jo0nas has quit [Read error: Connection reset by peer]
jo0nas has joined #linux-sunxi
Mangy_Dog has joined #linux-sunxi
megi1 has quit [Quit: WeeChat 2.8]
megi has joined #linux-sunxi
jbrown has joined #linux-sunxi
[VroomAutism] has quit [Quit: growth in middle of cactis, my rancor.]
romanstratiienko has joined #linux-sunxi
[MasterPiece] has joined #linux-sunxi
rsglobal has quit [Ping timeout: 256 seconds]
pmp-p has quit [Ping timeout: 258 seconds]
pmp-p has joined #linux-sunxi
cnxsoft1 has joined #linux-sunxi
cnxsoft has quit [Read error: Connection reset by peer]
rsglobal has joined #linux-sunxi
romanstratiienko has quit [Ping timeout: 256 seconds]
totalgaara has joined #linux-sunxi
warpme_ has quit [Quit: Connection closed for inactivity]
totalgaara has quit [Remote host closed the connection]
Asara has quit [Quit: leaving]
Asara has joined #linux-sunxi
abelvesa has quit [Quit: leaving]
abelvesa has joined #linux-sunxi
abelvesa has joined #linux-sunxi
rsglobal has quit [Remote host closed the connection]
reinforce has quit [Quit: Leaving.]
Wizzup has quit [Quit: brb]
Wizzup has joined #linux-sunxi
cnxsoft1 has quit [Quit: cnxsoft1]
TEKrantz has quit [Ping timeout: 265 seconds]
gaston1980 has joined #linux-sunxi
<Ashleee>
jernej, I've had one more idea... dd the android that is on eMMC onto a sd card and then hexedit the part where is the "press any key to stop boot: ..." delay from 0 seconds to something when you can actually press any key and see if there's a widget for memory dumping :V
<Ashleee>
that's if I can disassemble enough parts to see where it is called from
<Ashleee>
but ghydra :P
<jernej>
hm... why can't you try to drop to U-Boot prompt from eMMC?
<Ashleee>
the countdown has 0 seconds and no matter how much I bash any key it doesn't stop the boot
<jernej>
I successfully loaded your dump to ghydra and I managed to locate all DRAM code
<jernej>
now tedious task of comparing the code begins...
<Ashleee>
:)
<Ashleee>
if you can get it working I'll pass some beer or whatever you drink if I ever come in your country or vice versa :P
<jernej>
anyway, even if delay is set to 0, there should be method to drop to prompt...
<Ashleee>
hm I wonder if busybox/android that's on there has uboot env stuff
swiftgeek has quit [Read error: Connection reset by peer]
swiftgeek has joined #linux-sunxi
<Ashleee>
I've even tried contacting allwinner about libdram 2.78 but no response so far :D
<mirko>
i'm unfortunately still having the issue with H6 that power-cycling the attached AV-receiver and/or TV cuts any HDMI signal the H6 doesn't recover from (end devices say: "no signal" after being back on). would really like see this fixed. nothing in logs/dmesg. anything which might be useful narrowing it down?
TEKrantz has joined #linux-sunxi
<jernej>
mirko: does it work ok without AVR in between?
<mirko>
jernej: didn't try for some time, let me make sure..
<jernej>
I don't have AVR so I can't test it...
KotCzarny has quit [Ping timeout: 240 seconds]
<mirko>
jernej: same, direct connection to sony bravia tv - power-cycling tv -> no signal
<jernej>
btw, do you have HDMI CEC enabled on your TV?
<mirko>
i tested with megous' as well as vanilla kernel (5.7/5.8) - however both having (certain) libreelec patches applied
<jernej>
it's possible that your board gets shutdown signal when you turn off TV
<mirko>
jernej: i explicitly disabled/enabled on the AVR which didn't make a difference. i'm struggling with sony's nomenclature (TV side) though
<mirko>
jernej: the H6 is still reachable via network/SSH
<mirko>
everything is fine, except HDMI output
<jernej>
then I don't know
<mirko>
another observation: once it's in that state, a "reboot" takes quite a few /minutes/, while when HDMI is working fine, the reboot happens within an expected amount of time
<mirko>
reboot, as in the command `reboot`
<jernej>
I tested this in the past on my TV (LG B8)
<mirko>
jernej: switching off/on the TV by remote is fine (android tv, only goes into standby)
<jernej>
did you try different distro with mainline kernel?
<mirko>
jernej: no, just debian
<jernej>
can you try LE?
<mirko>
if that might help i'm happy to do so
<jernej>
I just want to rule out userspace
<jernej>
debugging HDMI is hard
<jernej>
that's why there is still no 4k @ 60 Hz support
<mirko>
because of hard to debug hdmi issues?
<jernej>
yes
<jernej>
it almost work (if one devmem write is executed after resolution is set)
<mirko>
just to maybe easily rule out another cause: my H6 is on an orange-pi-3 PCB - do/did you by chance test on that one, too? otherwise it might also be PCB related..
<jernej>
yes, I often test on OPi3
<mirko>
*do/did you "test"
<mirko>
ah, ok, so it's most likely not board related at least
<mirko>
and there's no hdmi related registers i can just dump before and after and compare i assume?
<jernej>
you can read status register to see if plug detection works ok
<jernej>
mirko: execute "devmem b 0x06003004" and check bit 1
<jernej>
note that you must use byte access to get correct value
<mirko>
by "devmem" you mean the busybox applet?
<mirko>
or the device node?
<jernej>
well, I have standalone devmem tool
<jernej>
but that which comes with busybox is also ok
<jernej>
on some distros it's devmem2
KotCzarny has joined #linux-sunxi
<jernej>
I don't have idea if it is present on debian or not
<mirko>
wasn't aware of a standalone util, will check..
<mirko>
# busybox devmem 0x06003004
<mirko>
0x000000F3
<mirko>
not sure what the 'b' does as 1st argument on your's, busybox's devmem only accepts the addr as 1st arg
<mirko>
jernej: not really, as a) i don't have another screen at hand and b) the HW is fairly well integrated by now (as the board also acts as ambilight)
<mirko>
i could order another H6 board, though
<jernej>
megi: that one is used only if CLK_RATE_FLAT_FACTORS flag is set and it's set only for sun8iw8 pll_cpu
<megi>
H6 is sun50iw3
<jernej>
I know, so that function is not used for H6
<megi>
good, but it may still fix other SoCs where this is broken, like H3
<megi>
I didn't try this method with my NKMP clock change tester, it's worth a try
<megi>
wouldn't mind if all the kludges that are mainline because of this, could be removed
<megi>
nevermind that the workarounds cause timer jumps
<megi>
and this might not
netlynx has quit [Quit: Ex-Chat]
<Ashleee>
timer jumps? :) *mumbles something about A64*
<megi>
on H3
<megi>
just small ones :)
<Ashleee>
ah
<Ashleee>
so not as bad as on A64 with 33 years or whatever it was? :)
<jernej>
megi: just one detail regarding clocks - mainline driver never disables lock before rate is changed
<jernej>
could be this the issue?
* psydread
is from the future, from the year 2115 to be exact
<jernej>
hm... more specifically, it seems mainline implements what BSP driver refers as PLL_LOCK_OLD_MODE
<jernej>
H6 BSP clk driver has all PLLs marked as "new" PLL lock mode
<jernej>
imo worth a try
TheSeven has quit [Remote host closed the connection]
TheSeven has joined #linux-sunxi
TheSeven has quit [Remote host closed the connection]
TheSeven has joined #linux-sunxi
<mirko>
jernej: one more observation: no matter if the opi3 is connected to AV-receiver or TV directly, if i suspend the TV (=switching off/on via remote (=not power-cycling)), the opi3 continue to work normal, but enters described state after a couple of minutes (if i had to guess i'd say 5-10mins). it feels like a timeout there. if the AV/TV stays on, though, everything is normal for hours
<megi>
interesting
<jernej>
mirko: out of curiousity - are you using Kodi or any other video app or other full screen app in gbm mode?
afaerber_ has joined #linux-sunxi
<jernej>
megi: another point - BSP clk driver waits for lock inside spinlock
<jernej>
so any time clock is changed, it's sure it's stable
<jernej>
technically you could queue two clk rate changes one after another and second could be executed before first one is stable
TheSeven has quit [Ping timeout: 244 seconds]
TheSeven has joined #linux-sunxi
<jernej>
but BSP clk driver also does other strange things, like writing read only bits (lock), at least according to datasheet
<mirko>
jernej: i'm running kodi, however also have librespot running (through pulse) - neither librespot nor kodi output anything on HDMI in this state
<mirko>
kodi on gbm, as far as i can see pretty much how LE does it
<mirko>
(using panfrost)
<jernej>
ok, I added workaround in LE kernel not long ago (about two weeks) which prevent some screen corruptions
JohnDoe_71Rus has quit [Quit: KVIrc KVIrc Aria 5.0.1, revision: 5.0.1+git-7433-0df9f22f2, build type: debug, sources date: 20160102, built on: 2019-12-08 19:19:20 UTC 5.0.1+git-7433-0df9f22f2 http://www.kvirc.net/]
<jernej>
in theory it could also cause that Display Engine stops providing image to HDMI
jstein has joined #linux-sunxi
<mirko>
jernej: but then audio would still work, no?
<jernej>
not sure
<mirko>
despite, i applied all linux patches from LE i could find from master only a couple of days ago