00:16 | _dab_ | cbxbiker61: what version of cubox is your cuboxi5? I get exactly the same results as m_anish with your 3.14.29 |
00:17 | _dab_ | i4pro-300-D |
00:18 | _dab_ | cbxbiker61: which brcm firmware did you use? |
00:18 | cbxbiker61 | it's the high end i4 pro |
00:18 | _dab_ | meaning where/how did you get it? |
00:19 | _dab_ | recent i4pro? |
00:19 | cbxbiker61 | it was one of the early ones |
00:20 | cbxbiker61 | my firmware comes from debian firmware nonfree 0.43 |
00:20 | _dab_ | mine is early too, it must be firmware |
00:22 | _dab_ | hmm, mine is the latest brcm firmware from https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git |
00:22 | _dab_ | i will check on the debian firmware |
00:29 | _dab_ | now using same firmware and http://pastebin.com/sGwj3w7L |
00:32 | cbxbiker61 | grep the kernel source for "HT Avail", i don't see that here |
00:32 | cbxbiker61 | i'm kinda in the middle of a soldering job |
00:38 | _dab_ | hd /lib/modules/3.14.29/kernel/drivers/net/wireless/brcm80211/brcmfmac/brcmfmac.ko|grep Avail |
00:38 | _dab_ | ok will leave you alone |
00:51 | staylor_ | Anyone know if the Vivante X11 drivers support EGLImage correctly? |
01:40 | kgp | In https://github.com/SolidRun/ignition-imx6/blob/master/Debian-Wheezy/install.sh |
01:40 | kgp | curl -L -k http://mirror.igorpecovnik.com/ is dead |
01:55 | mk01 | _dab_: still here ? |
01:56 | _dab_ | yep |
01:57 | _dab_ | mk01: hello |
01:58 | mk01 | constant problems wifi not starting up ? |
01:58 | _dab_ | yes |
01:58 | mk01 | try to compile the two FW files to kernel |
01:58 | mk01 | we had the same |
01:59 | mk01 | CONFIG_EXTRA_FIRMWARE="bcm4329-fullmac-4.bin brcmfmac4329-sdio.bin" |
01:59 | mk01 | and put them into ./firmware |
02:00 | mk01 | make zImage and retest |
02:01 | mk01 | _dab_: it was you running kernel 3.18+ or I remember wrong ? |
02:03 | _dab_ | I have tried a few different kernels. 3.18 I will use for this. where do I get bcm4329-fullmac-4.bin and friend? |
02:04 | _dab_ | ignore the last question |
02:06 | _dab_ | mk01: do you have any problems with slow ssh on ethernet on 3.18? |
02:06 | mk01 | 3.19 |
02:06 | mk01 | YES |
02:08 | mk01 | although this is not what stress me that much, ... 3.19 hangs on reboot. :-/ that I wanted to ask. |
02:10 | mk01 | so I fix something in kernel, recompile, copy new kernel to tftp, restart for test - ... |
02:10 | _dab_ | the slow ssh (clunky) started to grate on the nerves after a while 3.18.2, I have not tried 3.19 |
02:10 | mk01 | _dab_ so no solution for now, yes ? |
02:11 | mk01 | but tested throughput via iperf and all fine |
02:11 | _dab_ | no, will let you know how the firmware blobs go |
02:11 | _dab_ | yep iperf is ok |
02:12 | mk01 | too big buffers ? |
02:13 | mk01 | stack waiting for buffer to fill in ... what could add this "lagging" feeling ? |
02:13 | _dab_ | It is like doing ssh on a wifi connection that powersaves after 1 second. I was thinking powersave |
02:13 | mk01 | _dab_: never heard about the reset problem ? |
02:14 | _dab_ | tell me more |
02:16 | mk01 | I just yesterday rebooted into 3.19 and wasn't at home to see it in detail so i'm 100% not sure what is it (at what stage) |
02:16 | mk01 | but it looks like the problem with u-boot was already and pci |
02:17 | mk01 | (so after soft reset uboot going through pci froze because some clocks (or something) was already hot-running) |
02:17 | mk01 | until now I managed few reboots via calling people to hit reset |
02:18 | _dab_ | I have had troubles with soft resets, sometimes. hard to reproduce |
02:18 | mk01 | this is each time |
02:18 | mk01 | reboot -> dead. |
02:19 | mk01 | and have the feeling it is not on going down as last reboot tried via "sync && echo b > /proc/sysrq_trigger |
02:19 | mk01 | and the same |
02:19 | mk01 | _dab_: let me know about the wifi |
02:19 | _dab_ | will do compiling now |
02:22 | mk01 | _dab_: I was very angry on the wifi, specially HB2x was doing problems (maybe CB4p too, but that one is rebooted only on occasions so don't know). google was hitting still the same discussions - something like wifi on early start has no FW yet (ok, no root mounted), then it hits some kind of state (for other reasons) so never returns to FW loading ... so compiled it in and got those issiues loose |
02:53 | _dab_ | mk01: blobs did not help |
02:53 | mk01 | ok |
02:53 | mk01 | and error the same ? |
02:53 | _dab_ | all the same, do I need to delete /lib/firmware/brcm dir? |
02:54 | mk01 | _dab_: no no ... |
02:55 | _dab_ | mk01: thanks |
02:55 | mk01 | ehm. for waster 15 minutes? :) |
02:57 | mk01 | _dab_: and you have the same brcmfull revision (as myself) - so those FW were correct, yes ? |
02:57 | _dab_ | yes |
02:58 | mk01 | ok |
02:58 | _dab_ | how old is your cubox (s)? |
02:58 | _dab_ | and model |
03:05 | mk01 | got it maybe exactly 1h ago |
03:06 | mk01 | how to check the details (from system?) -> don't have the devices at hand |
03:11 | mk01 | _dab_: on 3.19 wlan0 not working completely. maybe just missing in dtb? |
03:14 | _dab_ | wow |
03:14 | m_anish | _dab_, welcome to the (wifi-issues) gang :-) |
03:15 | m_anish | _dab_, fwiw, using jmontleon's f21 image, wifi worked ... but without nat, so somewhat useless for my purpose. |
03:16 | _dab_ | m_anish: I tried about a year ago and gave up. trying again. What I did find then was that disabling everything to with bluetooth was required for wifi. I could not get the two to work together. 3.10 |
03:17 | m_anish | _dab_, i don't really care about bluetooth, just need the wifi working :-) but don't want to spend too many hours on it .. atleast now .. will just get a usb-wifi dongle |
03:18 | m_anish | _dab_, one thing which does work is hostapd (AP mode on the wifi). but without natting you cannot route any eth traffic. |
03:18 | cbxbiker61 | the dongles will perform much better from a performance standpoint, if that is important |
03:19 | cbxbiker61 | for the most part you only want to use the internal wifi if you don't need much bandwidth |
03:19 | m_anish | cbxbiker61, hi! .. I just need something demoable at this point. in a real use case, we'll probably end up using routers. connected through eth (or usb-> eth) |
03:19 | mk01 | m_anish: I remember a commit removing CONFIG_JUMP commenting this as reason for bridging not working |
03:19 | _dab_ | yes, I have rt dongles working fine |
03:21 | m_anish | mk01, 01, _dab_ +1. I also have an early prototype hummingboard (aside from the i4pro), and it def works there. |
03:21 | m_anish | mk01, *ok |
03:22 | m_anish | cbxbiker61, fwiw, I was surprised to learn that the radio is a dual band and supports 5GHz channels. quite impressive :) |
03:23 | cbxbiker61 | the embedded wifi? |
03:23 | mk01 | m_anish: ??? |
03:23 | m_anish | cbxbiker61, yeah, (or am I mistaken?) I remember doing iw list and seeing those 5GHz freqs |
03:24 | mk01 | OH right |
03:25 | mk01 | ehm no |
03:25 | mk01 | Frequencies: |
03:25 | mk01 | * 2412 MHz [1] (20.0 dBm) |
03:25 | mk01 | * 2417 MHz [2] (20.0 dBm) |
03:25 | mk01 | * 2422 MHz [3] (20.0 dBm) |
03:25 | mk01 | * 2427 MHz [4] (20.0 dBm) |
03:25 | mk01 | * 2432 MHz [5] (20.0 dBm) |
03:25 | mk01 | * 2437 MHz [6] (20.0 dBm) |
03:25 | mk01 | * 2442 MHz [7] (20.0 dBm) |
03:25 | mk01 | * 2447 MHz [8] (20.0 dBm) |
03:25 | mk01 | * 2452 MHz [9] (20.0 dBm) |
03:25 | mk01 | * 2457 MHz [10] (20.0 dBm) |
03:25 | mk01 | * 2462 MHz [11] (20.0 dBm) |
03:25 | mk01 | * 2467 MHz [12] (20.0 dBm) |
03:25 | cbxbiker61 | that would just be the regulatory domain that you saw |
03:25 | mk01 | * 2472 MHz [13] (20.0 dBm) |
03:25 | mk01 | * 2484 MHz [14] (20.0 dBm) |
03:25 | mk01 | nothing more |
03:25 | m_anish | oh |
03:25 | cbxbiker61 | it sets limits, but not capability |
03:26 | mk01 | it has N though, HT20/40. up to 75mbps |
03:39 | mk01 | _dab_: 3.19rc5 is rebooting again |
03:40 | mk01 | _deb_ : wasn't it you trying to make imxthermal start ? |
03:40 | mk01 | 3.19 also not working |
03:41 | _dab_ | imxthermal did not work for on 3.18.2 |
12:21 | Humpelstilzchen | joelbaby: for now install libegl1-mesa and manually rm the libegl.so |
12:22 | joelbaby | lol hello |
12:22 | Humpelstilzchen | joelbaby: iirc gst-plugins-bad did not depend on libegl1-mesa when it was written |
12:24 | Humpelstilzchen | joelbaby: when this works you probably want to use equivs to satisfy the dependency |
12:24 | joelbaby | which is the libegl.so ? find -name libegl.so returns nothing. |
12:24 | Humpelstilzchen | try -iname |
12:25 | Humpelstilzchen | debian usually puts them in /usr/lib/arm-linux-gnueabihf/ |
12:25 | joelbaby | that's interesting these are also in the /hfp that you ask us to copy. |
12:26 | Humpelstilzchen | joelbaby: the hfp is the one that you should use, its patched from freescale |
12:26 | Humpelstilzchen | joelbaby: the debian one in /usr/lib/arm-linux-gnueabihf/ is not patched and will not work |
12:26 | Humpelstilzchen | unfortunately the one in /usr/lib/arm-linux-gnueabihf/ is preferred |
12:26 | joelbaby | so maybe i should just recopy the hfp one over? |
12:27 | Humpelstilzchen | joelbaby: it will get overwritten on the next apt-get upgrade |
12:27 | Humpelstilzchen | afk eating |
12:31 | joelbaby | i moved /usr/lib/libEGL.so to libEGL.so.old and /usr/local/lib/libEGL.so to so.old, and now gst-inspect-1.0 libgstimxeglvivsink.so gives the fb.createwindow error again. |
12:32 | joelbaby | looking at what dv_ said in an earlier post "The "undefined symbol: fbCreateWindow" error indicates you were building gstreamer-imx with the --egl-platform=fb , but not using the framebuffer libraries." |
12:32 | joelbaby | so maybe your instructions are building gstreamer-imx at the wrong step. maybe build after some libraries are copied over ? |
12:35 | dv_ | unfortunately the EGL/GLES situation is messy |
12:35 | dv_ | often, mesa also gets mixed in |
12:35 | dv_ | even though it shouldnt |
12:41 | joelbaby | so why does uninstalling egl1-mesa also uninstall gstreamer-1.0 good, bad ? |
12:43 | joelbaby | apt-get remove libegl1-mesa also removes: gstreamer1.0-plugins-bad libegl1-mesa libgstreamer-plugins-bad1.0-0 |
12:43 | joelbaby | and so therefore gstreamer1.0 won't play |
13:01 | Humpelstilzchen | joelbaby: can't follow you there... you should have installed libegl1-mesa and rm /usr/lib/arm-linux-gnueabihf/libEGL.so.* |
13:03 | joelbaby | there is no such file in arm-linux-gnueabihf |
13:03 | Humpelstilzchen | joelbaby: is libegl1-mesa installed? |
13:04 | joelbaby | there is libGL.so and libGLU.so |
13:04 | joelbaby | yes. libegl1-mesa is isntalled. |
13:04 | Humpelstilzchen | ok so you already did rm this file? |
13:04 | joelbaby | it never appeared in arm-linux-gneuabihf |
13:05 | Humpelstilzchen | uhm that would be strange |
13:05 | joelbaby | this command installed libegl1-mesa: apt-get install gstreamer1.0-plugins-base gstreamer1.0-plugins-good gstreamer1.0-plugins-bad gstreamer1.0-plugins-ugly gstreamer1.0-tools libgstreamer-plugins-base1.0-dev gstreamer1.0-alsa |
13:06 | joelbaby | on debian jessie |
13:06 | joelbaby | now when I run gstreamer it complains about no x display |
13:06 | Humpelstilzchen | full pastebin of gst-launch -v |
13:07 | Humpelstilzchen | its probably libGLESv2.so again |
13:08 | joelbaby | ERROR: pipeline could not be constructed: empty pipeline not allowed. |
13:08 | joelbaby | gst-launch1.0 -v |
13:08 | Humpelstilzchen | ... the full pipeline |
13:09 | Humpelstilzchen | gst-launch -v playbin uri=... |
13:10 | joelbaby | http://pastebin.com/we1yxmdj |
13:13 | Humpelstilzchen | its using glimagesink...what does gst-inspec-1.0 /usr/lib/arm-linux-gnueabihf/gstreamer-1.0/libgstimxeglvivsink.so say? |
13:14 | joelbaby | (gst-inspect-1.0:7557): GStreamer-WARNING **: Failed to load plugin '/usr/lib/arm-linux-gnueabihf/gstreamer-1.0/libgstimxeglvivsink.so': /usr/lib/arm-linux-gnueabihf/gstreamer-1.0/libgstimxeglvivsink.so: undefined symbol: fbCreateWindow |
13:16 | Humpelstilzchen | joelbaby: ok pastebin ldconfig -v please |
13:18 | joelbaby | http://pastebin.com/FA63CDTb |
13:20 | Humpelstilzchen | joelbaby: it says you have a libEGL under /usr/lib/arm-linux-gnueabihf |
13:23 | joelbaby | hm. ok, i didn't see that before. so i delete it now. |
13:24 | Humpelstilzchen | run ldconfig again after that |
13:27 | Humpelstilzchen | after that try the gst-launch command again |
13:29 | joelbaby | gst-inspect now works |
13:29 | joelbaby | but gstlaunch :ERROR: from element /GstPlayBin:playbin0/GstPlaySink:playsink/GstBin:vbin/GstGLImageSink:glimagesink0: Failed to connect to X display server |
13:31 | joelbaby | new pastebin for gstlaunch:http://pastebin.com/szZ8rc9V |
13:32 | Humpelstilzchen | pastebin gst-inspect-1.0 | grep imx please |
13:33 | joelbaby | new ldconfig -v http://pastebin.com/38L3ndCH |
13:34 | joelbaby | gst-inspect | grep imx: http://pastebin.com/sXyScguj |
13:35 | Humpelstilzchen | you sure that |
13:35 | Humpelstilzchen | gst-inspect-1.0 /usr/lib/arm-linux-gnueabihf/gstreamer-1.0/libgstimxeglvivsink.so |
13:35 | Humpelstilzchen | does work now? |
13:36 | joelbaby | imxeglvivsink: http://pastebin.com/rnkCQKMC |
13:37 | Humpelstilzchen | why is it not in the first pastebin? |
13:38 | Humpelstilzchen | hmm try rm -rf ~/.cache/gstreamer-1.0 |
13:38 | Humpelstilzchen | and gst-inpect-1.0 again |
13:40 | joelbaby | yup, that works same output. |
13:43 | joelbaby | but a different gst-launch error now: http://pastebin.com/QHqCLQVK |
13:43 | joelbaby | exec gdb failed: No such file or directory |
15:05 | Humpelstilzchen | hmm sorry no idea there |
20:18 | Humpelstilzchen | metrem: seems you have two libGALs |
22:50 | _dab_ | Good morning all. Anyone having any luck with wifi on the i4pro? |
22:51 | wbx | _dab_: as client or access point? |
22:56 | forest | _dab_, I haven't done much with it, so I can't vouch for connection stability or speed, but my i4pro was able to connect to my wireless-g network when running OpenELEC 5. |
23:08 | _dab_ | wbx: client. I dont see a device with 3.14.14SR |
23:23 | wbx | _dab_: i actually have no problem in client mode. it is my router for my lab. |
23:23 | _dab_ | thanks wbx |
23:23 | wbx | _dab_: using 3.14.28 with solidrun patches on cubox-i2ultra |
23:24 | wbx | _dab_: may be you have not installed firmware files? show us your dmesg |
23:29 | _dab_ | I have installed firmware and even tried embedding it in the kernel as mk01 advised. |
23:30 | _dab_ | brcmfmac: brcmf_sdio_drivestrengthinit: No SDIO Drive strength init done for chip 4329 rev 3 pmurev 6 |
23:30 | _dab_ | and brcmfmac: brcmf_sdio_htclk: HT Avail timeout (1000000): clkctl 0x50 |
23:30 | _dab_ | [ 6.977296] brcmfmac: brcmf_sdio_htclk: HT Avail timeout (1000000): clkctl 0x50 |