**** BEGIN LOGGING AT Tue Aug 06 02:59:58 2013 Aug 06 04:31:24 <_av500_> ds2: the issue is the usb babble error Aug 06 04:31:29 <_av500_> babble meands a protocol violation Aug 06 04:31:33 <_av500_> on the HW level Aug 06 04:31:58 <_av500_> why one usb class protocol triggers that and another does not is beyond me Aug 06 04:32:16 <_av500_> the kernel oops is most likely fallout of the error handling Aug 06 04:32:25 <_av500_> but usually babble means game over for the bus Aug 06 08:03:46 _av500_: yes and no... getting rid of the babble error would avoid the problem but this still points to the error handling on the musb stuff being not quite right Aug 06 08:04:13 it isn't so much different device as much as it is a userland driver answering stuff. Aug 06 08:04:35 could be as simple as we not responding quick enough and MUSB getting confused when the host resends Aug 06 08:05:02 * vvu watches from the shadows Aug 06 08:06:41 so it would be useless to start debugging in usb.c code? Aug 06 08:13:23 vvu: when using a PC, you see the gadgetfs interface come up OK? Aug 06 08:13:31 compared to the android phone Aug 06 08:13:32 ? Aug 06 08:13:35 did not try Aug 06 08:13:45 will do it ~10 min, just woke up Aug 06 08:13:49 ? Aug 06 08:13:55 so what did you try? Aug 06 08:14:05 only gadgetfs with my nexus 7 Aug 06 08:14:09 BBB hooked up with the android Aug 06 08:14:14 http://pastebin.com/Rigpsjqg Aug 06 08:14:26 this is from android Aug 06 08:14:29 ah Aug 06 08:14:30 android side Aug 06 08:14:39 and at the same time error on the linux side Aug 06 08:14:41 ic Aug 06 08:14:53 yep, BBB freezes after this Aug 06 08:15:09 i was thinking to try and put some printfs in the usb.c code to see where it dies Aug 06 08:15:18 maybe it helps somehow to track down the error Aug 06 08:16:29 but this babble error is with g_serial too, when i have g_serial active and connect BBB to my android i see a couple of times the babble error but it works Aug 06 08:16:36 ah Aug 06 08:16:39 interesting Aug 06 08:17:10 i don`t think the babble is the source of the problem Aug 06 08:18:10 well, the babble should not occur Aug 06 08:18:29 i dont recall us having to fight that on omap3 musb Aug 06 08:18:33 we had it on ehci Aug 06 08:18:58 vvu: http://lkml.indiana.edu/hypermail/linux/kernel/1305.3/02332.html Aug 06 08:19:05 can you see if that patch is in your kernel? Aug 06 08:19:36 i`m using the 3.8 kernel from koen`s github Aug 06 08:19:56 even better, then you have the source :) Aug 06 08:19:57 how can i check if i have the patch applied? some smart way or just search for bits of code? Aug 06 08:20:32 git log Aug 06 08:22:00 i did random git log with no filename Aug 06 08:22:10 and i have a commit on my name...that is awkward Aug 06 08:22:32 normal Aug 06 08:22:37 you applied the patches Aug 06 08:22:46 so you are the commiter Aug 06 08:22:50 patch.sh Aug 06 08:22:55 ah that script applies stuff Aug 06 08:23:00 well yes Aug 06 08:25:06 av500: nop, the patches are not there Aug 06 08:44:03 av500: applied the patches but some issues when building Aug 06 08:44:06 drivers/usb/musb/musb_dsps.c: In function 'dsps_musb_restart': Aug 06 08:44:06 drivers/usb/musb/musb_dsps.c:512:2: error: implicit declaration of function 'usb_phy_vbus_off' [-Werror=implicit-function-declaration] Aug 06 08:44:06 drivers/usb/musb/musb_dsps.c:515:2: error: implicit declaration of function 'usb_phy_vbus_on' [-Werror=implicit-function-declaration] Aug 06 08:44:07 cc1: some warnings being treated as errors Aug 06 09:27:00 av500: i think koen knows some stuff about the usb gadget. https://github.com/beagleboard/kernel/tree/3.8 at status says usb gadget : not working Aug 06 09:27:48 vvu: well, gadget serial was working, no? Aug 06 09:27:54 as does gadget ethernet Aug 06 09:27:58 true Aug 06 09:29:07 the patch you gave me breaks building Aug 06 09:29:16 pasted above the error Aug 06 09:31:49 ah well Aug 06 09:33:41 i`ll continue working on the android app Aug 06 09:33:44 meanwhile Aug 06 09:35:10 http://permalink.gmane.org/gmane.linux.usb.general/82587 Aug 06 09:38:54 vvu: lets concetrate for now to make it work Aug 06 09:39:09 make gadgetfs to work? Aug 06 09:39:09 lets worry about make it faster later Aug 06 09:39:12 no Aug 06 09:39:16 the app? Aug 06 09:39:20 boot BBB from the internet Aug 06 09:39:23 with everything Aug 06 09:39:23 ok Aug 06 09:39:41 until 2morrow will have a demo and a progress report Aug 06 09:39:46 ok Aug 06 11:40:46 av500: ping Aug 06 11:41:37 oj Aug 06 11:42:01 every used any of TI`s DaVinci SoC? Aug 06 11:42:46 sadly yes Aug 06 11:42:58 DM6446 by any chance? Aug 06 11:44:04 dm6441 Aug 06 11:44:07 the same Aug 06 11:44:08 just slower Aug 06 11:44:56 my dad has a video-server on which you connect an analog camera and it streams over tcp/ip Aug 06 11:45:01 and has this SoC Aug 06 11:45:25 by any chance know the serial port baudrate? Aug 06 11:45:33 whatever you make it to be Aug 06 11:45:38 or something to get access to it Aug 06 11:45:48 ? Aug 06 11:45:56 solder pins to it Aug 06 11:46:03 we are kinda trying to hack into it mainly Aug 06 11:46:14 it`s BGA on the board Aug 06 11:46:18 I know Aug 06 11:46:32 try to find the serial pins Aug 06 11:46:37 might be on resistors too Aug 06 11:46:43 pins->pads Aug 06 11:46:47 test pads or so Aug 06 11:47:04 3.3v or 5v? any idae? Aug 06 11:47:05 me knowing the SoC does not help you there Aug 06 11:47:06 idea Aug 06 11:47:09 1.8 Aug 06 11:47:19 or was it 3.3 Aug 06 11:47:21 i forgot Aug 06 11:47:51 what do you plan to do with it? Aug 06 11:48:07 this davinci has an autofocus unit in it Aug 06 11:48:23 unlikely Aug 06 11:48:36 it might have 3A Aug 06 11:48:41 the TRM says so Aug 06 11:49:38 DM6446AZWT Aug 06 11:49:39 yes Aug 06 11:49:40 this is my chip Aug 06 11:49:42 3A as I said Aug 06 11:49:52 it means it has some HW units that do some math for you Aug 06 11:50:06 you still have to implement the actual focusing Aug 06 11:50:16 we are trying to build a long range camera Aug 06 11:50:22 all it does it to calculate the contrast values for you Aug 06 11:50:30 yes, that is a good thingie for me Aug 06 11:50:40 because i will drive some servos afterwards based on that contrast value Aug 06 11:50:45 ok Aug 06 11:51:18 a long range ip camera mainly black and white for night Aug 06 11:51:28 with a laser illuminator i think my dad told me Aug 06 11:54:40 av500: this is my videoserver https://picasaweb.google.com/116488451283079970817/ScrapbookPhotos#5908973280582439714 Aug 06 11:55:26 near the sdcard slot has some pins but i think it`s usb there, it is 5V Aug 06 11:56:59 J3 Aug 06 11:57:00 J5 Aug 06 11:57:32 the 4 pin header looks also promising Aug 06 11:57:51 that thingie it`s GND, pin, pin, 5V Aug 06 11:57:53 backside? Aug 06 11:57:54 the pinout Aug 06 11:57:57 yes Aug 06 11:58:01 can easily be serial Aug 06 11:58:10 backside nothing no pins, no pads Aug 06 11:58:20 what is on the back? Aug 06 11:58:35 a second to remove screws from it Aug 06 11:58:37 and pic Aug 06 12:03:01 it`s a 2 board device, stacked onto each other. the lower board from the pic is stacked onto the top one. the lower board has only power source related parts Aug 06 12:03:01 https://picasaweb.google.com/116488451283079970817/ScrapbookPhotos# Aug 06 12:04:33 too blurry Aug 06 12:04:43 but as said Aug 06 12:04:46 go hunt Aug 06 12:05:04 chances are there's an open console spitting out boot messages Aug 06 12:05:23 have a scope? Aug 06 12:05:50 yep tried a bit with it Aug 06 12:05:57 to check where i have signal Aug 06 12:07:42 i know some of their devices use armboot as bootloader Aug 06 12:10:43 this poor device will get so many resets until we eliminate all those pins for the serial Aug 06 12:12:45 who makes it? Aug 06 12:13:33 dahuasecurity.com Aug 06 12:13:56 taiwan company or something Aug 06 12:41:09 av500: i can telnet to it but needs login and pass :) Aug 06 12:46:45 :) Aug 06 12:47:33 passed the telnet Aug 06 12:47:37 it`s a busybox inside :D Aug 06 12:47:49 what pins? Aug 06 12:48:20 over ethernet Aug 06 12:48:28 oh Aug 06 12:48:30 http://pastebin.com/uJMN84kQ Aug 06 12:48:31 is there a getty? Aug 06 12:49:07 ps Aug 06 12:49:32 nop Aug 06 12:51:21 with uboot can a dump of the nand be done? Aug 06 13:11:47 yes Aug 06 13:13:45 A few ways Aug 06 13:13:46 first: Aug 06 13:13:53 nand read.raw - addr off|partition [count] Aug 06 13:13:53 nand write.raw - addr off|partition [count] Aug 06 13:13:53 Use read.raw/write.raw to avoid ECC and access the flash as-is. Aug 06 13:14:21 i have sdcard on the board Aug 06 13:14:22 This is a little complicated to use at first since you have to take into account the size of the OOB too, in memory Aug 06 13:14:35 if uboot likes it, can the nand be dumped on a sdcard Aug 06 13:14:49 Sure Aug 06 13:14:53 fatwrite or ext4write Aug 06 13:15:01 or just raw write with mmc write Aug 06 13:15:09 and then dd it off the card in the linux host later Aug 06 13:15:19 yep Aug 06 13:15:21 If you just want to examine a page Aug 06 13:15:26 thx for the hint Aug 06 13:15:26 nand dump might be easier Aug 06 13:15:31 since that'll just print to screen Aug 06 13:15:38 and for a page, is good enough usually Aug 06 13:15:49 i want to do a backup, i`ll mess around with the board and maybe make it unbootable Aug 06 13:25:34 k Aug 06 13:27:52 vvu: hook your telnet up to the internet Aug 06 13:27:57 I want to poke around in it :) Aug 06 13:28:07 one second, will forward the port from my router Aug 06 13:43:19 av500: do you remember what dsplink did to break ctrl-c? Aug 06 13:43:38 av500: victor has similar problems with a userspace i2c-dev "driver" Aug 06 13:45:05 signal handling Aug 06 13:45:10 heh Aug 06 13:45:14 speak of the devil :) Aug 06 13:45:19 they grabbed the signal Aug 06 13:45:24 and did not care for existing handlers Aug 06 13:45:25 en> av500: victor has similar problems with a userspace i2c-dev "driver" Aug 06 13:45:25 15:45 < av500> signal handling Aug 06 13:45:26 hi everybody Aug 06 13:45:26 15:45 -!- vmayoral [~victor@70.202.217.87.dynamic.jazztel.es] has joined #beagle- Aug 06 13:45:35 I think Aug 06 13:45:37 vmayoral: I was just replying to your email :) Aug 06 13:45:44 av500: that's what I remember as well Aug 06 13:46:36 koen: thanks! i'll try to be a bit more active in the IRC channels Aug 06 13:51:34 koen: btw, vvu has a dm6446 based ip cam thing that has a dsp stack that seems to predate DSPlink... Aug 06 13:52:18 googling random symbols yields nothing Aug 06 13:52:22 it must predate the internet too Aug 06 13:54:34 spox? Aug 06 13:55:20 it might not predate it, it could be a custom non-TI thing Aug 06 13:55:35 iirc TI bought the spox IP, they didn't design it Aug 06 13:56:04 ah Aug 06 13:56:10 didnt see a spox ref yes Aug 06 13:56:48 but all I have is "cat" to look at files :) Aug 06 13:57:13 timanDSPOpenExit2 Aug 06 13:57:22 imanSyncObjectDestroy Aug 06 13:57:22 'strings' also works wonders Aug 06 13:57:26 koen Aug 06 13:57:31 koen: yes if you have it Aug 06 13:57:36 not on a busybox Aug 06 13:57:40 ah Aug 06 13:57:48 im telneting into that thing Aug 06 13:57:55 mvl 2.6.10 Aug 06 13:58:00 the bestest linux ever Aug 06 13:58:12 nah Aug 06 13:58:14 it's 2.6 Aug 06 13:58:18 2.4 was best linux evah Aug 06 13:58:48 RMK promised to sort out the arm serial port naming mess in 2.5 Aug 06 14:04:27 spox..any relation to fex? Aug 06 14:05:36 all the same Aug 06 14:09:02 sounds spoxing great Aug 06 15:10:47 av500: still logged in the videoserver? Aug 06 15:11:45 i can log out Aug 06 15:11:54 if you want to stay more no problem Aug 06 15:12:01 no Aug 06 15:12:06 btw, what does that thing cost? Aug 06 15:12:24 220$ in Romania Aug 06 15:13:01 the price we buy from local distributor Aug 06 15:29:23 panto, ping Aug 06 15:39:45 jkridner|work: have you gotten my hardware e-mails? I sent one Wednesday and re-sent it Sunday when you asked if I had sent it. Aug 06 15:54:18 Thanks for the ping. I started to respond earlier. I'm all interrupt driven these days. Aug 06 15:55:05 ok, thanks. Aug 06 16:26:07 i've been observing that if i upgrade packages through "opkg upgrade ", after a while, i loose the connection because the file /etc/resolv.conf gets overwritten by the Connection Manager Aug 06 16:26:13 Has anybody experienced this? Aug 06 16:46:32 tcort: ping Aug 06 16:48:00 tcort: I don't understand why your address isn't on the spreadsheet. Aug 06 16:48:11 keesj, beng-nl: ^^^ Aug 06 16:59:44 jkridner: I don't have a business address, so a professor receives packages for me... Dr Matt Lemire, Department of Mathematics and Statistics, Carleton University, 1125 Colonel By Drive, Ottawa, ON, K1S 5B6, Canada Aug 06 16:59:53 it's in the spreadsheet Aug 06 17:39:56 mluckham-away, pong **** ENDING LOGGING AT Wed Aug 07 02:59:58 2013