00:11
Elpaulo has quit [Quit: Elpaulo]
01:10
angelsl has quit [Read error: Connection reset by peer]
01:11
angelsl_ has joined #linux-amlogic
01:11
angelsl_ is now known as angelsl
01:29
vagrantc has joined #linux-amlogic
02:38
vagrantc has quit [Quit: leaving]
04:50
Tonymac32 has quit [Ping timeout: 256 seconds]
05:04
<
Necrosporus >
chewitt, but mine is 8726M not MX, is it important?
05:04
<
chewitt >
same thing
05:04
<
chewitt >
unless it's even older
05:05
<
chewitt >
you called it Meson 3 .. I understand 8726MX as Meson 6
05:06
<
chewitt >
either way, this is 10+ years old and comprehensively outperformed by a Raspberry Pi 2 :)
05:08
Tonymac32 has joined #linux-amlogic
05:12
Tony_mac32 has joined #linux-amlogic
05:15
Tonymac32 has quit [Ping timeout: 256 seconds]
05:16
<
chewitt >
only 9 years old
05:16
<
chewitt >
and the MX variant comes later, only 8 years old
05:16
Tonymac_32 has joined #linux-amlogic
05:16
<
chewitt >
(but both still outperformed by a Raspberry Pi Zero)
05:18
<
Tonymac_32 >
chewitt which is? I missed it
05:19
<
Tonymac_32 >
it's a dual core though isn't it?
05:20
Tony_mac32 has quit [Ping timeout: 256 seconds]
05:20
<
chewitt >
nope, looks like 8726MX is the dual version
05:20
<
chewitt >
8726M is single
05:20
<
chewitt >
M and M3 are single
05:21
<
Tonymac_32 >
but a cortex-anything vs an arm11 shouldn't be a competition
05:22
<
chewitt >
the convo originally started at software
05:23
<
chewitt >
so pi wins hands down
05:23
<
Tonymac_32 >
ah well in that case. ;)
05:24
<
Necrosporus >
AMLOGIC MESON3 8726M SKT SH
05:24
<
Necrosporus >
Whatever it means
05:25
<
Necrosporus >
So I should look for 8726MX kernel since it's the same thing but dual core
05:26
<
Necrosporus >
chewitt, maybe it is, but there is an important point. The Amlogic thing I do already have but I do not have RPi0 or whatever
05:27
<
chewitt >
If you are retired or unemployed or just really determined (aka bloody minded) .. it's a nice project to get it working
05:27
<
chewitt >
If "my time is money" or anything vaguely like that factors into your life, bin it and get a Pi0
05:28
<
chewitt >
or a similarly cheap (but much newer) ARM thing
05:28
* Tonymac_32
has a Mini2440, will stay quiet
05:29
<
chewitt >
Necrosporus what do you want to do with it?
05:30
<
Necrosporus >
I want to turn it into a simple game console so it can run stuff like SuperTux on TV
05:30
<
Necrosporus >
or other native open source GNU/Linux games
05:31
<
Tonymac_32 >
my much bigger boards don't hold up well against games, I don't think that one will pull it off
05:31
<
chewitt >
mainline Linux support for Meson 9 is good, Meson 8 is increasingly good - via @xdarklight sources
05:31
<
Necrosporus >
Tonymac_32, even if that's 2D games?
05:31
<
Necrosporus >
Like logic puzzle
05:31
<
chewitt >
Meson 6 is bare-bones, not usable for anything beyond basic board boot
05:32
<
chewitt >
Meson 3 .. non existent
05:32
<
chewitt >
so you'd be dependent on BSP kernels
05:32
<
Necrosporus >
Another idea, maybe it can be used to watch youtube?
05:33
<
Tonymac_32 >
ok 2D sure, I misread and thought you put tuxcart ;)
05:33
<
chewitt >
(and u-boot)
05:33
<
Necrosporus >
Well, I can use u-boot burned onto the chip
05:33
<
Necrosporus >
But default android distro is terrible
05:35
<
Necrosporus >
chewitt, alright, would Orange Pi 4 be a good investment? Or RPi 4b 8G?
05:35
<
chewitt >
these days I'm not aware of anyone anywhere working on anything below Meson 8
05:35
<
Necrosporus >
Or which is the strongest ARM board in existance with good mainline support?
05:35
<
chewitt >
for the gaming use-case Pi hardware is a no brainer because it's so well supported by all the gaming distos
05:36
<
Tonymac_32 >
I use an XU4
05:36
<
chewitt >
pi hardware is never the best for raw performance, but the software life is simple
05:37
<
Tonymac_32 >
but for native games, an S912 has a Midgard GPU correct?
05:37
<
chewitt >
XU4 is old hardware these days
05:37
<
Tonymac_32 >
lol and yet, software ;)
05:38
<
Necrosporus >
Well, I guess I can try to compile something old for this board real quick and see if it works like the link to 3.10.y tree
05:38
<
chewitt >
Linux XU4 5.10.0-rc4 #1 SMP PREEMPT Sat Nov 21 08:57:09 UTC 2020 armv7l GNU/Linux
05:38
<
chewitt >
still can't get panfrost to run tho :(
05:39
<
chewitt >
kmscube works, but not Kodi
05:39
<
Tonymac_32 >
my retro gaming setup is perfect, can't remember the kernel though
05:39
<
Tonymac_32 >
the rest are headless
05:39
<
Tonymac_32 >
my RK hardware panfrost is no problem on
05:40
<
chewitt >
4.14 or 5.4 prob.
05:40
<
chewitt >
most of memeka's hacks aren't valid in the kernel these days either
05:40
<
chewitt >
so media support is rather regressed, while lots of other bugs are resolved
05:41
<
chewitt >
(sounds a bit like amlogic)
05:41
<
Tonymac_32 >
have you had luck on the S912? I never got my hands on one of those.
05:41
<
chewitt >
I have a few of them
05:41
<
Tonymac_32 >
I only have a Vim1
05:42
<
chewitt >
panfrost runs well enough that some of the gaming distros are now taking interest
05:42
<
chewitt >
(also because they track my kernels a bit)
05:42
<
chewitt >
but I suspect the issue will be mali blobs (as always)
05:43
<
chewitt >
I think they're still using the libhybris workaround
05:43
<
chewitt >
and Android blobs
05:45
<
Necrosporus >
Are there good boards which work with no blobs?
05:45
<
Necrosporus >
At least no blobs in kernel space
05:45
<
Necrosporus >
Like wi-fi firmware doesn't count
05:46
<
chewitt >
mali hardware is now very well supported by mesa (lima/panfrost)
05:47
<
chewitt >
but retro games push the boundaries a litttle
05:47
<
Tonymac_32 >
just a wee
05:47
<
Tonymac_32 >
native games though shouldn'
05:47
<
Tonymac_32 >
t be too bad
05:48
<
chewitt >
I suspect half the issue with retro games is the code quality of the emulators
05:48
<
Tonymac_32 >
I played some minetest on my Tinker board actually just to see if it would work
05:48
<
Tonymac_32 >
that's definitely it, especially since the hardware being emulated often had some wonky ways of doing graphics
05:48
<
chewitt >
the fragmentation in the retro-gaming ecosystem makes Android look amazingly conformant
05:49
<
chewitt >
the wheel is frequently reinvented .. it's a PITA from a maintainer perspective
05:50
<
chewitt >
for retro gaming i'd get a pi simply because i'd rather play games than trace driver weirdness to make things work
05:51
<
chewitt >
pi3/pi4 will be great for that use-case
05:51
<
Tonymac_32 >
performance wise I'd say Pi 3 < XU4 < Pi 4
05:51
<
Tonymac_32 >
I was horribly disappointed in the Pi3
05:52
<
Tonymac_32 >
and can only hope the Pi4 was worth the e-waste
05:52
<
chewitt >
it's a compromise like all pi designs
05:52
<
chewitt >
but it moves the needle in a forwards direction
05:52
<
Tonymac_32 >
I use a Pi for my music player with DAC and one for a LOAR gateway
05:53
<
chewitt >
and since it's a very different SoC to the previous ones, they're taking upstreaming seriously this time
05:53
<
Tonymac_32 >
Is it still like a goofy fiddler crab with a GPU leading the ARM Cores around?
05:54
<
Tonymac_32 >
The engineering catastrophe of the 3 kind of poisoned my opinion forever I'm afraid
05:55
<
chewitt >
it evolved
05:55
<
chewitt >
but still a compromise
05:56
<
chewitt >
at the end of the day its primary purpose is teaching kids to code, not be the best device in every category
05:56
<
Tonymac_32 >
well, maybe tell the foundation that, they seem confused
05:56
<
Tonymac_32 >
It is good for that purpose
05:57
<
chewitt >
they can be a bit schizophrenic sometimes
05:58
<
chewitt >
I do enjoy chats with them from time to time :)
06:05
angelsl has quit [Read error: Connection reset by peer]
06:06
angelsl has joined #linux-amlogic
06:19
buzzmarshall has quit [Remote host closed the connection]
06:25
angelsl has quit [Read error: Connection reset by peer]
06:26
angelsl has joined #linux-amlogic
06:33
chewitt has quit [Read error: Connection reset by peer]
06:33
chewitt_ has joined #linux-amlogic
06:40
angelsl has quit [Read error: Connection reset by peer]
06:41
angelsl_ has joined #linux-amlogic
06:41
angelsl_ is now known as angelsl
06:50
angelsl has quit [Read error: Connection reset by peer]
06:51
angelsl has joined #linux-amlogic
07:05
angelsl has quit [Read error: Connection reset by peer]
07:06
angelsl has joined #linux-amlogic
07:45
cmeerw has joined #linux-amlogic
08:53
asdf28 has joined #linux-amlogic
09:57
nexysno has joined #linux-amlogic
09:58
<
asdf28 >
i have compiled the kernel and loaded the meson drm driver, but there is no /dev/dri/card0, also nothing in dmesg
09:58
<
asdf28 >
has anyone made this work?
11:46
asdf28 has quit [Ping timeout: 260 seconds]
11:47
asdf28 has joined #linux-amlogic
11:57
sputnik_ has quit [Ping timeout: 265 seconds]
12:16
<
narmstrong >
asdf28: why would you use this version ? It’s the mainline 4.10 kernel driver I wrote back ported to their 3.14 tree
12:23
_whitelogger has joined #linux-amlogic
12:26
nexysno has quit [Remote host closed the connection]
12:27
nexysno has joined #linux-amlogic
12:29
<
asdf28 >
i just found it interesting and wanted to see if it works
12:30
<
asdf28 >
by the way, i finally got the mali framebuffer driver to work! but not on an amlogic box, it was on an allwinner board
12:30
<
asdf28 >
i used the same kernel and mali driver version on both devices
12:31
<
asdf28 >
without leak_smem, both boards showed a black screen when running a graphical app
12:32
<
asdf28 >
then i enabled it and the allwinner board worked, but the amlogic box had a segmentation fault
12:32
<
asdf28 >
so at least this seems to be the way to go, even if it didn't work out yet
12:33
<
asdf28 >
is the DRM driver supposed to create a /dev/dri device? or was that functionality not implemented yet?
12:34
<
asdf28 >
there were some lines commented out, but when i enabled them, i realized that they were probably meant for a newer kernel version
14:03
<
chewitt_ >
^ was working on 5.6/5.7 afaik .. not sure it's been tested since as LE switched to mesa/lima
14:04
chewitt_ is now known as chewitt
14:10
angelsl has quit [Read error: Connection reset by peer]
14:11
angelsl has joined #linux-amlogic
14:30
angelsl has quit [Read error: Connection reset by peer]
14:30
angelsl has joined #linux-amlogic
14:52
<
asdf28 >
chewitt thank you. do you mean it works with the mali fbdev blob? because i thought libreelec used the wayland/kms one
14:53
<
chewitt >
not tested with fbdev
14:53
<
chewitt >
but afaik we didn't disable anything in the driver to make it not-work with others
14:54
<
chewitt >
but mali is all black magic so who knows :)
15:00
<
asdf28 >
thank you, will try that. i think i haven't tried that version yet
15:02
<
chewitt >
what board are you trying to use? .. and for which use-case?
15:11
<
asdf28 >
an s905x box
15:12
<
asdf28 >
the use case is, i wanted to help to update the box to a newer kernel in lakka
15:12
<
asdf28 >
but i cannot find a working graphics driver
15:13
<
asdf28 >
i mean, they all work but not smoothly
15:21
nexysno has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
15:45
angelsl has quit [Read error: Connection reset by peer]
15:45
angelsl has joined #linux-amlogic
15:58
<
chewitt >
how far off is mesa/lima?
15:59
<
chewitt >
I ask because, at least with lima there are developers you can report issues to
15:59
<
chewitt >
with blobs it's all on you
16:00
<
chewitt >
and is it "everything" slow or just specific emulators?
16:02
<
asdf28 >
well the biggest issue right now is some kind of memory leak
16:02
<
asdf28 >
that's probably not even related to the graphics driver
16:04
<
asdf28 >
lima does not work smoothly on my allwinner device, it has a little hiccup every second, no matter what's running, it's jerky
16:04
<
asdf28 >
it works better on the amlogic box but it crashes from time to time and 3d games will mostly not work
16:05
<
asdf28 >
i'm trying to find out if any of that is even related to the graphics driver, i couldn't find out yet
16:05
<
asdf28 >
but none of this did happen with the legacy allwinner/amlogic kernels
16:07
<
asdf28 >
the mali wayland blob worked best so far but i get memory leak errors and i can see the memory filling up in task manager
16:07
nexysno has joined #linux-amlogic
16:07
<
asdf28 >
but that's probably something else as it happens with both mesa and mali blobs
16:10
<
asdf28 >
i'll try to install glmark to see if it happens there too
17:22
asdf28 has quit [Ping timeout: 240 seconds]
17:35
angelsl has quit [Read error: Connection reset by peer]
17:36
angelsl has joined #linux-amlogic
17:50
angelsl has quit [Read error: Connection reset by peer]
17:50
angelsl_ has joined #linux-amlogic
17:51
angelsl_ is now known as angelsl
18:20
angelsl has quit [Read error: Connection reset by peer]
18:20
angelsl has joined #linux-amlogic
18:23
asdf28 has joined #linux-amlogic
19:34
ldevulder_ has joined #linux-amlogic
19:38
ldevulder__ has quit [Ping timeout: 264 seconds]
19:55
angelsl has quit [Read error: Connection reset by peer]
19:56
angelsl_ has joined #linux-amlogic
19:56
angelsl_ is now known as angelsl
20:10
angelsl has quit [Read error: Connection reset by peer]
20:10
angelsl has joined #linux-amlogic
20:11
TheAssassin has quit [Ping timeout: 240 seconds]
20:16
TheAssassin has joined #linux-amlogic
20:22
angelsl has quit [*.net *.split]
20:22
chewitt has quit [*.net *.split]
20:22
AUser has quit [*.net *.split]
20:22
ndufresne has quit [*.net *.split]
20:25
angelsl has joined #linux-amlogic
20:25
chewitt has joined #linux-amlogic
20:25
AUser has joined #linux-amlogic
20:25
ndufresne has joined #linux-amlogic
21:00
sputnik_ has joined #linux-amlogic
21:31
yann has joined #linux-amlogic
21:46
vagrantc has joined #linux-amlogic
22:05
angelsl has quit [Read error: Connection reset by peer]
22:05
angelsl has joined #linux-amlogic
22:25
angelsl has quit [Read error: Connection reset by peer]
22:26
angelsl_ has joined #linux-amlogic
22:26
angelsl_ is now known as angelsl
22:29
asdf28 has quit [Ping timeout: 240 seconds]
23:35
angelsl has quit [Read error: Connection reset by peer]
23:35
angelsl has joined #linux-amlogic