**** BEGIN LOGGING AT Mon Feb 13 02:59:58 2012 Feb 13 04:22:06 nbd * r30489 /packages/net/samba36/ (15 files in 3 dirs): samba36: add a stripped-down multicall samba 3.6, including smbd, nmbd and smbpasswd (partially based on a package by Todd Richmond from #9992) Feb 13 05:09:13 nbd: revert-05 failed after 24 hours, so your 0004-ath5k-Calibration-re-work.patch looking like the culprit, as you anticipated Feb 13 05:09:48 i have 3 (soon 4) devices testing your patch Feb 13 10:03:07 build #104 of iop32x is complete: Failure [failed compile_10] Build details are at http://buildbot.openwrt.org:8010/builders/iop32x/builds/104 Feb 13 12:24:49 nbd by the way,regarding those tickets of users with tp-link wr1043nd wireless low throughput, i tested these days on a android tablet, got like 20-25 Mbit/sec, it seems stable but maybe should be a bit more, the tablet has "n", the wr1043nd has "n" too.. 4 meters between me and the router.. Feb 13 12:25:31 just my wireless in the area and a neighboor on another channel Feb 13 12:25:41 how did you measure throughput? Feb 13 12:26:08 speedtest several servers in my isp that i know they have very large results via usual LAN computers test Feb 13 12:26:36 wireless does add a bit of extra latency Feb 13 12:26:38 i guess its their flash applet that measures stuff, you probably know it Feb 13 12:26:44 and that tends to distort the speedtest results Feb 13 12:27:00 anyways its goddam stable, Feb 13 12:27:09 i think 20-25 mbit/s sounds pretty decent for a speedtest result with a single-stream N device Feb 13 12:27:22 especially with a mobile device Feb 13 12:27:36 i was just saying, im very satisfied, yes, it has a /mnt/sdcard that i suspect of low rw Feb 13 12:27:39 those tend to produce much less throughput than normal PCI based cards anywhay Feb 13 12:31:56 nbd * r30490 /packages/net/samba36/patches/ (3 files): samba36: add a few more size reduction patches Feb 13 12:46:54 nbd * r30491 /packages/net/samba36/patches/ (7 files): samba36: remove SAMR support and save a bit more space Feb 13 13:06:41 florian * r30492 /trunk/target/linux/malta/patches-3.2/ (. 001-remove_unused_perf_function.patch): [malta] add support for 3.2 Feb 13 13:06:44 florian * r30493 /trunk/package/kexec-tools/Makefile: [package] kexec-tools: update to 2.0.3 (fixes #9846) Feb 13 13:10:32 nbd * r30494 /packages/net/samba36/patches/190-no_srvsvc.patch: samba36: remove more parts of srvsvc Feb 13 15:07:21 hi everyone Feb 13 15:08:01 as far as i can see, openwrt does not support dhcpv6, am i right? Feb 13 15:09:54 it doesn't provide a nice uci way of configuring it, no Feb 13 15:10:20 if you need a DHCPv6 /client/ on the box, wide-dhcpv6-client is hotplugged Feb 13 15:10:24 i can't see any traces of dhcpv6 in base-files-network neither Feb 13 15:10:47 if you want something more fruity, I have a package that will build the latest ISC DHCP daemons Feb 13 15:10:51 client, server and relay Feb 13 15:11:03 oh! wide-dhcpv6 Feb 13 15:12:36 ya-hoo! it supports rfc3633! Feb 13 15:12:57 there's also Dibbler Feb 13 15:16:23 i wouldn't trust something named "dibbler" to do ANYTHING with my network configuration... Feb 13 15:18:35 then you have a markedly idiotic policy when it comes to deciding what software to trust Feb 13 15:18:48 Dibbler is actively developed and maintained, Wide is not Feb 13 15:38:08 raorn wide dhcp6 works quite nice Feb 13 15:38:38 i like how wide is packaged in openwrt Feb 13 15:44:15 build #137 of s3c24xx is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/s3c24xx/builds/137 Feb 13 15:48:06 juhosg * r30495 /trunk/target/linux/ramips/ (45 files in 12 dirs): ramips: add preliminary support for the RT3662/RT3883 SoCs Feb 13 15:48:08 juhosg * r30496 /trunk/package/kernel/modules/usb.mk: package/modules: enable ECHI/OHCI for RT3883 Feb 13 15:48:09 juhosg * r30497 /trunk/target/linux/ramips/files/drivers/net/ramips.c: ramips: raeth: enable MDIO code for RT3883 Feb 13 15:48:11 juhosg * r30498 /trunk/target/linux/ramips/ (11 files in 8 dirs): ramips: add preliminary support for the Asus RT-N56U board Feb 13 15:48:12 juhosg * r30499 /trunk/target/linux/ramips/rt3883/profiles/00-default.mk: ramips: rt3883: update package list in the default profile Feb 13 16:01:36 nbd * r30500 /trunk/include/quilt.mk: quilt: add a NO_RECONFIGURE override for compiling a package without re-running Build/Configure Feb 13 16:01:48 nbd * r30501 /trunk/package/block-mount/files/40-mount: block-mount: do not auto-mount anonymous mtdblock devices (fixes #10960, #10850) Feb 13 16:35:34 florian * r30502 /trunk/target/linux/ar7/patches-2.6.37/100-titan-gpio-size.patch: [ar7] fix titan gpio register size Feb 13 16:35:37 florian * r30503 /trunk/target/linux/malta/Makefile: [malta] update to 3.2 Feb 13 16:37:20 build #136 of at91 is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/at91/builds/136 Feb 13 16:38:02 hcg * r30504 /trunk/package/hostapd/files/wpa_supplicant.sh: [package] hostapd: correctly pass 64 char hex PSK to wpa_supplicant Feb 13 16:40:02 build #134 of ubicom32 is complete: Failure [failed compile_1] Build details are at http://buildbot.openwrt.org:8010/builders/ubicom32/builds/134 Feb 13 16:54:11 meh, just recovered my "bricked" wr703n with a very nice serial done by my father Feb 13 16:54:31 gosh it would be so cool if you guys let wireless on by default.. Feb 13 16:54:46 anyway, i see eth0 in /etc/config/network but no eth0 interface Feb 13 16:55:14 so i guess, reconfigure wireless in client mode and try trunk firmware Feb 13 17:18:22 guys, pleace chek out my patches for tayga. just sent to -devel@ Feb 13 17:18:32 *check Feb 13 17:21:24 nbd * r30505 /packages/net/samba36/patches/ (20 files): samba36: reorganize patches, fix windows 7 compatibility Feb 13 17:21:26 nbd * r30506 /packages/net/samba36/Makefile: samba36: install smbpasswd symlink Feb 13 17:42:05 juhosg * r30507 /trunk/target/linux/ar71xx/base-files/lib/ar71xx.sh: ar71xx: add model detection for TL-WR842N/ND Feb 13 17:42:06 juhosg * r30508 /trunk/target/linux/ar71xx/ (generic/profiles/tp-link.mk image/Makefile): ar71xx: add profile and build image for the TP-Link TL-WR842N/ND v1 boards Feb 13 17:42:43 new router has been baptized in holy openwrt :) Feb 13 18:58:09 just don't get it too wet Feb 13 19:53:26 how well does 2.4GHz propagate under water, anyway? Feb 13 19:54:38 jow_laptop: I'd like to finish up those dnsmasq changes we were making to allow set/tag pairs. have to say, though, I don't understand why the dnsmasq script creates a list of args instead of just writing out a /tmp config file. Feb 13 19:55:02 a config file is easier to stare at and debug... and that's the way most other services work, anyway. Feb 13 19:57:59 philipp64|laptop: are you planing an under water wifi ? Feb 13 19:58:36 just the apocalypse. it is 2012, after all. Feb 13 19:59:06 I'm 160km away from a dam that's seismically unsound. :-( Feb 13 20:11:59 http://g.co/maps/8mweh Feb 13 20:16:08 big lake, it would end pretty bad for all the people in near the river if the dam bursts Feb 13 20:16:32 RF underwater propagates pretty much not at all Feb 13 20:19:48 so that's the reason submarines are using very low frequencies to communicate? Feb 13 20:19:54 yes Feb 13 20:20:39 water could hardly be a worse medium for RF. Feb 13 20:20:49 even over very short range Feb 13 20:23:25 I know that submarines use LIDAR instead of Radar to avoid terrain collisions... but probably also because it's harder to detect at great range... so you can do it without giving away your location. Feb 13 20:49:29 I am trying alix.3d3 x86 openwrt build in vmware workstation, and i get: This kernel requires the following features not present on the CPU: 3dnow . Unable to boot - please use a kernel appropriate for your CPU." Feb 13 20:49:57 was thinking this old stuff are not that important in .vmdk... Feb 13 20:51:56 crow: you're trying to boot on the wrong CPU type. Feb 13 20:52:30 Geode/LX has 3dnow 3dnowext sse and sseext instructions, and the compiled code is generated using those instructions. Feb 13 20:53:04 philipp64|laptop i see, well i was hoping to try this build before doing dd to CF from alix :) Feb 13 20:53:57 you should be able to reconfigure your VM to the correct CPU type. Feb 13 20:56:36 nbd * r30509 /packages/net/samba36/ (6 files in 2 dirs): samba36: add even more size optimizations, binary size is now down to about 2.7M uncompressed Feb 13 20:57:50 meh, shame we cant separate kmod-ipt-nat and every nat thingie from a simple firewall iptables package Feb 13 20:58:12 im force to include useless nat stuff in this wr703n bridge firmware :( Feb 13 21:11:01 crow: you can always change the CFLAGS in target/linux/x86/alix2/target.mk and change the -march=i586 and take out the specific instruction subset extensions (3dnow and sse). Feb 13 21:11:44 I have Parallels on my Macbook and VMplayer on my Fedora box... but neither allows very fine-grained selection of CPU types. Feb 13 21:11:54 philipp64|laptop well it actually doesnt matter i was hope i can test it before getting alix.3d3 down for maintrance. and for that i would like to use "as it is" Feb 13 21:12:24 i also checked if there is way to change these in cpu, but couldnt find anything regarding that Feb 13 21:12:26 not sure we understand each other. you can build an image using the i586 instruction set and run that as a VM. Feb 13 21:12:40 yes i understand you. Feb 13 21:12:40 you just need to recompile. Feb 13 21:13:14 also worth noting that the Alix platform driver looks for a specific string in the ROM address space... Feb 13 21:13:33 which a VM probably wouldn't emulate... Feb 13 21:13:36 btw. are you doing all that alix2 stuff in openwrt? Feb 13 21:13:55 (although I know that Qemu lets to fine-tune the emulation of hardware including copying BIOS images, etc.) Feb 13 21:14:09 sorry, all of which alix2 stuff? Feb 13 21:14:27 http://www.mail-archive.com/platform-driver-x86@vger.kernel.org/msg02851.html Feb 13 21:14:33 I own the platforms and wrote the various Geode definitions and platform drivers... Feb 13 21:14:49 yup, I am him. Feb 13 21:15:29 ok, than you for work with Geode definitions and platform drivers :) Feb 13 21:15:42 no problem. glad someone is able to make use of them. Feb 13 21:16:56 (i will try), to see how it work, was planing to use openwrt from first, but then i tried gentoo i486 and leaved it as that, but now i need new chalanges .. Feb 13 21:27:47 nbd * r30510 /packages/net/samba36/ (Makefile patches/250-remove_domain_logon.patch): samba36: disable some unused modules Feb 13 21:27:48 nbd * r30511 /packages/net/samba36/patches/110-multicall.patch: samba36: fix some linking issues Feb 13 21:37:09 This isn't openwrt related, but I have a quick question regarding flash technologies.. Feb 13 21:37:20 I'm working with a flash chip (http://download.siliconexpert.com/pdfs/2005/05/15/semi_ap/3/sam/flash%20memory/ds_k9f2808u0c_rev29.pdf) Feb 13 21:37:45 how can I tell if it supports sub-pages? I'm assuming that it does not, but I'd like to understand what to look for. Feb 13 21:41:49 woo . Feb 13 21:42:13 Austrailian cows today Feb 13 22:00:31 nbd * r30512 /packages/net/samba36/patches/300-assert_debug_level.patch: samba36: adjust the debug level of asserts to save some more space Feb 13 23:07:18 nbd: ping Feb 13 23:08:13 nbd: sucessfully trolled in the cyanogemod touchpad channel about putting OpenWrt on it. Remind me why that was a good idea? ;-) Feb 13 23:10:49 heh Feb 13 23:11:13 the idea is to use the openwrt build system and patch structure to be able to make bigger changes with less maintenance effort Feb 13 23:11:35 well, at least after the initial porting effort is complete, which will still take some time Feb 13 23:13:16 right, they started ranting about optware and native builds Feb 13 23:13:32 build #95 of mpc52xx is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/mpc52xx/builds/95 Feb 13 23:13:33 optware I take it is the dd-wrt packaging system Feb 13 23:13:54 optware is something else Feb 13 23:13:56 and naturally they were focussed on openwrt being a "router OS" Feb 13 23:14:00 not sure what build system it's based on these days Feb 13 23:14:34 much of the point was missed in any case. But the setup for the CM9 build seems insanely complex Feb 13 23:14:44 yeah Feb 13 23:14:58 it's still going to be a lot of work to port this to openwrt though Feb 13 23:15:16 for a pure openwrt build, why? Feb 13 23:15:31 it's all ARM packages, the partitioing is a little tricky,but still Feb 13 23:15:45 the kernel seems kinda closed though. Feb 13 23:15:48 3.x anyway Feb 13 23:15:52 well, at least the way i planned it it's going to be some work Feb 13 23:15:58 because i want android to just be some packages on top of openwrt Feb 13 23:16:07 for sure Feb 13 23:16:41 what system provides the underlying Linux in CM9 et al? Is it a Google custom thing? Feb 13 23:18:03 google has their custom build system, custom libc, custom user space crap, etc. Feb 13 23:18:14 it doesn't really look like a normal linux system Feb 13 23:18:24 yuck Feb 13 23:20:35 there could be a lot of comedy e.g. in their libc for Android Feb 13 23:20:57 their libc sucks Feb 13 23:21:12 it's very hackish and only supports what android needs and virtually nothing else Feb 13 23:21:18 aw Feb 13 23:21:37 but many of the android apps could be based upon such assunmptions Feb 13 23:22:02 most android apps do not have native components Feb 13 23:22:09 so they'll run just fine in a different environment Feb 13 23:22:34 as for the others, i was planning on integrating the android libc as an optional compatibility layer package Feb 13 23:34:05 hm Feb 13 23:34:31 and the processor on the phone you are targetting is what? Feb 13 23:36:48 and is closest to what in openwrt platform wise? Feb 13 23:40:29 samsung galaxy s2 Feb 13 23:40:33 no openwrt platform for that yet Feb 13 23:40:37 but i'll create one soon Feb 13 23:41:23 ok, ARM A9 Feb 13 23:41:47 1 GHz ARM Cortex-A8 based CPU core http://en.wikipedia.org/wiki/Samsung_Galaxy_S Feb 13 23:41:58 wonder how to find the specs for my gt-b5510 :S Feb 13 23:42:02 he said S2 though Feb 13 23:42:05 aka galaxy y pro Feb 13 23:42:12 but touchpad is A8 Feb 13 23:42:17 s2 dual core variant i guess ? Feb 13 23:42:31 let's not guess please. Feb 13 23:43:03 http://en.wikipedia.org/wiki/Samsung_Galaxy_S_II Feb 13 23:43:22 assume that wikipedia is correct :) Feb 13 23:44:11 nbd: looking at getting the HTC Sensation, that's different processor again Feb 13 23:44:30 does it have an unlocked boot loader? Feb 13 23:44:32 but I'm more interested in the Touchpad Feb 13 23:44:55 it has convoluted instructions for install cyanogenmod, so I assume for some value of "unlocked" Feb 13 23:49:16 i wonder how far is android devices going to be used with openwrt if that is ever planned :) Feb 13 23:50:40 own creative idea -> portible luci with touchscreen as remote interface for a working wrt device kinda like a active photo frame on the desk Feb 13 23:52:29 get going Feb 13 23:52:58 nbd: and thank you for pushing libc stuff finally. I'll complain after I check the precise changes :o Feb 14 00:09:44 * Chocks checks and complains Feb 14 00:29:14 nbd * r30513 /packages/net/samba36/ (3 files in 2 dirs): samba36: more size optimizations and configurable debug level Feb 14 00:44:03 Chocks: [x] send patches Feb 14 00:44:12 did Feb 14 00:44:35 well, they are less relevant now, since it's 2.14 not 2.13 Feb 14 00:45:16 maybe a resend with 2.14? Feb 14 00:45:29 aw..., just for you. Feb 14 00:45:32 I do it special Feb 14 00:47:06 nbd: lol sgs2 Feb 14 00:47:15 nbd: miui not good enough for you? Feb 14 00:48:37 nbd: ever booted gateworks boards? i think i just softbricked mine using cns3xxx image Feb 14 00:48:57 I hacked on those years ago Feb 14 00:48:58 or do we have a cns3xxx guru Feb 14 00:49:02 Avila et al Feb 14 00:49:23 SmugLeaf: what's miui? Feb 14 00:49:25 but it's been too long, so don't ask me about them Feb 14 00:50:00 SmugLeaf: haven't done any tests with cns3xxx in a while Feb 14 00:50:21 http://weedy.ca/miuiandroid/2.2.10/miuiandroid_SGS2-2.2.10.zip Feb 14 00:50:30 ah Feb 14 00:50:57 well, i'm interested in bending the android system to my will, and making it easy to maintain bigger changes Feb 14 00:51:04 that's what i want to use the openwrt build system for Feb 14 00:51:17 Chocks: http://dpaste.com/702823/ :/ Feb 14 00:51:20 frozen Feb 14 00:51:33 when it comes to low-level stuff and platform porting, android is a steaming pile of poo Feb 14 00:51:55 nbd: https://github.com/MiCode Feb 14 00:52:09 chinas GB scripts and shit Feb 14 00:52:45 i'm not really into the whole 'modding' stuff Feb 14 00:52:53 i like to rebuild stuff from scratch Feb 14 00:52:57 yes Feb 14 00:53:04 using better tools than what the original was made of Feb 14 00:53:12 miui is based on CM7/9 Feb 14 00:53:12 cyanogenmod does seem to be a bit cowboy Feb 14 00:53:20 which is based off ASOP Feb 14 00:53:22 ok, a lot cowboy Feb 14 00:53:32 they throw in whatever they can make work Feb 14 00:53:38 Chocks: yeah they are trying to fix that in CM9 Feb 14 00:53:50 since MIUI spanked them on polish Feb 14 00:53:51 one thing i also don't like about android is that it's so monolithic Feb 14 00:53:52 I know, but that kind of culture takes ages to fix Feb 14 00:54:06 it expects to be compiled as a big ball of mud Feb 14 00:54:12 yep Feb 14 00:54:20 with openwrt i'm changing that ;) Feb 14 00:54:32 i've ported dalvik already Feb 14 00:54:43 split up into several packages, which are compiled separately Feb 14 00:54:52 it even runs on a mips router ;) Feb 14 00:54:59 \o/ Feb 14 00:55:06 \o/ Feb 14 00:55:20 so the next step is to port the framework code and gui stuff Feb 14 00:55:21 app launcher over serial console Feb 14 00:55:25 :V Feb 14 00:55:32 shouldn't be very hard, but still takes some time Feb 14 00:55:36 Chocks: any idea with gateworks shit? Feb 14 00:55:38 and after that it gets interesting Feb 14 00:55:55 smug: sorry no. like I said, it's been a long time. but I'm pretty sure it can be recovered Feb 14 00:56:01 then i want to rip out crappy network abstraction code from android, and make android talk to the openwrt components Feb 14 00:59:41 Chocks: well it's a ramdisk Feb 14 00:59:45 i just need to power cycle Feb 14 00:59:55 but i'm at home 10km away Feb 14 01:00:16 just wondered if there was any tips i should know Feb 14 01:00:29 nbd: since you around i have a build system question Feb 14 01:01:39 gateworks has a big ass patch Feb 14 01:01:41 http://svn.gateworks.com/bsp/generic/trunk/targets.patch Feb 14 01:01:55 in it they have a buch of trunk/target/linux/cns3xxx/config-default shit Feb 14 01:02:04 how should i be doing this? make a target? Feb 14 01:02:45 just the config-default http://dpaste.com/702828/ Feb 14 01:03:23 how should you be doing what? Feb 14 01:04:00 like intergrate all of there shit Feb 14 01:04:21 no idea, does the patch still apply properly? Feb 14 01:04:22 since i'm guessing some thing they have changed cause me to lockup the cpu Feb 14 01:04:44 maybe, i honestly just tried to use vanilla openwrt at first Feb 14 01:05:30 there are some difference in memory layout, which i think haven't been integrated in openwrt yet Feb 14 01:05:50 Kaloz maintains that platform, he probably knows more about that Feb 14 01:05:59 http://dpaste.com/702831/ Feb 14 01:06:01 mostly Feb 14 01:06:07 Kaloz? cool Feb 14 01:06:24 Kaloz: sup, i'm going to be a pain in your ass for the next week ~_~ Feb 14 01:09:05 nbd: I brought home a few "broken" G300NH routers and set up wds a few places around the house. I couldn't believe how simple it was to configure. Feb 14 01:09:24 by broken, their usb ports were physically damaged, but the radios are fine. Feb 14 01:09:39 so, um, thanks for the ath9k work. :) Feb 14 01:10:24 fullstop: nah, ath9k worked before nbd poked it Feb 14 01:10:33 he just made it FAST AS FUCK Feb 14 01:10:37 and stable Feb 14 01:10:45 it Feb 14 01:10:48 has been rock solid here. Feb 14 01:11:03 great on my wndr3700 and rspros Feb 14 01:11:16 and tew-someting Feb 14 01:12:28 brp542 or something like that. Feb 14 01:12:45 I had one of those, but the radio died. Feb 14 01:15:45 mine has a six in it i think Feb 14 01:16:04 SmugLeaf: ath9k worked before i poked it? that's news to me ;) Feb 14 01:16:49 it was in pretty bad shape when i started working on it Feb 14 01:17:02 SmugLeaf: 652? Feb 14 01:17:19 I forget the number, but it was a cheap little thing. I think I got it for $9 USD Feb 14 01:18:53 Also, I used ath9k early in its life. It was, as nbd stated, in pretty bad shape. Feb 14 01:21:37 heh, i just checked the diffstat of all my linux-wireless ath9k commits Feb 14 01:21:42 i still have more deletions than insertions ;) Feb 14 01:21:45 64 files changed, 21918 insertions(+), 23882 deletions(-) Feb 14 01:23:38 nbd: loling Feb 14 01:26:17 nbd, I know you are more of wireless, but how familiar are you with nand flash and ubifs? Feb 14 01:27:24 played with it for a very short time a few years ago Feb 14 01:27:34 on my nand flash, jffs2 works great. ubifs is proving to be far more difficult Feb 14 01:27:36 still remember some things, but don't really have much experience with that stuff Feb 14 01:27:43 just wondering. Thanks. Feb 14 01:27:44 i've never used ubifs Feb 14 01:27:51 i've just played with ubi+jffs2 once Feb 14 01:28:09 what issues do you have with ubifs? Feb 14 01:28:18 as flash gets larger, you might see more ubifs on new routers in a year or so. Feb 14 01:28:36 I've not been able to mount a volume yet, which should be pretty simple. Feb 14 01:28:57 However, I'm not 100% certain that I have the flash layout correct and I don't know if my flash can do sub-page writes or not. Feb 14 01:29:19 last time i used ubi was in an emulated environment Feb 14 01:29:31 The flash layout looks good and everything was in the right place in /dev/mtdX Feb 14 01:29:31 not sure if it needed sub-page writes Feb 14 01:29:37 but it definitely did need sub-page reads Feb 14 01:29:41 and the driver that i played with didn't have that Feb 14 01:29:44 so i had to add it myself Feb 14 01:30:01 it was whining about lots of read failures Feb 14 01:30:05 did you see something like that as well? Feb 14 01:30:07 Maybe I'll use the simulator to test what I'm doing. I suppose that it's possible that the flash chip I have is not compatible with ubi Feb 14 01:30:10 Yes. Feb 14 01:30:26 maybe the flash chip is compatible, but the driver isn't Feb 14 01:30:29 I could create the volumes, but if I ubidetach / ubiattach again, it wouldn't scan properly. Feb 14 01:30:37 you should add some debug code to the driver to see how it fails Feb 14 01:30:42 and if it's something that can be dealt with in software Feb 14 01:30:43 I'll check the lower level driver. Feb 14 01:31:07 It's probably not the flash that I'm going to end up with long term, so it might be okay to stick with jffs2 for now. Feb 14 01:31:22 It detects sub-page write capability, but that might be mistaken. Feb 14 01:31:52 yeah, the issues that i had were definitely sub-page read issues Feb 14 01:31:57 check if the driver is able to do that Feb 14 01:32:03 Will do. Thanks again. Feb 14 01:42:30 inb4kaloz is on vacation Feb 14 01:48:01 swalker * r30514 /packages/libs/libsigc++/ (Makefile patches/001-no_doc_tests_examples.patch): [packages] libsigc++: update to 2.2.10 (#10967, thanks jeff), refresh patches Feb 14 02:11:23 nbd: update the pre 0004- ath5k patches are all still running, some over 2 days uptime now. 4 instances of your patched version of r30427 are still running with about 24 hours of uptime. i'll be 98% confident you fixed it in another 24 hours. Feb 14 02:12:20 all the versions starting with 0004 have failed now Feb 14 02:28:34 nice Feb 14 02:28:52 any gain calibration messages in those 4 with my patch? Feb 14 02:29:06 in the kernel log **** ENDING LOGGING AT Tue Feb 14 02:59:56 2012