<gediz0x539>
you're welcome. datasheet of a50 is uploaded by either libv or yoq, i think.
camus has joined #linux-sunxi
faruk has quit [Ping timeout: 240 seconds]
kaspter has quit [Ping timeout: 256 seconds]
camus is now known as kaspter
reinforce has joined #linux-sunxi
asdf28 has joined #linux-sunxi
hexdump0815 has joined #linux-sunxi
<hexdump0815>
jernej: when you are getting into upstreaming the h616 u-boot changes, can you maybe also upstream the rank detection stuff we tested some months ago too - looks like its not yet in mainline u-boot: https://freenode.irclog.whitequark.org/linux-sunxi/2020-09-23#27979633
<hexdump0815>
jernej: for h6 that was
<hexdump0815>
jernej: apritzel: and when you are getting ready to upstream the opi zero 2 dts to u-boot it might be nice to also shift over the h6 tanix-tx6 one from the kernel too and a defconfig for it as there is still no such ddr3 h6 config in the tree
hexdump0815 has quit [Remote host closed the connection]
ldevulder has joined #linux-sunxi
AneoX has joined #linux-sunxi
faruk has joined #linux-sunxi
xes has quit [Ping timeout: 260 seconds]
xes has joined #linux-sunxi
AneoX has quit [Ping timeout: 264 seconds]
AneoX has joined #linux-sunxi
ldevulder has quit [Quit: Leaving]
ldevulder has joined #linux-sunxi
kaspter has quit [Ping timeout: 260 seconds]
kaspter has joined #linux-sunxi
cmeerw has joined #linux-sunxi
tnovotny has joined #linux-sunxi
<Tooniis>
does the lima driver work with fbturbo? or do I need something else?
matthias_bgg has joined #linux-sunxi
pion3k has joined #linux-sunxi
apritzel has joined #linux-sunxi
faruk has quit [Quit: Leaving]
faruk has joined #linux-sunxi
cnxsoft has quit [Read error: Connection reset by peer]
cnxsoft1 has joined #linux-sunxi
victhor has joined #linux-sunxi
<Tooniis>
is hardware acceleration possible at all on mainline on A13 currently?
<insep_>
it did work for me on a10
<insep_>
not sure how much of gpu was involved though, but at least it was showing lima for me
<Tooniis>
with simple fb?
<insep_>
is hw accel even possible with simplefb?
<Tooniis>
idk
<Tooniis>
that's why im asking
<Tooniis>
afaik it isnt
<Tooniis>
this q8 tablet i have uses simple fb by default
<Tooniis>
idk if anything else is possible
faruk_ has joined #linux-sunxi
damex has quit [Ping timeout: 272 seconds]
netlynx has joined #linux-sunxi
faruk has quit [Ping timeout: 256 seconds]
<Tooniis>
insep: which device did you use?
<insep_>
cubieboard
<insep_>
also that was with 5.9.9 kernel, so not exactly mainline
faruk__ has joined #linux-sunxi
faruk_ has quit [Ping timeout: 260 seconds]
ldevulder has quit [Quit: Leaving]
ldevulder has joined #linux-sunxi
matthias_bgg has quit [Ping timeout: 264 seconds]
camus has joined #linux-sunxi
kaspter has quit [Ping timeout: 256 seconds]
merbanan has quit [Ping timeout: 256 seconds]
camus is now known as kaspter
pion3k has quit [Quit: Leaving.]
random_yanek has quit [Quit: random_yanek]
random_yanek has joined #linux-sunxi
andyp73 has joined #linux-sunxi
andyp73 has left #linux-sunxi [#linux-sunxi]
matthias_bgg has joined #linux-sunxi
merbanan has joined #linux-sunxi
kaspter has quit [Ping timeout: 256 seconds]
kaspter has joined #linux-sunxi
<apritzel>
clementp[m]: boy, there were quite some more clocks to fix
<apritzel>
clementp[m]: up to the point where I am no longer sure sharing the file is helpful, but I will leave this up for the list to decide
Mangy_Dog has joined #linux-sunxi
camus has joined #linux-sunxi
kaspter has quit [Ping timeout: 272 seconds]
camus is now known as kaspter
damex has joined #linux-sunxi
pion3k has joined #linux-sunxi
<apritzel>
so it seems like G2D is back (A50, A100, H616, ...), is that something useful we can make use of in the Linux driver?
<buZz>
to not have the default heartbeat stuff illuminating half your room :P
<jernej>
that's was already discussed for all board in general
<buZz>
ah, np, just thought you'd like to know
<jernej>
but I forgot what was result of that
<jernej>
but yes, I can imagine it can be annoying
<buZz>
:) easy enough to fix, i added those lines to /storage/.config/autostart.sh
vagrantc has joined #linux-sunxi
lucascastro has quit [Ping timeout: 256 seconds]
lucascastro has joined #linux-sunxi
popolon has joined #linux-sunxi
lucas_ has joined #linux-sunxi
lucascastro has quit [Ping timeout: 240 seconds]
lucas_ has quit [Ping timeout: 265 seconds]
lucascastro has joined #linux-sunxi
pion3k has joined #linux-sunxi
pion3k has quit [Client Quit]
lucascastro has quit [Ping timeout: 256 seconds]
matthias_bgg has quit [Remote host closed the connection]
lucascastro has joined #linux-sunxi
lucas_ has joined #linux-sunxi
lucascastro has quit [Ping timeout: 256 seconds]
kaspter has quit [Quit: kaspter]
hlauer has quit [Ping timeout: 240 seconds]
tnovotny has quit [Quit: Leaving]
<jernej>
apritzel: I can't get second emac working
<apritzel>
do you request the PortA pins, function 2?
lucas_ has quit [Ping timeout: 265 seconds]
<jernej>
of course
<jernej>
I checked with reg dump
<jernej>
issue is that PHY can't be found
<jernej>
but I don't think it's power issue
<jernej>
there is another possibility
<jernej>
AC200 seems to really exist on port A
<jernej>
at least on H6, integrated EPHY is on AC200 die
<apritzel>
in the H6 AC200 there was one I2C register bit to enable the PHY
<apritzel>
yes
<apritzel>
I just wasn't able to talk to that I2C device on the H6
<jernej>
why not? I wrote a driver for it and it worked ok
<jernej>
the only issue why it's not upstream is due to DT issue
damex has quit [Ping timeout: 240 seconds]
<apritzel>
dunno, didn't investigate much, just tried some I2C debug tool
<jernej>
hm... I can go other way around - disable pwm5 on Android and see if ethernet still works
<apritzel>
right, there was this weird thing where they (ab)used a PWM pin for a clock?
<jernej>
yes
<apritzel>
the PWM module on the H616 is not the same as on the H6, it's a newer one (more channels) as used in the R40
<apritzel>
there was some attempt for a driver a while ago, which I commented on
<apritzel>
newer one means very different from the A13 version
damex has joined #linux-sunxi
<jernej>
apritzel: As soon as I changed PA10-PA12 to 7, I got "sunxi-gmac gmac1 eth0: Link is Down"
mru has quit [Ping timeout: 246 seconds]
<jernej>
so yes, AC200 is there with EPHY and it will be pain to support it
damex has quit [Ping timeout: 265 seconds]
<jernej>
apritzel: and yes, syscon location 0x34 is also important
<jernej>
writing 0 there also breaks ethernet
mru has joined #linux-sunxi
damex has joined #linux-sunxi
luke-jr has quit [Read error: Connection reset by peer]
luke-jr has joined #linux-sunxi
afaerber has quit [Ping timeout: 240 seconds]
<Ashleee>
oh I just remembred, that rgmii-txid fix for Pine boards, it needs to be adedd for SOpine as well (at least in my case). Where do I send the patch? :) I'm used only to github PRs but linux kernel uses mailing lists?
<apritzel>
jernej: so did I get this correctly: there is the EMAC, but also the PWM and I2C pins on PortA that we need to talk to the AC200? So they should be in the pinctrl driver?
<apritzel>
I have added just the EMAC pins (as in the H6) so far
<jernej>
apritzel: that's correct
<apritzel>
and the AC200 also handles the audio then, so we need the audio pinmux as well?
<jernej>
you mean on port A?
<apritzel>
yes
<apritzel>
the BSP mention them
<jernej>
I don't think so, because most of the pins are occupied with rmii
<apritzel>
right ...
<apritzel>
so it's 10 pins for Ethernet, two for I2C3 and the PWM5
<jernej>
correct
<apritzel>
thanks, makes sense, I will add this
<apritzel>
jernej: nice testing method, btw (taking away pins/clearing registers and seeing it stop working)
<apritzel>
so what was the problem with the AC200 support, exactly? I dimly remember seeing your patches back then. Uncertainty about how it should be described in the DT?
<jernej>
yeah, driver loading sequence
<jernej>
ac200 should probe before ethernet, but I'm not sure how to describe that
<jernej>
technically, ac200-ephy subnode belongs in mdio and in ac200 mfd nodes at the same time...
<apritzel>
do we need the AC200 for something else than the PHY?
<jernej>
on H6 it is also audio codec
<apritzel>
IIRC all we need is to set the PHY enable bit in the I2C registers?
<jernej>
you can also set led polarity there
<jernej>
for H6
<jernej>
and there is master enable bit for all AC200 subdevices
<apritzel>
we could just set that in firmware (U-Boot) ...
<jernej>
no...
<jernej>
it works already, only issue is how to properly represent everything in DT
<jernej>
and that won't go away
<apritzel>
yeah, and I meant just not exposing all those details in the DT in the first place
<jernej>
what about analog audio on H6?
Mangy_Dog has quit [Ping timeout: 240 seconds]
<jernej>
you can't just hide everything...
<apritzel>
but anyway, the PWM pin is providing the clock for what? The AC200 or the PHY?
<jernej>
oh, also CVBS is on AC200 - for H6
eduardas has quit [Quit: Konversation terminated!]
<jernej>
PWM is master clock source for AC200, it can be used also for TVE
<jernej>
and even audio, I think
<jernej>
but I'm not sure if I2C interface needs by itself
night199uk has quit [Ping timeout: 272 seconds]
<apritzel>
jernej: and if you model this PHY enable bit as an I2C controlled regulator?
<jernej>
as I said, it's not just one bit
night199uk has joined #linux-sunxi
<jernej>
and imo it would be hack and you know how such things are accepted upstream
<jernej>
hm... for some reason I can't probe AC200 from U-Boot on H616
<jernej>
either it's not powered on or something else is missing
<apritzel>
did you set up the PWM?
<apritzel>
there is one PWM pin on the headers, so you could test this there (it's another channel, but just to verify that the PWM works)
<apritzel>
(I mean on the Orange Pi Zero2)
<apritzel>
jernej: oh, btw: there is an EMAC clock gate in the CCU
<apritzel>
actually two, one bus clock gate, and one EMAC_25M clock gate
<apritzel>
+0x970 and +0x97c
<jernej>
I dumped pwm regs and programmed them in the same way
<asdf28>
is it somehow possible to see which /dev/ttyS device corresponds to a serial entry in the device tree file?
<jernej>
yes, of course, gates are enabled
<jernej>
afaik EMAC_25M can be used for external phys on emac0 for clock source instead of external crystal
<apritzel>
asdf28: /sys/class/tty/ttyS*/device
lurchi__ is now known as lurchi_
<apritzel>
asdf28: and you can enforce a mapping through the /aliases node in the DT
<asdf28>
thanks apritzel
<asdf28>
would this lead to something like "serial@c81004c0"?
<asdf28>
i found "/sys/class/tty/ttyAML1/device/of_node/name" but it only outputs "serial"
<apritzel>
it points to the (platform) device name, which is derived from the DT node, so it's "1c28400.serial", for instance
<apritzel>
asdf28: traitor! That's the Amlogic UART ;-)
<asdf28>
yes i meant to ask this in #linux-amlogic
<apritzel>
no need, it's a Linux thing, so it's the same
<asdf28>
haha... sorry
<asdf28>
what i'm trying to do is, the bluetooth module is connected to /dev/ttyS1 in the android kernel, and i'm trying to figure out the equivalent in the mainline kernel
<asdf28>
but i could not find it yet in /sys/class/tty/ttyS*/device
<apritzel>
this /sys/class/tty/ttyAML1/device is a symlink, the target is the device name
<apritzel>
asdf28: can you access the DT? maybe it's there in the /aliases node
<jernej>
apritzel: I made mistake, apparently there is no I2C3 on port A
<jernej>
it's not configured in pinctrl
gaston1980 has quit [Quit: Konversation terminated!]
<jernej>
but PWM5 is definitively important for EPHY
<jernej>
albeit, there are some signs that i2c3 is used... this is very confusing...
<jernej>
ah, probably it's just in suspend
Mangy_Dog has joined #linux-sunxi
hlauer has quit [Ping timeout: 256 seconds]
cmeerw has quit [Ping timeout: 268 seconds]
jstein has joined #linux-sunxi
dev1990 has quit [Excess Flood]
dev1990 has joined #linux-sunxi
hexdump0815 has joined #linux-sunxi
Asara has quit [Read error: Connection reset by peer]
Asara_ has joined #linux-sunxi
<hexdump0815>
jernej: apritzel: i think i cannot be of a great help here, but maybe its useful anyway - a few weeks back i tried to get the bsp kernel working on a h313 tv box which i think has the other eth phy as well and some things you mentioned sounded familiar
<jernej>
I think I found root cause - pull ups weren't enabled on i2c3 port
<hexdump0815>
jernej: apritzel: for instance i had to enable the ac200 node on the tv box which is disabled in the opi2 dts to get ethernet working and some other things were different too
<jernej>
interestingly, in BSP DT they are not, but if you decompile DT from box, they are
<hexdump0815>
also the box dts had some extra ""PI15", "PI16"" on the gmac0 and gmac1 nodes which i think were reuired ... but as said, i'm not fully understanding what i'm saying here :)
<hexdump0815>
with the above mentioned x96q dts i had ethernet finally working on the box
swiftgeek has quit [Ping timeout: 246 seconds]
swiftgeek has joined #linux-sunxi
<apritzel>
hexdump0815: PI15 is MDIO, which is always required
<apritzel>
PI16 is EPHY_25M, as jernej mentioned this might be the clock for an *external* *100Mbit* PHY
<apritzel>
but I doubt they use this, it might only make some sense for really cheap dual 100Mbit Ethernet, as you could save the crystal for the (external) PHY
jstein has quit [Quit: quit]
<jernej>
apritzel: why only fast ethernet? OPi also uses 25 MHz crystal for Realtek gigabit PHY
<apritzel>
don't you need 125 MHz for GBit?
<jernej>
sure, but there are PLLs...
<jernej>
apritzel: I can now probe AC200 over I2C in U-Boot
<jernej>
however, it's very sensitive and can lock up quickly
<jernej>
only power cycle helps then
<apritzel>
jernej: very nice! so you need PWM (24 MHz bypass?) and I2C3 with pull ups?
<jernej>
but, communicating doesn't make any sense
<hexdump0815>
oh yes - my fault - it was only PI16 ... that one is there in the box dts and not in the opi dts - and there is a use_ephy25m flag in the dts which is set to 0 in the opi dts and not required for the box
<jernej>
I can only read register 0, which is ID
<jernej>
everything else is sometimes all 0, sometimes all ff
<jernej>
and write does nothing
<apritzel>
jernej: what's the bus frequency?
<apritzel>
I wouldn't expect this to be a problem for a device such close, but anyway ...
<jernej>
DT has it set to 200k, I set it to 100k
<jernej>
datasheet says 200k max.
hexdump0815 has quit [Remote host closed the connection]
<jernej>
it's a bit annoying, because it's paged access
<jernej>
and reg width is 8, while data width is 16 bit
<apritzel>
so can you enable the PHY? then claim it's just there in the DT and boot Linux with it? Just for checking ...