| 00:11 | fataku | rabeeh: https://paste.null-life.com/#/vIUI2tkDj/sKvLuCvUpI6/alkfQwuSulgvuziYh5z9CnlbIzIk3lGlA0 |
| 00:13 | rabeeh | ata1: SATA link down (SStatus 0 SControl F300) |
| 00:13 | rabeeh | SStatus = 0 means that the sata controller is not seeing any signal |
| 00:13 | rabeeh | even if it was trying to connect you would see a non zero value |
| 00:19 | fataku | im unsing an orico 2598us3, using the usb3 interface works perfectly in the cubox, but i wanna use the esata interface, somehow it doesnt work, it doesnt even start the hdd neither the lights |
| 00:20 | rabeeh | how are you powering it? |
| 00:24 | cbxbiker61 | rabeeh, hi, my hb has the same sd rate as the cubox-i, is there a bit of code that has to be patched to trigger the higher speed SD mode? |
| 00:24 | rabeeh | yes |
| 00:24 | rabeeh | it's not out there yet |
| 00:24 | rabeeh | actually support ddr50 and sdr104 modes |
| 00:25 | rabeeh | if you 'dmesg | grep mmc1' you should be seeing SDXC |
| 00:25 | rabeeh | which card are you using? |
| 00:25 | rabeeh | cbxbiker61: i actually want to ask if you already have an img for xilka |
| 00:25 | rabeeh | for imx6? |
| 00:26 | rabeeh | i'd be happy to add it to Ignition - https://github.com/SolidRun/ignition-imx6 |
| 00:26 | cbxbiker61 | my sd card is a sandisk ultra plus, quite a speedy card |
| 00:26 | rabeeh | so those are ddr50 |
| 00:26 | rabeeh | the gray colored; right? |
| 00:27 | cbxbiker61 | gold/red |
| 00:27 | rabeeh | those are the sandisk extreme then |
| 00:27 | cbxbiker61 | extreme plus, i should say |
| 00:27 | rabeeh | now those are real speedy devices - sdr104 |
| 00:27 | rabeeh | so it's a 200mhz interface |
| 00:28 | cbxbiker61 | yes, they test out very high with a good usb3 adapter |
| 00:29 | rabeeh | so HB (even hb-i1) supports uhs-1 |
| 00:29 | rabeeh | they have the power cycle and the 3.3v to 1.8v switch over circuitry to accomodate those speeds |
| 00:29 | cbxbiker61 | is the patch good for device tree, so it will work properly for both devices with common code? |
| 00:29 | rabeeh | there are some patches to enable that |
| 00:29 | rabeeh | the device tree is already patched and ready to go |
| 00:30 | rabeeh | there are patches for the i.mx6 sd interface itself; i can't recall where jnettlet put those |
| 00:30 | rabeeh | i'm having a black out now; need to get some rest |
| 00:30 | rabeeh | will discuss this tomorrow with you |
| 00:30 | cbxbiker61 | ok |
| 00:30 | rabeeh | i recall measuring >= 50MB/sec on those |
| 00:31 | rabeeh | potentially the bust interface can reach ~90MB/sec |
| 00:31 | rabeeh | i can't recall if it was read or write. |
| 00:31 | rabeeh | anyhow; ttyl |
| 01:08 | Exaga | rabeeh: fixed it. i've also renamed the image - humbo-slackwarearm-install_BETA_02Nov14_fd.img.xz |
| 09:23 | topi` | how/where do I get the most recent imx-lib and libfslvpuwrap for debian/jessie and kernel 3.14.14+ ? |
| 09:23 | topi` | and is there something else that I need to get gstreamer-1.0 working? |
| 09:24 | topi` | I have imx-lib-3.10.17-1.0.0.tar.gz from freescale's Yocto packages, but this probably won't work with 3.14.x |
| 09:26 | cbxbiker61 | the existing 3.10 libs should work |
| 09:26 | cbxbiker61 | i haven't had any issues |
| 09:27 | topi` | ok |
| 09:27 | topi` | so I'll proceed to compile imx-lib for 3.10 and libfspvpuwrap which depends on the latter... |
| 09:28 | topi` | where did you get the existing 3.10 libs? My jessie image did not come with those |
| 09:29 | cbxbiker61 | sources.openbricks.org/3.1 |
| 09:30 | topi` | hm, interesting "repository" |
| 09:30 | topi` | lacks SHA1 hashes to ensure authenticity :) |
| 09:32 | topi` | what is libcec? |
| 09:32 | topi` | I saw it in some README, do I need it to get gstreamer accelerated? |
| 09:33 | R0nd | pretty sure you don't |
| 09:33 | R0nd | it's for remote control over hdmi |
| 09:35 | topi` | I think my Hummingboard Quad overheats (despite the red heatsink) - I was compiling the kernel with make -j6 and it reset itself in the middle of it |
| 09:36 | topi` | and the heatsink feels unbearably hot. Maybe I'll try forced air cooling to see if it helps. |
| 09:37 | R0nd | have you tried monitoring the temperature sensor? |
| 09:37 | R0nd | gotta be more reliable than sticking your finger into it |
| 09:38 | topi` | let's see if I find any temp |
| 09:39 | topi` | IIRC the temp sensor did not work in 3.14.14 |
| 09:39 | R0nd | works on my cubox |
| 09:39 | cbxbiker61 | it works with my 3.14.23 kernel |
| 09:40 | topi` | hm, it seems to work, I got values between 57-52 |
| 09:40 | topi` | it dropped when I blew a little on the heatsink |
| 09:40 | topi` | let's see what happens with make -j6 |
| 09:40 | cbxbiker61 | my box idling with kodi/xbmc is at 33.4C |
| 09:41 | topi` | cbxbiker61: where did you get 3.14.23 kernel? I cloned the repo in github/linux4kix/ |
| 09:41 | cbxbiker61 | xilka.com |
| 09:42 | cbxbiker61 | the directions are in the solid-run forum kernel threads |
| 09:48 | topi` | hmm, lots of fixes fit between .14 and .23 |
| 09:49 | topi` | is it possible to have the Hummingboard keep correct time? it's annoying that after reboot, date is back in 1970 and recompiling kernel complains about timestamps |
| 09:49 | topi` | I saw a mention of RTC but the solidrun specs don't mention it |
| 09:50 | cbxbiker61 | there's a small header for a battery |
| 09:50 | cbxbiker61 | probably uses a pretty standard pin configuration, although i haven't seen a spec |
| 09:50 | topi` | there are several headers underneath the mPCIe slot |
| 09:50 | topi` | 6 of them at least |
| 09:51 | cbxbiker61 | in the meantime, use ntp |
| 09:51 | topi` | then I need to get the damn thing on the 'net :/ |
| 09:51 | cbxbiker61 | it's in the corner between the usb plugs and the mono jack |
| 09:51 | topi` | possibly by ethernet, needs a bit of work |
| 09:51 | topi` | oh, yes, there! |
| 09:51 | topi` | looks like a solid header. |
| 09:53 | topi` | hmm, the temperature during compilation has reached 80C |
| 09:53 | topi` | is it hot enough to cause reboots? |
| 09:57 | cbxbiker61 | that sounds damn hot to me, but i'm not sure of the spec, 80 or 85C is a common limit in circuits |
| 09:58 | jnettlet | the thermal shutdown limit is 100C |
| 09:58 | topi` | right, so kernel compilations might need some lower MHz setting |
| 09:59 | jnettlet | topi`, I will be pushing thermal limiting to my repo soon |
| 09:59 | topi` | jnettlet: does that force a hardware reboot? |
| 09:59 | jnettlet | topi`, yes it is a forced shutdown |
| 09:59 | cbxbiker61 | compiling code really shouldn't be a problem in any case |
| 09:59 | topi` | ok |
| 09:59 | cbxbiker61 | i do it all of the time with lots of threads |
| 09:59 | topi` | I'm in a normal office environment and the thing just rebooted with make -j6 |
| 10:00 | jnettlet | how warm is your office? |
| 10:00 | topi` | previously I finished a kernel make successfully, but that was with the SD card |
| 10:00 | topi` | and the SD card causes a lot of IO latency |
| 10:00 | topi` | now I'm compiling on a mSATA |
| 10:01 | cbxbiker61 | it's probably as likely to be a kernel bug |
| 10:01 | topi` | no oops, nothing, just reboot |
| 10:01 | cbxbiker61 | connecting a serial console, so you can see the messages just before reboot would be good |
| 10:13 | topi` | hmm, it seems I have a kernel vs kernelheaders mismatch, the debian package linux-libc-dev is for version 3.11 and my kernel is 3.14.14 from jnettlet |
| 10:13 | topi` | how do I "update" the /usr/include/linux subdir without messing the debian system? |
| 10:14 | TrojanHorse | ping jmontleon |
| 10:15 | KsaRedFx | topi`: I got a question for you |
| 10:15 | KsaRedFx | How many cores are on that little HummingBoard you're dong -j6 on |
| 10:16 | topi` | wtf, compiling libfslvpuwrap fails at config phase, config.status executing libtool commands that fail |
| 10:16 | topi` | KsaRedFx: it's the Quad model |
| 10:16 | KsaRedFx | So why are you running -j6 |
| 10:16 | KsaRedFx | That means "Six Threads" there are 4 Cores on a Quad |
| 10:16 | KsaRedFx | You're overusing the system |
| 10:16 | topi` | it's just a convention to multiply your core count with 1.5 when make'ing |
| 10:16 | TrojanHorse | its more than that |
| 10:16 | KsaRedFx | Only when you have Hyper Threading |
| 10:16 | TrojanHorse | compile jobs dont use 100% cpu |
| 10:16 | TrojanHorse | and certainly not at all time |
| 10:16 | KsaRedFx | They *can* |
| 10:16 | topi` | there are I/O latencies etc |
| 10:17 | TrojanHorse | with j4 it will be idle every now and then |
| 10:17 | TrojanHorse | yes, I/O |
| 10:17 | KsaRedFx | -j4 would compile it fine and it probably wouldn't overheat |
| 10:17 | topi` | I bet the temperatures are quite similar with -j4 as well |
| 10:17 | TrojanHorse | indeed |
| 10:17 | KsaRedFx | Give it a try? |
| 10:17 | topi` | trying |
| 10:17 | TrojanHorse | its not like more jobs will overheat |
| 10:18 | KsaRedFx | If you have a heat problem with 6 threads, then you bring the load down |
| 10:18 | KsaRedFx | Less load is less heat |
| 10:18 | TrojanHorse | no |
| 10:18 | topi` | the whole point of 4 cores is to load them up to the max |
| 10:18 | TrojanHorse | it depends on teh frquency scheduler |
| 10:18 | topi` | to maximise speed :) |
| 10:18 | KsaRedFx | It depends on a lot of things. |
| 10:19 | cbxbiker61 | what you should probably do if you think cpu load is causing the reset is run a cpu bound tester, if it reboots then it validates your assumption |
| 10:19 | topi` | the amount of heat depends on the exact sequence of instructions |
| 10:19 | topi` | so a simple loop won't do |
| 10:19 | cbxbiker61 | there's a utility called stress that should do the trick |
| 10:20 | TrojanHorse | topi` so lets analyse teh whole isntruction set then :) |
| 10:20 | cbxbiker61 | math computations have a tendency to heat the cpu the most |
| 10:22 | topi` | I'd wager that to get the most heat, you'd need to stress the load-store units as well (but by hitting L1 cache, to get rid of mem latency) |
| 10:25 | TrojanHorse | I heard you just have to put the GC2000 to work |
| 10:25 | topi` | wow, libfslvpuwrapper compiled |
| 10:25 | TrojanHorse | and you get temperatures in no time |
| 10:25 | topi` | let's see... |
| 10:25 | topi` | is it a good idea to make install it to /usr/local ? |
| 10:25 | TrojanHorse | sure |
| 10:25 | KsaRedFx | What a surprise |
| 10:26 | topi` | grr, imx-lib doesn't compile: ERROR, macro names must be identifiers |
| 10:26 | topi` | oh, it's in hdmi-cec (do I need it??) |
| 10:29 | topi` | took away the whole dir, now it complains about missing linux/videodev.h ... there's only videodev2.h , what's the deal? |
| 10:29 | topi` | probably videodev.h is obsolete... |
| 10:30 | cbxbiker61 | there is a headers-install target in the linux kernel source |
| 10:39 | topi` | I don't have linux kernel source, just a git clone'd repo from jnettlet |
| 10:40 | TrojanHorse | how is that any different? |
| 10:42 | topi` | oh, you meant the *linux* makefile |
| 10:42 | topi` | has a target |
| 10:42 | topi` | sorry, was thinking of some debian packaging schemes... |
| 11:23 | topi` | grr, imx-lib installs stuff to /usr, not /usr/local ... |
| 11:28 | topi` | waf thinks I have pxp installed, although I specifically removed it from the imx-lib build... how to get rid of it? --without-pxp ? |
| 11:29 | cbxbiker61 | old /usr/lib/*.la files or /usr/lib/pkgconfig files |
| 11:30 | topi` | I have libfslvpuwrap.pc in /usr/lib/pkgconfig, maybe that's it |
| 11:32 | topi` | how do I direct waf to search for .pc and .la files in /usr/local ? |
| 11:35 | cbxbiker61 | not sure, but there is a package config env variable that affects package config |
| 11:36 | topi` | I got the stuff to compile when I commented out conf.recurse('pxp') |
| 11:41 | topi` | what's the difference between pxp and ipu? I hear ipu is only in IMX6Q |
| 11:42 | topi` | if I want to have my videos working also on the DualLite and Single models, i need to get this pxp compiling, right? |
| 12:15 | topi` | YES! video playback works :) at least on the imx6Q |
| 12:21 | cbxbiker61 | hmmm, i've got hb that reboots immediately after kodi/xbmc flashes the logo, u-boot indicates that it rebooted due to POR, so I guess i'll conclude that the power supply isn't handling the load when it starts graphics |
| 12:25 | jnettlet | cbxbiker61, which SOC? |
| 12:25 | cbxbiker61 | 2ex |
| 12:25 | jnettlet | extra peripherals? |
| 12:26 | cbxbiker61 | i have a usb keyboard plugged in, but it was doing it without the keyboard but with a usb wifi adapter also |
| 12:28 | jnettlet | that shouldn't be enough draw, but if you are getting dirty power it could be just enough of a peak load that the adapter can't keep up |
| 12:32 | cbxbiker61 | turns out that it was a 1A power supply that worked well with r-pi, i'll just have to get a 2A supply |
| 12:34 | jnettlet | oh 1A...haha yeah that is a bit low |
| 12:34 | topi` | it seems gstreamer works by decoding the video straight to the framebuffer pixels, so no overlays and stuff like on Raspberry |
| 12:35 | jnettlet | it depends on what sink you are using |
| 12:35 | topi` | I tried the ipusink |
| 12:35 | topi` | eglsink is no go since I haven't yet gotten X11 running |
| 12:35 | jnettlet | you don't need X11 for egl |
| 12:35 | topi` | oh... |
| 12:39 | cbxbiker61 | jnettlet, do you have the patches for the higher speed SD for hummingboard? |
| 12:41 | jnettlet | cbxbiker61, they are all patched. upstream broke something in the driver and I haven't had time to debug it yet |
| 12:41 | jnettlet | the mmc subsystem is a mess at this point. |
| 12:41 | cbxbiker61 | so these are kernel patches that should already be in the solid-run kernel tree? |
| 12:42 | jnettlet | all the patches that used to work have been integrated yes |
| 12:42 | cbxbiker61 | ok, it's really not a pressing issue for me |
| 12:46 | jnettlet | one sec, my usb system has just gone awol again |
| 12:56 | cbxbiker61 | topi, just in case you missed it...my reboot problem on the hummingboard is almost certainly because of a weak power supply, so your reboot problem may be the same |
| 12:59 | jnettlet | cbxbiker61, I have some notes here about things to look at. I will try to get to it this week |
| 12:59 | cbxbiker61 | sounds good |
| 13:28 | Artox | ping: jmontleon |
| 14:42 | jmontleon | Artox, pong - thanks, I saw your message. I will take a look, maybe this evening and see if I can use it to build |
| 21:23 | rabeeh | Ignition now supports wireless - http://www.solid-run.com/community/topic1833.html |
| 21:23 | rabeeh | http://download.solid-run.com/pub/solidrun/Ignition-imx/6-nov-2014/ignition.img |
| 21:25 | Exaga | cool rabeeh :> |