**** BEGIN LOGGING AT Mon Jan 25 02:59:57 2010 Jan 25 03:01:02 cshore: I did some prototyping of a configfile format, but I have some doubts about it ... Jan 25 03:01:04 cshore: http://pastebin.com/d230c2afd Jan 25 03:01:29 cshore: hm, for some reason it's not loading the gpio and led modules on ar71xx Jan 25 03:01:43 so hotplug via button doesn't work Jan 25 03:02:15 i don't see the bug at the moment, theoretically it should load it, but it doesn't Jan 25 03:02:30 the hook is loaded in a 05_* file and added to preinit_main Jan 25 03:03:24 should it maybe be added to preinit_essential? Jan 25 03:03:38 nbd: hmm...I'll take a look Jan 25 03:05:23 when does it first get used? 10_indicate_preinit, right? Jan 25 03:05:57 Is there anything for insmod that is defined in . /etc/functions.sh? Jan 25 03:06:47 nothing in functions.sh Jan 25 03:07:26 yes, it gets used first in 10_indicate_preinit Jan 25 03:07:32 but if i enter failsafe, the modules are still not loaded Jan 25 03:07:36 and the file is in build_dir/target_xxx/lib/preinit? Jan 25 03:07:37 so something prevents them from being loaded at all Jan 25 03:07:40 it's not an ordering issue Jan 25 03:07:45 the file is in there correctly Jan 25 03:07:55 and if i execute its callback manually, that works as well Jan 25 03:08:19 i'll add some debug code to print the list of callbacks Jan 25 03:08:26 if you turn off the supression of stderr, does it say anything (under Image Options) Jan 25 03:09:40 will try Jan 25 03:11:40 if you echo $preinit_main_hook, what does it say? Jan 25 03:12:13 in failsafe I mean Jan 25 03:12:21 nv Jan 25 03:12:22 currently building the new image with stderr enabled Jan 25 03:12:29 will let you know in a minute Jan 25 03:12:52 ok...I realized that I don't export variables so echo from failsafe won't tell you anything Jan 25 03:15:00 would load ar71xx.sh in each of the ar71xx files do anything bad? Jan 25 03:15:03 sourcing I mean Jan 25 03:15:25 no Jan 25 03:15:28 that only defines one function Jan 25 03:15:38 btw. i don't get any error messages on stderr Jan 25 03:16:08 are the interfaces defined? Jan 25 03:16:52 it did bring up eth0 Jan 25 03:16:57 so i assume they are Jan 25 03:17:09 does it have the mac assigned in preinit? Jan 25 03:17:26 no Jan 25 03:17:42 it comes up with the right one by itself Jan 25 03:17:59 nbd: oh, I was looking at 05_set_iface_mac_ar71xx Jan 25 03:18:10 ah Jan 25 03:18:10 found it Jan 25 03:18:17 it's a typo in the ar71xx script Jan 25 03:18:25 heh Jan 25 03:18:44 a typo that is present on multiple files Jan 25 03:18:52 book_hook_add instead of boot_hook_add Jan 25 03:18:54 ;) Jan 25 03:19:02 ack Jan 25 03:19:15 I'd best grep Jan 25 03:20:06 i did Jan 25 03:20:11 brcm-2.4 and ar71xx affected Jan 25 03:20:22 will test again Jan 25 03:20:48 my brain got ahead of my fingers apparently Jan 25 03:21:06 happens... Jan 25 03:21:28 I'm glad you found it...you can go blind trying to find that sort of thing Jan 25 03:23:46 shouldn't madwifi.sh use wifi_fixup_hwmode for https://dev.openwrt.org/ticket/6139? Jan 25 03:24:04 rtz2: will hotplug2_coldplug work with brcm-2.4? Jan 25 03:24:17 cshore: nope Jan 25 03:24:29 cshore: there is also devfs stuff, but I didn't post it Jan 25 03:24:49 cshore: it's only meant as an example to get a feeling for the config file format Jan 25 03:25:45 I can't think of any glaring problems with it Jan 25 03:25:50 huh, now i'm getting even weirder stuff Jan 25 03:25:55 /etc/preinit: line 1: syntax error: unterminated quoted string Jan 25 03:25:55 Kernel panic - not syncing: Attempted to kill init! Jan 25 03:26:38 but only if i attempt to get into failsafe via button Jan 25 03:26:44 cshore: the init program would be 20-30kb Jan 25 03:26:54 cshore: so, should I go ahead? Jan 25 03:27:13 nbd: that will be a problem with failsafe_wait Jan 25 03:27:26 nbd: probably a bad trap statement Jan 25 03:28:01 nbd: which is why it probably doesn't affect anything I was able to test Jan 25 03:28:30 maybe it's a race condition Jan 25 03:28:40 the trap must be run before button events are possible Jan 25 03:28:43 otherwise it'll kill preinit Jan 25 03:28:54 nbd: hmm...that's probably it then Jan 25 03:28:58 so we probably need to trap USR1 earlier Jan 25 03:29:06 i just kept repeatedly hitting that button for testing Jan 25 03:29:12 nbd: or add buttons later Jan 25 03:29:24 i think i'll go to bed now Jan 25 03:29:30 can you test and fix this and send followup patches? Jan 25 03:29:39 nbd: yep Jan 25 03:29:42 i'll keep your stuff in my build tree then and apply the rest tomorrow Jan 25 03:29:55 i want to get this in soon Jan 25 03:30:00 nbd: I can't do anything about testing the buttons though. I have no buttons Jan 25 03:30:12 oh Jan 25 03:30:21 i thought you had some broadcom hw with buttons Jan 25 03:30:23 nbd: or rather they don't work yet Jan 25 03:30:33 cshore: if you have time till tomorrow, I can test it Jan 25 03:31:10 cshore: didn't bring my router with me for the weekend Jan 25 03:31:26 rtz: ah, time off...sort of Jan 25 03:32:10 good night Jan 25 03:32:14 i'll do some more testing tomorrow Jan 25 03:32:25 nbd: I think I'll pull out the trap stuff from failsafe_wait... Jan 25 03:32:45 nbd: good night Jan 25 03:32:56 and please look for the unterminated string bug Jan 25 03:32:57 nbd: I'll post a patch tonight Jan 25 03:33:05 nbd: will do Jan 25 03:33:36 cshore: well, the serial cable is kinda fragil, so it's hard to transport Jan 25 03:33:40 thanks Jan 25 03:33:49 cshore: and I had too much stuff and not enough hands Jan 25 03:34:47 rtz: heh, yeah - visting home from school or you own place? Jan 25 03:35:29 cshore: something of both Jan 25 03:36:25 cshore: I'm going to a university and sharing an flat with some other students Jan 25 03:36:26 rtz: ooohhhh, I see a problem Jan 25 03:36:39 rtz: yeah, I remember those days Jan 25 03:36:52 rtz: variables....setting and accessing in scripts Jan 25 03:37:33 cshore: enviroment variables? Jan 25 03:37:49 rtz: yes Jan 25 03:38:08 cshore: I can add a command for that, but it didn't appear to be neccesary yet Jan 25 03:38:16 but it would be rather easy Jan 25 03:38:50 rtz: will the scripts be able to set variables for other scripts? Jan 25 03:39:00 something like 'env-set variable content' Jan 25 03:39:28 rtz: programattically I mean...e.g. based on what they see during execution Jan 25 03:39:51 cshore: what do you mean by script, the jobs or the whole file? Jan 25 03:40:06 the command (could be a script) Jan 25 03:40:58 rtz: for example where cut and sed are needed to extract content from commands Jan 25 03:41:05 rtz: and/or awk, etc Jan 25 03:42:08 cshore: it would make things harder, but not impossible Jan 25 03:44:07 rtz: how well do some of the longer scripts fit into this framework, and trapping SIGUSR1 for button events? Jan 25 03:46:17 rtz: also how do you handle condition execution (e.g. jffs2 vs ext2 vs ramdisk)? Jan 25 03:46:34 cshore: as for the failsafe button, i would send the signal to the init process Jan 25 03:47:06 rtz: I suspect everything can be done, but I'm trying to think of things that will need to be handled Jan 25 03:47:15 cshore: and handle it with the enter_failsafe condition Jan 25 03:47:51 rtz: maybe a failsafe_button event? Jan 25 03:48:11 cshore: which longer scripts? this pretty much replaces preinit and mount_root Jan 25 03:49:06 cshore: I would mash it together with the press key on console stuff to a singe failsafe conditon, but not sure yet Jan 25 03:49:53 rtz: would it be possible to access sequences of presses and wait times (that something I'm going to want to add to the current system) Jan 25 03:50:08 cshore: as for jff2 vs ext2 vs ramdisk, I hope to handle it with conditions directly Jan 25 03:50:18 rtz: ok Jan 25 03:50:26 cshore: for example for jffs2, there will be a rootfs_data mtd partition Jan 25 03:50:40 rtz: the longer scripts are not current...there some older devel posts with them Jan 25 03:50:48 rtz: yes Jan 25 03:51:59 cshore: what exactly do you mean by "access sequences of presses and wait times"? Jan 25 03:52:06 cshore: could you give an example? Jan 25 03:52:30 rtz: e.g. if I press for 1 second wait for 1 second and press again for 1 second Jan 25 03:52:32 of what you want to do? Jan 25 03:52:46 rtz: multiple uses for failsafe Jan 25 03:53:01 rtz: one mode is formatting usb disk Jan 25 03:53:11 rtz: another is a provider provisioning Jan 25 03:53:18 rtz: and another is current failsafe Jan 25 03:53:21 hmmm Jan 25 03:54:11 I will have to think about this Jan 25 03:54:30 well, I'm off to bed now Jan 25 03:54:35 good night Jan 25 03:54:36 rtz: ok Jan 25 03:55:03 rtz: good night Jan 25 04:10:41 is it possible to use those define Download/ things for svn checkouts as well? Jan 25 04:11:38 or what would have to be done to check out multiple svn repositories for one package? Jan 25 04:26:09 latest trunk blows building utils/collectd -- hmmmmm Jan 25 04:54:03 Ascendion: pastebin the error somewhere Jan 25 04:54:25 any mjpg-streamer users around? Jan 25 04:57:27 making verbose and playing WoW :) Jan 25 05:03:03 http://pastebin.com/m41fc51f6 Jan 25 05:28:30 hmm, ah, found an svn define Download/ example and it seems to work Jan 25 05:47:13 somebody want to change https://dev.openwrt.org/browser/trunk/scripts/download.pl#L138-139 to ftp.kernel.org or ftp.geo.kernel.org to use the geodns replies kernel.org has now? Jan 25 10:41:39 can someone dupe #5350 to #5779 and close #6436 after r19302? Jan 25 10:47:44 I have a ar71xx/ag71xx specific question/problem: Jan 25 10:51:14 I'm trying to register a phy driver to eth1 Jan 25 10:51:52 but config_init never gets called for any eth1 connected phy, only eth0 (it is connected to both, but the cpu port is at eth1) Jan 25 10:52:26 kaloz * r19324 /trunk/toolchain/gcc/ (6 files in 2 dirs): add preliminary support for the latest CodeSourcery toolchain Jan 25 12:21:52 build #23 of atheros-trunk is complete: Failure [failed compile_12] Build details are at http://tksite.gotdns.org:8010/builders/atheros-trunk/builds/23 Jan 25 14:24:51 juhosg * r19325 /trunk/target/linux/ar71xx/ (7 files in 2 dirs): ar71xx: merge generic machine code into setup.c Jan 25 14:24:54 juhosg * r19326 /trunk/target/linux/ar71xx/files/drivers/net/phy/rtl8366s.c: rtl8366s: simplify mib counters handling Jan 25 14:24:57 juhosg * r19327 /trunk/target/linux/generic-2.6/config-2.6.32: kernel: add missing config symbol Jan 25 14:25:00 juhosg * r19328 /trunk/target/linux/ar71xx/ (config-2.6.31 config-2.6.32): ar71xx: sync kernel configs Jan 25 15:10:59 hello Jan 25 15:11:18 is /etc/init.d/network restart enough to reload /etc/config/network? Jan 25 15:13:39 juhosg * r19329 /trunk/target/linux/ (20 files in 11 dirs): rtl8366: make it available on all platforms Jan 25 15:38:24 hi [florian] Jan 25 15:38:56 are you the one who is in charge of the target ar7? Jan 25 17:11:01 nbd * r19330 /trunk/package/opkg/patches/012-cleanup-tmpdir.patch: opkg: remove an empty patch Jan 25 17:11:20 nbd * r19331 /trunk/ (98 files in 33 dirs): add preinit modularization work by Daniel Dickinson (cshore) Jan 25 21:03:48 cshore: ping Jan 25 21:04:11 [florian]: ping Jan 25 21:09:06 just figured out in my little mind that developing on top of old version is going to be pain with SVN Jan 25 21:09:20 no local branches or anything... Jan 25 21:10:10 dolphin: there's a git tree available Jan 25 21:10:32 how long history does it have? Jan 25 21:11:35 I bet it's at least few weeks, that'd satisfy... Jan 25 21:12:24 dolphin: it's pretty much the complete svn history afaik Jan 25 21:13:43 does it keep up with trunk well? Jan 25 21:14:53 dolphin: it's automatically synced Jan 25 21:15:08 dolphin: not sure, if with every commit, but at least every few minutes Jan 25 21:15:08 I meant like, per minute/hourly/daily? Jan 25 21:15:15 ah, ok Jan 25 21:15:22 well, that would be good enough Jan 25 21:16:20 I could also try and see what broke ramdisk images between 18th and yesterday Jan 25 21:17:01 dolphin: maybe the new preinit stuff Jan 25 21:17:04 every two minutes Jan 25 21:17:13 dolphin: not sure, when exactly this was commit Jan 25 21:17:23 could have been today Jan 25 21:17:46 rtz2: my only clue (is by now) is that 18th it worked with no modifications, and yesterday it didn't work at all Jan 25 21:18:13 and I was patching the rt288x port which had not bee touched in 7 months, so I figured it would be something more generic that go broken... Jan 25 21:18:38 and now I typo way too much, I should go sleeping... Jan 25 21:20:08 g'nights Jan 25 21:43:39 Bartman007: May I do some question about buildbot? Jan 25 21:45:12 nbd * r19332 /trunk/include/toplevel.mk: unset user overrides for quilt settings to prevent them from interfering with the build process (#6553) Jan 25 21:48:38 nbd * r19333 /trunk/package/hostapd/Makefile: wpa_supplicant: add missing dependencies Jan 25 22:11:22 nbd * r19334 /trunk/package/mac80211/files/lib/wifi/mac80211.sh: mac80211: restrict the killing of wpa_supplicant processes to the right interface (#5991) Jan 25 23:03:55 nbd * r19335 /trunk/package/base-files/image-config.in: fix missing quote (thx, puchu) Jan 25 23:06:48 nbd * r19336 /trunk/target/linux/brcm-2.4/base-files/lib/preinit/ (12 files): rename -brcm-2.4 preinit files to -brcm, as they're shared with brcm47xx Jan 25 23:18:44 ping cshore Jan 25 23:18:46 hi Jan 25 23:32:39 ping cshore Jan 25 23:50:51 pong puchu Jan 25 23:53:06 cshore i it seems your modulariced base-files doesnt work on brcm47xx Jan 25 23:53:21 what happens? Jan 25 23:53:36 i dont have a serial attached but the device isnt reachable ... wiht the old base-files everything is fine Jan 25 23:53:44 eth is dead Jan 25 23:53:53 seems the switch isnt initaliced right ... Jan 25 23:54:02 the devide isnt reachable Jan 25 23:54:04 hmm....I'll look at it now Jan 25 23:54:34 may im the pebkac .. but i dont think so Jan 25 23:54:43 maybe Jan 25 23:56:16 is this with your .32 stuff or normal? Jan 25 23:56:47 2.6.32 Jan 25 23:56:56 but i can test it with 2.6.30 too Jan 25 23:57:08 2.6.32 works nice with the old base-files Jan 25 23:57:32 ok....right, but you'll have to change it for the new one (you use root on usb right?) Jan 25 23:57:53 you need to initialize the usb before hotplug Jan 25 23:58:05 if you're doing usb on root Jan 25 23:58:38 because hotplug causes uses /etc/hotplug.d/block/10-mount Jan 25 23:58:57 when it detects a device is added Jan 25 23:59:56 cshore no i didnt modify your init Jan 25 23:59:59 is vanilla Jan 26 00:00:03 i boot from flash Jan 26 00:00:14 ok, so it's not that then Jan 26 00:00:38 I'm grabbing the current tree now Jan 26 00:01:16 cshore i removed the /etc/hotplug.d/block stuff ... because i dont like the automount feature Jan 26 00:01:48 is brcm47xx a 2.4 kernel? Jan 26 00:03:47 if not, then the preinit needs to be modified...it looks like it's using the preinit stuff from brcm-2.4 Jan 26 00:04:12 cshore no its 2.6 Jan 26 00:04:35 ok, we'll have to find a dev who works on 47xx Jan 26 00:04:44 cshore: there are 2 versions available ... one for 2.6 and one for 2.4 Jan 26 00:04:52 it uses base-files from brcm-2.4 Jan 26 00:04:57 it's just a symlink Jan 26 00:06:07 yeah Jan 26 00:06:15 so it needs its own? Jan 26 00:06:19 yes Jan 26 00:06:51 but i think 2.6 uses the 2.4 image to load its compressed kernel Jan 26 00:07:08 i never understood the exact way how this works Jan 26 00:07:18 puchu: which kernel is it using for preinit? Jan 26 00:07:24 2.6 Jan 26 00:07:28 puchu: it simply uses the same code to generate images Jan 26 00:07:43 okay Jan 26 00:08:07 so the preinit files have to be replaced for 2.6? Jan 26 00:08:25 I believe so...right now it's doing the 2.4 hotplug stuff Jan 26 00:08:42 nbd what do u think? Jan 26 00:09:22 a simple kernel version check in the one place that needs to be treated differently would suffice Jan 26 00:10:28 I think it's 03_init_hotplug_failsafe_brcm that's the problem Jan 26 00:11:16 maybe Jan 26 00:11:28 i can post the old preinit that works on brcm47xx Jan 26 00:11:29 under target/linux/brcm47xx/base-files/lib/preinit Jan 26 00:11:40 or the deconfig of the interface after preinit is done Jan 26 00:12:14 normally somewhere in preinit the led was turned off ... this doesnt happen now Jan 26 00:12:27 the led is always on and the device unreachable Jan 26 00:12:46 puchu: it sounds likes it's not getting out of preinit Jan 26 00:13:10 yeah .. it seems it is stucking somewhere at the beginning Jan 26 00:13:28 nbd: but it works on brcm-2.4, right? Jan 26 00:13:41 puchu: hmm....could be the vlan stuff too Jan 26 00:13:51 cshore: haven't tested Jan 26 00:13:54 but vlan is configured by the kernel module Jan 26 00:14:06 or am i wrong Jan 26 00:14:13 puchu: it requires userspace as well Jan 26 00:14:27 puchu: to actually configure the vlan Jan 26 00:14:46 nbd * r19337 /trunk/package/mac80211/ (23 files in 2 dirs): mac80211: update to wireless-testing 2010-01-25, includes fixes for crypto and retransmissions Jan 26 00:14:51 nbd * r19338 /trunk/package/carl9170/ (5 files in 2 dirs): carl9170: update to 0.9.9.1 Jan 26 00:15:04 cshore: yeah preinit.arch in the old base files had this Jan 26 00:15:06 echo "0 1 2 3 4 ${cpu_port:-5u*}" > /proc/switch/eth0/vlan/0/ports Jan 26 00:15:20 also Jan 26 00:15:21 insmod diag Jan 26 00:15:22 set_state preinit Jan 26 00:15:22 echo '/sbin/hotplug.failsafe' > /proc/sys/kernel/hotplug Jan 26 00:15:32 in /etc/preinit.arch Jan 26 00:15:47 puchu, yes: I probably got the vlan stuff wrong Jan 26 00:15:58 puchu: no way to test it here Jan 26 00:16:16 ican test it here ... but i dont know where to start ... Jan 26 00:16:31 puchu: it's really help if you had a serial console Jan 26 00:16:36 the new modulariced init is a bit complicated to understand Jan 26 00:16:55 yeah i know ... with a serial console i could try to fix it on my on Jan 26 00:17:00 on=own Jan 26 00:17:20 puchu: I've got some docs on the wiki in Inbox Jan 26 00:17:41 but like this ... its like serching for a needling in a complete dark room :) Jan 26 00:17:46 or http://cshore.is-a-geek.com/openwrt/preinit_mount.html Jan 26 00:17:49 needling=needle Jan 26 00:18:18 cshore how long did u work on it? Jan 26 00:19:09 puchu: including the docs a week or two Jan 26 00:19:18 nice work Jan 26 00:19:20 cshore: could you please move your docs to the openwrt wiki at some point? Jan 26 00:19:30 nbd: already there Jan 26 00:19:33 ah Jan 26 00:19:34 ok Jan 26 00:20:03 nbd: It's in inbox Jan 26 00:20:20 ah Jan 26 00:20:24 so it needs to be moved Jan 26 00:21:15 nbd: yes Jan 26 00:22:04 cshore so do u have an idea what should be changed? Jan 26 00:22:23 does someone in here have a brcm47xx device with a serial attached? Jan 26 00:22:26 puchu: well I think it's either the failsafe or vlan or both on 2.6 Jan 26 00:23:15 puchu: I really wish I knew of 2.4 was affected as well, though Jan 26 00:23:35 cshore: are you interested in moderator access? Jan 26 00:23:46 nbd: sure Jan 26 00:24:10 My account is cshore Jan 26 00:24:47 cshore i never tested 2.4 ... the best would be to have a serial attached ... Jan 26 00:25:04 about 2 years ago the last time Jan 26 00:25:18 cshore: done Jan 26 00:32:30 nbd: how do you move a page in the namespace? Jan 26 00:34:13 cshore will u look at it now or maybe at a later point of time? Jan 26 00:34:25 puchu: shortly Jan 26 00:34:43 puchu: but I might not figure it out Jan 26 00:34:47 okay Jan 26 00:35:33 the only problem is ... the device i flash i my connection to the inet ... so when i doesnt work ... i have to reflash it ... takes some minutes ... so its no big deal ... but im not reachable for this time Jan 26 00:35:51 i flash i/i flash is Jan 26 00:43:02 http://wiki.openwrt.org/doc/techref/preinit_mount Jan 26 00:44:31 build #22 of iop32x-trunk is complete: Failure [failed compile_7] Build details are at http://tksite.gotdns.org:8010/builders/iop32x-trunk/builds/22 Jan 26 00:54:39 puchu: I think I found the problem: I had a typo Jan 26 00:54:58 ohh Jan 26 00:55:00 where? Jan 26 00:55:14 15_set_preinit_interface_brcm Jan 26 00:55:27 port_net_echo Jan 26 00:55:39 I was missing && before { Jan 26 00:55:54 after /proc/net/dev > /dev/null ] Jan 26 00:56:35 hmm Jan 26 00:56:36 puchu: it might not be a kernel version thing after all Jan 26 00:56:40 this one is correct Jan 26 00:56:42 [ -n "$pi_ifname" ] && grep "$pi_ifname" /proc/net/dev >/dev/null ] Jan 26 00:56:45 closed Jan 26 00:56:47 ] Jan 26 00:56:52 but but not opend? Jan 26 00:56:56 ] && { Jan 26 00:57:12 no i mean Jan 26 00:57:16 shouldnt [ -n "$pi_ifname" ] && grep "$pi_ifname" /proc/net/dev >/dev/null ] Jan 26 00:57:17 be Jan 26 00:57:21 [ -n "$pi_ifname" ] && [ grep "$pi_ifname" /proc/net/dev >/dev/null ] Jan 26 00:57:31 yes Jan 26 00:57:33 that too Jan 26 00:58:15 should be : Jan 26 00:58:15 [ grep "$pi_ifname" /proc/net/dev >/dev/null ] && { Jan 26 00:58:56 [ -n "$pi_ifname" ] && [ grep "$pi_ifname" /proc/net/dev >/dev/null ] && { Jan 26 00:59:04 yes Jan 26 00:59:12 no Jan 26 00:59:18 except; let me test; I'm not sur ethe [] belong Jan 26 00:59:18 grep shouldn't be enclosed by [ ] Jan 26 00:59:42 nbd: right; I was wondering about that Jan 26 00:59:55 [ -n "$pi_ifname" ] && grep "$pi_ifname" /proc/net/dev >/dev/null && { Jan 26 01:00:55 - [ -n "$pi_ifname" ] && grep "$pi_ifname" /proc/net/dev >/dev/null ] { Jan 26 01:00:55 + [ -n "$pi_ifname" ] && grep "$pi_ifname" /proc/net/dev >/dev/null && { Jan 26 01:01:09 yes Jan 26 01:01:59 damn, I really need more hardware to test on ... so easy to make dumb mistakes Jan 26 01:02:08 i hope it works Jan 26 01:02:13 i rebuild the image and flash it Jan 26 01:02:24 puchu: let me know Jan 26 01:02:56 cshore: yeah more hardware would be nice ... every hardware with a serial attached :) Jan 26 01:03:08 puchu: that too :) Jan 26 01:03:23 puchu: better yet jtag too Jan 26 01:03:27 cshore: another ... okay maybe i should ask when it works :) Jan 26 01:03:30 yeah thats true Jan 26 01:06:47 jtag is only necessary in rare cases Jan 26 01:07:04 nbd: true Jan 26 01:07:19 nbd: but I've used it to recover from a bad brick before Jan 26 01:07:34 nbd: couldn't even get in the CFE Jan 26 01:08:42 nbd: when you're working on the image flashing stuff it happen sometimes (e.g. on brcm63xx having too big an image flashed via CFE will brick the board) Jan 26 01:08:49 ouch Jan 26 01:08:53 that sucks Jan 26 01:09:22 so the image is ready Jan 26 01:09:26 i flash it Jan 26 01:09:33 puchu: we cross our fingers Jan 26 01:09:39 me too: :) Jan 26 01:21:18 hmm Jan 26 01:21:20 didnt work Jan 26 01:21:25 same as befor Jan 26 01:21:31 @cshore Jan 26 01:21:55 cshore: I'm back with hardware Jan 26 01:22:02 cshore: you want me to test something? Jan 26 01:24:04 puchu: back Jan 26 01:24:15 cshore: it didnt work Jan 26 01:24:17 rtz2: could you try latest trunk? Jan 26 01:24:53 puch: ok let me look see if I can find another other issues Jan 26 01:25:06 cshore: maybe i should checkout the old base-files and try if it works with them Jan 26 01:25:19 puchu: sure; I'll be a bit Jan 26 01:26:10 okay im going back to opkg: remove an empty patch Jan 26 01:30:20 cshore: if I want to load the input modules for the reset button, where should I do it? Jan 26 01:30:39 before failsafe wait Jan 26 01:31:12 cshore: yes, but how should I name the hook function? Jan 26 01:31:26 rtz2: I'd add a file that is before failsafe wait and make a function and then Jan 26 01:31:26 boot_hook_add failsafe name_of_function Jan 26 01:31:41 sorry Jan 26 01:31:49 boot_hook_add preinit_main name_of_function Jan 26 01:32:18 rtz2: it's not the name that matters, it's adding to to the preinit_main hook before failsafe_wait Jan 26 01:33:14 rtz2: if it doesn't do SIGUSR1 for the reset button then you'll need to modify failsafe_wait Jan 26 01:34:49 cshore did u find another error? Jan 26 01:35:10 puchu: not yet Jan 26 01:35:33 cshore: sow below 30? Jan 26 01:35:42 rtz2: yes Jan 26 01:41:36 ok Jan 26 01:41:46 this may take some time Jan 26 01:43:17 pucu: could you try replacing preinit_net_echo with an empty function; also port_net_echo? Jan 26 01:43:46 wait; I see another error Jan 26 01:44:38 i think i get now a complete rebuild of the toolchain because of the downgrade Jan 26 01:44:46 if [ "$pi_preinit_net_messages" = "y" ] || [ "$pi_failsafe_net_message" = "true"] && [ "$pi_preinit_no_failsafe_netmsg" != "y" ]; then Jan 26 01:44:46 should be Jan 26 01:44:46 if [ "$pi_preinit_net_messages" = "y" ] || [ "$pi_failsafe_net_message" = "true"] && [ "$pi_preinit_no_failsafe_netmsg" != "y" ]; then Jan 26 01:44:54