| 01:35 | heap_ | hey, i setup cubox1 as nas, but performance laptopt->eth 1gbit switch -> cubox ->ext usb disk is around 15mb/s while cpu is 100% utilised all the time. |
| 02:04 | cbxbiker61_ | heap_ usb2 isn't the best choice for nas, i tested cubox-i with e-sata, much better performance/lower cpu uti |
| 08:41 | ttl | hi, did anyone get BT working under ubuntu ? i under patchram to upload the .hcd file |
| 08:41 | ttl | now i have hci0 showing up, but it doesnt really send anything |
| 08:41 | ttl | using hcidump i see packets sent out but no answer |
| 08:42 | ttl | in dmesg i see "Bluetooth: hci0 command 0x0401 tx timeout" |
| 10:22 | heap_ | cbxbiker61_: yeah i know, i thought that usb2 should be one of the reason |
| 10:23 | heap_ | cbxbiker61_: and did you make any comparison usb2/esata? |
| 13:05 | cbxbiker61_ | heap_, pretty sure i was doing 40MB/sec with e-sata |
| 13:24 | _rmk_ | we're debating the names of the various audio ports which are found on the solid run hardware for mainline kernels. Currently, I have "DW-HDMI", "Hummingboard Codec" and "Hummingboard SPDIF". Do you feel that they are user friendly names for each port (the HDMI, the codec via the 3.5mm socket on the hummingboard, and the SPDIF output). How about "Integrated HDMI", "Integrated SPDIF", "Integrated Codec" ? Or maybe "Built-in" or "On-board" ? |
| 13:26 | _rmk_ | (bear in mind that tools such as pauvcontrol display this name to the user, and adding a suffix such as "Analogue stereo", "Digital Stereo (HDMI)") |
| 13:30 | cbxbiker61_ | Hummingboard seems too long, and really i think you would go towards a more interchangeable name like the later options |
| 13:38 | kgp | it's just 3 letters less than raspberry ;) |
| 13:43 | Coolgeek | On-board is self explanatory I think |
| 13:54 | kgp | Hummingboard Pro Extended 2015 On-Board Arm10 |
| 13:54 | kgp | :) |
| 13:55 | _rmk_ | kgp: you missed "Integrated" in that :) |
| 13:56 | kgp | That's amazing ;) |
| 13:57 | kgp | You can even have special versions - Hummingboard Pro Integrated Extended 2015 On-Board Arm10 Extreme Edition Summer Edition (The CPU is 10% faster) |
| 13:57 | kgp | Hum forgot the red |
| 13:58 | kgp | Hummingboard Pro Integrated Extended 2015 On-Board Arm10 Extreme Summer Edition - With the red board. |
| 14:17 | heap_ | cbxbiker61_: and with usb? |
| 14:52 | AndrewBeam | hi |
| 14:53 | AndrewBeam | what country is box sent from to europe? |
| 14:58 | Artox | AndrewBeam: what exactly are you asking? |
| 14:58 | Artox | I only know that the cubox-i comes from israel |
| 15:02 | AndrewBeam | Artox: thats what i asking to.. ;) thx |
| 15:02 | AndrewBeam | so i have to pay taxes |
| 15:07 | Artox | yeah |
| 15:17 | jmontleon | _rmk_, are you able to mount swap partitions on mmcblk0 on the cubox-i with 3.16? |
| 15:18 | jmontleon | i am seeing some real weirdness; was talking to jnettlet about it; I can do a mkswap and blkid will not report a UUID and doing, say swapon /dev/mmcblk0p2, will just give me errors. Same host if i plug in a thumb drive or whatever I can create and mount swap on /dev/sda1 or whatever |
| 15:18 | jmontleon | very odd |
| 15:19 | jmontleon | that was with your hummingboard-cubox-i-v3.16-rc1-20140616.diff patch applied |
| 16:23 | _rmk_ | jmontleon: I've not tried it on MMC - I avoid putting swap on flash based devices, preferring to run without swap in such scenarios |
| 16:25 | jmontleon | _rmk_, ya - not that I see it get used that much or that I tend to push the system very hard anyway, but I've not been able to even get it to mount with 3.16-rc1 or rc2 whereas with 3.15.1/3.14.8 and lower it's been fine |
| 16:56 | jnettlet | jmontleon, I would recommend you enable zswap |
| 16:58 | jmontleon | I have CONFIG_ZSWAP=y |
| 16:59 | jnettlet | jmontleon, you need to add zswap.enable=1 to your kernel cmdline as well |
| 16:59 | jmontleon | ah |
| 16:59 | jnettlet | zswap.enabled=1 |
| 16:59 | jmontleon | did not know that |
| 17:00 | jnettlet | but it might be better to use zram. let me check. I remember one of them got yanked from the kernel recently |
| 17:01 | jnettlet | hmmm can't find that email now |
| 17:02 | jnettlet | so zswap is a compressed cache if you already have swap. If you want to run with no swap you want to enable and setup zram |
| 18:37 | heap_ | anyone using backuppc on cubox? im occuring strange behavior |
| 18:51 | heap_ | any http://sourceforge.net/p/backuppc/mailman/message/32499870/ ? |
| 19:05 | jmontleon | jnettlet, well - I'd want to enable swap, but again I'm always just greeted with 'swapon: /dev/mmcblk0p2: read swap header failed' |
| 19:05 | jmontleon | with 3.16 anyway |
| 19:07 | jmontleon | and only with mmcblk0; very odd |
| 19:50 | heap_ | what do you guys use to monitor cubox resources like ram/cpu/ |
| 19:51 | rabeeh | top/htop |
| 19:51 | rabeeh | if you are running an intensive I/O then a good way to average things is 'time dd if=X of=Y' |
| 19:51 | rabeeh | that will show how much it takes to perform I/O jobs and how much system and user space took to perform that |
| 19:53 | rabeeh | for instance on HB i'm running on mSata with ~140MB/sec |
| 19:53 | rabeeh | debian@arm:~/build-test/linux-linaro-stable-mx6-linux-linaro-lsk-mx6$ time sudo dd if=/dev/sda of=/dev/null bs=1M count=1000 |
| 19:53 | rabeeh | [sudo] password for debian: |
| 19:53 | rabeeh | 1000+0 records in |
| 19:53 | rabeeh | 1000+0 records out |
| 19:53 | rabeeh | 1048576000 bytes (1.0 GB) copied, 7.61565 s, 138 MB/s |
| 19:53 | rabeeh | real 0m9.932s |
| 19:53 | rabeeh | user 0m0.109s |
| 19:53 | rabeeh | sys 0m5.543s |
| 19:53 | heap_ | hb? |
| 20:22 | TrevorH | anyone know if or why the dts/dtb file for dove is broken in kernel 3.15.1? |
| 20:23 | TrevorH | went back to 3.14.8 and that works fine, the one from 3.15.1 will not "compile" (or whatever you call what dtc does) |
| 20:24 | cbxbiker61_ | TrevorH...you have to switch to multi arch v7 with 3.15+ |
| 20:25 | TrevorH | does that mean use a different dts file? |
| 20:26 | cbxbiker61_ | i don't remember if they changed the filename |
| 20:27 | cbxbiker61_ | TrevorH, i compiled 3.15 for dove, but it didn't boot, it may have been because i had multi-core selected so it could be used across a wide swath of devices |
| 20:28 | cbxbiker61_ | i may have to do a single core version for dove |
| 20:29 | TrevorH | I'm on the original cubox-pro so there's not much point in enabled multi-core ;) |
| 20:29 | cbxbiker61_ | right...but i'm trying to support as many devices as possible with a single kernel, that's the whole advantage of .dts |
| 20:29 | TrevorH | could you explain more about what you mean by switching to multi-arch v7? it's not something I've heard of before |
| 20:30 | cbxbiker61_ | grep the .config for V7 you'll see the option, it has to be turned on |
| 20:30 | cbxbiker61_ | then you can select dove as a sub-arch |
| 20:31 | TrevorH | does that no longer need a dts file? |
| 20:31 | cbxbiker61_ | yes |
| 20:31 | cbxbiker61_ | yes, it is still required |
| 20:32 | TrevorH | irght but I get compilation errors with it |
| 20:32 | cbxbiker61_ | mine compiled but didn't boot, your .config probably has some dependencies that are wrong |
| 20:33 | cbxbiker61_ | they don't always get the inter-dependencies right |
| 20:34 | TrevorH | mm, not sure mine hung at uncompressing vmlinuz and when I went back to look the dtb file was not created because dove-cubox.dts has syntax errors |
| 20:35 | cbxbiker61_ | oh, yeah...i saw that, google it, someone has probably already patched that...or just wait for a 3.15 that is fixed |
| 20:58 | jmontleon | hmm, similar comment about swap with 3.16 here: https://plus.google.com/104877287288155269055/posts/JTNvi2VH9qq |
| 21:35 | jenkins101 | cbxbiker61_: you build for a lot of devices, what do you do with u-boot? |
| 21:35 | jenkins101 | do you have any solution... |
| 21:36 | cbxbiker61_ | for the old devices...i.e. original cubox and sheevaplug i use .dtb compatibility mode where it appends the dts to the end of the kernel, so no updated u-boot is required |
| 21:36 | cbxbiker61_ | that's the easiest, instead of requiring users to flash a new u-boot |
| 21:37 | cbxbiker61_ | in general i don't mess with u-boot unless i need a new feature that isn't supported with the original |
| 21:38 | jenkins101 | cbxbiker61_: ok but for say the cubox-i, wandboard and utilite |
| 21:38 | jenkins101 | and other imx6 boards... |
| 21:38 | cbxbiker61_ | cubox-i is a piece of cake, since u-boot resides on sd...you can play as much as you want |
| 21:38 | jenkins101 | yes |
| 21:39 | jenkins101 | but they all use different u-boot versions and types.. |
| 21:39 | cbxbiker61_ | i don't have a wandboard or utilite, so i'm not sure of the best advice there |
| 21:39 | jenkins101 | hum... ok |
| 21:39 | jenkins101 | are you using kernel 3.16 ? |
| 21:39 | jenkins101 | on cubox-i |
| 21:40 | cbxbiker61_ | should be able to support almost anything with either the .dtb compatibility mode or the native mode |
| 21:40 | cbxbiker61_ | my install script appends the .dtb when necessary (unless i'm unfamiliar with the platform...and then it's up to the user) |
| 21:41 | cbxbiker61_ | jenkins101, i'm doing a stable 3.10.xx....since i want full video support, i haven't tested 3.15+ yet |
| 21:42 | jenkins101 | cbxbiker61_: what kind of releases are you building |
| 21:42 | jenkins101 | ? |
| 21:43 | cbxbiker61_ | kirkwood - latest kernels, dove up to 3.14.8, cubox-i currently 3.10.44 |
| 21:43 | cbxbiker61_ | alsa latest for rpi |
| 21:43 | cbxbiker61_ | also |
| 21:44 | jenkins101 | have you got suspend resume to work? |
| 21:45 | cbxbiker61_ | on cubox-i???? are you suggesting that the ir receiver can wake the device? |
| 21:45 | jenkins101 | no it cant.. |
| 21:45 | jenkins101 | bu a usb device can |
| 21:46 | cbxbiker61_ | ic, i haven't tested that |
| 21:46 | jenkins101 | however if i plugg one in it wakes up directly |
| 21:47 | jenkins101 | internal ir should be able to wakeit though... but something is broken there... |
| 21:48 | cbxbiker61_ | the problem with ir, is that you really need a receiver that has hardware support to wake the device, otherwise even if it can do it....it would do it for any ir signal |
| 21:49 | jenkins101 | yes |
| 21:49 | cbxbiker61_ | for example zotac built-in ir receiver can do that (remote has a special button) |
| 21:50 | jenkins101 | probably the problem with the internal ir |
| 21:51 | jenkins101 | most usb receivers can wake from suspend |
| 21:51 | jenkins101 | but not from power off |