scream has quit [Remote host closed the connection]
fire219 has quit [Read error: Connection reset by peer]
IgorPec has quit [Ping timeout: 244 seconds]
tkaiser has joined #linux-sunxi
<tkaiser>
Can anyone familiar with what happens now in Pine64 forum explain to me why the average users are so keen to play with 'Mali blobs'?
<tkaiser>
Especially when it's about video?
dearfibonacci has joined #linux-sunxi
IgorPec has joined #linux-sunxi
<plaes>
tkaiser: because everyone wants to try out Android?
<tkaiser>
plaes: No, it's about Linux and (most of the times) video. And users are talking about these blobs as if getting them would solve a problem that does not even exists (there is Cedrus accelerated video decoding on A64 since months)
<plaes>
oh.. so they think Mali stuff is for media decoding?
<tkaiser>
Seems so. And I wonder who told them or is responsible for the confusion :)
<Net147>
mripard: seems sun4i drm driver is missing calls to drm_panel_prepare/drm_panel_unprepare - http://pastebin.com/6GPYBqYb
<Net147>
mripard: so the GPIO listed in enable-gpios of panel in DT is never enabled/disabled
<plaes>
Net147: nice catch \o/
<mripard>
Net147: yep, nice catch
<mripard>
can you send a patch?
<Net147>
mripard: no, I can send one if you like
<Net147>
mripard: preparing patch...
<plaes>
mripard: do you have any info about the Saturday event in XDG2016?
<mripard>
plaes: I don't
<mripard>
you'll be there?
<plaes>
yeah..
<plaes>
buying ferry tickets
mzki has joined #linux-sunxi
<Net147>
mripard: patch sent
<wens>
wonder if there are prepare/unprepare functions for drm_bridge as well
<wens>
nope
IgorPec18 has joined #linux-sunxi
Andy-D has quit [Remote host closed the connection]
bfree has quit [Ping timeout: 240 seconds]
IgorPec18 has quit [Ping timeout: 252 seconds]
florianH has joined #linux-sunxi
ricardocrudo has joined #linux-sunxi
premoboss has joined #linux-sunxi
bfree has joined #linux-sunxi
Andy-D has joined #linux-sunxi
<longsleep>
tkaiser: i am trying to counter the "i want mali" phenomenon :)
<plaes>
though, the need for it might send a signal to ARM
<longsleep>
plaes: what need for it - people who want it can do it as the blobs are there (without EULA), and everyone else does not need it
<longsleep>
someone was telling people that having the mali blobs will solve everything - which of course we know here that it is not true - but the average pine64 dude has no clue ..
IgorPec has quit [Ping timeout: 265 seconds]
IgorPec18 has joined #linux-sunxi
dearfibonacci has quit [Ping timeout: 244 seconds]
dearfibonacci has joined #linux-sunxi
<rellla>
tkaiser: this mali/ve/g2d common-user-knowledge-errata is there for years already...
<plaes>
yeah, but there are lots of newcomers with Pine64
<Net147>
OTG port not working for me with 4.8-rc4. hmm....
<rellla>
tkaiser, plaes: It needs to be clarified at prominent place, the wiki. So someone simply can link to it... They most likely won't search for it themselves.
premoboss has quit [Ping timeout: 276 seconds]
hansg has joined #linux-sunxi
hansg has quit [Client Quit]
<Net147>
OTG port works with 4.7. looks like something broke between 4.7 and 4.8-rc4...
reinforce has joined #linux-sunxi
<tkaiser>
rellla: It seems this Mali BS is spread even by moderators there, so 'Mali will cure everything' became the truth in Pine64 land :)
<tkaiser>
And they have a new moderator that blocked my account already and recommending destroying CAT6 cables to workaround the GbE negotiation issue some boards show instead of using ethtool. Well, what happens over there is not so important anyway... but poor Pine64 users relying on this mess :)
<mripard>
Net147: please send a mail to hans, chen yu and I about hthis
iaglium has quit [Ping timeout: 240 seconds]
iaglium has joined #linux-sunxi
<rellla>
tkaiser: omg
<tkaiser>
rellla: Doesn't really matter :) This specific GbE issue longsleep and me tried to isolate months ago when other moderators always sabotaged all attempts to get a clue what's going on by recommending workarounds instead of asking users for details. Now another one from team 'Dunning-Kruger' joined the party and continues this 'work'...
<Net147>
mripard: done
IgorPec18 has quit [Ping timeout: 255 seconds]
<wens>
and hans replied
<wens>
that was quick
IgorPec has joined #linux-sunxi
enrico_ has joined #linux-sunxi
<Net147>
mripard: any update on mali-400 r6p0 blobs or whether mali-450 blobs will work on mali-400?
OverCR has joined #linux-sunxi
<Net147>
all good, OTG port working again with Han's fix
<tkaiser>
KotCzarny: Since some are pretty dumb? As soon as they read 'Allwinner' they comment with 'GPL violators! Never again!' and so on... at least that's what normally happens at CNX in the comments section ;)
<plaes>
let's see
<KotCzarny>
isnt including in mainline requires having gpl requirement?
<KotCzarny>
or its out of tree?
<plaes>
what is out of tree?
<KotCzarny>
thing that's not on kernel.org but is patchable with one
<plaes>
source is released
<tkaiser>
KotCzarny: That's the nice thing this time. It turns into an intelligence test :)
<KotCzarny>
hehe
<plaes>
original issue was that Allwinner released bunch of binary library files which seemed to include some ffmpeg codef
<oliv3r_>
tkaiser: hey dude :) I just booted your armbian image, after boot i get to log in with a forced password change, but all passwords yield authentication token manipulation error?Authentication token manipulation error
<oliv3r_>
tkaiser: ah, i have to wait for it to finish stuff :p
oliv3r_ is now known as oliv3r
<tkaiser>
oliv3r: Maybe an automated reboot happens after the first boot. We added an appropriate warning just the day before yesterday so... sorry :)
<tkaiser>
oliv3r: Authentication errors you get also when rootfs is read-only (due to FS errors due to bad SD cards, at least that's the case when people come up with this problem in Armbian forum)
<oliv3r>
tkaiser: hmm, how long does the 'init' usually last?
<plaes>
a while
<oliv3r>
actually, while its doing that, i have a few lime2 with RTL8211E PHY's but they only work for 2 minutes and then go into a fail loop
<oliv3r>
that ring any bell?
<KotCzarny>
connect screen/uart fordebug?
<oliv3r>
KotCzarny: i has :p
<oliv3r>
but since i cannot log in, i can't see my logs :(
<oliv3r>
and armbian doesn't seem to put it on the console
<tkaiser>
oliv3r: Any chance that the SD card is corrupted?
<oliv3r>
tkaiser: highly-unlikly
<oliv3r>
i'll power-off the board and fsck it
<tkaiser>
oliv3r: Great, and while you're at it you could adjust /boot/boot.cmd to send log to serial console
<KotCzarny>
oliv3r, put that card in pc then hack your root pass/account ?
<oliv3r>
tkaiser: sure :)
<oliv3r>
KotCzarny: that's a work-around, not a fix :p
<KotCzarny>
ie. add root2 account in /etc/passwd
<KotCzarny>
but then you can login and debug further
<oliv3r>
tkaiser: disk is not corrupted
<KotCzarny>
its not meant as permanent fix
<oliv3r>
but i'll disable the forced password change
<KotCzarny>
you can also add your ssh key to root account so you can login
<oliv3r>
true
<KotCzarny>
there are multiple ways to debug it, butthing is, you want to see what's in /var/log/
<oliv3r>
yeah
<KotCzarny>
Also make sure you are mounting the file system read/write.
<KotCzarny>
one of google hits
<oliv3r>
well i only dded the raw and booted it :p
<oliv3r>
so i'm going at it as a 'user'
<KotCzarny>
add root2 just to see what's happening
<oliv3r>
i'm trying to boot the armbian image to see a) if the lime12's clock freq for the ram is to high by default, and if the phy works properly with the 3.4 kernel
ninolein has quit [Remote host closed the connection]
formruga has quit [Quit: Konversation terminated!]
formruga has joined #linux-sunxi
OverCR has quit [Quit: Leaving.]
<plaes>
umm.. are Olimex guys really that misinformed?
<KotCzarny>
maybe,or its just a tiny product positioning
<plaes>
what fork are they talking about? 3.4 one?
<KotCzarny>
linux-sunxi git
<Net147>
I wouldn't mind a RK3288-based Olimex board for 4K support
OverCR has joined #linux-sunxi
<KotCzarny>
hmm, regarding possible mali uses, wouldnt it speed up operations as resize etc?
<plaes>
yes, if you use opengl
<plaes>
well, maybe ;)
<KotCzarny>
can gles do it?
<KotCzarny>
ie. render to texture etc
foxx__ has joined #linux-sunxi
<plaes>
well, there's also something called g2d in Allwinner SoCs
<KotCzarny>
only in pre-h3 ones
<plaes>
and not in A13 for example
<wens>
Simos quoted Linus' reply to GPL enforcement from the kernel summit ML
paulk-collins has joined #linux-sunxi
tkaiser has quit [Read error: No route to host]
<MoeIcenowy>
A33 musb seems to have many bugs on USB 2.0 devices...
<MoeIcenowy>
It cannot connect to a USB CDC ACM (the gadget is a CHIP)
<MoeIcenowy>
When it's connect to a USB Sound Card
<MoeIcenowy>
It will cause an oops
<MoeIcenowy>
linux-next 仍然没有更新。
<MoeIcenowy>
(oh I sent a message to a wrong channel...
<MoeIcenowy>
sorry
JohnDoe_71Rus has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org]
<_mamalala>
regarding the mali frenzy: my guess would be that many people are confused because of the way stuff is done on desktop machines ...
<_mamalala>
on a normal pc you usually need the proper drivers for the graphics card to get 2d, 3d and video acceleration
<KotCzarny>
one gpu to do it all, one gpu to amaze them all
<KotCzarny>
_mamalala: the difference is gpu on pc doing 2d AND 3d
OverCR has quit [Read error: Connection reset by peer]
<MoeIcenowy>
yes people is fuzzing PC Graphics Card and mobile GPU
<_mamalala>
so my guess is that people assume it is the same on these small boards, i.e. a single driver giving you all this, and are simply unaware that on our soc's, there are actually independent sections that do these things seperately
<KotCzarny>
there was an early time of 3d ACCELERATORS on pc
<KotCzarny>
voodoo anyone? ;)
<_mamalala>
yea
<MoeIcenowy>
Voodoo... historical...
<MoeIcenowy>
It's before I'm born...
<KotCzarny>
hehe
<KotCzarny>
A typical Voodoo Graphics PCI expansion card consisted of a DAC, a frame buffer processor and a texture mapping unit, along with 4 MB of EDO DRAM. The RAM and graphics processors operated at 50 MHz. It provided only 3D acceleration and as such the computer also needed a traditional video controller for conventional 2D software
<_mamalala>
i still remember when i used to work with softimage on a raptor workstation ... had some expensive open-gl graphics card, like 40k bucks or so ....
<KotCzarny>
whopping 4mb of ram baby
<_mamalala>
which, compared to a cheap sub-200 bucks card from nowdays just sucked, but at the time was one of the fastest thing available
OverCR has joined #linux-sunxi
<wens>
before that there was only 2d acceleration
<KotCzarny>
wens, and earlier there were no computers
<KotCzarny>
;)
<MoeIcenowy>
And in-graphics-card video decode is a more new thing...
<KotCzarny>
amazingly powervr is still alive, while 3dfxwent booboo
<MoeIcenowy>
PVR lives as an embedded solution
<KotCzarny>
embedded evil incarnate
kaspter has quit [Ping timeout: 265 seconds]
OverCR has quit [Ping timeout: 255 seconds]
<plaes>
two weeks ago, I set up Voodoo-based "old" machine
<wens>
wonder what the reddit thread he mentioned was
jbrown has quit [Ping timeout: 258 seconds]
<ssvb>
"Allwinner was a pain for a very long time. But as developers, and through the efforts of a lot of people at the Linux Foundation and Linaro, Allwinner is now a contributor to the kernel, and actively sponsors developers to write GPLv2 code for their chips."
<ssvb>
^ really?
<KotCzarny>
"talk is cheap; show me the code"
<_mamalala>
isn't allwinner a member of the linaro stuff for some time now?
<jelle>
KotCzarny: it will not replace mount don't worry
<KotCzarny>
we will see
jbrown has joined #linux-sunxi
paulk-collins has quit [Quit: Leaving]
nove has joined #linux-sunxi
<Keziolio>
is it normal that the legacy 3.4 kernel doesn't work with mainline u-boot on H3?
<_mamalala>
"doesn't work" how?
<_mamalala>
i have the 3.4.112 kernel and use mainline u-boot
<Keziolio>
doesn't start, no dmesg output on the serial port
<Keziolio>
uh, ok, i'll have to try again then
<_mamalala>
let me guess, the kernel doesnÄt even start?
<topi`>
does the mainline uboot understand about .fex files?
<_mamalala>
in the u-boot config, one important thing is that under "ARM Architecture" you enable "Enable support for booting in non-secure mode" and then "Boot in secure mode by default"
<_mamalala>
also, there is an issue with passing the system type from u-boot to the kernel, so you have to add that manually in the boot.scr
<Keziolio>
the kernel doesn't even start, now i reflashed everything to make it work, thanks i'll make sure to check that next time
<_mamalala>
namely, you have to add "machid=1029" there
<_mamalala>
i have "U-Boot 2016.09-rc2-dirty" on right now, before that i had the -rc1 version
<_mamalala>
there are orangepi_*_defconfig files in there as well, so getting a suitable base config should be easy
<Keziolio>
ok now i selected the right options
bonbons has joined #linux-sunxi
<_mamalala>
but my first guess, if u-boot starts but not the kernel, are the two options mentioned, as well as the machid=1029 setting in boot.scr
<_mamalala>
btw, i made a small patch for u-bbot, so that it will use uEnv.txt instead of the binary boot.scr (or was it boot.cmd) ... one less textfile-to-binary conversion to be done if one wants to edit these parameters
ninolein has quit [Remote host closed the connection]
ninolein has joined #linux-sunxi
ninolein has quit [Remote host closed the connection]
<nove>
so many things happening today, but is true the wiki should have more information to counter all the misunderstanding
ninolein has joined #linux-sunxi
<nove>
there is really the need to have all the license issues cataloged, for informational reasons only (of course)
<_mamalala>
lol, the systemd/mount-tool thread on phoronix ...
<_mamalala>
"If systemd is so bad, why do all major distributions use it instead of moving on without it ?"
<_mamalala>
i wonder if people even think twice before typing such stuff
MiLeon has quit [Quit: Leaving]
<nove>
and i will (try) to that, (even knowing that by doing so i will make no friends)
<nove>
there appears that are many people, that aren't still aware of the hdmi driver gpl violations in H3 and A64 vendor kernel
IgorPec has quit [Ping timeout: 252 seconds]
reinforce has quit [Quit: Leaving.]
scream has quit [Remote host closed the connection]
hansg has quit [Quit: Leaving]
ninolein has quit [Remote host closed the connection]
ninolein has joined #linux-sunxi
ninolein has quit [Remote host closed the connection]
Putti has quit [Ping timeout: 244 seconds]
ninolein has joined #linux-sunxi
\\Mr_C\\ has joined #linux-sunxi
foudubassan has quit [Ping timeout: 260 seconds]
curtis22 has joined #linux-sunxi
ninolein has quit [Remote host closed the connection]
Putti has joined #linux-sunxi
ninolein has joined #linux-sunxi
lemonzest has quit [Quit: Leaving]
bonbons has quit [Quit: Leaving]
Putti has quit [Ping timeout: 250 seconds]
a|3x has left #linux-sunxi ["Leaving"]
<Keziolio>
ok got mainline u-boot to work :D
<_mamalala>
Keziolio: good ... so, what was the issue that it didnt work before?
<Keziolio>
i have no idea, I did what you suggested and now it works
nove has quit [Quit: nove]
<_mamalala>
hehe, ok
Nacho has joined #linux-sunxi
Nacho_ has quit [Ping timeout: 244 seconds]
<Keziolio>
one of the two config options was missing (maybe that should be patched mainline ?) and then i had to edit a file because of too big uImage
<_mamalala>
hmm, maybe too many options were enabled? there are some *_defconfig for orangepi boards, some of which use the h3 ...
Guest13037 has quit [Read error: Connection reset by peer]
<_mamalala>
but then, in my config i disabled basically everything that's not needed to get a functioning u-boot that basically falls through to boot the kernel
<_mamalala>
that is, i disabled the networking stuff, usb, etc.... no wait time either
<Keziolio>
I used orangepi_pc_defconfig even though it's not an orange pi
<Keziolio>
yeah that seems a good idea, there's a lot of stuff
<Keziolio>
have you ever tried to get mainline linux on H3? or it's too early?
<_mamalala>
not yet, no ...
<_mamalala>
withsome luck, mainline support for h3 is ready once i'm finished with my current project, so i can switch to that then
<_mamalala>
but then, no problem in using 3.4 for that particular thing, anyways
<Keziolio>
uh cool
<Keziolio>
that kernel doesn't even support my mouse, and has security holes
<_mamalala>
hu? usb mice and keyboards are prety generic, the standard usb hid driver should work for the vast majority ....
<_mamalala>
unless it's some rather special and exotic mouse
<Keziolio>
I remember back in 2013 I had to patch a line to get it work, it got mainlined pretty quick, in 2013...
<_mamalala>
i think the lack of decent docu for their chips, plus often poorly written code by them, is the main reason that mainlining efforts for allwinner stuff are going so slow
<_mamalala>
all we have are some datasheets that barely give a description of the registers of the chip's peripherials, and some code blobs with chinglish messages ...
<_mamalala>
... plus some #ifdef sunsomething sprinkeld into the main linux sources of the 3.4 sunxi kernel
<_mamalala>
sure, still way better than having to reverse engineer all that stuff, but not that helpful either
<Keziolio>
yeah i know, i often "fight that battle" on other places, arm devices are just bad
Andy-D_ has joined #linux-sunxi
<Keziolio>
but i've never contributed, even if I'm fluent in C kernel programming seems on another level
<_mamalala>
well, i guess that's just what happens when you have a gazillion of shops implementing the arm ip in their own chips, instead of havong a single source like intel ...
<Keziolio>
...and another nazi entity like MS that sets the standards
<_mamalala>
well, programming on the kernel level is not _that_ hard ... however, just don't expect you code ending up in some mainline kernel, since they have rather strict requirements (understandably so)
<Keziolio>
back in the days there was a lot of competition on x86 chips
<Keziolio>
yeah sure
<_mamalala>
no, there really wasn't much competition in the x86 field ... amd being the only ones (of note) that produced their own x86 compatible chips
<_mamalala>
the majority of other x86 chips from other brands were just that: x86 chips produced by others
<Keziolio>
well "a lot" compared to today's monopoly
<_mamalala>
on arm it is completely different, because arm does not sell chips, but only ip ... everyone integrates that ip differently
Andy-D has quit [Ping timeout: 260 seconds]
<_mamalala>
back then you could use an x86 from nec, and have it run the same software that ran on intel x86, because it was basically the same (ok, there have been a few exceptions, but generally)
<_mamalala>
on arm, although a lot of stuff is "the same", everyone implemets it a tad differently, when it comes to register addresses, etc.
<_mamalala>
(registers for peripherials, i mean)
<Keziolio>
yeah, coprocessors
<Keziolio>
not sure if they are called that way in the arm manual
<_mamalala>
or just look at the start address for the external ram .... pretty much every vendor puts it at a different address, although the cores used are the same, etc...
Da_Coynul has joined #linux-sunxi
<Keziolio>
what are you going to work on for the H3?
<Keziolio>
what's missing? I see DRM and ethernet for the important stuff
<_mamalala>
i'm going to make a controller board for 3d printers, where the h3 also directly controls the stepper motors ... because i'm sick of all that 8-bit arduino based crap out there
unixdork is now known as nixdork
<Keziolio>
lol
<_mamalala>
so i spent some time to get the preempt_rt patches put into the 3.4.112 sunxi kernel, and did a nasty kernel hack to get a better latency/less jitter for specific interrupts i use, so that the steppers will move smoothly
<_mamalala>
now i'm writing the kernel modules that control all the required parts, starting with the core stepper driver module
<Keziolio>
ah so pretty domain specific work, seems interesting btw, good luck
<_mamalala>
yea, pretty much ... thanks
<_mamalala>
so far it looks very good
paulk-nyan-big has joined #linux-sunxi
curtis22 has quit [Quit: Leaving]
ricardocrudo has quit [Remote host closed the connection]
afaerber has quit [Quit: Ex-Chat]
paulk-nyan-big has quit [Ping timeout: 260 seconds]
leio has quit [Ping timeout: 276 seconds]
jernej has quit [Ping timeout: 240 seconds]
<_mamalala>
Keziolio: but it's rather interresting how a little bit of kernel hacking suddenly makes things possible that were not possible before ...
<Keziolio>
yeah sure, considering that was a device for shitty home entertainment devices
<_mamalala>
one of the main reason why i use linux since decades, even on my desktop ... i just can change whatever i want, if i need to
<Keziolio>
s/devices/systems/
<Keziolio>
i usually write long blog posts on this matter
<_mamalala>
i just hope that once h3 mainlining has been good enough, that i can implement the "pseudo-fiq" hacks and thus improve things for me just a bit more ...
<_mamalala>
long blog posts on such matter? nice, have a link?
tkaiser has quit [Ping timeout: 258 seconds]
<Keziolio>
they are in Italian so you probably won't understand a word
<_mamalala>
ha, ok ...
<Keziolio>
lol
<Keziolio>
they are only here for me to spam and raise some awareness, most people are misinformed
leio has joined #linux-sunxi
<Keziolio>
especially android fanboys
<_mamalala>
yea, that's the cost that comes with the ubiquity of small computers
<Keziolio>
they can't figure out what they are losing, for us leet haxors it's easy to compile kernels and stuff
<_mamalala>
well, to be fair, one must make a difference between john doe users, makers and their ilk, and kernel/distribution folks
<_mamalala>
like, as much as i hate that arduino stuff, i still recognize that it helped a lot to bring microcontroller stuff and programming to a lot of people
Da_Coynul has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<Keziolio>
well, the fact that arm chips are mostly closed affects the general public with lack (and low quality) of roms and alternative os's like ubuntu touch
<Keziolio>
and all the planned obsolescence of those 3.4.* kernels
<_mamalala>
true ... i could rant for hours about the silly practice of having to sign nda's to even get a datasheet for chips, etc...
<_mamalala>
or look at the mali stuff ... why binary blobs? what on earth do they think is so precious about some code to use some piece of hardware that they have to keep it closed?
<_mamalala>
personally i like to think that they are just extremely shitty coders and are ashamed to make that code public ... :P
<Keziolio>
yeah that drives me nuts
<_mamalala>
same goes for binary blob drivers on desktop machines too, of course ...
<Keziolio>
i read on the dolphin drivers page that an opensource implementation outrunned the blob in some benchmarks
<_mamalala>
i _hate_ all that secrecy and crap around graphic card drivers, for example
<Keziolio>
sure, i'll buy AMD once the price is settled
<Keziolio>
i'm tired of kernel panics when upgrading nvidia blobs
<_mamalala>
frankly, i don't expect them to open up the latest and greatest, or provide exquisite docu on their code ... heck, i don't even expect them to maintain it once they opened up ...
<_mamalala>
but at least open it up, damn ... so that others can pick it up and continue to work on it, fix bugs, etc
<Keziolio>
and they don't expect my money
<Keziolio>
they share most of the code between OS's, they probably have other kinds of IP (game specific optimization?)
<Keziolio>
but yeah, there are no reasons in hell on why they souldn't document stuff or help with nouveau
<Keziolio>
they are going to reverse that anyway, nouveau devs or amd
<_mamalala>
i don't care for specific bordserline cases/optimizations ... just giving enough to enable other to produce an open-source implementation would be a very good start already
<_mamalala>
just look at the mali debacle for arm64 allwinner .... that's just annoying