lifwl has quit [Killed (Sigyn (Spam is off topic on freenode.))]
vpeter has quit [Remote host closed the connection]
matthias_bgg has joined #linux-sunxi
elros1 has joined #linux-sunxi
wtw_ has joined #linux-sunxi
wtw_ has quit [Remote host closed the connection]
Jackie365 has joined #linux-sunxi
<Jackie365>
Where can I get more information about F1c600?
<Jackie365>
the wiki does not provide much
Kwiboo has quit [Ping timeout: 245 seconds]
<karlp>
probably because it's an archaic legacy part, that predates ~all sunxi, and this dude just found a roll or three of them and is flogging them off?
<karlp>
you're talking about this lichee thing I presume?
<Jackie365>
Ha! yes!
<Jackie365>
I have some difficulty understanding the user manual of f1c600
<Jackie365>
Currently I am trying to finger out the boot system
yann has quit [Ping timeout: 268 seconds]
<karlp>
they seem like curious toys honestly.
<Jackie365>
yes, I agree
<Jackie365>
I'd like to develop a product using f1c600
<Jackie365>
so I bought the lichee board
<karlp>
have you got supply chain stbaility verified for that part?
<karlp>
or is this a product of ~20?
<Jackie365>
sorry, what is "~20"
<karlp>
was curious at the number of units you want to do for your product.
<karlp>
~ == circa/about
<Jackie365>
I don't know if Allwinner will supply that kind chip in the future
<Jackie365>
And it's not going to be "a product of ~20"
<Jackie365>
You're saying that I should consider choosing another chip?
<Jackie365>
And I'm really curious about how you get to know the lichee board.
tnovotny has joined #linux-sunxi
vpeter has joined #linux-sunxi
<Jackie365>
Is the F serials going to be end of line?
<karlp>
as I understood it, they were EoL'd years ago.
<Jackie365>
Oh my God
<KotCzarny>
werent they quite recently new models?
<karlp>
but maybe speak to your allwinner sales rep
<KotCzarny>
ie. few new eink ereaders
<Jackie365>
And the Allwinner website has F serials
Everspace has joined #linux-sunxi
Everspace has quit [Remote host closed the connection]
<Jackie365>
So,were they really EOL?
* karlp
shrugs it's still on their website, no idea what their plans are,
<Jackie365>
agree
<Jackie365>
so I'd better change my plan
<karlp>
well, speak to them.
<Jackie365>
ok, I'll contact the Allwinner soon
<Jackie365>
thank you, karlp
<willmore>
I thought the A13 was the 'someone just found a spool of it and was trying to sell them off.' looks like they found other spools. ;)
<karlp>
were you goint to integrate the part directly or try and integrate the lichee part?
<Jackie365>
use the soc directly
dddddd has joined #linux-sunxi
<Jackie365>
I think it's powerful enough for my product
<Jackie365>
peripheral may cost a lot if I use other chips
<Jackie365>
Products based on F serials can be extremely cheap,right?
<KotCzarny>
as any other allwinner soc based product
<Jackie365>
low power consumption
<Jackie365>
F serials has both,right?
<karlp>
both what?
<Jackie365>
it's cheap and consume less power
<Jackie365>
am i getting it right?
<Jackie365>
I am new to allwinner chips, so after searching the web for some time,I came to that conclusion
<Jackie365>
So any other choice?
<karlp>
f1c600 product brief is from 2015-2016, so *shrugs*
<karlp>
how could anyone possibly know what other choices there are, you've not metnioned anything about your requirements :)
Andy-D has joined #linux-sunxi
<Jackie365>
oh,performance is not needed. cost and power are the only factors
<Jackie365>
peripheral does not need to be very rich
<Jackie365>
SD card and USB and RGB output would be enough
<Jackie365>
all others are irrelevant
<Jackie365>
I think the V serials is too much
<Jackie365>
so is other serials
leviathanch has joined #linux-sunxi
yann has joined #linux-sunxi
jazkarB has joined #linux-sunxi
<Jackie365>
I am leaving . Thank you guys, thank you for your help.
Jackie365 has left #linux-sunxi ["WeeChat 1.6"]
jazkarB has quit [Ping timeout: 252 seconds]
Network2501uB has joined #linux-sunxi
Network2501uB has quit [Remote host closed the connection]
afaerber has quit [Quit: Leaving]
<paulk-leonov>
mripard, wens: hey there! I've been working on bringing up the bananapi 3.5" LCD (used by the FOSDEM video boxes) with mainline and found that the LCD DE line must be configured as a GPIO held low (the panel most likely takes an inverted state, which our TCON apparently can't do so DE has to be asserted low all the time instead)
qdickonYZ has quit [Remote host closed the connection]
indy has joined #linux-sunxi
IgorPec has quit [Ping timeout: 252 seconds]
afaerber has joined #linux-sunxi
<lvrp16>
allwinner will manufacture old chips if you want 50k or more
<lvrp16>
so no chip us really eol, just masks are shelved
<karlp>
kinda why we suggested to talk to allwinner not us :)
<mripard>
paulk-leonov: I'm not really sure that it's a safe thing to do on all displays
<paulk-leonov>
mripard, indeed, I think it's only relevant as a workaround for this particular display
<mripard>
paulk-leonov: the display engine repeats the last pixel when not in the active window
<mripard>
so it can lead to really weird panel behaviours and side effects
<paulk-leonov>
either way, if the DE pin is not always asserted low, I get a black screen
<paulk-leonov>
would it make sense to have a property for the de node in device-tree to indicate that this quirk is required?
<paulk-leonov>
display-engine node*
<paulk-leonov>
this seems more related to the display engine not being able to invert the DE pin polarity than a panel-side issue, IMO
agraf` is now known as agraf
<f11f12>
paulk-leonov: which SoC is that?
<paulk-leonov>
f11f12, A20
<f11f12>
paulk-leonov: I have a custom board with A20 + LCD that works (without patch). I can test your patch on my LCD to see what it does.
lurchi_ is now known as lurchi__
<paulk-leonov>
f11f12, well what I'm talking about only concerns the very specific panel I'm dealing with, it does not affect other panels in any way
<paulk-leonov>
so that won't really be necessary :)
<f11f12>
paulk-leonov: ok. If you keep DE low, youuse h/v sync?
<paulk-leonov>
I suppose so
<paulk-leonov>
I'm still unsure whether it's an inverted DE or just that the pin that was routed to something else that needs to be kept low
<mripard>
paulk-leonov: that should be a panel I guess?
<mripard>
+property
IgorPec has joined #linux-sunxi
d__ed has joined #linux-sunxi
d_ed has quit [Excess Flood]
The_Loko has joined #linux-sunxi
timdorohin has joined #linux-sunxi
EnorMOZxc has joined #linux-sunxi
libv has quit [Ping timeout: 246 seconds]
libv has joined #linux-sunxi
EnorMOZxc has quit [Remote host closed the connection]
russ_ has joined #linux-sunxi
<rellla>
wens, mripard: i don't want to nag you :) but i have pushed my log when executing v4l2-request-test on A10, which includes cma and drm debug infos know. my issue seems to be drm related, so maybe you can find any suspicious things. the same kernel runs fine on A20.
<rellla>
kernel is 4.19-rc4 with your cedrus/4.18/drm (rebased) and the latest h265 tree. dts nodes for sun4i drm frontend and cedrus are added. v4l2-request-test reports everything decoded and displayed correctly.
<mripard>
have you tested the display without v4l2-request-test?
<rellla>
mripard: a few days ago i did it with vlc which seems to do correct iirc
<rellla>
there has to be a small bit which is still different from sun4i-a10 to sun7i-a20, because on the cubieboard2 it is working fine.
<mripard>
I guess a good first test would be to test whether the display works fine or not using modetest
<mripard>
including with multiple planes
<rellla>
mripard: can you give me hint for "with multiple planes"?
wasutton3 has joined #linux-sunxi
Kwiboo has joined #linux-sunxi
m4ssi has joined #linux-sunxi
epictetusYm has joined #linux-sunxi
maz has quit [Ping timeout: 252 seconds]
epictetusYm has quit [Remote host closed the connection]
JohnDoe_71Rus has joined #linux-sunxi
<mripard>
rellla: you have multiple planes on the display engine, each of them being able to display a different framebuffer, and the engine will do the hardware composition
<mripard>
the framebuffer console takes one
<mripard>
and thene v4l2-request-test takes a second one, and put it on top of fbdev's
<mripard>
so you want to test that those two planes are working properly
<rellla>
with modeset i have 4 planes for the crtc-id, doing "modetest -M sun4i-drm -P 28:1280x720+0+0@NV12 for example works for all plane ids.
<f11f12>
mripard: is the mali output renderd to a different plane than the default /dev/fb0 output? (on R40/V40)
<mripard>
rellla: and if you try two -P options, one being like half the first one ?
<mripard>
so something like -P 28:1280x720@XR24 -P 29:640x360@NV12
<mripard>
(adjust for actual IDs)
talyzqI has joined #linux-sunxi
<mripard>
f11f12: it will draw to fb0 if you're using the fbdev blob
<f11f12>
mripard: on V40 I can run qt5 demos but the screen remains blank, fbv works fine (fb0==HDMI out). I use r6p2/fbdev
nuuuciano has joined #linux-sunxi
<f11f12>
mripard: proc/interrupts shows that interrupts are generated when I run a qt5/opengl demo (e.g. hellowindow)
talyzqI has quit [Remote host closed the connection]
<mripard>
f11f12: I never tested the R40, or heard of anyone that tested the mali on that SoC
<mripard>
so I don't really know, sorry
ashnurte has joined #linux-sunxi
reinforce has quit [Quit: Leaving.]
ashnurte has quit [Remote host closed the connection]
<rellla>
let me check the same with A20 to see a difference ...
jbrown has quit [Ping timeout: 252 seconds]
msimpson has quit [Quit: Leaving]
jbrown has joined #linux-sunxi
maz has joined #linux-sunxi
leviathanch has quit [Ping timeout: 252 seconds]
xpete has joined #linux-sunxi
<rellla>
mripard: drm debug log while executing v4l2-request-test seems to be the same on a10 and a20. modetest also runs fine on both. i have no idea where to search for the issue anymore. can i dump the output of v4l2-request-test into a file somewhat?
Mandragor has quit [Read error: Connection reset by peer]
Mandragor has joined #linux-sunxi
<rellla>
*dump the decoded frame of course
murr4yXi has joined #linux-sunxi
murr4yXi has quit [Remote host closed the connection]
Mandragor has quit [Remote host closed the connection]
Putti has joined #linux-sunxi
Mandragor has joined #linux-sunxi
leviathanch has joined #linux-sunxi
tnovotny has quit [Quit: Leaving]
wasutton3 has quit [Ping timeout: 252 seconds]
leviathanch has quit [Ping timeout: 268 seconds]
okapia has joined #linux-sunxi
wasutton3 has joined #linux-sunxi
okapia has quit [Remote host closed the connection]
s_frit has quit [Remote host closed the connection]
s_frit has joined #linux-sunxi
i__ronMi has joined #linux-sunxi
aib has quit [Ping timeout: 268 seconds]
aib has joined #linux-sunxi
i__ronMi has quit [Remote host closed the connection]
jaganteki has joined #linux-sunxi
daamienvH has joined #linux-sunxi
daamienvH has quit [Remote host closed the connection]
netlynx has joined #linux-sunxi
<rellla>
paulk-leonov, mripard: found it!!
FisiuNZ has joined #linux-sunxi
<rellla>
the issue is with scaling. if i drop calculation of scaled_height and scaled_width and just take the original values of width and height and feed coomit_atomic_mode with them, everything is shown nice
<paulk-leonov>
hmm
<paulk-leonov>
I see
<paulk-leonov>
rellla, I guess I'll try to setup an A10 tablet soonish and see what I can do