<macromorgan>
you'll need to use something like interception to join the adc joystick with the gpio buttons so that retroarch or anything can see it as a single unified gamepad
<macromorgan>
otherwise, nothing weird in userspace at all... in fact since it's using mainline everything from Mesa to U-Boot it should be less weird than stock OS and the current crop of custom firmware
<macromorgan>
I'm in constant discussions with the JelOS folks about dropping all the BSP crap and going to mainline on all these devices
<macromorgan>
6.4 is possibly the earliest that can be done, more likely 6.5 since we're still waiting on wifi to be mainlined and the v2 panel (which all new Anbernic RGxx3 devices likely have) driver is still pending in drm's mailbox
<macromorgan>
mainline Linux also gets the benefit of the touchscreen, which is something custom firmware Linux doesn't have :-)
<macromorgan>
the one last bug I've been slowly building up the willpower to attack is with the DSI subsystem. Basically when you turn off the DSI controller (either because you're going down for a suspend or shutting down the system) it tries to deactivate the display panel twice. This causes an unbalanced regulator disable error
<macromorgan>
it's harmless as best I can tell, but still a bug.
<macromorgan>
The bug is somewhere in the combination of rockchip_dw_mipi_dsi.c and dw_mipi_dsi.c drivers. Upon deactivation they both try to deactivate any associated panels, however only one of these two should be doing that...
<macromorgan>
but any questions on the anbernic stuff, feel free to ping me or macc24 as we've been leading the mainlining efforts
lurchi__ has quit [Ping timeout: 240 seconds]
lurchi__ has joined #linux-rockchip
kevery has joined #linux-rockchip
rajkosto has joined #linux-rockchip
<amstan>
macromorgan: thanks for the details, sounds pretty good!
kevery1 has joined #linux-rockchip
<amstan>
heh, i'm not surprised macc24 is involved too
<dlg>
it's good enough to boot and root off emmc or an sdcard at least
rajkosto has quit [Read error: Connection reset by peer]
<naoki>
ah I forgot to put dtb...where? ;)
<dlg>
ive been using the one inside u-boot and it's been good enough
<dlg>
havent had to use a separate dtb
<naoki>
hmm... I'll try -current snapshot
Tenkawa has joined #linux-rockchip
<naoki>
rge0 at pci1 dev 0 function 0 "Realtek RTL8125" rev 0x04: msi, address 00:e0:4c:02:00:6f
<naoki>
I can see only 1 rge on -current ;)
<dlg>
on what board?
<naoki>
e25
<dlg>
can you send me the dmesg?
<Tenkawa>
naoki: the Radxa E25 right?
<naoki>
right
<Tenkawa>
Three of us have been working with that unit and have run into all kinds of random "glitches"
<Tenkawa>
Ours have mostly been rooted in u-boot being the cause we think
<Tenkawa>
The devicetree timings were off on a few things
<Tenkawa>
I'd have to ask what the other fixes were
<Tenkawa>
Only one of us has the physical board so its a remote effort trying to fix 1 board
<Tenkawa>
We are in the process of trying to get two more of the units
<dlg>
once i got the code right on the e25 it's been pretty robust for me
<Tenkawa>
dlg: I need to ask them what the one last nagging bug is they are working on right now (I'm mostly a theory person on this project)
<Tenkawa>
I will ask once they are online today
<dlg>
npz
<dlg>
im off for the night
<dlg>
the dts im using is in the uboot branch above
kevery has joined #linux-rockchip
<dlg>
and i think all the openbsd changes that the e25 uses are committed now
<dlg>
which might be useful for comparing to
<Tenkawa>
oh... that could be good to compare with.. yeah thanks
kevery has quit [Quit: kevery]
rajkosto has joined #linux-rockchip
<Tenkawa>
dlg: I just found out it is one of the ethernet ports is not working is the problem we are running into currently... the sd slot can be a bit flaky but we have put in some dts changes to help that
<naoki>
dlg: it seems some regulator(s) need to be on at u-boot. I modified defconfig, then both rge are attached
<Tenkawa>
naoki: rge9 is the ethernet isnt it?
<Tenkawa>
er rge0
<Tenkawa>
Yeah you are having similar hiccups in BSD it would seem
<naoki>
rge0 and rge1 on e25 are rtl8125b
<naoki>
regulator is not enabled in defconfig in my u-boot tree
<Tenkawa>
I wonder if the regulator is causing our problem on the linux side
<Tenkawa>
in a similar way
<naoki>
I guess it's not "similar" thing
<Tenkawa>
is the Realtek RTL8125 not the onboard NIC on that board?
<Tenkawa>
that's what you output above
<Tenkawa>
naoki: rge0 at pci1 dev 0 function 0 "Realtek RTL8125" rev 0x04: msi, address 00:e0:4c:02:00:6f
<naoki>
it's on-board NIC on E25
<Tenkawa>
yes.. I'm talking about the E25
<naoki>
I'm talking about OpenBSD and u-boot defconfig for E25
<Tenkawa>
Yes... we are seeing an "exact" same problem in Linux
<Tenkawa>
with its kernel and u-boot
<naoki>
regulators with "regulator-boot-on" should be enabled at u-boot
<naoki>
what is your problem?
<naoki>
both NIC are working on Linux for me
Stat_headcrabed has joined #linux-rockchip
<Tenkawa>
what kernel?
<naoki>
now vendor 4.19 kernel is running, and it works with mainline kernel when I tried some time ago
<naoki>
(I wrote dts for E25 from scratch)
<Tenkawa>
We are trying "not" to use the vendor and it currently doesn't work with 6.1-6.2
<Tenkawa>
only one port shows up
<Tenkawa>
We're farly certain its just dts
<Tenkawa>
er fairly
<CounterPillow>
RK356x PCIe ranges on mainline currently have some problems which could be causing an issue
ldevulder has quit [Quit: Leaving]
<Tenkawa>
CounterPillow: I wondered about that too... my Odroid-M1 had that problem
<Tenkawa>
I think we wont see the person with the unit until this evening so I won't know if it worked until tomorrow.. however thank you in advance either way.
<Tenkawa>
I think they are about 12 hours timezone diff
<naoki>
hmm
<naoki>
it's late night here
<Tenkawa>
Oh.. they are online..
<naoki>
I cannot try myself
<naoki>
for now
<Tenkawa>
I was wrong.. its our other person who has one
<Tenkawa>
we are prepping now
<naoki>
that typo exists in mainline, so you should apply that fix on 6.2/6.3 too
<naoki>
your 6.2/6.3 tree
<Tenkawa>
Ok. relayed
<Tenkawa>
We have our own Discord server for all of this...
<Tenkawa>
heheh
<Tenkawa>
and we use #arm-img-builder on Libera
<Tenkawa>
This is for a custom SBC builder project.
<Tenkawa>
It "has" been just for ARM/ARM64.... I'm hoping eventually to add RISC-V to the mix.
Stat_headcrabed has quit [Ping timeout: 240 seconds]
Kwiboo has quit [Quit: .]
Kwiboo has joined #linux-rockchip
Rathann has joined #linux-rockchip
Guest9425 has quit [Ping timeout: 248 seconds]
<Kwiboo>
CounterPillow: regarding the issue with pcie ranges, do you have more details?
<Kwiboo>
using the config reg as cgf_base like linux driver make it work, I think bsp use cfg_base in the <4gb 32mb area addr instead of the 4gb+ 1gb area addr
<diederik>
Whether that's the latest version is (still) not entirely clear to me
<Kwiboo>
diederik: thanks, looks like similar values I picked independent of that when I first tried to get nvme working in u-boot
<Kwiboo>
looks like the io range may be 1mib to large
<Kwiboo>
sorry, the smaller mem area
<diederik>
I _think_ that 'neggles' has a more recent version, but if that is pushed somewhere is not clear to me.
<Kwiboo>
initial 1mib for config, next 1mib for io and remaining 30 mib for mem
<diederik>
CP could possibly clarify it as he better understands this material then I do (it was a discussion that took place on #quartz64 on pine64's IRC)
<Kwiboo>
great, at least it does not look like such dts change would affect anything for u-boot patches, will continue :-)
vagrantc has joined #linux-rockchip
Net147 has joined #linux-rockchip
Net147 has quit [Changing host]
Net147 has joined #linux-rockchip
stikonas has joined #linux-rockchip
vagrantc has quit [Quit: leaving]
rajkosto has quit [Read error: Connection reset by peer]
lurchi__ has quit [Read error: Connection reset by peer]