**** BEGIN LOGGING AT Thu Sep 29 02:59:56 2011 Sep 29 03:23:42 i guess manufacturers could sign anything. doesn't mean it's encrypted though. you can sign plaintext. it just verifies auth Sep 29 03:24:11 you change it though, it fails, and bootloader says f- this Sep 29 03:24:20 where is this? Sep 29 03:25:39 In a firmware iso file of a Vizio device i was planning on purchasing. Sep 29 09:40:01 jmccrohan: well, if they care so much to have the source only oavailable over ssl, hopefully they'll fix their certs promptly then :) Only other choice I can see would be to rehost it yourself for a while, but that's well ugly. Sep 29 10:04:15 [florian]: I'm trying to find this gpio buttons code in bcm63xx to compare, and move the ar231x reset button code to that style, but I can't find it anywhere in arch/mips/bcm63xx, where should I be looking? Sep 29 10:10:01 <[florian]> karlp: the registration of the gpio-buttons is actually done in arch/mips/bcm63xx/boards/board_bcm963xx.c Sep 29 10:10:26 <[florian]> karlp: and the corresponding platform driver is in drivers/input/misc/gpio_buttons.c Sep 29 10:11:04 <[florian]> karlp: basically you have to supply the platform_data with a gpio, debounce delay and keycode Sep 29 10:12:00 <[florian]> karlp: ar71xx can also be used as an example for this Sep 29 10:12:29 how can I test this is working before hand even? on my board, pressing the reset button doesn't even work Sep 29 10:13:11 <[florian]> you can use evtest Sep 29 10:13:28 <[florian]> otherwise, once you have gpiosysfs Sep 29 10:13:33 <[florian]> if you mount debugfs Sep 29 10:13:44 I've got the gpio sysfs running Sep 29 10:14:04 that's working fine, I just had to comment out some of this reset button setup things, that I wasn't convinced was working. Sep 29 10:14:22 <[florian]> you can access the gpio pins status using cat /sys/kernel/debug/gpio Sep 29 10:14:37 <[florian]> press the reset button, and see which gpio pin changes its status Sep 29 10:15:28 <[florian]> note that the gpio pins must be exported using echo N > /sys/class/gpio/export Sep 29 10:16:33 ok. back to board_bcm963xx.c, where in that is the gpio button, debounace and delay set? Sep 29 10:16:41 I can see leds being configured, but no buttons? Sep 29 10:17:00 <[florian]> .buttons = { .. } Sep 29 10:17:23 there's no button in my version of that file at all? Sep 29 10:17:26 <[florian]> the patch to look for is target/linux/brcm63xx/patches-3.1/100-reset_buttons.patch Sep 29 10:17:33 oh, Sep 29 10:17:43 because I'm on atheros, the file hasn't been patched Sep 29 10:17:50 now thinks make more sense. Sep 29 10:17:55 <[florian]> yeah, forgot to mention that ;) Sep 29 10:18:01 I've been looking at the standard kernel versions Sep 29 10:18:12 <[florian]> yeah, the standard kernel has not yet been patched to accept gpio-buttons Sep 29 10:18:22 no wonder it didn't make sense to look at that :) Sep 29 10:18:54 but gpio-buttons is being used for all the owrt boards, and is definitely "the right thing to do" ? Sep 29 10:19:04 <[florian]> it is, yes Sep 29 10:19:19 <[florian]> this is the standard interface for triggering events based on button status Sep 29 10:19:23 ok good good. Sep 29 10:20:04 I'll look through those patches, see if I can verify that the reset button was ever working on this platform, and try and get it all together. Sep 29 10:20:15 marek * r28315 /packages/net/batman-adv/ (3 files in 2 dirs): Sep 29 10:20:15 batman-adv: fix upstream latency bug Sep 29 10:20:15 Signed-off-by: Marek Lindner Sep 29 10:20:24 <[florian]> karlp: I strongly suggest using /sys/kernel/debug/gpio for testing purposes Sep 29 10:20:42 <[florian]> karlp: it provides a very easy way to see which gpio on a system is being changing Sep 29 10:20:49 that shows all exported gpios, and their names and states right? Sep 29 10:20:54 <[florian]> indeed Sep 29 10:21:02 <[florian]> well, only those exported for sysfs use Sep 29 10:21:41 <[florian]> so you need not to register gpio-buttons first to allow that specific gpio to be used for sysfs Sep 29 10:22:09 <[florian]> let's say that sysfs gpio is useful for trial and error until you can map a gpio to its corresponding driver Sep 29 10:22:17 hmm, ar231x isn't using these board_info structures at all though Sep 29 10:22:34 <[florian]> so how is the reset button implemented? Sep 29 10:22:51 static irqreturn_t Sep 29 10:22:51 button_handler(int irq, void *dev_id) Sep 29 10:23:10 <[florian]> oh, it's using an irq generated by a gpio Sep 29 10:23:13 and it was using this pseudo global custom gpiodev object to read and check themselves Sep 29 10:23:19 <[florian]> I do not know how that can fit in gpio-buttons Sep 29 10:23:27 <[florian]> most of the time gpio-buttons is polling the gpio Sep 29 10:23:41 well, I can just replace the old object access with calls from right? Sep 29 10:23:55 <[florian]> let me have a look at this Sep 29 10:24:08 arch/mips/ar231x/reset.c Sep 29 10:24:18 probably in atheros/100-board.patch? Sep 29 10:25:44 <[florian]> it's in 200-reset_button.patch Sep 29 10:26:19 <[florian]> I think you should be able not to request the irq, and get the same functionnality by just polling the gpio Sep 29 10:38:46 acoul * r28316 /packages/libs/libpam/ (4 files in 2 dirs): libs/libpam: update to version 1.1.4 Sep 29 12:05:36 acoul * r28317 /packages/libs/libpam/patches/005-fix_ruserok.patch: libs/libpam: make KanjiMonster happy Sep 29 12:06:30 \o/ Sep 29 12:39:39 acinonyx * r28318 /trunk/package/hostap-driver/files/lib/wifi/hostap.sh: [package] hostap-driver: introduce "disabled" option for wifi-iface section to shutdown single networks on a radio Sep 29 16:13:11 build #86 of at91 is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/at91/builds/86 Sep 29 16:30:21 build #84 of ubicom32 is complete: Failure [failed compile_3] Build details are at http://buildbot.openwrt.org:8010/builders/ubicom32/builds/84 Sep 29 19:48:21 could this still be libpam problem or more openssh? http://paste.pocoo.org/show/s3ddGcTIvLalHIlLbHgt/ Sep 29 23:46:33 acinonyx * r28319 /packages/net/quagga/ (9 files in 2 dirs): [packages] quagga: Update to v0.99.20 Sep 29 23:52:27 acinonyx * r28320 /branches/backfire/package/hostap-driver/files/lib/wifi/hostap.sh: [backfire] backport: "[package] hostap-driver: introduce "disabled" option for wifi-iface section to shutdown single networks on a radio" **** ENDING LOGGING AT Fri Sep 30 02:59:57 2011