IRC log of #cubox of Tue 18 Mar 2014. All times are in CET < Back to index

23:59 satellit timm HDMI
23:59 satellit timm_
00:01 satellit_e timm__ I am using HDMI
00:33 timm__ satellit_e: jmontleo can probably tell us how to get rid of the console on hdmi if you want to be rid of the nuisance messages. You need to edit /boot/uEnv.txt
00:37 satellit_e ok do you know what to edit?
00:54 timm__ I think it is one of the console= phrases in mmcargs=setenv bootargs root=/dev/mmcblk0p3 rootfstype=ext4 rootwait console=ttymxc0,115200n8 console=tty1 video=mxcfb0:dev=hdmi,1920x1080M@60,if=RGB24,bpp=32 ahci_imx.hotplug=1 cma=64M
00:56 timm__ the baudrate on console=ttymxc0,115200n8 makes me think that is the serial port, which would make tty1 hdmi, but it is a funny device for hdmi
00:56 timm__ I would get the ability to ssh in before I edit it, so you can get back in there
01:00 satellit_e timm__ I installed vi and can see the file...not sure to edit
01:00 timm__ be aware that these are boot arguments, so messing it up might mean it won't boot.
01:01 satellit_e took picture and am trying to save and decipher the args. Do not see anyting relative to the script
01:02 timm__ do you have the ability to mount the sdcard on another linux machine or vm?
01:02 satellit_e my fedora laptop has broken SD slot will try on 6 in one reader and adapter....
01:05 timm__ I find the serial console works really well
01:05 satellit_e mssege it brcmf mac: bcrmf_cfg80221_sched_scan_start Invalid sched scan req!! n_ssids:0
01:05 satellit_e message is*
01:05 timm__ I use putty to access it. the hardest part is finding a micro usb that fits
01:06 timm__ yes, I see those messages constantly in the console, but not when I ssh in
01:07 timm__ just plug an ethernet wire into the connector and it is already set up for dhcp
01:08 timm__ use ifconfig in the console to get the ip address and away you go
01:08 satellit_e I believe I have ethernet wire and wireless here I can use internet for yum
01:10 satellit_e wlan0 is on per iwconfig
01:10 timm__ what does ifconfig say about eth0?
01:12 satellit_e IEEE 802.11bgn ESSIDoff/any Mode: managed Acesspoin: Not_Asscociated Tx-Power=1496 dbm Retry short limit 7 RTS thr:off Fragment thr;OFF
01:13 satellit_e Encription key:off Power Management:off
01:13 satellit_e this is wlan0
01:13 satellit_e eth0 no wireless extensions
01:13 timm__ nmcli dev wifi connect password will connect to wifi
01:14 satellit_e ok will try later as script interrups any long entrys
01:15 timm__ ok talk to you later. I got to leave.
01:15 satellit_e hop you next build works better...
01:15 satellit_e bye
03:25 satellit http://wiki.sugarlabs.org/go/Sugar_Creation_Kit/sck/Advanced_Topics#FC20forXSCE.zip screenshots of CuBox-i4
03:25 satelli 03:25 * satellit F20
06:08 _dab_ Can someone articulate the switch/router failure with waits on low end routers? I have netgear asking (they might fix it).
06:14 jnettlet _dab_, it has to do with hardware flow negotation that is specified as part of the gigabit spec.
06:19 _dab_ thanks jnettlet: I will try that description and see how I go.
06:27 _dab_ jnettlet: Can you name a switch/router brand/model that works?
06:28 jnettlet _dab_, I can't name one that doesn't work. But my Trendnet tew-639gr doesn't have a problem. It negotiates to off but doesn't exhibit the pause/hang that has been reported.
06:28 _dab_ tank you
06:30 _dab_ s/t/th/
06:33 _dab_ FYI: Netgear FVS318N fails
06:40 jnettlet _dab_, interesting. can you check what the interface reports with ethtool?
06:42 jnettlet ethtool -a eth0 will tell you what is going on. You can also disable it if you need to.
06:43 _dab_ 1 minute
07:10 mk01 guten morgen
07:15 jnettlet god morgen
07:38 _dab_ Autonegotiate: on
07:38 _dab_ RX: off
07:38 _dab_ TX: off
07:43 _dab_ jnettlet: ethtool will not let me change the rx parameteer
07:43 jnettlet _dab_, both rx and tx have to be the same, so you need to change both of them
09:34 rjc hi all
09:34 rjc I'm trying to get my original CuBox back into life
09:35 _dab_ jnettlet: did that, no change in iperf, but less overruns
09:35 rjc but when I connect the serial cable I can see:
09:35 rjc Hit any key to stop autoboot: 1
09:36 rjc first of all, it starts with '1', and second of all, it doesn't respond to keys being pressed
09:36 rjc has anyone came across that issue?
09:36 rjc how can I get the U-Boot prompt back?
09:36 mk01 _dab_ jnettlet what is the discussion? slow eth ?
09:43 _dab_ overruns on receive, iperf maxes out at 13mbit/sec receive and 44mbit/sec send. Presumption is that it is caused by the switch
09:44 _dab_ 130mbit/sec
09:44 _dab_ one colleague report a max of 30mbit/sec receive
09:45 _dab_ something wrong with zero key. 130/440
09:49 mk01 switch with bandwidth of 130mbit/s ???
09:51 _dab_ the switch is supposedly gigbait, i4pro does no better than 130/440. Other boxes with e100 ethernet get 850/900
09:51 rjc again, I will answer myself - disable Hardware Flow Control
09:51 rjc :P
09:51 mk01 ok, understand
09:53 _dab_ I would settle for 400/400
09:55 jnettlet _dab_, you won't get 400/400 the transmit speed is topping out in the 300's receive is up to about 530 or so
09:56 mk01 jnettlet: 530 ? isn't eth0 on usb?
09:56 jnettlet mk01, not at all. I don't know why people keep spreading that disinformation.
09:57 mk01 becuase it was on Solidrun shop page :))
09:57 mk01 currently USB is taken out and this is still there: (*) 1000Mbps link is limited to 470Mbps actual bandwidth due to internal chip busses limitation
09:57 cbxbiker61 jnettlet, i think they're just assuming it is so, since other arm device do it that way
09:58 jnettlet mk01, yes that is what Freescale rates it at.
09:58 jnettlet although their driver and the upstream one were horrible and rmk overhauled them and got a lot more performance out of them.
09:59 jnettlet cbxbiker61, yep...the downfall of the internet
10:01 mk01 guys, I'm 90% sure it was on Solidrun's product page. actually the same text only *busses was *USB2
10:01 cbxbiker61 those 64bit amd arm chips look technically interesting with 10G support
10:04 mk01 cbxbiker61: are there any real products on market with arm64 ? I remember there was producer of server-line machines but closed business recently - only info I have.
10:04 cbxbiker61 latest apple is 64bit, amd's should be 2nd half of this year
10:05 mk01 yes, forgot put exclude on that
10:06 mk01 they have own design on top ARM definitions. and of course supplying just own device group.
10:06 cbxbiker61 amd looks to want to do heavy network traffic with relatively light processors
10:10 mk01 cbxbiker61: to be honest and speaking only pure my opinion, ... I don't feel currently the limitations on speed / bus bandwidth - the other way around - don't have usage for more. only what bothers me is the memory addressing / vmalloc
10:13 cbxbiker61 mk01, i have a server process that needs to pass messages/data back and forth between kernel and 2 userspace programs, it's currently using low-level primatives, but it occassionally glitches
10:13 cbxbiker61 i'm just getting started on rewriting it to use kdbus
10:14 cbxbiker61 kdbus is just stabilizing now, but its potential for my type of scenario is huge
10:16 jnettlet cbxbiker61, yeah I was using the old implementation.
10:19 mk01 colleagues, jnettlet: should the spdif not be working on your stable ?
10:20 jnettlet mk01, I don't have spdif right now to test it, but I am pretty sure the geexbox guys have tested it and it is working
10:21 mk01 jnettlet: I don't have eq currently at hand too - so can't confirm. but after migration I got reports that SPDIF stopped working.
10:21 mk01 jnettlet: and as I read forums (any - xbmc/geek/solid). SPDIF is nada
10:21 mk01 jnettlet: don't want to believe that
10:23 jnettlet mk01, it is very possible something isn't working properly. Do you have posts?
10:24 jnettlet the thing is I don't have time to keep track of every users problem in the forums. I heavily rely on the distro maintainers to get back to me.
10:24 mk01 jnettlet: why I don't want to believe is because dmesg shows ok mappings, alsa correct devices, pcm settings, speaker test shows playing on the device. simply no indication to anything
10:25 jnettlet yep I agree, but it could be a clock problem
10:25 mk01 but all say that it's just silence
10:26 mk01 jnettlet: just wanted to point that I have nothing to provide to you. beside clean logs of anything and the "not working".
10:28 jnettlet mk01, actually a log could be useful. spdif is using sdma to transmit the audio. Perhaps it is a problem with the sdma firmware loading.
10:28 mk01 nope
10:28 mk01 double checked
10:28 mk01 [ 1.923243] imx-sdma 20ec000.sdma: loaded firmware 1.1
10:28 mk01 [ 1.925478] imx-sdma 20ec000.sdma: initialized
10:28 mk01 but ok, will send you log. XBMC ?
10:30 mk01 http://imx.solid-run.com/forums/viewtopic.php?f=10&t=847&p=6004&hilit=spdif#p6004 the "general" claims are not so important, but Martin's feedback. actually comments only SMB.
10:31 mk01 jnettlet: for other builds / distros all about the same.
10:32 mk01 jnettlet: but wanted to ask you because for 3.0.35 it was more or less mostly the same but I it worked for our system with no issues. ...
10:32 jnettlet mk01, well that list of things that don't work are mostly hardware limitations.
10:32 mk01 so claims are like public psychosis sometimes
10:33 jnettlet The micro-sd is limited to class 10, or 50mhz HS
10:33 jnettlet the SPDIF light is a gpio problem, which I know about.
10:33 mk01 jnettlet: the list as is is not the point. I told you. user's talking is like my Grandma
10:34 jnettlet No standby for cubox-i is correct. There is no standby, just low power mode, which does work.
10:34 mk01 Just posted this one as Martin is doing the Geekbox builds daily
10:34 jnettlet although I plan on making it better
10:34 jnettlet This one is interesting. http://imx.solid-run.com/forums/viewtopic.php?f=7&t=716
10:34 jnettlet pass-through is broken and normal mode only supports 2-channel
10:35 jnettlet I believe the xbmc guys were looking at the latter mode.
10:35 mk01 I'm aware of this
10:36 mk01 but speaker test sending noise is doing mono(I run -c2 stereo) mode.
10:36 jnettlet so spdif does work, but pass-through mode does not. That I can believe
10:37 mk01 ok, I will tripple ask the users the same and ask even for part of guisetting
10:43 mk01 jnettlet: isn't the 5.1 problem the issue with 24bits are not supported above 2channels ?
10:44 jnettlet I have not looked into it because I was told the xbmc guys are.
10:46 mk01 yes. it is in IEC958 docs
10:46 mk01 clearly stated
10:46 jnettlet mk01, the spdif code is really basic. which makes sense it isn't doing much.
10:47 jnettlet but the big thing I notice between the 3.0.35 kernel and the 3.10 kernel is the spdif codec has added the formats SNDRV_PCM_FMTBIT_S20_3LE and SNDRV_PCM_FMTBIT_S24_LE
10:47 satellit_ CuBox-i4P:I tested ubuntu-oneiric-freescale: It has very basic screen resolution and overruns my 20" HDMI monitor? Is there a fix coming for higher res?
10:47 jnettlet perhaps one of those is being used instead of SNDRV_PCM_FMTBIT_S16_LE
10:48 jnettlet satellit_, what is the native resolution of your monitor?
10:49 satellit_ not sure: Dell 20" 1680x1050 @60
10:50 jnettlet satellit_, hmmm odd. hold off until I land my graphics patches.
10:51 satellit_ ok other wise it works nicely
10:51 jnettlet you could also try and force 720p and see what taht does
10:51 satellit_ k
10:51 jnettlet satellit_, also a lot of the dell monitors have an option to switch between, computer and video modes. You want to choose video if your monitor has that setting
10:52 satellit 10:52 * satellit_ also debian has no networking...
10:52 satellit_ thanks
10:53 mk01 jnettlet: that's why I opened the 5.1 passthrough and 24 bits ... because this is even in IEC958 standard described as 20bit max (24 optional). as there is no bandwidth for it
10:54 mk01 that's why I patched the drivers in kernel to support only S16_LE
10:55 jnettlet mk01, adding #define DEBUG to sound/soc/fsl/fsl_spdif.c should give some useful debug info
10:55 jnettlet mk01, okay so your kernel already carries a patch for that.
10:55 mk01 yes
10:55 jnettlet Really though, you can just do that in an asoundrc.
10:55 jnettlet no reason to patch the kernel
10:56 mk01 jnettlet: what should I say. wanted to be sure with the users
10:56 mk01 :-.
10:56 mk01 :-/
10:56 jnettlet if you could enable debugging in that driver for your builds it would be useful for collecting information
10:58 mk01 will export with dyn debug right now. I see the code is updated to dev_dbg and _inf calls
11:00 mk01 jnettlet: thanks for directions.
11:00 jnettlet mk01, np ttyl
11:07 _dab_ jnettlet: I get 440 tx
11:09 linaro_e I get 640x480@60 for Ubuntu
11:15 mk01 satellit_: 1680x1050 is not CEA resolution and driver is cutting it out from list of res
11:16 linaro_e thanks
11:16 linaro_ 11:16 * linaro_e old Dell E207WFP : /
11:18 mk01 linaro_e: the "old" is not a problem :)) only driver is too picky.
11:18 linaro_e ^ nice to know anyway...
11:18 mk01 it is possible to strip that condition out, then (if EDID information is provided correctly by monitor) you would see wider list of resolutions
11:18 satellit http://wiki.sugarlabs.org/go/Sugar_Creation_Kit/sck/Advanced_Topics#CuBox-i4
11:19 linaro_e my testing ...
11:19 linaro_e want to use it for XFCE school server so testing
11:20 linaro_e xsce*
11:20 linaro_ 11:20 * linaro_e too early
11:24 mk01 linaro_e: for TESTING try this http://xbian.brantje.com/jnetker+/, don't tell jnettlet and use it for the resolutions only
11:25 linaro_e zimage?
11:25 mk01 yes
11:25 linaro_e thanks
11:25 mk01 (fw for vpu,sdma,wlan and BT) is not included
11:27 mk01 just out of interest report back the resolutions.
11:27 linaro_e k
11:29 linaro_ 11:29 * linaro_e xsce is a server so this is tangential issue but thanks
11:30 linaro_ 11:30 * linaro_e fedora 20
11:36 satellit http://web.dodds.net/~vorlon/wiki/blog/My_CuBox-i_has_arrived/ HDMI looks ok for debian but no networking....
11:36 satelli 11:36 * satellit here
11:54 jnettlet mk01, the driver isn't too picky. But there is reasons that it is limited to CEA resolutions. The first and foremost being audio support. To support those other resolutions you need to switch to DVI mode and disable audio which isn't in the driver yet.
11:55 mk01 jnettlet: "picky" on pure CEA
11:55 jnettlet there should be nothing keeping networking on debian from working. It works everywhere else
11:55 mk01 we had this discussions
11:55 jnettlet mk01, yes, but like I said audio is only supported on CEA resolutions
12:13 mk01 jnettlet: :) yes, but only because is disabled if CEA mode not in effect / found
12:17 mk01 (when the timings are properly updated too of course)
12:22 jnettlet it may work but there is no guarantee you can make the audio work at the proper frequency for non CEA modes. There are reports already that people have enabled it for the MX6 and audio does not play properly for non-CEA modes. Which is what I would expect
12:23 mk01 yes, it is off correct clock
13:33 MarcusVinter join ##security
13:33 MarcusVinter lol
15:00 rjc hi all
15:00 rjc is the U-Boot set up with serial console on new CuBox-i4?
15:01 rjc I connected the serial console to it, then booted without any micro SD card but didn't get any output
15:02 rjc do I need to boot to a working image first in order to activate serial console?
15:02 timm_ rjc: I think u-boot is on the sd card, so no card, no boot
15:02 rjc oh
15:02 timm_ there is a u-boot and spi only image
15:02 timm_ for me the serial console works
15:03 rjc so U-Boot is not in flash?
15:03 rjc like it was in the original CuBox?
15:03 timm_ rjc: no
15:03 rjc right
15:03 rjc good to know
15:03 rjc :)
15:04 timm_ I found SPL-U-Boot.img.xz, but don't remember where
15:05 timm_ there's a very lite busybox as well
15:06 timm_ http://dl.dropboxusercontent.com/u/72661517/tbr/spl/u-boot.img
15:06 rjc ta
15:08 rjc I preferred to have U-Boot in the SPI - that way I didn't need the SD card to boot of TFTP or USB
15:09 rjc oh well
15:09 rjc you win some, you lose some
15:09 rjc :)
15:12 satellit_e timm__ tested 2nd .img http://wiki.sugarlabs.org/go/Sugar_Creation_Kit/sck/Advanced_Topics#CuBox-i4
15:12 jnettlet rjc, there is no onboard storage for u-boot. This basically makes it impossible to brick the device.
15:14 satellit_e how do I install to thinternal storage? I think my CuBox-i4P-300-D MAC has it
15:15 satellit_e android microSD had the option (came with it)
17:13 MarcusVinter is bt working on geexbox?
18:25 MarcusVinter i stole the geexbox uImage, i have hte patchram by broadcom, and the driver from the internet. Yet i still get "cant get hci protocol" when i run the patchram on ubuntu. Any ideas anyone? Getting pretty desperate.
18:28 rabeeh MarcusVinter: i think i saw posts that Xbian has BT working (Debian based)
18:29 MarcusVinter :O as default?
18:30 MarcusVinter or kernel-wise?
18:30 MarcusVinter Either way, ill give it a shot right now. Thanks a lot!
18:31 rabeeh notice that you need to install Xbian, then update it few times until it gets the latest kernel
18:32 MarcusVinter Thank you, I'll do that right now. :)