<davidebeatrici>
"While being open source, some of the Allwinner's code in the kernel has either no license boilerplate or an "all rights reserved" license notice. Most notably this includes HDMI support."
<davidebeatrici>
Ah, here's why HDMI audio is not working.
<anarsoul>
davidebeatrici: not really
<anarsoul>
:)
<davidebeatrici>
Is that another issue? I didn't investigate much into it, I just noticed that the HDMI output doesn't appear as an ALSA device, only SPDIF is present.
<anarsoul>
I have HDMI audio working in my tree (2 channels only)
<anarsoul>
for A64
<davidebeatrici>
Better than nothing.
<anarsoul>
as per jernej there's no consensus on implementation details for hdmi audio
<anarsoul>
and he has up to 8 channels support in his tree
<davidebeatrici>
Oh, I see. Thank you.
<wrmyrx_>
is this the wrong place to discuss integration of upstream kernel features into android?
<anarsoul>
well, if you find anyone to discuss it :)
<wrmyrx_>
yeah, i know =/ as mentioned above, relatively simple objective; just enabling a weird custom display mode. i'm doing my best to patch the sunxi/disp2 drivers & fex files, as well as to comb through logcat/dmesg to figure out what's not working
<wrmyrx_>
for my application i do need hw accelerated video playback so as far as i can tell i basically have to use the manufacturer provided android builds, is that not the case? obviously if it weren't i'd switch to upstreamier/gpl-compliant stuff right away
SergiusUA has quit [Remote host closed the connection]
aalm has quit [Ping timeout: 250 seconds]
<[TheBug]>
wrmyrx_: what board?
<[TheBug]>
ahh k1
random_yanek has quit [Ping timeout: 246 seconds]
random_yanek has joined #linux-sunxi
agraf has quit [Ping timeout: 244 seconds]
agraf has joined #linux-sunxi
aalm has joined #linux-sunxi
specing has joined #linux-sunxi
tuxillo has quit [Disconnected by services]
tuxillo has joined #linux-sunxi
victhor has quit [Ping timeout: 257 seconds]
lynxis has quit [Quit: No Ping reply in 180 seconds.]
lynxis has joined #linux-sunxi
<wens>
plaes: yeah, and someone hacked it lol. no suprise there
<wrmyrx_>
TheBug: yeah. not really sure what to do beyond adding modelines in hdmi_core, fooling around with the fex and praying
<wrmyrx_>
but if there are any upstream kernels with mali drivers that might be safe to try and pull down into the default android build... that'd be good to know
ganbold has quit [Ping timeout: 255 seconds]
vagrantc has quit [Ping timeout: 250 seconds]
dddddd has quit [Remote host closed the connection]
Andy-D has quit [Ping timeout: 255 seconds]
egbert has quit [Ping timeout: 250 seconds]
Andy-D has joined #linux-sunxi
egbert has joined #linux-sunxi
aalm has quit [Read error: Connection reset by peer]
aalm has joined #linux-sunxi
[7] has quit [Ping timeout: 250 seconds]
TheSeven has joined #linux-sunxi
tllim has quit [Read error: Connection reset by peer]
<jernej>
and also note that you have to set newly added mode in fex
reinforce has joined #linux-sunxi
<montjoie>
wens: he forget the offload task
leviathanch has joined #linux-sunxi
selfbg has joined #linux-sunxi
Putti has joined #linux-sunxi
airwind has joined #linux-sunxi
reinforce has quit [Quit: Leaving.]
<wrmyrx_>
jernej: thanks, i've been working from patches like that and others. managed to dump a lot more debug info into dmesg which is really helpful.
Putti has quit [Ping timeout: 255 seconds]
<wrmyrx_>
jernej: if it isn't in fex, that just means it won't be the default display at boot, right? i.e., the different display mode is still hotpluggable?
<wrmyrx_>
jernej: i only ask because i'm not sure where the ordering of the display modes in fex comes from
<jernej>
what do you mean by hotpluggable?
<wrmyrx_>
like if I plug the display into my device after it's already booted... it should detect the EDID and set the mode accordingly, no?
<wrmyrx_>
assuming the support for that mode has been patched into the kernel?
<jernej>
sadly, no
<wrmyrx_>
ah. well. that changes my debugging heuristic somewhat
<libv>
montjoie: only if you move out the changelog to a separate page
<libv>
because that is not useful for users
<libv>
to have that right next to supported decivices, compile, configure
<libv>
3/5th of the page is changelog
<libv>
which is not sensible
<montjoie>
the move could be in second time, when the matrix will be full of "all info in changelog"
<libv>
montjoie: if you are doing a table, you might as well have links to anchors in the right changelog page
<libv>
otherwise you will be doing twice the work
<montjoie>
but the changelog speak about boards, the matrix about soc
<montjoie>
I think both are needed
<libv>
you're undertaking major work to that page, start with moving out the changelog to a separate page
<libv>
it's not an insane task, it's not an unreasonable demand, and it will vastly improve what is now a rather useless and overwhelming behemoth
<montjoie>
any idea on the new page(changelog) name ?
<libv>
append _Changelog :)
nashpa has quit [Ping timeout: 250 seconds]
Rafael1980 has joined #linux-sunxi
nashpa has joined #linux-sunxi
lynxis has quit [Quit: No Ping reply in 180 seconds.]
lynxis has joined #linux-sunxi
lynxis has quit [Quit: No Ping reply in 180 seconds.]
lynxis has joined #linux-sunxi
lynxis has quit [Quit: No Ping reply in 180 seconds.]
lynxis has joined #linux-sunxi
<libv>
it's amazing just how 2014 the mainline kernel howto is
<libv>
how the f did people allow that to happen?
<libv>
don't you want users?
<libv>
is the wiki really just a glorified changelog to you guys?
<libv>
if you really just want to feel good about yourself, and do not care about producing something others can use as well, go rub one out
<libv>
if you want mainline to require a masonary handshake, then go place this changelog on an obscure server somewhere else.
Nakaori has quit [Remote host closed the connection]
Nakaori has joined #linux-sunxi
<KotCzarny>
wiki is as good as users editing it, i guess
<libv>
s/users/people/
<KotCzarny>
most people stop using it once they got whatever they needed
<libv>
users specifically excludes mainline developers, who seem to have no interest in seeing their code used
<KotCzarny>
isnt mainline code reason being in the state of 'it's mainlined and just works' ?
<libv>
oh?
<KotCzarny>
ie. once mainlined, it works
<libv>
really?
<KotCzarny>
then distros embrace it and users use it
<libv>
when did you last throw kernel source on an sd card and boot it?
<KotCzarny>
um, yesterday?
<libv>
_actual_ kernel source?
<KotCzarny>
yes, compiled it on-device
<libv>
ah
<libv>
compiled.
<KotCzarny>
then used the resulting binary
<libv>
how did you compile it?
<libv>
where did you place what?
<KotCzarny>
i use armbian mostly
<libv>
ok
<KotCzarny>
but prefer my own kernel compiled
<libv>
so i should go and rm -Rf the wiki, except for the two mainline changelogs
<libv>
and then see how far tkaiser gets with that
<KotCzarny>
tkaiser isnt interested in allwinner anymore
<KotCzarny>
he went onto more powerfuls socs
<libv>
ok
<libv>
so what about board makers?
<libv>
like ntc
<KotCzarny>
what is ntc?
<libv>
how did that story go?
<libv>
where is their forums and wiki now?
<libv>
next thing co
<KotCzarny>
i think they went bankrupt anyway
<KotCzarny>
and did some scams
<libv>
so how does one use the 9usd board nowadays?
<KotCzarny>
regular joe or you ask me specifically?
<libv>
anyone
<KotCzarny>
hard to guess, i bet on using distro's images
<libv>
without this wiki, how do you get to an sd card which works?
<libv>
how did the people who made those images get to making those images?
lynxis has quit [Quit: No Ping reply in 180 seconds.]
<KotCzarny>
by putting available pieces together
<KotCzarny>
banging them into place if necessary
<libv>
from lookin at a crystal ball, no doubt
<KotCzarny>
which includes info available on wiki too
<libv>
that really would be a fun experiment
<libv>
reduce the wiki to the changelog and see how the world reacts
lynxis has joined #linux-sunxi
<KotCzarny>
most likely it would be not fun
<KotCzarny>
wiki IS the main source of hw info
<KotCzarny>
and some cream on top
<libv>
KotCzarny: wasn't that what i was trying to tell you for the last 10 minutes?
<KotCzarny>
depends, i was answering the questions asked, i know what you want to convey via this
<KotCzarny>
but i'm not the right person i guess, since im closer to regular joe
<KotCzarny>
wiki is used as a scratchpad and place to store reliable info about allwinner socs
aalm has quit [Ping timeout: 255 seconds]
<KotCzarny>
so new people (both new users and devs interested in new features) can use to get familiar with the topic
<KotCzarny>
and add to it too
aalm has joined #linux-sunxi
leviathanch has joined #linux-sunxi
AneoX_ has joined #linux-sunxi
AneoX has quit [Ping timeout: 272 seconds]
<karlp>
fwiw, I actually _liked_ the uboot page ahving all the chagnelog, seleected and curated as it applies to sunxi boards in one place, it was nice and easy to ctrl-f for things on.
<karlp>
also, I don't see how anyone ever suggested keeping only a changelog.
dddddd has joined #linux-sunxi
<libv>
karlp: suggested, no, in practice, this is what linux-sunxi has turned into these last few years
<libv>
karlp: you can ctrl-f the other page
JohnDoe_71Rus has joined #linux-sunxi
<karlp>
yes, without doubt
<karlp>
just trying to say that the sunxi curated uboot changelog is useful.
<libv>
it is, but there is more to life
<libv>
and there is no intention of throwing that away
<karlp>
so let's drop the whole "see what happens when there's just a changelog?!?"!?"!"#!!" ranting then maybe?
<karlp>
noone other than you was suggesting that.
<libv>
as a way of proving how valuable the rest of the wiki is, or should be
<libv>
the way the documentation to get mainline u-boot and kernel running on your device has horribly rotted in the past 5 years
<KotCzarny>
but current changlogs are still on mainlining page
<libv>
KotCzarny: that's just stupid
<KotCzarny>
wasnt me
<libv>
and deliberate
<rellla>
libv: heh
<libv>
true, but it still is stupid and shows just how some people are against providing others with useful information
<karlp>
as someone who came to sunxi int eh last six moths, I found it quite clear what to do if I only wanted mainline uboot and linux.
<karlp>
it can always be better
<libv>
karlp: look at the history of those bits that were useful to you
<libv>
and see who added what, when
<libv>
that will clearly show exactly what i am complaining about
<karlp>
no, I was only interestined in productive living, not blamign and vitriol
<libv>
sure, it probably is still better than getting mainline going on rpi, but that's a sign of defeat
<libv>
karlp: check the history and see the reason for the vitriol
lynxis has quit [Quit: No Ping reply in 180 seconds.]
lynxis has joined #linux-sunxi
elros1 has quit [Remote host closed the connection]
<montjoie>
libv: for the uboot changelog move, does the wiki have a good way to do it instead of copy/paste ?
lynxis has quit [Client Quit]
lynxis has joined #linux-sunxi
<libv>
montjoie: not without a mediawiki extension it seems, and i seriously doubt that we have it installed
afaerber has quit [Quit: Leaving]
IgorPec has joined #linux-sunxi
IgorPec has quit [Ping timeout: 255 seconds]
afaerber has joined #linux-sunxi
airwind has quit [Quit: airwind]
ashleyk_ has quit [Ping timeout: 256 seconds]
nixdork has quit [Ping timeout: 252 seconds]
popolon has joined #linux-sunxi
fkluknav has quit [Remote host closed the connection]
lurchi_ is now known as lurchi__
fkluknav has joined #linux-sunxi
nixdork has joined #linux-sunxi
wrmyrx_ has joined #linux-sunxi
selfbg has quit [Remote host closed the connection]
Putti has joined #linux-sunxi
wrmyrx_ has quit [Remote host closed the connection]
ashleyk_ has joined #linux-sunxi
fkluknav has quit [Remote host closed the connection]
lynxis has quit [Quit: No Ping reply in 180 seconds.]
lynxis has joined #linux-sunxi
fkluknav has joined #linux-sunxi
vagrantc has joined #linux-sunxi
lynxis has quit [Quit: No Ping reply in 180 seconds.]
lynxis has joined #linux-sunxi
samrose has quit [Quit: Konversation terminated!]
samrose has joined #linux-sunxi
JohnDoe_71Rus has quit [Ping timeout: 255 seconds]
lynxis has quit [Quit: No Ping reply in 180 seconds.]
lynxis has joined #linux-sunxi
mpmc has quit [Quit: ZNC 1.6.5+deb1+deb9u1 - http://znc.in]
lynxis has quit [Quit: No Ping reply in 180 seconds.]
<willmore>
montjoie, I think Linus covered offload in his comments.
lynxis has joined #linux-sunxi
<willmore>
(ie a "network
<willmore>
card that does WG") <----this comment
leviathanch has quit [Remote host closed the connection]
<buZz>
what does WG stand for in that context?
<willmore>
Wire Gurard
MoeIcenowy has quit [Quit: ZNC 1.6.5+deb1+deb9u1 - http://znc.in]
<willmore>
Guard, rather..
<willmore>
My finders are running async today...
<buZz>
ah
<buZz>
that hw crypto story
MoeIcenowy has joined #linux-sunxi
<willmore>
WRT wen's link to Linus taking about crypto accelerators and zinc.
<willmore>
He makes a lot of very good points showing how useless most crypto accelerators are to actual practical applications.
<buZz>
uhuh
<willmore>
His arguement is that the cost of syncing to an external processor (cache issues, DMA issues, IRQ overhead) outweigh the benefits--unless the CPU is horribly slow.
<buZz>
yeah totally, in-cpu crypto makes most sense
mpmc has joined #linux-sunxi
<willmore>
And that external accelerators only make sense if they implement the whole thing or they are necessary for the transfer anyway--like a SATA controller that did crypto on the data it was moving and it took care of all the work.
<KotCzarny>
on marvel hwcrypto helps quite nicely with luks
<willmore>
buZz, yep.
<KotCzarny>
~50MB/s achievable with cpu free to do other tasks
<KotCzarny>
in aes256-cbc
vpeter has quit [Ping timeout: 245 seconds]
<buZz>
willmore: well that kinda already happens since first IDE disks?
<willmore>
KotCzarny, only if you don't have to sit around and busy wait while it's processing.
<buZz>
iirc even superold PATA/IDE disks had inbuilt hw crypto
<willmore>
buZz, in this context we're talking about the host side of the interface.
JohnDoe_71Rus has joined #linux-sunxi
<buZz>
well, a 'crypto networkcard' wouldnt be host side, would it?
<willmore>
Like having your network card run the whole VPN stack.
<KotCzarny>
keys are in app often, no?
JohnDoe1 has joined #linux-sunxi
<willmore>
KotCzarny, not sure what you mean.
<KotCzarny>
or kernel
<buZz>
they can be yeah, or in TPM
<willmore>
Which he addresses. :)
<buZz>
or on your yubikey :P
<KotCzarny>
before network starts encryption
<KotCzarny>
and in case of ssh, you cant rely on network card to do it
<KotCzarny>
etc
<willmore>
KotCzarny, why not?
<buZz>
why not though?
<buZz>
i think thats do-able
<willmore>
It's only doing the symetric part.
<KotCzarny>
for multiple users?
<buZz>
sure
<willmore>
Sure, each network descriptor has the key for that stream.
<KotCzarny>
but how would ssh pass that key to network card?
<willmore>
IOCTL?
<willmore>
Adding it into some crypto socket setup call.
<KotCzarny>
would require whole new layer of setup, no?
<willmore>
Sure.
<willmore>
No one said it would be free.
<KotCzarny>
..and now we have 16 standards
<KotCzarny>
:)
<willmore>
KotCzarny, well, 14 of those are *BSD and no one cares about those.
tnovotny has quit [Quit: Leaving]
JohnDoe_71Rus has quit [Ping timeout: 245 seconds]
<KotCzarny>
would you trust network cards producers with your crypto keys tho?
<willmore>
Why trust a TPM? And, yeah, the high end crypto cards have always been trusted to hold the keys--even when the CPU wasn't.
<willmore>
So, that's not an unusual trust model.
vpeter has joined #linux-sunxi
lynxis has quit [Quit: No Ping reply in 180 seconds.]