Does GameShell Actually Use FBTurbo?

[r043v]

from my old history file :

libump deps of fbturbo

git clone --depth 1 https://github.com/linux-sunxi/libump.git
cd libump/
autoreconf -i
./configure --prefix=/usr
make -j4
sudo make install

fbturbo

git clone --depth 1 https://github.com/ssvb/xf86-video-fbturbo.git
cd xf86-video-fbturbo/
./autogen.sh --prefix=/usr
make -j4
sudo make install

and if need, fbdev :

git clone --depth 1 https://gitlab.freedesktop.org/xorg/driver/xf86-video-fbdev.git
cd xf86-video-fbdev/
./autogen.sh --prefix=/usr
make -j4
sudo make install

[Joao_Manoel]

This is very useful and valuable information @r043v.

I think we could make a wiki page with this procedures and how to compile from source some important libraries like this one that you just posted.

[adcockm]

Interesting and thanks, @Joao_Manoel and @r043v! I’ll definitely try this soon (probably tomorrow). I didn’t really have much time tonight to spend on this. I removed the card from my Gameshell for now, but I’ll see what we can learn from it.

I put a different card in, with a fresh install of deot_V2+ 200128.img.bz2. Haven’t done much with it, but it works. (I swapped my physical A/B, X/Y buttons though since the default mappings were different. I’d been using the XBOX defaults until now because they felt right to me.)

It seems that something good might be coming from my troubles after all, so I’m glad about that. I’ll share what I find and then hopefully we’ll have a better picture of what needs to be done to support proper upgrading. Thanks!

[guu]

@slock

sudo usermod -aG render cpi

Does this work for you?

[javelinface]

From what I’ve gathered, hopefully we can get buster working correctly with the correct graphic drivers, doing an apt-get full-upgrade, and then running the scripts mentioned here by @r043v

Does GameShell Actually Use FBTurbo?

from my old history file : libump deps of fbturbo git clone --depth 1 https://github.com/linux-sunxi/libump.git cd libump/ autoreconf -i ./configure --prefix=/usr make -j4 sudo make install fbturbo git clone --depth 1 https://github.com/ssvb/xf86-video-fbturbo.git cd xf86-video-fbturbo/ ./autogen.sh --prefix=/usr make -j4 sudo make install and if need, fbdev : git clone --depth 1 https://gitlab.freedesktop.org/xorg/driver/xf86-video-fbdev.git cd xf86-video-fbdev/ ./autogen.sh --prefix…

And then running

@slock

sudo usermod -aG render cpi

I’ll give it a shot. @adcockm as @Joao_Manoel, did you end up having any luck?

Thinking back to this post, perhaps having a kernel that doesn’t have the -/_ change, and a community image that has the file not requiring the symbolic link or file name change could be a way forward. The community image was the way forward in the past. Naturally, I would include this with my own DEOT image.

It may fragment releases, having a kernel only applicable to a fork, but as it stands now, there has been a lot of kernel development that isn’t included with the current official image anyway.

Edit: Looks like it works, at least functionally. Here’s my dmesg | grep output.

cpi@clockworkpi:~$ dmesg | grep lima
[    0.940178] lima 1c40000.gpu: IRQ ppmmu2 not found
[    0.944971] lima 1c40000.gpu: IRQ ppmmu3 not found
[    0.949795] lima 1c40000.gpu: gp - mali400 version major 1 minor 1
[    0.956009] lima 1c40000.gpu: pp0 - mali400 version major 1 minor 1
[    0.962331] lima 1c40000.gpu: pp1 - mali400 version major 1 minor 1
[    0.968629] lima 1c40000.gpu: IRQ pp2 not found
[    0.973156] lima 1c40000.gpu: IRQ pp3 not found
[    0.977685] lima 1c40000.gpu: l2 cache 64K, 4-way, 64byte cache line, 64bit external bus
[    0.986195] lima 1c40000.gpu: bus rate = 200000000
[    0.991006] lima 1c40000.gpu: mod rate = 384000000
[    0.996073] [drm] Initialized lima 1.0.0 20190217 for 1c40000.gpu on minor 0
cpi@clockworkpi:~$ dmesg | grep sun4i
[    1.082607] sun4i-ss 1c15000.crypto-engine: Die ID 5
[    1.161971] sun4i-drm display-engine: bound 1e00000.display-frontend (ops 0xc0849db0)
[    1.170061] sun4i-drm display-engine: bound 1e60000.display-backend (ops 0xc08495f4)
[    1.177817] sun4i-drm display-engine: bound 1e70000.drc (ops 0xc0849124)
[    1.184906] sun4i-drm display-engine: bound 1c0c000.lcd-controller (ops 0xc084831c)
[    1.205116] fb0: switching to sun4i-drm-fb from simple
[    1.216506] [drm] Initialized sun4i-drm 1.0.0 20150629 for display-engine on minor 1
[    1.328705] [<c043bbe8>] (drm_fbdev_generic_setup) from [<c046fbb0>] (sun4i_drv_bind+0x130/0x160)
[    1.328719] [<c046fbb0>] (sun4i_drv_bind) from [<c0480fe4>] (try_to_bring_up_master+0x15c/0x198)
[    1.328742] [<c04810b4>] (__component_add) from [<c04712d8>] (sun4i_tcon_probe+0x70/0x8c)
[    1.328754] [<c04712d8>] (sun4i_tcon_probe) from [<c0488c80>] (platform_drv_probe+0x48/0x98)
[    1.818467] sun4i-drm display-engine: fb0: sun4i-drmdrmfb frame buffer device

[Joao_Manoel]

@javelinface

s the way forward in the past. Naturally, I would include this with my own DEOT image.

It may fragment releases, having a kernel only applicable to a fork, but as it stands now, there has be

I managed to upgrade, not to Buster, but to bullseye/sid

Apt full-upgrade | Getting the new software updates from Debian

Hello, I couldn’t work on the device for a few months, but yesterday I tried do finally upgrade to the new Debian. I tried it before and I remember that it didn’t go well. I got a lot of problems and I didn’t get the lima driver working even compiling everything myself. In the end, I gave up. So I invested a lot of time yesterday, and today to understand what was made to get the lima driver working with X11. I know that getting everything working was difficult but if only a fe…

The good news is that you shouldn’t need to compile anything yourself to get lima. Everything is in the repository. Didn’t try fbturbo, maybe for it we have to follow what @r043v suggested