**** BEGIN LOGGING AT Sun Jun 23 02:59:58 2013 Jun 23 09:19:00 av500: ping Jun 23 09:27:03 av500: when you have time please take a look at https://github.com/ungureanuvladvictor/BeagleDroid/blob/master/BeagleDroid/src/main/java/com/vvu/beagledroid/MainActivity.java Jun 23 09:27:41 i implemented device_filter, when the BBB is attached to the android device the app starts itself and permissions are granted automatically. Jun 23 09:28:47 the only problem is at line 60, at the bulk transfer. it just hangs there, like the board is not sending anything. i checked the endpoint to see the direction of the transfer and it is device->host(IN directon). Jun 23 10:17:56 alo tried to do it in a separate thread than main one same result. Jun 23 14:08:40 * vvu tries to find some android people around Jun 23 17:09:56 doing IO in the main loop is not great indeed. you need to create a "service" or thread isn't there some known sample code? Jun 23 17:20:11 keesj: did in a separate thread Jun 23 17:20:13 same result Jun 23 17:20:43 keesj: http://pastebin.com/V1Bf1sjr Jun 23 17:22:45 it is the same for send or receive *changed endpoint*, just stalls there. and i checked the dmesg the BBB is seen correctly. Jun 23 18:05:56 ZubairLK: now I am, get it all working again? Jun 23 18:06:19 yup. thankfully :) Jun 23 18:06:29 ZubairLK: nice. Jun 23 18:06:43 is there a better way than tftp the kernel uImage file? Jun 23 18:09:14 ZubairLK: i don't know of one, sorry. Jun 23 18:09:25 its ok. seems to work pretty well Jun 23 18:11:29 gregkh: I can't seem to find a way to remove a device overlay. do i need to reboot everytime and apply another? Jun 23 18:14:06 ZubairLK: I think so, I have no experience with them. People are working on dynamic overlays, but that code hasn't been merged yet. Jun 23 18:14:35 I see. I'll stick with restarting for now. emmc makes it boot much faster Jun 23 18:15:59 vvu: line 55 is lacking a "=" Jun 23 18:16:38 no Jun 23 18:18:55 did you read http://developer.android.com/guide/topics/connectivity/usb/host.html (the part about permissions ?) Jun 23 18:20:10 I did something like https://github.com/keesj/android-rtlsdr/blob/master/src/rtlsdr/android/MainActivity.java#54 Jun 23 18:24:45 note that this code used the same Thread approach (I didn't take the time to make it work properly at the signals obtained where looking really weird Jun 23 18:28:30 keesj when we have a device filter aren't permissions granted automatically? Jun 23 18:31:49 try it Jun 23 18:32:14 I can open the device so i have permissions Jun 23 18:34:34 Also i have read the host documentation and followed those steps Jun 23 18:52:39 ka6sox: any idea of reasons why the pru driver would fail to open/ Jun 23 18:52:56 is loading uio_pruss via 'modprobe uio_pruss' Jun 23 18:52:59 the only thing you need to do? Jun 23 18:53:34 koen: Jun 23 18:53:37 gregkh Jun 23 18:53:39 or anyone else :) Jun 23 18:53:41 Any kernel I compile and tftpboot myself results in missing usb0 entries 192.168.7.2 .. Jun 23 18:53:45 If i load the uImage-dtb-am335x-boneblack file. I manage to tftp boot. but no usb0 (dont need it. but i'm trying to replicate the stock kernel functionality first) Jun 23 18:53:49 Even if I try to load the uImage file that came with the bbb. Jun 23 18:53:51 TFTP boot doesnt work and gives Error: unrecognized/unsupported machine ID (r1 = 0x00000e05). Jun 23 18:53:55 uEnv is Jun 23 18:53:57 optargs=quiet drm.debug=7 Jun 23 18:54:01 ipaddr=192.168.0.250 Jun 23 18:54:03 serverip=192.168.0.1 Jun 23 18:54:05 #run tftpboot 0x80200000 uImage Jun 23 18:54:07 #uenvcmd=run loaduimage run loadfdt run mmcargs bootz 0x802800000 - 0x80F80000 Jun 23 18:54:11 bootargs=console=ttyO0,115200n8 root=/dev/mmcblk0p2 ro rootfstype=ext4 rootwait Jun 23 18:54:13 uenvcmd=echo "Booting from Network";tftpboot 0x80200000 uImage; bootm 0x80200000 Jun 23 18:54:17 Toolchain could be code sourcery or the one OE generates Jun 23 18:54:19 but i'm not sure as i have confused them up and dont remember which folder I compiled with which toolchain Jun 23 18:56:20 ZubairLK: I don't know, sorry. Jun 23 18:56:39 its more of a BBB question than a kernel question :) Jun 23 20:02:57 <_av500_> ZubairLK: that error seems to be toolchain related Jun 23 20:03:04 <_av500_> koen mentioned something like that Jun 23 20:03:09 <_av500_> so please use the angstrom toolchain Jun 23 20:03:12 <_av500_> node code sourcery Jun 23 20:03:40 i see. but I'm unable to boot the uImage that came with the BBB via tftpboot. Jun 23 20:03:50 the ones i compile myself can be booted via tftp.. Jun 23 20:04:15 or are we talking about the missing usb0 stuff? :s Jun 23 20:04:29 _av500_: Jun 23 20:06:01 <_av500_> hmm Jun 23 20:06:09 <_av500_> so what do you get with the supplied uimgae? Jun 23 20:06:12 <_av500_> what error? Jun 23 20:31:17 Error: unrecognized/unsupported machine ID (r1 = 0x00000e05) Jun 23 20:31:38 _av500_: I get the above if I use the stock uImage3.8.13 with the stock uboot and try to tftp it Jun 23 20:31:52 uEnv is Jun 23 20:31:54 optargs=quiet drm.debug=7 Jun 23 20:31:58 ipaddr=192.168.0.250 Jun 23 20:32:00 serverip=192.168.0.1 Jun 23 20:32:02 #run tftpboot 0x80200000 uImage Jun 23 20:32:04 #uenvcmd=run loaduimage run loadfdt run mmcargs bootz 0x802800000 - 0x80F80000 Jun 23 20:32:07 <_av500_> and the dt? Jun 23 20:32:08 bootargs=console=ttyO0,115200n8 root=/dev/mmcblk0p2 ro rootfstype=ext4 rootwait Jun 23 20:32:10 uenvcmd=echo "Booting from Network";tftpboot 0x80200000 uImage; bootm 0x80200000 Jun 23 20:32:18 <_av500_> you need to supply the DT Jun 23 20:32:25 aha Jun 23 20:32:29 <_av500_> you just load uimage over tftp Jun 23 20:32:30 and what/where do i do that? Jun 23 20:32:31 <_av500_> not the DT Jun 23 20:32:39 <_av500_> same as uimgae Jun 23 20:32:40 <_av500_> sec Jun 23 20:32:57 so the uImage-dt-beagleblack that I compile myself already has a dt in it? Jun 23 20:33:04 <_av500_> loadtftp=tftpboot $fdtaddr /boot/am335x-bonelt.dtb ; tftpboot $loadaddr /boot/uImage Jun 23 20:33:07 <_av500_> ZubairLK: maybe Jun 23 20:33:12 <_av500_> appended DT is one way Jun 23 20:33:22 thanks! Jun 23 20:33:25 lemme try this Jun 23 20:33:37 <_av500_> you need to adapt to your setup Jun 23 20:33:43 <_av500_> but basically your tftp both Jun 23 20:33:46 <_av500_> DT and kernel Jun 23 20:33:52 <_av500_> to two different addresses Jun 23 20:33:57 <_av500_> that you both pass to bootm Jun 23 20:34:08 <_av500_> but that is same for SD or tftp Jun 23 20:34:10 how to pass second address to bootm Jun 23 20:34:17 <_av500_> see the stock Jun 23 20:34:18 <_av500_> script Jun 23 20:34:36 <_av500_> bootcmd=echo TFTP/NFS boot ...; run loadtftp; run nfsargs; bootm ${loadaddr} - ${fdtaddr} Jun 23 20:34:50 <_av500_> see ^^^ Jun 23 20:34:53 ah Jun 23 20:35:04 i'd see but not observe :p Jun 23 20:35:09 now i understand :) Jun 23 20:35:11 thank-you very much Jun 23 20:36:06 <_av500_> np Jun 23 20:40:35 _av500_: it works! Jun 23 20:41:18 _av500_: now for the missing usb0 on my kernels.. Jun 23 20:41:23 is it toolchains? Jun 23 20:41:35 which toolchain does everybody use Jun 23 20:42:43 http://www.angstrom-distribution.org/toolchains/angstrom-eglibc-i686-armv7a-vfp-neon-v2012.12-toolchain.gz Jun 23 20:42:46 ? Jun 23 20:43:09 or some linaro one? Jun 23 20:44:17 or MACHINE=beaglebone ./oebb.sh bitbake meta-toolchain Jun 23 20:45:59 keesj: tried with the permissions usbmanager.haspermissions and i have them Jun 23 20:46:12 now i totally don`t have any clue what to do anymore Jun 23 20:46:43 _av500_: any idea how i can debug more the app? Jun 23 21:08:01 koen: Does the following make the best toolchain? Jun 23 21:08:03 MACHINE=beaglebone ./oebb.sh bitbake meta-toolchain Jun 23 21:08:17 kernel/uboot/applications. I want to be toolchain worry free. which one do you use? Jun 23 21:11:19 <_av500_> vvu: tomorrow Jun 23 21:11:24 okei Jun 23 21:27:20 i did MAGIC and it worked! Jun 23 21:28:49 vvu: +1 :) Jun 23 21:29:48 thx! Jun 23 21:32:36 keesj: _av500_ i always tried to read using timeout 0, now changed like http://pastebin.com/FehnfpKE and it works Jun 23 21:32:37 <_av500_> vvu: please tell Jun 23 21:32:53 <_av500_> aha Jun 23 21:33:33 can`t explin why, but happy i can read Jun 23 21:33:47 <_av500_> good Jun 23 22:11:32 back Jun 23 23:41:41 Hmm Jun 24 02:49:57 So, What is a good way to debug a kernel module **** ENDING LOGGING AT Mon Jun 24 02:59:58 2013