| 03:03 | dbsx | I have a broadcomm bluetooth dongle BCM2046 that works in everything including sheevaplug/dreamplug/. Bluetooth (OBEX) is super unstable on rabeeh kernel and armhf debian |
| 03:06 | dbsx | it worked ok on the original cubox 2.6.xx kernel and ubuntu armel |
| 03:21 | dung | how do you compile the kernel?+ |
| 03:21 | dung | gnueabi(hf)? |
| 03:22 | dung | i think it's bluez |
| 03:23 | Punkley_chillin | dung, the issue I had appeared to be caused by galcore, I changed galcore to be a module, and problem solved. Very strange. |
| 03:23 | dbsx | bluez is junk, but it does work ok on sheevaplug/armel |
| 03:24 | dung | How did you make galcore modular? |
| 03:25 | Punkley_chillin | changed the Makefiles |
| 03:32 | dung | armel is ready with Linux 2.6 and first GL release. next (not so far away) for armel is to run Linux 3 and 2nd release. i'm near to xorg-dovefb for armhf |
| 03:35 | dung | dbsx, do you run armel on 3.5? |
| 03:36 | dbsx | yes |
| 03:36 | dbsx | headless |
| 03:44 | dbsx | I repeated my obexpushd test with sheeva/armel linux 3.1.10 works with no errors. so either armhf or the 3.5 kernel |
| 03:45 | dung | i guess bluez armhf |
| 03:48 | dbsx | dung, I will test and see if I can prove you wrong. I bet on the kernel |
| 03:50 | dung | :) |
| 03:52 | dung | bluetooth does crypt (vfp?) |
| 03:54 | dbsx | good guess, give me 15 minutes and I will let you know |
| 04:31 | dung | diff doesn't work cause of white-space |
| 04:35 | dung | diff -w (--ignore-all-space) |
| 05:22 | dbsx | dung its the kernel, happens with armel wheezy on cubox |
| 05:50 | dung | has anybody looked into xorg-dovefb? |
| 05:50 | dung | I get: AddScreen/ScreenInit failed for driver |
| 05:50 | dung | can this be configured in xorg.conf? |
| 05:53 | dung | dbsx: aha |
| 05:55 | dung | that's bad |
| 05:56 | dbsx | not good |
| 05:56 | dbsx | hard to isolate cause |
| 05:58 | dbsx | note that obexfs works ok |
| 05:58 | dung | which version? |
| 05:58 | dbsx | of obexpushd? |
| 05:58 | dung | kernel |
| 05:59 | dbsx | rabeehs git with vfp patch 3.5.0-25145-gdb82aaf-dirty |
| 06:00 | dbsx | I guess I could try bikers kernel and see if it has the same problem |
| 06:22 | rabeeh | dbsx: kernel is still unstable for me too on armel/armhf distro. lots of seg faults on boot where the old kernel is ok. |
| 06:22 | rabeeh | i previously thought it was mis configuration of the processor (there was one feature difference between the two kernel configs); but after fixing that the seg faults continued to appear. |
| 06:23 | rabeeh | when porting the galcore driver over from 2.6.32.9 to 3.x.x series; there was only one difference in the driver pgoff mapping; maybe someone can review it? |
| 06:24 | rabeeh | dung? |
| 06:24 | dbsx | its my first segfault, so given that I am headless its means that most of the no gui stuff is ok |
| 06:28 | dbsx | if its any help the problem exists in bikers 3.5.2 binary download |
| 06:31 | dung | yes? |
| 06:36 | dbsx | last debug message from obexpushd is 0.1: OBEX_EV_REQDONE, OBEX_CMD_DISCONNECT, a quick look at the code makes bet on memory deallocation so I am wondering about kernel memory config. |
| 06:57 | dung | 3 times changed do_mmap_pgoff to do_mmap. shall i change that? |
| 06:59 | rabeeh | i'm not sure if that's the reason why it's segfaulting. |
| 06:59 | rabeeh | want to try? |
| 07:00 | dung | sed 's@do_mmap error@do_mmap_pgoff error@' -i gc_hal_kernel_os.c |
| 07:02 | dung | my X didn't segfault |
| 07:03 | rabeeh | ? |
| 07:03 | rabeeh | and you got dovefb driver to use galcore back again? |
| 07:03 | dung | i didn't see - only the ScreenInit failed |
| 07:05 | dung | galcore from kernel, patched all header |
| 07:06 | dung | no more errors |
| 07:07 | dung | do i just have to configure X? |
| 07:08 | dung | the same config than i had didn't work |
| 07:09 | rabeeh | the same config should have worked. |
| 07:10 | dung | Err: AddScreen/ScreenInit failed for driver |
| 07:10 | dung | no segfault, is this correct? |
| 07:10 | dung | i further experiment with armel |
| 07:11 | rabeeh | armel / armhf is the same in my case |
| 07:11 | rabeeh | hmm... or maybe i missed porting some cache operations? |
| 09:01 | Thirsty | rabeeh: I send you a pull request on github to send a v3.5.2 kernel merge with your v3.5.0 kernel. v3.5.2 also incluids the VFP patch. |
| 09:10 | tomlohave | @Thirsty : don't you have a problem with dove_crypto_init ? |
| 09:20 | Thirsty | tomlohave: no, you have to apply a patch to enable it. |
| 09:20 | Thirsty | but I didn't added yet to the kernel |
| 09:21 | tomlohave | ok |
| 09:21 | tomlohave | i have reverted this part on my local build |
| 09:22 | Thirsty | I am working on CESA driver of Phil Sutter, He did added TDMA support to the CESA driver. |
| 09:22 | tomlohave | so right now with the 3.5.2 branch , we can't build it |
| 09:22 | Thirsty | what error do you get? |
| 09:23 | Thirsty | I build the kernels on the Cubox itseld |
| 09:23 | tomlohave | first decalartion of CRYPTO_PHYS_BASE, DOVE_SRAM_PHYS_BASE, |
| 09:23 | tomlohave | - DOVE_SRAM_SIZE in common.c |
| 09:23 | Thirsty | I am running Debian hardfp |
| 09:23 | tomlohave | something like that |
| 09:23 | Thirsty | ah |
| 09:24 | Thirsty | that is from my github? |
| 09:24 | tomlohave | yes |
| 09:24 | tomlohave | clean build |
| 09:24 | Thirsty | oke, then I messed up, with the commits :( |
| 09:25 | tomlohave | with or without CONFIG_CRYPTO_DEV_MV_CESA |
| 09:25 | tomlohave | defined |
| 09:27 | Thirsty | this is how it had to be. |
| 10:01 | dung | looks like a hot crash :) |
| 10:02 | dung | So why does anybody wants this crypto complication? |
| 10:34 | Thirsty | to speedup SSH & VPN connections but also encrypted partitions |
| 10:38 | dung | ssh is speedy enough and i won't stress my machine with encrypted fs |
| 10:38 | tomlohave | @Thirsty : forget what i said, don't know what i have done with git, now i have the good sources, will build now |
| 10:40 | Coolgeek | dung: you're not the only one using it :) |
| 10:45 | dung | it :) |
| 10:46 | Coolgeek | ? |
| 10:49 | dung | i'm still guessing what's IT |
| 10:53 | Coolgeek | crypto :) |
| 10:53 | Coolgeek | you said you don't need it, but other might need it |
| 10:54 | Coolgeek | maybe I should said "maybe you don't need it but other might" |
| 13:23 | _loom_ | kwer |
| 15:25 | Thirsty | My git repository seems to be broken. If I look on github on the Branch "CuBox-v3.5.2", i can see that the last commit is the merge between linux v3.5.2 kernel en cubox items. |
| 15:26 | Thirsty | But when I clone the git to my Cubox I see that an extra patch is applied to the branch CuBox-v3.5.2 |
| 15:32 | Thirsty | I allready tried to fetch or push the items |
| 15:33 | Thirsty | But I am allready in sync |
| 15:33 | Thirsty | I think I have to start all over again to fix it... Any ideas? |
| 17:03 | tomlohave | @Thirsty : i use : |
| 17:03 | tomlohave | git checkout -b Cubox-v3.5.2 origin/Cubox-v3.5.2 |
| 17:04 | tomlohave | but i'm not a specialist of git |
| 17:49 | dash | hey guys. bit of a long shot here - I grabbed the wrong cable by accident and plugged a 12V power supply into my cubox. |
| 17:50 | dash | I can't get it to boot or respond to anything now, even after finding my actual 5V power supply. Did I murder it? or should I be checking for other stuff |
| 18:00 | Coolgeek | did it smell weird when plugged the 12v power supply ? |
| 18:05 | dash | i don't think so, but i hardly have any sense of smell :) |
| 18:05 | dash | it doesn't seem to now... |
| 18:51 | shesselba | dash: If you are lucky you just killed your dc-dc-converter(s) that generate the required voltages.. but cubox is very integrated and without good soldering skills I'd rather announce it dead already |
| 18:51 | shesselba | does /dev/ttyUSBx still come up when you plug it into usb? |
| 18:52 | dash | good question, haven't tried that |
| 18:52 | dash | the power brick has an LED on it |
| 18:52 | dash | it's solid on when not plugged into the cubox, but blinks after I plug it in |
| 18:52 | dash | that seems bad. |
| 18:53 | shesselba | I'd guess that a broken dc-dc will draw to much power out of the brick, it's complaining about it |
| 18:54 | shesselba | If /dev/ttyUSBx also gone you have at least 2 broken chips. |
| 18:54 | shesselba | btw ttyUSBx also works without power attached to the cubox |
| 18:54 | shesselba | BUT |
| 18:54 | shesselba | wait please |
| 18:56 | shesselba | how much knowledge do you have with insides of electronic stuff? |
| 18:57 | dash | very little :) i am a software guy |
| 18:57 | shesselba | well, honestly your chances of repairing the cubox are near zero. you really want to try? |
| 18:58 | dash | well, I don't really want to throw it away |
| 18:59 | dash | hm, /dev/ttyUSB0 shows up |
| 18:59 | dash | Bus 006 Device 003: ID 067b:2303 Prolific Technology, Inc. PL2303 Serial Port |
| 18:59 | shesselba | ok, then you didn't kill the pl2303 ;) |
| 19:00 | dash | i may have to make friends with some more soldering-iron-savvy people... :) |
| 19:01 | shesselba | dash: I you really want to go that road, inside the cubox there are some black little chips ~2x2mm. that could be your dc-dc converters. |
| 19:02 | shesselba | but finding the broken one, getting a new one, and resolder it .. leave it |
| 19:02 | shesselba | well for finding: it's the one that gets _very_ hot |
| 19:02 | dash | shesselba: well, i can throw it away now, or throw it away after trying to fix it. :) |
| 19:03 | dash | unless you know somebody that would like a free slightly-broken ARM computer... :) |
| 19:03 | shesselba | nope |
| 19:03 | dash | me neither |
| 20:07 | Thirsty | tomlohave: I am making a new one. This time it's better |
| 20:08 | tomlohave | well, your kernel work ;-) |
| 20:08 | tomlohave | works |
| 20:08 | Thirsty | I know |
| 20:08 | Thirsty | but git was a mess |
| 20:09 | tomlohave | sure |
| 20:09 | Thirsty | I can't commit anything any more. |
| 20:09 | tomlohave | ? |
| 20:10 | Thirsty | he was complaining about a lot of things |
| 20:10 | Thirsty | I could not figure it out was it was so I maked a new one |
| 20:10 | tomlohave | link |
| 20:11 | tomlohave | ? |
| 20:12 | Thirsty | I renamed the old one to linux-broken. |
| 20:12 | tomlohave | ok |
| 20:12 | Thirsty | I made a new one with the same name. But uploading doesn't go so well... |
| 20:13 | Thirsty | fatal: The remote end hung up unexpectedly |
| 20:14 | Thirsty | I am also working on CESA enabled version which using TDMA. |
| 20:15 | tomlohave | me only on the overlay problem on geexbox |
| 20:16 | Thirsty | Do bad that is not so easy to fix. |
| 20:16 | Thirsty | Do the other XBMC distos have the same problem? |
| 20:17 | tomlohave | well, that's not a problem with xbmc |
| 20:17 | tomlohave | don't know |
| 20:17 | tomlohave | xilka no |
| 20:18 | tomlohave | tested latest patches (rev9), same problem for me |
| 20:18 | Thirsty | I think that vmeta and opengl are writing to the same videolayer |
| 20:19 | Thirsty | That is what I guess whats wrong |
| 20:20 | tomlohave | maybe |
| 20:20 | tomlohave | or something is bad with kernel / libs |
| 20:21 | tomlohave | will test with gstreamer |
| 21:55 | Thirsty | tomlohave: finally, i am able to push to github. too bad my girlfriend internet connection is not so fast as my Full-duplex 100mbit fiber connection at home. |
| 22:38 | cmair__ | tomlohave: have you ever compared Xorg logs between your system and Xilka? Maybe this could give you a hint? |
| 22:42 | cmair__ | tomlohave: I'm running Xilka right now and fuser tells me that /dev/fb0 is used by xbmc and Xorg. |
| 22:43 | cmair__ | my xorg log also says that the GFX layer is 0. (http://pastie.org/4558282 line 118) |
| 22:45 | cmair__ | /dev/fb1 is only used by xorg until I start playing a movie. Then xbmc also opens this device. |
| 22:47 | cmair__ | which is the "Video layer 0" (xorg.log:235) |