cnxsoft has quit [Read error: Connection reset by peer]
linkmauve has joined #linux-sunxi
linkmauve has quit [Ping timeout: 246 seconds]
Mangy_Dog has quit [Ping timeout: 240 seconds]
cnxsoft has joined #linux-sunxi
victhor has joined #linux-sunxi
choozy has joined #linux-sunxi
Mangy_Dog has joined #linux-sunxi
Mangy_Dog has quit [Changing host]
Mangy_Dog has joined #linux-sunxi
apritzel has quit [Ping timeout: 240 seconds]
jstefanop has joined #linux-sunxi
jstefanop has quit [Ping timeout: 268 seconds]
specing_ has joined #linux-sunxi
specing has quit [Ping timeout: 268 seconds]
specing_ is now known as specing
jstefanop has joined #linux-sunxi
jstefanop has quit [Ping timeout: 246 seconds]
apritzel has joined #linux-sunxi
cnxsoft1 has joined #linux-sunxi
cnxsoft has quit [Ping timeout: 246 seconds]
elros1 has joined #linux-sunxi
<apritzel>
jernej: so in summary dumping the pinctrl registers on that H6 TV box was much more insightful than trying to parse that Android DT
<apritzel>
for instance that PL5 in their DT: it was mentioned both as a power key, and also as driving VBUS
<apritzel>
but actually it was neither: there is no power key on that box, and VBUS is always on, toggling PL5 does nothing to the voltage on the USB ports
<apritzel>
same with PL2 ("recovery key"): there is a FEL switch on the board, accessible through a small hole, but this is just the normal FEL switch, not a GPIO
elros1 has quit [Read error: Connection reset by peer]
<apritzel>
jernej: no, change, without forcing it on the command line I get no picture
<apritzel>
I could at least verify that PH2 enables DDC, without it I get only 1024x768 on my WUXGA screen
buzzmarshall has joined #linux-sunxi
Throwawayname has joined #linux-sunxi
<megi>
there's some func cyclic dependency issue in 5.13-rc1
Throwawayname has quit [Max SendQ exceeded]
<megi>
platform 1ca0000.dsi: Fixing up cyclic dependency with 1c0c000.lcd-controller
<megi>
and the display fails to work :)
danboid has joined #linux-sunxi
danboid has left #linux-sunxi [#linux-sunxi]
danboid has joined #linux-sunxi
<danboid>
Hi all
<danboid>
I recently got a H6 based T95 Max box. The newest kernel I've got booting on it so far is 5.9 but its got a bad HDMI bug. If you disconnect then reconnect the HDMI, the display stops working until you power cycle the box. Android on the same box doesn't do this.
<danboid>
Is this a known prob (thats fixed in newer kernels)?
<apritzel>
danboid: which 5.9 kernel? vanilla mainline?
<danboid>
I'm forced into rebuilding the kernel (5.10) anyway as I want to try getting the cedrus video decoder stuff working under armbian
<danboid>
I'm running an unofficial build of armbian from the forums
<danboid>
but I've not been able to get any of those to boot on my box
<danboid>
They have 5.10.9
<apritzel>
sorry, this is all too downstream for me, but I would try to apply the patch on 5.12 first
warpme_ has quit [Quit: Connection closed for inactivity]
<apritzel>
and if you need to go back to 5.10, use the latest, it's at v5.10.35 already
<danboid>
Makes sense
<apritzel>
does your box show a picture with a specific video=HDMI... command line?
<apritzel>
(that's the problem I am facing on my Eachlink H6 Mini)
<jernej>
those patches won't apply as-is, at least not all of them
warpme_ has joined #linux-sunxi
<jernej>
on 5.12 I mean
<danboid>
jernej, Yeah I thought that'd be likely
<danboid>
apritzel, video=HDMI... being kernel argument?
<danboid>
I've not experimented with any kernel asrguments yet
<apritzel>
danboid: yeah, that's a hack because apparently the "hot plug detection" doesn't work on this Eachlink box
<danboid>
Ah yes, thats the phrase I'm bitching about
<danboid>
hpd
<apritzel>
exactly, you can override this by "video=HDMI-A-1:e" on the kernel command line
<apritzel>
it doesn't work for me at all without, but apparently you get a picture initially, and it's just reconnecting that fails?
<apritzel>
what is your DT? I don't see the box having one mainline?
<apritzel>
danboid: and it that box otherwise decent? I see one quite cheap on eBay here ...
<danboid>
Yes I get a picture initially and its reconnecting that fails
<danboid>
Under Android its called MBOX
<danboid>
its dtb is tanix-tx6
<danboid>
It only has 100 Mb ethernet but you can buy a USB 3 gig ethernet adapter and hub for $10
<danboid>
The analogue audio isn't supported by Linux yet
<danboid>
and you need to use a USB fan on it (another $10) to avoid thermal throttling when using all cores
<apritzel>
yeah, I am after the H6 killer feature: USB 3.0
<danboid>
Yes the USB 3 is pretty good. I get about 290 MB/s read speeds off my SATA3 SSD
<apritzel>
my box gets killed when I plug a SATA adapter plus SSD in (due to voltage drop and poor power supply), does that work for you?
<apritzel>
I see that the Tanix TX6 DT doesn't describe WiFi and the eMMC, did you fix this?
<danboid>
I have successfully used a mechanical 5 TB USB 3 drive with it, no probs
<danboid>
using the bus power
<danboid>
no powered hub
<apritzel>
it's a barrel plug?
<danboid>
Yeah
<danboid>
Cost me £30
<danboid>
It includes a PSU, HDMI cable and a remote but I failed to get the remote to work under the stock Android and an update stock. Maybe my controller is bust?
<danboid>
Not bothered
<apritzel>
the IR controller worked out of the box for me, but the TX6 remote probably uses different codes
<apritzel>
try "ir-keytable -t"
<danboid>
I've not tried it under armbian yer
<danboid>
Is ir-keytable an android command?
<apritzel>
no, it's part of v4l-utils
<apritzel>
it's packaged by Ubuntu
<danboid>
will it work with mpv?
<danboid>
Should it?
<danboid>
or MATE
<danboid>
Maybe control the master volume?
<danboid>
dunno what I'd use ir for under normal linux :)
<danboid>
mpv
<danboid>
Don't suppose it might work as a pointer would it?
<danboid>
I can't imagine that working too great
<megi>
hmm, I guess that dependency fixup is unrelated to my display issue, display works with pure mainline
<megi>
there are just extra messages about dependencies in 5.13-(almost)rc1
<megi>
fw_devlink=on is default in 5.13 again, so the probe order is different
<megi>
maybe devlink heuristics fail when having both tcons enabled
<megi>
in DT
<megi>
using fw_devlink=permissive on kernel command line makes it work again
<megi>
and probe correctly
jstefanop has joined #linux-sunxi
jstefanop has quit [Ping timeout: 252 seconds]
<danboid>
apritzel, If you get one and have USB power issues, just run:
<danboid>
echo on > '/sys/bus/usb/devices/1-1/power/control
<danboid>
I needed to run that to get it working with my SATA disks
<danboid>
apritzel, Not fully tested the wifi under armbian but it was showing SSIDs so it seemed to be at least partially working
<danboid>
lsblk is showing my eMMC no and I've not tried to fix that
<danboid>
It ISN'T showing my eMMC sorry
<danboid>
So no eMMC, no analogue audio output and HDMI hotplug are only current issues I'm aware of with this box
<jernej>
danboid, apritzel: I did send Tanix TX6 eMMC patches, but was too lazy to follow up
<jernej>
I guess I could resend it this cycle :D
<danboid>
jernej, Oh cool! Have you got a link to that patch pls?