<Esmil>
This patch did the trick to turn on the RTL8188CU
<Esmil>
It's probably not quite right to call it a regulator for vbus, but it works :)
<Esmil>
Oh, and thanks for the link to the uboot sources. It works much more reliable than the sources from https://github.com/imx6-dongle I was using earlier
codinho has joined #imx6-dev
codinho has quit [Changing host]
codinho has joined #imx6-dev
<hste>
Esmil: Nice. So now u have the kernel up running with usb and wifi?
<Esmil>
hste: Yes, both USB ports are working. It turned out that because I was powering the dongle from my computer it would fight over who got to control the device I was plugging into the dongles big host port
<Esmil>
Apparently the micro-USB and the big USB host port are on the same USB hub
<Esmil>
..and then reason I didn't see the wifi on the other usb port was that it wasn't turned on
<Esmil>
Until that patch
cnxsoft has quit [Quit: cnxsoft]
<hste>
Maybe I should take the dust out of mine to test :)
codinho has quit [Remote host closed the connection]
<Esmil>
This is the imx6q-gk802.dts I'm using with Freescale's 3.10 GA kernel
<Esmil>
It's a mix of upstream, the file you linked me and the patch
<hste>
What sort of rootfs are u running?
<Esmil>
I wrote a script that prepares an SD card with u-boot, partitions and such and then does yum --installroot=/mnt ... to install a minimal fedora20
<Esmil>
Oh yeah, and it's run from a wandboard
<hste>
Nice. If u have some links or scripts I could put it on the imx6-dongle wiki :)
unsolo has quit [Remote host closed the connection]
jk_ has joined #imx6-dev
<jk_>
Hey dv_, if you are there could you tell me if gstreamer-imx uses anything from the gst-fsl-plugins package in Yocto?
datagutt has joined #imx6-dev
jas-hacks has joined #imx6-dev
<hste>
jas-hacks: hi :) Done some hacking lately?
<jas-hacks>
hste: hi, not on imx6
<jas-hacks>
hste: but did some testing on SPARC T5 boxes
<hste>
jas-hacks: havn't used sparc servers for may years :)
<dv_>
jk_: it doesnt
<jas-hacks>
hste: yes, don't think they sell as many
<obinou>
damn
<obinou>
is someone remember which is the 802.11 chip inside the cubox-i & the gk802 ? I'm making a 3.14 kernel now
<Esmil>
obinou: On the gk802 you need the RTL8192CU USB Wifi driver
<hste>
I think the cubox-i uses BCM4329
<jk_>
dv: Thanks. What about package ordering? Does it make a difference when gstreamer-imx is built (after gstreamer and gstreamer-base of course).
<obinou>
thanks, Esmil & hste
<jk_>
dv_: I ask because I am seeing weirdness with the imxeglvivsink plugin with the FrameBuffer backend. The decoded image is all in grayscale and distorted so that part of the right side of the image is getting shown on the left. The imxipusink element still looks ok. Trying to chase down what I did to cause that.
<dv_>
it doesnt happen with the x11 and wayland backends?
<dv_>
also, what are you playing? h264? mpeg4?
<jk_>
dv_: I have not tried in x11. And I did not think wayland was supported yet, so I have only tested ipusink on wayland. The distortions are occuring for all media, mpeg2, mpeg4 and h264. I have also sent streaming video (h264) over udp and rtp and seen the same distortions.
<jk_>
dv_: I don't think the issue is so much with gstreamer-imx as it is with how I have been configuring my project
<jk_>
dv: Oh, I have one other question about some weirdness. In my gstreamer application, I am looping video from a local file using playbin. When the EOS signal is reached on the bus, I set the pipeline to NULL then unreference it. However, this does not close down gstreamer correctly and I am left with exactly 4 open socket connections each time I loop. I have run the same code on x86 and did not have any trouble, so it has to
<jk_>
be something on the imx6 side. Have you had any problems like this before?
<dv_>
hmmm no.
<dv_>
what socker connections?
<dv_>
*sockect
<dv_>
damnit
diego_r has quit [Ping timeout: 264 seconds]
<jk_>
dv_: Thats a good question isn't it. I don't know a good way of debugging the issue, but when I use either imxipusink or imxeglsink, a handful (6) unix socket connections are opened. I'm not sure what they are opened too. If I replace those sinks with fakesink, the issue goes away. I have been debugging to see if it is a.) me, b.) gstreamer-imx, c.) some underlying vpu or kernel issue
<dv_>
I did not add any socket code anywhere
<jk_>
so that eliminates gstreamer-imx from the list :)
<dv_>
I just open the mxc_ipu device in the ipu elements
<dv_>
and mxc_vpu is opened by the imx-vpu library
<dv_>
but thats regular I/O with ioctls
<jk_>
have you noticed mxc_vpu doing anything like what I had described?
bard3307 has quit [Quit: Page closed]
tumdedum has quit [Ping timeout: 264 seconds]
tumdedum has joined #imx6-dev
<dv_>
no
<dv_>
well I never looked. I never had any of these issues.
<jk_>
thanks. Maybe it is me. Weird it runs on my build machine using the same gst version. I will have to take another look.
tumdedum has quit [Ping timeout: 255 seconds]
tumdedum has joined #imx6-dev
<obinou>
ok kernel compiled . I just have to create/find a DTS file for gk802 cubox-i since it seems they are not in the latest vanilla kernel yet