Er0l has quit [Read error: Connection reset by peer]
Er0l has joined #imx6-dev
Er0l has quit [Ping timeout: 245 seconds]
Er0l has joined #imx6-dev
victhor has quit [Ping timeout: 245 seconds]
jnettlet has quit [Ping timeout: 240 seconds]
projectg1s has joined #imx6-dev
projectgus has quit [Ping timeout: 252 seconds]
projectg1s is now known as projectgus
obinou_ has joined #imx6-dev
cnxsoft has quit [Ping timeout: 255 seconds]
epl has quit [Ping timeout: 252 seconds]
epl has joined #imx6-dev
silviof1 is now known as silviof
kroon has joined #imx6-dev
jnettlet has joined #imx6-dev
applepi has quit [Ping timeout: 240 seconds]
cnxsoft has joined #imx6-dev
rz2k has joined #imx6-dev
diego_r has joined #imx6-dev
tasslehoff has joined #imx6-dev
fossxplorer has joined #imx6-dev
<tasslehoff>
anyone know what the issue with edid is? is it not read correctly? my debug prints show that it discards (almost) all modes because the timings don't match the mxc_cea_modes.
<tasslehoff>
as mentioned yesterday 720p works fine on my system, but if I boot with hdmi connected it is discarded because the timing parameters don't match.
<simmel80___>
tasslehoff: maybe you can try to relax the requirements a little in the code, so that it does not decide to dicard?
<simmel80___>
did you find out in which function the discard actually happens?
<simmel80___>
most monitors should be just fine with the timing parameters a little of
<simmel80___>
some caveats apply though, eg. trying to set 1366x768 in LVDS would not work, 1360x768 is just fine, the same may be true for HDMI
<simmel80___>
an the parameter space for the pixel clock is pretty restricted anyways
<tasslehoff>
simmel80___: it discards stuff in mxc_edid_fb_mode_is_equal, since that demands everything to be exact
<simmel80___>
so i guess freescale and the display vendor just have differnt interpretation of what the exact cea mode should be
<simmel80___>
does it allways discard regarding a specific parameter (eg h_width, v_width, pixclock etc.) or all over the place
<tasslehoff>
simmel80___: all over the place. left/right/upper/lowe margin, hsync and vsync
<simmel80___>
maybe just code a generic function, that accepts a value if it is in a say 2% margin around the expected value
<tasslehoff>
simmel80___: yeah. I actually consider disabling edid instead, since I know "my resolution".
cnxsoft has quit [Ping timeout: 255 seconds]
<simmel80___>
yes, i thought to do the same when it would have come to "fix" this issue for an actual customer project
<simmel80___>
so far i just lived with the brokenness
<simmel80___>
but thanks a lot, with your research it should be more easy to find a solution when i would need to
<tasslehoff>
simmel80___: np. my "fix" is to run mxc_hdmi_default_modelist on HDMI_EDID_SUCCESS in mxc_hdmi.c
<tasslehoff>
the commit message will be "look ma, no hands!" :p
<simmel80___>
;-)
<tasslehoff>
simmel80___: more info. from what I see in the kernel, it throws away everything that does not matches a cea timing.
<tasslehoff>
the edid of my display has the 720p dmt timing
<simmel80___>
maybe it was coded with the expectation of having an consumer/entertainment/tv display
<tasslehoff>
cea = consumer electronics association (TVs and stuff), dmt = display monitor timing (Computer displays)
<tasslehoff>
simmel80___: my thought exactly
<simmel80___>
so they made assumptions that maybe not hold for productive work displays
<simmel80___>
hehe, we come to the same conclusion...
<tasslehoff>
yup. I wonder if this is intended. perhaps I should ask on some mailing list or on the freescale forums.
<simmel80___>
could be, desktop experience is maybe not high up on the list of thing to do with the imx6, consumer video consumption more so
<simmel80___>
i think it is just lazyness/coding accoring only to limited requirements
<simmel80___>
which is not a bad approach per se, why do more than is needed
<tasslehoff>
simmel80___: I also suspect that ok/good monitors support cea timings
<tasslehoff>
that's a guess :)
<simmel80___>
i do not remember from the edid standard, but it should be possible
<simmel80___>
eg. my our projector in the office displays just fine, while several older monitors seem to have the issue as described
<simmel80___>
its probably just a matter of declaring both a list of dmt and a list of cea modes in the same edid
<tasslehoff>
simmel80___: yeah. anyway, I guess that concludes the research. now I/we know how to fix it if needed. thanks for sparring.
<simmel80___>
next time when i have a look at it, i will dump both edids and compare them regarding this aspect
<simmel80___>
you are welcome, thanks for doing all the hard work ;-), i was just sitting there watching
<simmel80___>
sorry i could not do more atm, my test setup for imx6 is currently not running
<tasslehoff>
np :)
cnxsoft has joined #imx6-dev
cnxsoft has quit [Ping timeout: 255 seconds]
tasslehoff has quit [Quit: WeeChat 0.4.2]
tasslehoff has joined #imx6-dev
cnxsoft has joined #imx6-dev
cnxsoft has quit [Ping timeout: 255 seconds]
victhor has joined #imx6-dev
ballock has quit [Quit: Changing server]
ballock has joined #imx6-dev
rz2k has quit []
cnxsoft has joined #imx6-dev
tasslehoff has quit [Quit: WeeChat 0.4.2]
kroon has quit [Quit: Leaving]
cnxsoft has quit [Quit: cnxsoft]
<codinho>
dv_, hi, I'm observing error in imxvpudec on master, <imxvpudec0> error: loading VPU failed: failure is it known issue?
<codinho>
[ERR] Error in IOSystemInit()[ERR] IOSystemInit() failure.
<dv_>
no, never seen this
<dv_>
details. kernel version, gstreamer-imx version, what hardware you are using
<codinho>
Linux imx6qsabrelite 3.10.17-1.0.0_ga+yocto+gcfb334d #1 SMP Tue Jun 24 23:34:16 NOVT 2014 armv7l GNU/Linux