbot.openwrt.org:8010/builders/ramips/builds/42 Jun 02 13:45:27 florian * r27090 /trunk/target/linux/au1000/ (7 files in 4 dirs): [au1000] update to 2.6.39 Jun 02 14:05:46 <_lore_> ping nbd Jun 02 14:09:57 Has any develeopers here gotten some spare time to check out this ticket #9310 (https://dev.openwrt.org/ticket/9310 )? Jun 02 14:14:23 mazilo: provide a patch for the issue and it will be fixed much faster Jun 02 14:14:57 build #40 of x86 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/x86/builds/40 Jun 02 14:16:49 _lore_: pong Jun 02 14:16:59 <_lore_> hi Jun 02 14:17:02 hi Jun 02 14:17:04 <_lore_> how are you? Jun 02 14:17:08 fine, how are you? Jun 02 14:17:23 <_lore_> fine thanks...today is vacation in Italy Jun 02 14:17:25 <_lore_> :) Jun 02 14:17:33 :) Jun 02 14:18:33 <_lore_> I have done some improvements in my project...I fixed some bugs and the system works quite well Jun 02 14:19:02 cool Jun 02 14:19:07 <_lore_> one more issue yet not fixed Jun 02 14:20:19 <_lore_> when the system is under heavy load I am not able to guarantee a limited round duration Jun 02 14:20:28 <_lore_> I mean Jun 02 14:21:19 <_lore_> sometimes the token frame is delayed up to 200ms but I would mantain this value under 30ms Jun 02 14:21:58 <_lore_> I suspect this is due to the lack of multiple queue in rx path Jun 02 14:23:45 <_lore_> I mean, I have just one rx queue and the token frame is sunken in a huge number of data frame Jun 02 14:23:57 why would you need multiqueue in the rx path? is the cpu loaded enough to produce a significant delay there? Jun 02 14:24:05 have you measured the delay there? Jun 02 14:24:10 <_lore_> yes Jun 02 14:24:24 florian * r27091 /packages/net/tcpdump/ (6 files in 2 dirs): Jun 02 14:24:24 [package] tcpdump: update to 4.1.1 Jun 02 14:24:24 Corrected patch (last try) to update tcpdump to version 4.1.1. Jun 02 14:24:24 Signed-off-by: Christoph König Jun 02 14:24:26 florian * r27092 /packages/net/strongswan4/ (Makefile files/ipsec.conf files/strongswan.conf): (log message trimmed) Jun 02 14:24:26 [package] update strongswan4 to 4.5.2 Jun 02 14:24:26 Update version of strongswan4 package from 4.5.1 to 4.5.2. Jun 02 14:24:26 Add new strongswan4 plugins: coupling, duplicheck, whitelist. Jun 02 14:24:27 Add strongswan4-libfast package and make build dependencies on Jun 02 14:24:27 clearsilver and fcgi conditional. Previously libfast was being built, Jun 02 14:24:28 but not packaged. Now libfast will only be built when the it's package Jun 02 14:24:28 florian * r27093 /packages/utils/mg/Makefile: Jun 02 14:24:29 [package] mg: update to 20110120 Jun 02 14:24:29 patch from Justus Winter Jun 02 14:24:30 florian * r27094 /packages/libs/libao/Makefile: Jun 02 14:24:30 [package] libao: update to 1.1.0 Jun 02 14:24:40 <_lore_> I compute the cdf of the token round Jun 02 14:24:51 well, i still don't think you'd need multiqueue to fix it Jun 02 14:24:56 you just need to restructure the rx path Jun 02 14:25:00 to defer passing packets to the stack Jun 02 14:25:19 until you've looked for time critical frames in the entire queue of pending frames Jun 02 14:25:45 <_lore_> I have some fixed constraints on token frame Jun 02 14:26:45 <_lore_> you mean put the token frames in the top of rx queue? Jun 02 14:27:39 yes Jun 02 14:27:50 <_lore_> ok Jun 02 14:27:59 <_lore_> but this search is not O(n)? Jun 02 14:28:46 <_lore_> I mean Jun 02 14:29:33 <_lore_> if the system is under havy load it will take a lot of time Jun 02 14:29:43 <_lore_> *heavy Jun 02 14:29:50 there's no separate search Jun 02 14:29:58 it's simply a change in the order in which processing is done Jun 02 14:30:01 like this: Jun 02 14:30:08 you make a new skb queue on the stack Jun 02 14:30:33 in the rx function you process frames, but instead of releasing them to the stack, you stick them in the skb queue Jun 02 14:30:46 and while you're doing that, you check the packet contents for the token frame Jun 02 14:30:51 and process that separately Jun 02 14:31:04 after you're done, you go through the skb queue and release all frames to mac80211 Jun 02 14:31:28 (and every few packets re-check for new incoming frames in the queue) Jun 02 14:34:13 <_lore_> so you mean to accumulate all data frame in ath9k rx taskelet Jun 02 14:34:31 yes Jun 02 14:34:50 <_lore_> and after have consumed the token frame release the data burst to the mac80211 stack Jun 02 14:35:02 yes Jun 02 14:35:12 well, or after reaching the end of the queue Jun 02 14:35:21 <_lore_> yes Jun 02 14:35:44 and while you're releasing the frames, occasionally re-check for new frames Jun 02 14:35:54 to avoid long delays there Jun 02 14:35:58 <_lore_> ok Jun 02 14:36:11 <_lore_> tnx, good idea!! Jun 02 14:36:37 <_lore_> I am going to try in the next days Jun 02 14:37:55 <_lore_> how does your work going? Jun 02 14:38:02 florian * r27095 /trunk/package/acx-mac80211/patches/010-io-endianness.patch: Jun 02 14:38:02 [package] acx-mac80211: add io endianness patch Jun 02 14:38:02 Signed-off-by: Alexey Torkhov Jun 02 14:38:07 florian * r27096 /trunk/package/soloscli/ (Makefile patches/001-newline.patch): Jun 02 14:38:07 [package] soloscli: fix new lines in soloscli Jun 02 14:38:07 If the string received from the ADSL firmware is missing \n, add it in. Jun 02 14:38:07 Signed-off-by: Nathan Williams Jun 02 15:19:18 Is cmake supported under OpenWRT? Jun 02 15:27:48 yes Jun 02 15:27:57 there are several packages using it Jun 02 15:28:04 e.g. uci, libubox, ubus, etc. Jun 02 15:46:16 nbd: Thanks for your response. Jun 02 19:32:15 is there a way to include multiple default scripts from uci-defaults in the built image? Jun 02 19:50:49 /rom/etc/uci-defaults is populated but /etc/uci-defaults is empty, is this correct? Jun 02 19:54:48 netprince: yes Jun 02 19:54:59 netprince: all uci-defaults scripts get deleted after they've been run Jun 02 19:56:24 Ok, where does this happen? (what script?) I seem to be missing some default settings, trying to hunt it down... Jun 02 19:56:55 /etc/init.d/boot Jun 02 19:59:29 ok, thanks! Jun 02 22:46:03 nice iPad spam on the forum Jun 02 22:47:35 what spam? :) Jun 02 22:48:06 nbd: I took the liberty and moved the the WZR-HP-AG300H from possible to supported in the TOH ;) Jun 02 22:48:57 :) Jun 02 22:49:39 nbd: unfortunately, the kind that keeps giving and giving :) Jun 02 23:08:09 nbd: hm, the patches in patches/ in compat-wireless are unused in openwrt, right? because mac80211-patches modifying these break quilt/refresh ;) Jun 02 23:12:07 yes, they're unused Jun 02 23:12:13 because they're already applied Jun 02 23:35:27 nbd: okay, good. btw, I created an openwrt "fork" at github so I have a place to put my multimegabyte patches ;) Jun 02 23:36:12 :) Jun 02 23:36:26 well, imho you will get svn access soon Jun 02 23:38:24 sure, but not everything is instantly ready for prime time; I plan to use it for RFC/RFT patches and work in progress stuff Jun 02 23:40:09 both "mark $xxx as broken" patches are now "add linux-3.0 compatibility" patches in v2 (I backported Hauke's compat(-wireless) patches Jun 03 01:02:05 build #41 of rb532 is complete: Failure [failed shell_4] Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/41 Jun 03 01:14:15 build #29 of octeon is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/octeon/builds/29 Jun 03 01:27:17 build #37 of ar7 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ar7/builds/37 Jun 03 02:05:34 build #32 of etrax is complete: Failure [failed shell_4 compile_3] Build details are at http://buildbot.openwrt.org:8010/builders/etrax/builds/32 Jun 03 02:22:34 build #37 of sibyte is complete: Failure [failed shell_4] Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/37 **** ENDING LOGGING AT Fri Jun 03 02:59:57 2011