TheSeven has quit [Remote host closed the connection]
TheSeven has joined #linux-sunxi
ErwinH has joined #linux-sunxi
ErwinH has quit [Ping timeout: 258 seconds]
popolon has quit [Quit: WeeChat 1.4]
apritzel has quit [Ping timeout: 272 seconds]
<wens>
KotCzarny: there is no 'master' control on these codecs
<wens>
KotCzarny: each output path can have its own volume control
Wizzup has quit [Ping timeout: 268 seconds]
<wens>
apritzel: are you going to do an h5 series next?
reinforce has quit [Quit: Leaving.]
cnxsoft has joined #linux-sunxi
Wizzup has joined #linux-sunxi
TheSeven has quit [Ping timeout: 245 seconds]
deskwizard has quit [Ping timeout: 256 seconds]
egbert has quit [Disconnected by services]
egbert has joined #linux-sunxi
TheSeven has joined #linux-sunxi
ErwinH has joined #linux-sunxi
ninolein has quit [Ping timeout: 258 seconds]
ninolein has joined #linux-sunxi
KB3VGW- has joined #linux-sunxi
KB3VGW has joined #linux-sunxi
ErwinH has quit [Ping timeout: 248 seconds]
KB3VGW- has quit [Client Quit]
mripard_ has quit [*.net *.split]
marvs has quit [*.net *.split]
Patsie has quit [*.net *.split]
gumblex has quit [*.net *.split]
fest has quit [*.net *.split]
critch has quit [*.net *.split]
pulser has quit [*.net *.split]
MoeIcenowy has quit [*.net *.split]
joedj has quit [*.net *.split]
vetkat has quit [*.net *.split]
datagutt has quit [*.net *.split]
Patsie_ has joined #linux-sunxi
critch_ has joined #linux-sunxi
mripard has joined #linux-sunxi
fest_ has joined #linux-sunxi
marvs_ has joined #linux-sunxi
MoeIcenowy has joined #linux-sunxi
gumblex has joined #linux-sunxi
pulser has joined #linux-sunxi
datagutt has joined #linux-sunxi
vetkat has joined #linux-sunxi
joedj has joined #linux-sunxi
fest_ is now known as fest
vetkat has quit [Changing host]
vetkat has joined #linux-sunxi
datagutt has quit [Changing host]
datagutt has joined #linux-sunxi
ganbold has quit [Ping timeout: 246 seconds]
KB3VGW has left #linux-sunxi ["Leaving"]
ganbold has joined #linux-sunxi
ErwinH has joined #linux-sunxi
ErwinH has quit [Ping timeout: 258 seconds]
victhor has quit [Ping timeout: 256 seconds]
my123 has quit [Ping timeout: 256 seconds]
my123 has joined #linux-sunxi
my123 has quit [Changing host]
my123 has joined #linux-sunxi
terra854 has joined #linux-sunxi
pg12 has quit [Ping timeout: 245 seconds]
pg12 has joined #linux-sunxi
ErwinH has joined #linux-sunxi
ErwinH has quit [Ping timeout: 240 seconds]
laj has quit [Quit: Page closed]
ganbold has quit [Ping timeout: 256 seconds]
ganbold has joined #linux-sunxi
leviathanch has joined #linux-sunxi
TheSeven has quit [Ping timeout: 240 seconds]
TheSeven has joined #linux-sunxi
ErwinH has joined #linux-sunxi
ErwinH has quit [Ping timeout: 240 seconds]
IgorPec has joined #linux-sunxi
foxx has joined #linux-sunxi
iamfrankenstein has joined #linux-sunxi
tsuggs has quit [Ping timeout: 248 seconds]
a1d3s has joined #linux-sunxi
<a1d3s>
good morning :)
andromeda-galaxy has joined #linux-sunxi
chomwitt4 has quit [Quit: WeeChat 1.0.1]
chomwitt has joined #linux-sunxi
<andromeda-galaxy>
I am working with a NextThingCo chip based on the R8 (A13-ish), and am curios about linux-sunxi powersaving support
chomwitt has quit [Quit: WeeChat 1.0.1]
<andromeda-galaxy>
Currently the device is based on the 4.4 mainline kernel and has no powersaving support at all
chomwitt has joined #linux-sunxi
<andromeda-galaxy>
I saw that on the wiki page there is a mention of a patch that enables a very basic psci suspend, but it seems like that still uses a fair bit of power. Are there plans for any more advanced powersaving on A13?
scream has joined #linux-sunxi
<andromeda-galaxy>
Also. are there any good resources for understanding the A13/R8 hardware pm implementations? The manuals seem to just list a bunch of opaquely named registers
<andromeda-galaxy>
Also, in particular, are there any plans for implementing deep sleep where the processor is fully powered off (and if there are, what order of resume time would be likely?)
<andromeda-galaxy>
Sorry for all the questions ;-). Finally. with the existing light sleep psci implementation, can some high-priority userspace processes be able to run occasionally? (for e.g. maintaining wifi connections in the background)?
ErwinH has joined #linux-sunxi
scream has quit [Remote host closed the connection]
lemonzest has joined #linux-sunxi
ErwinH has quit [Ping timeout: 240 seconds]
tsuggs has joined #linux-sunxi
ErwinH has joined #linux-sunxi
ErwinH has quit [Ping timeout: 255 seconds]
DullTube has joined #linux-sunxi
massi has joined #linux-sunxi
JohnDoe_71Rus has joined #linux-sunxi
cnxsoft has quit [Read error: Connection reset by peer]
cnxsoft has joined #linux-sunxi
ErwinH has joined #linux-sunxi
repka has joined #linux-sunxi
apritzel has joined #linux-sunxi
florianH has joined #linux-sunxi
giomba has joined #linux-sunxi
apritzel has quit [Ping timeout: 240 seconds]
Mr__Anderson has joined #linux-sunxi
repka has quit [Quit: Leaving]
apritzel has joined #linux-sunxi
msevwork has joined #linux-sunxi
<apritzel>
wens: do you mean H5 series for U-Boot?
<apritzel>
wens: since the patches are ready, I was thinking about it
<apritzel>
wens: but actually the FIT support is more important, I think
<wens>
well my r40 patches touches the same parts in the dramc driver
<wens>
so i was thinking you should go first :)
<apritzel>
wens: yeah, I can send them out
<tuxillo>
hi
giomba has quit [Ping timeout: 248 seconds]
giomba has joined #linux-sunxi
fkluknav has joined #linux-sunxi
giomba has quit [Ping timeout: 248 seconds]
giomba has joined #linux-sunxi
dgp has quit [Ping timeout: 256 seconds]
dgp has joined #linux-sunxi
popolon has joined #linux-sunxi
giomba has quit [Ping timeout: 240 seconds]
Leepty has quit [Remote host closed the connection]
The_Loko has joined #linux-sunxi
reinforce has joined #linux-sunxi
<montjoie>
updated my opipc to 4.10rc1 and get "sunxi-mmc 1c0f000.mmc: fatal err update clk timeout
<montjoie>
typed enter too early, seems that some people already get it
<premoboss>
hi, im doin modification to DTS file into source linux kernel 4.9 form my board (NanoPi NEO). Now i would like to compile the dts, but it can't be done with dtc utility because the lile hase some #include that five sintax errot. how i can compile the dts into dtb withour recompiling every time whole the souce linux tree?
<beeble>
premoboss: make <dtbfilename>
<beeble>
but since it is only calling dtc you will have to fix the errors?
<a1d3s>
make dtbs ?
DullTube has quit [Quit: Leaving]
marvs_ is now known as marvs
marvs has quit [Changing host]
marvs has joined #linux-sunxi
<premoboss>
i am into /arch/arm/boot/dts. i do make sun8i-h3-nanopi-neo.dts but make say "no rules"
<beeble>
no, top level linux source directory
<a1d3s>
go to folder before /arch
<beeble>
and make sun8i-h3-nanopi-neo.dtb
<a1d3s>
then make dtbs works
<a1d3s>
or as beeble say
<premoboss>
beeble, no way, i am in root od linux-4.9 (where i expasnded the compressed archive). i do make sun8i-h3-nanopi-neo.dtb but make reply 'no rule to make target sun8i-h3-nanopi-neo.dtb' stop
<premoboss>
do i miss some packeges to support compiling?
<premoboss>
i suspect pre-processor of C is used.
<premoboss>
so to do just make <dtbnamefile> can not be enough.
<a1d3s>
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- sun8i-h3-nanopi-neo.dtb
<a1d3s>
if you are not in the pi , you must cross compile
<beeble>
ARCH=arm make sun8i-h3-nanopi-neo.dtb
<beeble>
works for me
<beeble>
so, you probably don't have ARCH set
BenG83 has joined #linux-sunxi
<premoboss>
i am on native NEO, so no cors compiling in needed
<a1d3s>
hm and what happen if you try make dtbs?
victhor has joined #linux-sunxi
<premoboss>
dtbs command not found
<apritzel>
premoboss: works for me: $ make sun8i-h3-nanopi-neo.dtb
<premoboss>
if i do 'ls' i see Documentation,m fs, ipc, kernel ....etcetc
<premoboss>
yes i am in the source tree of expanded linux sources
<apritzel>
premoboss: so make apparently works now, you just have a bug in your changed DT?
<beeble>
then it should work
<beeble>
apritzel: no, i think he tries to call dtc himself
<premoboss>
apritzel, i am in the *origianl* dts, no chage at all at this moment,ù
<premoboss>
original measn 'just expanded from linux-4.9.0.xz
paulk-blaze has quit [Quit: Leaving]
<NiteHawk>
premoboss: what version of dtc is that? ("dtc -v")
<premoboss>
NiteHawk, 1.4.0
<apritzel>
premoboss: calling dtc directly will not work, I think we mentioned that already
<apritzel>
please type: make sun-8i-h3-nanopi-neo.dtb
<apritzel>
and paste the error message
<premoboss>
ok
corecode_ is now known as corecode
<apritzel>
make sun8i-nanopi-neo.dtb
<apritzel>
actually
<apritzel>
(without the "-" between "sun" and "8i")
<premoboss>
apritzel, it works now!
<premoboss>
make sun8i-nanopi-neo.dtb works. but i installed before libncurses5-dev.
<premoboss>
because i wish to do make menuconfig to see the .config.... does ncureses need also to compile? or maybe i was doing something wrong before when i try the previous make.
<premoboss>
mmm make sun8i-nanopi-neo.dtb works for a while but then stop with error. no rule to make target..... stop
<a1d3s>
for menuconfig you need dialog i think
<premoboss>
with or without "-" between sun and 8i
<beeble>
its sun8i-h3-nanopi-neo.dtb
<beeble>
the h3 part is missing
<premoboss>
ok i try wit no "h3"
<beeble>
like the dts file, but with the dtb suffix
fkluknav has quit [Ping timeout: 248 seconds]
<premoboss>
see error here pastebin.com/2tQ6GFGY
<beeble>
still the h3- part missing
<premoboss>
i redo with h3 and pastebin, wait,.
The_Loko has quit [Quit: Leaving]
<premoboss>
beeble, it works now.
<premoboss>
i got the .dtb into dts directory.
<premoboss>
now i add my dodification to dts and recompile to see if it works. many thanks to all :)
victhor has quit [Quit: Leaving]
victhor has joined #linux-sunxi
tkaiser has joined #linux-sunxi
my123 has quit [Ping timeout: 256 seconds]
paulk-collins has joined #linux-sunxi
a1d3s has quit [Quit: Leaving]
Pepe has joined #linux-sunxi
<premoboss>
bad news: uboot in NanoPi-NEO DOESNOT use sun8i-h3-nanopi-neo.dtb, it look for sun8i-h3-orangepi-one.dtb. it means that uboot need to be recompiled specifically for NanoPi-NEO
<tkaiser>
premoboss: Nope, you just need to add one line to boot.cmd (and you need to get a bit familiar with Armbian and stop to misuse this developer channel for distro topics)
<wens>
isn't there an armbian channel?
<tkaiser>
wens: no idea
<premoboss>
it exist #armbian on freenode, but i see linux-sunxi is the main cnallen about arm board, so i suppose here is the right place. sorry if i misuse, i will not do more.
deskwizard has joined #linux-sunxi
my123 has joined #linux-sunxi
my123 has quit [Changing host]
my123 has joined #linux-sunxi
Leepty has joined #linux-sunxi
<tuxillo>
there is a channel but just a bunch of people thre
<tuxillo>
I guess the forums is a better place (slower tho)
<premoboss>
tuxillo, yes, and no traffic.
Ntemis has joined #linux-sunxi
<tkaiser>
premoboss: Basically you add a line to the top of /etc/boot.cmd with the following contents: setenv fdtfile "sun8i-h3-nanopi-neo.dtb"
<tkaiser>
premoboss: Then 'compile' the file to boot.scr (see last line of boot.cmd for the syntax)
<premoboss>
good, so i can "overvrite" the $fdtfile. thanks!
<IgorPec>
wens: we have unoficial #armbian channel :)
<tkaiser>
IgorPec: Is there a log available?
<IgorPec>
don't know. have to ask admin to set it upž
<tkaiser>
Without a backlog it's just a stupid waste of time to join there.
<IgorPec>
i know
perr has joined #linux-sunxi
perr has joined #linux-sunxi
perr has quit [Changing host]
<beeble>
tkaiser: FUD
<beeble>
the resistance of the connector is low enough for even high currents. buy an awg 24 usb cable and be happy
<KotCzarny>
beeble, so you are saying those failing micro usb boards would work on thick usb cable?
<tkaiser>
beeble: People use AWK 28 cables and combine them with 'smart chargers' not providing more than 500 mA. And then complain about software problems.
interrobangd has joined #linux-sunxi
<beeble>
as long as they routed the connection from the connector in a useful way, yes
<KotCzarny>
The boards are amazingly fast. My laptop (dual core-i5 3.1 GHz) builds a full-featured kernel (4040 modules, 16456 compiled files) in 43 minutes. With 5 boards it's down to 11m50
<KotCzarny>
o.O
<KotCzarny>
what are those boards based on?
NiteHawk has quit [Quit: Bye.]
<tkaiser>
KotCzarny: RK3288. And his laptop obviously relies on a HDD ;)
<KotCzarny>
still, it's arm, and i was under impression x86 is much faster than arm mhz-to-mhz
<apritzel>
KotCzarny: the A17 is quite a beast in terms of performance
<KotCzarny>
a17, hum.
<KotCzarny>
looks like.
<apritzel>
rumour has it it's one of the fastest v7 silicons
<KotCzarny>
now i want it, damn you
<tkaiser>
apritzel: But the average user prefers slower A53 since... twice the bits!
<wens>
yup, that's the board i mentioned some months ago :p
<apritzel>
tkaiser: and twice the registers ;-)
<tkaiser>
apritzel: The average user doesn't know about registers. Only bits and MHz ;)
<KotCzarny>
is a15 faster than a17 or not?
<apritzel>
KotCzarny: presumably not
<apritzel>
also A17 can be clocked higher, because it's more efficient
<KotCzarny>
sad thing they didnt even set high goal
cnxsoft has quit [Quit: cnxsoft]
<beeble>
tkaiser: then they use bad connectors. you have 500mV of margin from the usb spec. so by ohms law that thats 250mOhms you should not exceed. a connector should have 30mohms worst case. take that by 4 (return path and jack), leaves you with 120mOhm. an awg24 cable has 80mOhm/m. so with a 50cm cable you are fine
<beeble>
if your power supply is not increasing its output voltage to compensate the voltage drop on the cable
fkluknav has joined #linux-sunxi
<beeble>
so get one with a fixed cable instead a usb plug
<tkaiser>
beeble: You're an engineer. That prevents you from understanding the real problem: Users and their expectations ;)
<tkaiser>
Users don't know Ohm's law, they use their fake cellphone charger and crappy cables and flood eg the Armbian forum complaining about an instable distro ;)
<beeble>
yeah, also i now user. they want to pay shit. so you put the cheapest connector there thats available
paulk-collins has quit [Quit: Leaving]
<beeble>
thats micro usb
<beeble>
you get what you pay for
<beeble>
(btw, my boards do have barrel connectors)
<tkaiser>
beeble: Yes, Micro USB is encouraging users to run into troubles. See Pine64 forums for example. So many people eg complaining the 'LCD OS images don't work'. But it's just the additional LCD backlight that causes voltage drops too high)
<beeble>
yeah i fully understand. but they are all sell and forget vendors
<beeble>
if you don't support your customers anyway you don't really care
<tkaiser>
beeble: True. But from my perspective (being dumb enough to do unpaid 1st level support from time to time) boards with Micro USB are simply support nightmares. Since people don't *want* to understand that the problems they run into are related to PSU or power cable.
<beeble>
and if it doesn't work that well he may buy the next version. because that may work, and hey its so cheap anyway! :)
NiteHawk has joined #linux-sunxi
mzki has quit [Ping timeout: 256 seconds]
<beeble>
ah, and if you don't use a micro usb connector user complain if there isn't a psu bundled. if you do that they complain about the price again. so whatever you do it's wrong
<beeble>
best not to talk with other people :)
<KotCzarny>
:)
<tkaiser>
beeble: Well, at least with Orange Pis it seems to work. All models use the same barrel plug and could not be powered through OTG port (OPi Zero being the exception).
<tkaiser>
And they bundle a 5V/3A PSU that's not that bad for $4 more.
<wens>
you get the PSU with set A or sth
<KotCzarny>
but at least it will do that A, unlike phone chargers
lamer14836261100 has joined #linux-sunxi
lamer14836261100 has quit [Client Quit]
IgorPec has joined #linux-sunxi
tkaiser has quit [Ping timeout: 246 seconds]
<Pepe>
Do we know when OPI Zero 2 or OPi 2E plus will be released?
tkaiser has joined #linux-sunxi
msevwork has quit [Quit: Leaving]
perr has quit [Quit: Leaving]
<tkaiser>
Pepe: OPi Plus 2E is already released but AFAIK currently out of stock (same with OPi Lite)
<Pepe>
Ohh.. my mistake. I meant Orange Pi PC 3 or Orange Pi PC2 plus
<tkaiser>
'PC2 Plus'?!
<Pepe>
yes same as Orange Pi PC2 with eMMC ?
premoboss has quit [Ping timeout: 248 seconds]
<Pepe>
or it wont be?
<tkaiser>
Pepe: no idea. If Xunlong is following their usual release schedule then we'll see something like that in a few months. Maybe not. And regarding OPi PC3 it seems they want to use the name now for an A64 based board :(
<Pepe>
nooooo another c**p :/
paulk-blaze has joined #linux-sunxi
<Pepe>
So I think I'm going to buy PC 2, I was waiting for Pi PC 3 or Pi PC2 plus.. or I m still interesting in OPI Zero 2
<Patsie_>
KotCzarny: and which android version might that be for the OPi PC+2E (and where can I download it?)
<KotCzarny>
4.4.2
Patsie_ is now known as Patsie
<apritzel>
montjoie: "made" as in: "it's done now"?
<apritzel>
montjoie: so do you have something that somehow works?
<Patsie>
KotCzarny: any donwload link available?
<montjoie>
apritzel: nothing for the moment
tlwoerner has quit [Quit: Leaving]
ErwinH has quit [Remote host closed the connection]
ErwinH has joined #linux-sunxi
ErwinH has quit [Ping timeout: 248 seconds]
jernej has joined #linux-sunxi
my123 has quit [Ping timeout: 272 seconds]
my123 has joined #linux-sunxi
my123 has quit [Changing host]
my123 has joined #linux-sunxi
gzamboni has quit [Read error: No route to host]
ErwinH has joined #linux-sunxi
\\Mr_C\\ has quit [Quit: .]
ErwinH has quit [Ping timeout: 240 seconds]
Pepek has joined #linux-sunxi
<Pepek>
KotCzarny: Sorry, you didnt scare me. But my connection dropped. I meant charger / adapter (it should be same)
<Pepek>
KotCzarny: OT: Android version will there be Android image with atleast version 5.x.x ?
<KotCzarny>
pepek: well, adapter could mean passive barrel-to-usb adapter, buf if you get power brick from xunlong its quite good
fkluknav has quit [Ping timeout: 258 seconds]
paulk-blaze has quit [Quit: Leaving]
<tkaiser>
Pepek: You could take an Android image for BPi M3 for example (5.1) and then replace u-boot, kernel and this sys_config.fex stuff and see whether that works.
<KotCzarny>
tkaiser: that assumes decoder libs working with older chips
<tkaiser>
KotCzarny: in other words: 4.4 will be highest Android version for H3 since Allwinner decided so?
<KotCzarny>
tkaiser: well, its easier to reuse things, than reimplement. dont remember which droid required kernel 3.10 at least
<tkaiser>
KotCzarny: BPi M3 is based on A83T, Allwinner's BU2 is reponsible for this SoC. Kernel 3.4.39 and Android 5.1. BU3 responsible for identical H8 SoC also uses kernel 3.4.39 but only 'allows' Android 4.4 (they check the chipid). No technical reasons involved. It's all just business unit madness
<beeble>
KotCzarny: lollipop was 3.10
<KotCzarny>
lollipop is which one?
<beeble>
4.4
<KotCzarny>
tkaiser: if the sources are available it should be possible to fix that
<beeble>
thats for aosp version
<tkaiser>
KotCzarny: There are no sources available. Remember Xunlong's Android github repo? DCMA takedown?
<beeble>
so you can have lower versions in android release. but 3.10 the aosp "requirement"
<beeble>
s/lower/lower and higher
<tkaiser>
beeble: While Allwinner used 3.4.39 at that time (and still with their tinalinux crap)
<dgp>
Kernel isn't the biggest issue with android.. surfaceflinger etc are harder to get working
<dgp>
the hal stuff for bluetooth, NFC is also a nightmare
fkluknav has joined #linux-sunxi
<tkaiser>
KotCzarny: Jon Smirl put the A64 Android sources online: https://github.com/pine64-android?tab=repositories (and that's the reason there exist now community Android builds with 6.0, 7.0 and 7.1 for A64 devices)
<KotCzarny>
i think moeicenowy managed to run 7.0 on a33
<KotCzarny>
anyway, my droid image for opi/h3 works nicely, even on emmc
<KotCzarny>
mainline uboot makes it easy to customize/move to new board/disk
<KotCzarny>
most important things is ability to run on almost any realtek/mediatek dongle and google store
<Pepek>
and Kodi ;-)
my123 has quit [Ping timeout: 245 seconds]
<KotCzarny>
yes
<KotCzarny>
so 4.4.2 is enough to have usable media os
<Pepek>
For Kodi 17 I think you need Android 5.x.x
<KotCzarny>
dont remember what version was installed, but the one available from google store installed and ran
<Pepek>
That's why I was asking about android version
deskwizard has quit [Quit: Leaving]
deskwizard has joined #linux-sunxi
paulk-collins has joined #linux-sunxi
<beeble>
btw, has anyone experience with register on baidu with an non mainland number? i know there is the overseas=1 option, but knowing it works at the end increases my willingness to enter a phone number :)
<montjoie>
apritzel: opipc, I have begin to bisect
<apritzel>
montjoie: I meant the compatible string on the MMC node ...
<montjoie>
apritzel: allwinner,sun7i-a20-mmc
ErwinH has quit [Ping timeout: 240 seconds]
<apritzel>
montjoie: and you are sure that it's not the SD card which is broken?
viktor has joined #linux-sunxi
<montjoie>
no reverting to last kernel work
<viktor>
Hi! I spent a lot of time working on some issues related to Allwinner updater. There is no source code for axf file responsible for firmware update. Can I somehow obtain source code? Or it's better to decompile axf file?
<KotCzarny>
viktor: you are talking about something different than allwinner ?
<KotCzarny>
wth is axf?
<viktor>
It's file format used by eGon Allwinner bootloader.
<apritzel>
montjoie: btw: I think the "update clk timeout" message is a follow up error, the actual culprit is probably the data error above
<viktor>
In fact it's ELF file.
<KotCzarny>
you might've missed the principle of this channel, discarding legacy allwinner stuff and rewrite in opensource (uboot, linux kernel, device drivers (cedrus))
<viktor>
Ok. Did someone rewrited Allwinner updater from eGon bootloader?
<andromeda-galaxy>
What is the status of power management on sun5i in mainline? (hard bound of >=4.4 kernel for my application)
<KotCzarny>
viktor: if your box is compatible with armbian just go for it (armbian.com). assuming you want linux
my123 has joined #linux-sunxi
my123 has quit [Changing host]
my123 has joined #linux-sunxi
<montjoie>
apritzel: probably:)
<viktor>
KotCzarny: We have some existing linux system with our software on top. Devices have to update in field. It is achieved by means of updater from eGon bootloader. However we have some issues with this updater. And structure of firmware produced by "dragon" is unknown,
<dgp>
viktor: I'm not sure if someone has done what you want but it seems allwinner ship a lot of the sources for the internal boot roms etc with their SDK drops.. if that stuff is included in there it should be possible to work out how it works
<viktor>
Actually it's not a ROM. It's executable file used by eGon bootloader. I inspected a few of SDKs on Github. This application has to be located in folder like this one
<dgp>
viktor: I think one way forward would be to suffer with the closed tools to put out an update that adds your own update method in
apritzel has quit [Ping timeout: 272 seconds]
<dgp>
viktor: so you can drop the closed tools after that point
<KotCzarny>
since you are using own platform, you would probably have to tweak some existing config
<viktor>
dpg: Nice point. I'm working on it. However, we already have a bunch of devices with old update scheme. So we are thinking on maintaining backward compatibility.
Guest11822 has quit [Remote host closed the connection]
russell-1 has joined #linux-sunxi
<dgp>
That's why you do your best with the closed tools to release on last update in the old format that enables the new update route
<KotCzarny>
maybe sdcard based updater
<viktor>
KotCzarny: I'll read these 3 articles. Thanks.
<dgp>
KotCzarny: it sounds like it updates the SD card from the SD card already
<KotCzarny>
dgp: he might want to use nand, which would require more trickery
<viktor>
dpg: Yes, it updates the SD card from the SD card.
<viktor>
No, I don't need NAND.
<KotCzarny>
viktor: if you just use sdcard, just release image to write with dd/some windoze equivalent
<dgp>
viktor: the filesystem etc on the SD card are normal after the update stuff has happened? It's just the update file itself that is a weird format?
<viktor>
dpg: So you suggest to leave old boot1 and perform further updates from U-boot?
<dgp>
I think doing the updates from within an initramfs would be easier
<KotCzarny>
viktor: think how normal linux distros update
deskwizard has quit [Ping timeout: 240 seconds]
<KotCzarny>
via either in-os tools, or via installer media
<KotCzarny>
or just writing ready made image
<viktor>
So it's just an update from running Linux? But how about recovery in case of failed update?
<KotCzarny>
backup the sdcard?
<dgp>
viktor: how does the current tool do it? do you have multiple copies of everything?
<viktor>
KotCzarny: we have closed housing and update has to be made without user intervention.
<dgp>
My own setup on SPI flash has two FIT images with a kernel, device tree and rootfs.. u-boot boots the newest one and passes in which it booted so the kernel can overwrite the older image and reboot into it
<KotCzarny>
viktor: there is also something calleed fel boot, in case something goes wrong you can load/run linux in ramdisk that way
<KotCzarny>
for my android image i've added ramdisk based second os (~16mb total) that can fix things/update/rewrite everything
<KotCzarny>
if you know linux its not hard to create such setup
<dgp>
KotCzarny: Recovery image?
<viktor>
Is it hard to use U-boot to handle updates? Because U-boot can leave unchanged and handle failures.
<KotCzarny>
dgp: more like rescue linux
<KotCzarny>
viktor: uboot is just grub/lilo (simplifying)
<KotCzarny>
from uboot you can load multiple operating systems
<dgp>
viktor: the problem with doing it in u-boot is that the drivers it has for hardware, filesystems etc are the bare minimum to load an OS
<KotCzarny>
it can be scripted to load reocvery when user keep pressing some button for example
<viktor>
KotCzarny: Like ramdisk, that you was talking about?
<KotCzarny>
yes
mzki has joined #linux-sunxi
<viktor>
> <dgp> viktor: the filesystem etc on the SD card are normal after the update stuff has happened? It's just the update file itself that is a weird format?
<viktor>
Filesystem has a few different volumes. Even separate one for boot1.
my123 has quit [Max SendQ exceeded]
<KotCzarny>
viktor, with mainline uboot/kernel you can just use 1-2 partitions
<KotCzarny>
and you dont need boot0/boot1 anymore
LiberalCarrot has joined #linux-sunxi
<KotCzarny>
but that's assuming you add your device to uboot/kernel
LiberalCarrot has left #linux-sunxi [#linux-sunxi]
<viktor>
I know. Previously I worked with Linux 4.1 for other device. SD card has only SPL, U-boot, DTS, Linux kernel and rootfs.
<KotCzarny>
for convenience you might use separate partition for /boot
<viktor>
What I need to add my device to uboot/kernel?
<KotCzarny>
pins/clocks configuration, try using some other a13 based device as a base
massi has quit [Quit: Leaving]
<KotCzarny>
if you are lucky it might even work unchanged
my123 has joined #linux-sunxi
<viktor>
Otherwise I will have to extract settings from Allwinner BSP, if I'm not wrong. Because datasheet for A13 is not available.
<KotCzarny>
you should read wiki more, it's a treasure trove of information
my123 has quit [Changing host]
my123 has joined #linux-sunxi
<viktor>
I saw this page previously. And the documents doesn't describe registers or periphery.
<KotCzarny>
as i've said, try basing on some other a13 device with similar config
deskwizard has joined #linux-sunxi
<viktor>
Ok. Thanks for suggestions. I will discuss the way you 2 proposed: to dismiss current update schematic and to move to open-source U-boot and kernel.
premoboss has joined #linux-sunxi
<KotCzarny>
and remember, no one here loves boot0/boot1
apritzel1 has quit [Ping timeout: 258 seconds]
<KotCzarny>
and phoenix tool is used only to grab new firmwares contents
<KotCzarny>
here are the mapping if you arent sure what you have
premoboss has quit [Ping timeout: 245 seconds]
<andromeda-galaxy>
KotCzarny: I had taken a quick look at those pages, but didn't see much about powersavin
<andromeda-galaxy>
KotCzarny: I am not even clear on what the hardware provides in terms of powersaving. and all that the mainlining effort links to is one u-boot/linux patch that implements a psci_suspend, but it doesn't talk that much about (1) what are other possibilities for power saving and (2) how much exactly is still on
<KotCzarny>
i think if something doesnt use particular device, its turned off
|Jeroen| has joined #linux-sunxi
<andromeda-galaxy>
so is intra-soc clock gating enabled?
<andromeda-galaxy>
and what levels of deep/light sleep are available?
<KotCzarny>
dont think there are any sleep states
<andromeda-galaxy>
since A13 ran android at one point, I would assume that there is some hardware at least supported way... And there is that psci_suspend patch, which disables PLLs, downclocks the soc and puts the ram in self-refresh
<andromeda-galaxy>
I think it should also be possible to put the dram in self-refresh and then fully power down the soc but I can't find any code to do that anywhere
<andromeda-galaxy>
I have some hardware available to test and would be up to try to help implementing better powersaving, also...
<KotCzarny>
android uses legacy uboot/kernel
<KotCzarny>
you might check the code there and port to mainline
<andromeda-galaxy>
right, I guess I have two questins (1), what does the hardware support, (2) what does the mainline kernel support, (3) is anyone working on making the mainline kernel support more of the hardware features, and how can I help
<andromeda-galaxy>
*three questions
<andromeda-galaxy>
I have trouble reading the A13 manuals, as they just seem to have lots of lists of PM registers without any description
<andromeda-galaxy>
so even knowing what thehardware supports is kind of difficult
<KotCzarny>
with newer socs one could also write controller app to run on openrisc core, which can reenable things (so you can turn off almost everything)
<andromeda-galaxy>
hmm, interesting. I am running on A13/R8 which I think does not have that support?
<KotCzarny>
yeah, it was introduced later
<jernej>
andromeda-galaxy: AFAIK one of the issues is that there is no code to put DRAM in self refresh mode which is one of the must have things for suspend to RAM mode
<andromeda-galaxy>
jernej: hmm, I believe I saw something about that floating around on the uboot mailing list
<jernej>
I guess "floating around" means not good enough? Also if you talk about U-Boot this could only mean psci interface?
<andromeda-galaxy>
yes, psci. is there a disadvantage to psci in this context?
<jernej>
not really sure, I don't know much about this mechanism
<jernej>
but I think that (almost) none of sunxi driver contains power management code
<andromeda-galaxy>
overall it seems like the old mach-sunxi specific pm support from allwinner hasn't realyl been ported to mainline-sunxi. Is that something that is being worked on?
<andromeda-galaxy>
hmm, the old allwinner one or the new mainline one?
<jernej>
you probably want support on mainline kernel, I guess?
kivutar has joined #linux-sunxi
<andromeda-galaxy>
yes, I need >4.4, but could maybe slowly try to help port some of the old code if it exists and ideally if there is any docs for it
<andromeda-galaxy>
but I wanted to make sure that I know what is going on so far in terms of that
<jernej>
I'm not sure if code could be ported
<andromeda-galaxy>
why not?
<jernej>
As I said, I'm not expert in power management, but from what I saw, PM hooks takes care for disabling/reenabling clocks and sometimes also for reinitializing registers
<jernej>
Most drivers are pretty different from those from BSP kernel
<KotCzarny>
andromeda-galaxy: if you manage to make even some devices pm aware, it would be great, as it's usually not the priority
<andromeda-galaxy>
jernej: I would assume that psci is a mechanism for doing that kind oft hing in a beter/more general way? like armish acpi?
<andromeda-galaxy>
KotCzarny:well, I will definitely take a look into it. I just want to make sure that I am not duplicating somebody eulse's work
<andromeda-galaxy>
KotCzarny: the openrisc that you mentioned is only on A31/H3 right?
<KotCzarny>
yeah, and above
<KotCzarny>
btw. if you want to register your work on something you use mailing list and/or mainlining page
<jernej>
andromeda-galaxy: I think PSCI is only for ARM cores, not for peripherals
<andromeda-galaxy>
jernej: ah, that may be true. I will maybe have to go find some of the upstream people to talk to
<KotCzarny>
there is psci-suspend/resume by antoine tenart
<andromeda-galaxy>
KotCzarny: cool, I will do that when I have time to start workign more on this. Maybe I will also emain Antoine Tenart who made the other slight psci changes
<andromeda-galaxy>
last thing, anybody knwo where I can find info on what the hardware supports, apart from reverse engineering the old 3.4 sunxi drivers?
<KotCzarny>
soc's documentation pdfs?
<jernej>
in the datasheets?
<jernej>
but of course, they are not complete
<andromeda-galaxy>
datasheets/manuals pms sections for a13/r8 just seem to have a long list of registers with acronym-heavy opaque descriptions
<KotCzarny>
yeah, and its all docs you can expect from allwinner
<andromeda-galaxy>
not like the intel/amd manual sthat I am used to where they actually say: the soc supports a, b, c, states. You do this to get into a, state, etc...
<andromeda-galaxy>
I was wondering if the community had already figured out what some of that was supposed to mean
<KotCzarny>
your best bet would be checking vendor provided software
<andromeda-galaxy>
like what does "DVFS Mode Select: mode 0, mode 1, mode 2" mean?
<andromeda-galaxy>
alright, will take a look at the old allwinner bsp sources
<andromeda-galaxy>
thanks for all the info! I may be back in here once I start trying to run pm code and have things fail to work ;-)
<jernej>
andromeda-galaxy: Did you actually try that U-Boot patch?
<andromeda-galaxy>
jernej: currently finishing getting a build env re-set-up and trying to build that, yeah. I think that the SoChas more power features than that though ~.5 W for idle without support for keeping some softwarse running during idle seems high to me...
<andromeda-galaxy>
btw what is the point of putting the dram in self-refresh ift he SoC is still powered on (albeit at a low clock frequency)?
<jernej>
If it works, it may be that just peripheral units are active. Currently they aren't disabled at suspend. Try to unload as many drivers as possible and try again.
<KotCzarny>
or add pm functions to the drivers
<andromeda-galaxy>
cool, will try that
[Awaxx] has quit [Quit: "*ButterFlyEffect*"]
<andromeda-galaxy>
btw what is the point of putting the dram in self-refresh ift he SoC is still powered on (albeit at a low clock frequency)?
<jernej>
andromeda-galaxy: In theory, cores should be completely halted during suspend
<jernej>
or not
<andromeda-galaxy>
jernej: right, but if the core isn't being halted, what does it add? the uboot patch disables plls (makes sense), puts the dram in self refresh (would make sense if not for the next bit), downclocks thec ore and leaves it powered on
<KotCzarny>
it might be preparation for some more patches
<andromeda-galaxy>
is loclk actually so slow that the soc can't refresh the dram?
<KotCzarny>
author should know
<andromeda-galaxy>
yeah, I'll go send him an email I guess.
<jernej>
best ask author
<andromeda-galaxy>
well thanks for all the info on status! To recap for the future: sounds like (1) hardware supports it, but nobody really knwos how much or what modes, (2) kernel doesn't support it at all and (3) there is not much "official" work inside linux-sunxi on it (basically I think nothing except those patches from Antoine Tenart)
<KotCzarny>
there isnt anyone doing work on linux-sunxi with any particular plan in mind
<KotCzarny>
everything starts by a patch from someone
<andromeda-galaxy>
true enough. I was curious if the community had sort of self-organized into major units working on mainlining things, thoguh, and it would seem that pm is not one of those groups, if they exist
<KotCzarny>
but once you get a driver done, you usually become a kind of maintainer
<KotCzarny>
for example monjoie is security engine expert
<andromeda-galaxy>
ah. I will keep that in mind.
<andromeda-galaxy>
thanks for all the info, KotCzarny and jernej!
<KotCzarny>
its not official, but such person knows more on particular topic than everyone else
apritzel has joined #linux-sunxi
<montjoie>
i bisect between 4.9 and 4.10rc1 and the bad commit is in 4.9rc2. I dont understand at all
<KotCzarny>
you cant understand bugs, often they just work, but break when fixed
interrobangd has quit [Ping timeout: 240 seconds]
p_rossak_ has joined #linux-sunxi
p_rossak has quit [Ping timeout: 260 seconds]
ErwinH has joined #linux-sunxi
yann-kaelig has quit [Quit: Leaving]
ErwinH_ has joined #linux-sunxi
[Awaxx] has joined #linux-sunxi
ErwinH has quit [Ping timeout: 240 seconds]
ErwinH_ has quit [Ping timeout: 240 seconds]
KB3VGW has joined #linux-sunxi
KB3VGW has left #linux-sunxi [#linux-sunxi]
KB3VGW has joined #linux-sunxi
KB3VGW- has joined #linux-sunxi
KB3VGW- has left #linux-sunxi [#linux-sunxi]
<apritzel>
bisecting within the merge window is weird sometimes
terra854 has quit [Quit: Connection closed for inactivity]
KB3VGW has quit [Ping timeout: 258 seconds]
ErwinH has joined #linux-sunxi
KB3VGW__ has joined #linux-sunxi
KB3VGW__ has quit [Client Quit]
ErwinH has quit [Ping timeout: 240 seconds]
lemonzest has quit [Quit: Leaving]
netlynx has quit [Quit: Ex-Chat]
foxx has quit [Ping timeout: 255 seconds]
ErwinH has joined #linux-sunxi
ErwinH has quit [Ping timeout: 256 seconds]
Pepe has joined #linux-sunxi
jernej has quit [Ping timeout: 248 seconds]
ErwinH has joined #linux-sunxi
russell-1 is now known as russell--
ErwinH has quit [Ping timeout: 240 seconds]
|Jeroen| has quit [Quit: dada]
deskwizard has quit [Changing host]
deskwizard has joined #linux-sunxi
jernej has joined #linux-sunxi
IgorPec has quit [Ping timeout: 240 seconds]
Ntemis has quit [Remote host closed the connection]
ErwinH has joined #linux-sunxi
ErwinH has quit [Ping timeout: 240 seconds]
likewise has joined #linux-sunxi
ErwinH has joined #linux-sunxi
ErwinH has quit [Ping timeout: 240 seconds]
interrobangd has joined #linux-sunxi
paulk-collins has quit [Remote host closed the connection]
ErwinH has joined #linux-sunxi
jernej has quit [Ping timeout: 248 seconds]
HeavyMetal has quit [Quit: BNC Services Provided by the ASoTnet IRC Network.]
HeavyMetal has joined #linux-sunxi
HeavyMetal has joined #linux-sunxi
HeavyMetal has quit [Changing host]
ErwinH has quit [Ping timeout: 240 seconds]
ErwinH has joined #linux-sunxi
ErwinH has quit [Ping timeout: 240 seconds]
AneoX has joined #linux-sunxi
likewise has quit [Quit: ChatZilla 0.9.93 [Firefox 50.0.2/20161201160049]]
utente_ has quit [Ping timeout: 248 seconds]
<AneoX>
HI! Please help! I working with A20 sunxi kernel, mali_r3p2-01rel2 and my Qt app. I have found some linux distr (3.4.79, but undefined kernel source), there my app works with 60 fps well. But when i am trying to build own kernel with mali_r3p2-01rel2, 3.4.75, 3.4.105 or any another version, qt app send warning on launch "QFbVtHandler: socketpair() failed (Address family not supported by protocol)" and works only 40-45 fps. What i am
Mr__Anderson has quit [Remote host closed the connection]
cptG has joined #linux-sunxi
cptG_ has quit [Ping timeout: 246 seconds]
<MoeIcenowy>
mripard: could you push your WIP A13 CSI code somewhere?
florianH has quit [Quit: Connection closed for inactivity]