**** BEGIN LOGGING AT Fri May 07 02:59:56 2010 May 07 06:46:12 blogic: re patchwork: so the one you have setup is going public? May 07 07:43:46 good morning May 07 07:45:09 jk-: yes May 07 07:45:22 jk-: my laptop died 2 dsays ago and i lost the link :) May 07 07:45:32 i will however find out what it was today :) May 07 09:01:40 build #40 of x86 is complete: Failure [failed shell_3 compile_6] Build details are at http://tksite.gotdns.org:8010/builders/x86/builds/40 May 07 11:24:24 jow * r21394 /trunk/package/broadcom-wl/files/lib/wifi/broadcom.sh: [package] broadcom-wl: check iface pid file when invoking nas, fixes starting nas for additional wireless networks (#7255) May 07 11:40:20 nbd: ping May 07 11:40:31 rtz2: pong May 07 11:42:35 nbd: two things May 07 11:42:49 nbd: firstly about the brcm cache patch May 07 11:42:55 nbd: could you commit it? May 07 11:43:06 sure May 07 11:43:08 nbd: because I don't have access to it May 07 11:43:34 nbd: secondly about the mtd patch May 07 11:43:40 nbd: about the question you had May 07 11:44:24 nbd: I did try to check it out as far as I could, but I don't see a problem May 07 11:44:33 nbd: did you have something specific in mind? May 07 11:47:04 well, doesn't it change the trx length to only cover one erase block? i wonder if that could cause issue with the kernel/rootfs offset check on brcm-2.4 or brcm47xx May 07 11:47:22 imho it might make more sense to just make it cover the kernel May 07 11:48:17 nbd * r21395 /trunk/target/linux/brcm47xx/ (6 files in 3 dirs): brcm47xx: adjust the brcm4710 cache workaround code so that it does not break bcm4716 (patch by Bernhard Loos) May 07 11:49:20 hmmm May 07 11:50:32 build #58 of brcm_2_4 is complete: Failure [failed compile_6] Build details are at http://tksite.gotdns.org:8010/builders/brcm_2_4/builds/58 May 07 11:59:21 nbd: the map driver does some magic, but as far as I cann tell, the size is only used to calculate a new checksum, not for the actual offset of the rootfs May 07 11:59:36 nbd: at least for 2.6 May 07 12:00:01 nbd: and I could change it to cover the kernel, but it will make things more complicated May 07 12:00:03 still, technically it violates the format ;) May 07 12:00:18 well, you can't avoid that May 07 12:00:39 len is supposed to cover the whole image and this will blow up with a jffs2 root May 07 12:00:47 why not? if size covers the kernel and the rootfs is not covered by the trx, it'd work May 07 12:01:00 and that's also the same result of what the kernel code does May 07 12:01:06 it moves the rootfs outside of the trx area May 07 12:04:00 nbd: take a look at an actual device May 07 12:04:23 nbd: i'm 90% sure, len will be 0x1c (the size of the trx header) May 07 12:05:09 nbd: and covering the whole kernel will be possible, but it will make things more complicated and for what? May 07 12:07:29 well, if we ever get boot loaders that check if any trx offset is beyond the trx size limit, it could break May 07 12:20:03 ahh crap, I was wrong May 07 12:20:16 nbd: it does change the len field to the kernel size May 07 12:20:30 nbd: but I still kinda doubt it's worth the effort May 07 12:21:23 nbd: and if broadcom ever changes cfe, it's still possible to do it later May 07 12:32:14 nbd: if I mmap the whole mtdblock device, will it end up completly in ram? May 07 12:32:20 nbd: or only dirty pages? May 07 12:38:43 only dirty pages May 07 12:39:09 rest will be loaded on demand and cached in a way that it can be freed when something else needs ram May 07 12:39:58 nbd: in this case, I will simply mmap the whole dev May 07 12:40:28 nbd: that way it's still easy to adjust the len field to the kernel size May 07 12:40:44 nbd: I will poste a new patch over the weekend May 07 12:41:01 ok, thanks May 07 13:55:30 build #68 of atheros is complete: Exception [exception failed slave lost shell_10 compile_12] Build details are at http://tksite.gotdns.org:8010/builders/atheros/builds/68 May 07 13:55:31 build #64 of ar71xx is complete: Exception [exception shell_3 failed slave lost shell_10 compile_12] Build details are at http://tksite.gotdns.org:8010/builders/ar71xx/builds/64 May 07 16:20:34 nbd * r21396 /trunk/target/linux/ar71xx/ (44 files in 6 dirs): ar71xx: add a new subtarget for nand based devices - saves around 44k on the lzma compressed generic kernel image May 07 16:20:39 nbd * r21397 /trunk/target/linux/ar71xx/ (config-2.6.32 config-2.6.33 config-2.6.34): ar71xx: remove some bogus kernel config overrides May 07 17:43:24 Hi, May 07 17:44:42 Did anyone get WiFi working on Linksys WRT160NL (ar71xx architecture, atheros radio)? May 07 17:46:23 quite a few people are using the wrt160nl with wifi May 07 17:46:27 (including myself) May 07 17:46:47 you should build a recent version of the backfire branch May 07 17:46:52 it now includes ath9k in the default profile May 07 17:46:58 (which was left out earlier) May 07 17:47:09 and it includes many fixes that aren't in the 10.03 point release yet May 07 17:47:10 I have added the ath9k package. May 07 17:47:19 you also need wpad-mini or hostapd-mini May 07 17:47:36 What do those do? May 07 17:48:11 bring up the ap interface and do some aspects of station management May 07 17:49:12 And what is the difference between the two? May 07 17:50:19 wpad-mini also includes wpa_supplicant-mini May 07 17:50:21 in the same binary May 07 17:53:20 Hmm, maybe the lack of one of those was the cause of my problems: the wifi worked in client mode but not in AP mode. May 07 17:53:41 Yepp. Thank you! May 07 17:55:14 After installing wpad-mini, everything seems to be allright. May 07 17:55:46 Well, except the LED. What controls that one? May 07 18:00:31 While we are at it, I have managed to get the device working with a Huawei E1752 cellular modem, but it's a bit of a hack. I'm wondering what the proper procedure would have been. Here's what I did: May 07 18:03:07 Added the check and initialization for the device to /etc/hotplug.d/usb/10-usb: if the device has been plugged in, I run usb_modeswitch to turn it from a CD rom (with the windows drivers) into a cellular modem, then wait for 10 seconds and run ifup wlan May 07 18:03:41 However, this only works if the usb dongle is being plugged in after the device has booted up. May 07 18:04:32 If it is turned on with the dongle already plugged in, I do the initialization in /etc/rc.local which I understand not to be the right way of doing it. May 07 18:11:16 can anybody explain to me the reason for this: https://dev.openwrt.org/browser/trunk/target/linux/brcm47xx/patches-2.6.32/900-disable_early_printk.patch May 07 18:11:46 epointsystem: the hotplug script should still get called, even if the device is plugged in during boot May 07 18:12:16 It doesn't. May 07 18:13:48 That is the behavior that I expected, but unfortunately in practice it does not get called. May 07 18:17:08 good evening May 07 18:36:08 rtz2: early printk prevented (some) buffalo devices from booting properly. https://dev.openwrt.org/changeset/15232 May 07 18:38:50 Bartman007: hmm, thanks May 07 18:45:48 hi Bartman007 May 07 18:52:35 Any idea why hotpug scipts may fail to run for devices that have been attached upon boot? May 07 21:18:30 not openwrt related (sorry!) but I know that there are some really good developers here, I was wondering if anyone could help me clear up a question I have about libusb May 07 21:19:02 the software is running on openwrt though May 07 21:20:01 whoops, two of me May 07 22:53:58 build #62 of avr32 is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/avr32/builds/62 May 07 23:47:28 why is e1000e not enabled by default for 10.03? wish it had **** ENDING LOGGING AT Sat May 08 02:59:56 2010