ChanServ changed the topic of #linux-sunxi to: Allwinner/sunxi development discussion - Don't ask to ask. Just ask! - See http://linux-sunxi.org | https://github.com/linux-sunxi/ | Logs at http://irclog.whitequark.org/linux-sunxi
hramrach has quit [Ping timeout: 276 seconds]
<mnemoc> -* Descript: power manager for allwinners chips platform.
<mnemoc> +* Descript: power manager for newbies chips platform.
<mnemoc> that's weirder than reuuimlla...
<mnemoc> arch/arm/mach-sun5i/pm/pm.c of a13_source_1.5
user135 has joined #linux-sunxi
hramrach has joined #linux-sunxi
<user135> Hi Turl, for a serial console, would I need to open up my a13_mid?
<user135> Or are there usb adaptors available?
<Turl> there's often an uart on the uSD slot
<Turl> you need a breakout board and some script.bin configuration
<mnemoc> isn't hipboi selling clones of allwinner's breakout?
<Turl> may be, idk
<user135> so how does that sd to microsd card plug into my pc?
<mnemoc> with a uart/usb adapter
<mnemoc> ttl
<user135> okay, let me look
<user135> ok, so can I emulate with this a microsd card & get console output?
<user135> if I get both a uart/usb adapter and card converter module
<user135> ?
<mnemoc> you can replace nanda's u-boot with one that configure the uSD pins as uart/jtag, connect the breakout and the uart/usb adapter, and get a serial console. yes
hramrach has quit [Ping timeout: 276 seconds]
<user135> hmm, that seems a bit excessive for what I'm trying to do maybe, I'll explain:
<user135> I have a recovery.img file which I compiled
<mnemoc> 2:30... need to go to sleep now. keep poking Turl
<mnemoc> good night
<user135> lol ok
<rz2k> good night mnemoc
<user135> Thanks anyway
<user135> good night
<rz2k> anyone worked with i2c or spi on sunxi'es?
<theOzzieRat> rz2k, i am right now
<theOzzieRat> spi that is
<rz2k> does it envolve ton of chinese shit-code?
<rz2k> or is it standard-wired to linux api for it?
<theOzzieRat> not a ton, but not real good
<theOzzieRat> it is a driver that is used by the standard linux spi subsystem
<rz2k> I have a guy who wants to connect an spi or i2c for data reading from external MCU
<rz2k> not really sure what to answer him
<rz2k> oh
<theOzzieRat> i have it working talking to an external atmel mcu
<rz2k> then its ok
<rz2k> cool
<theOzzieRat> at present it only supports half duplex
<theOzzieRat> i have found a couple of bugs with it and am putting together a couple of patches
<user135> Which kernel defconfig should I be using for a13_mid? Is it a13_nuclear?
<rz2k> great
<rz2k> we have only one available
<theOzzieRat> after that I will make it handle full duplex, I have done a quick hack that has it working full duplex for transfers <64 bytes
<rz2k> for it
<rz2k> theOzzieRat: that guy who contacted me is doing some sort of 'clever house' systems and wants to use cheap sunxi tablets as a GUI mounted in walls
<rz2k> as far as I understood
<theOzzieRat> that sounds like a good use. mine is being a web server connected to a board I built that collects data from my weather station
<rz2k> dont you mind if I ask your contact info in PM? he is money-interested in this, also I'm not a best developer in the world, so I might give up at some point :p
<theOzzieRat> no problem
hipboi has joined #linux-sunxi
<theOzzieRat> can anyone tell me how many CS's there are on each port of the A10 and A13?
bfree_ has quit [Ping timeout: 252 seconds]
bfree__ has joined #linux-sunxi
hramrach has joined #linux-sunxi
<Turl> theOzzieRat: on SPI?
<Turl> theOzzieRat: according to the docs, SPI0 has 4 chip selects and SPI1 has 1
<theOzzieRat> ahh, yes. missed that
<user135> Turl, is there anything specific I need to do for the a13_mid if I'm using this: https://github.com/allwinner-dev-team/android_device_allwinner_common?
<user135> I followed the instructions here: http://linux-sunxi.org/Starting_a_CyanogenMod_device_tree
<theOzzieRat> Turl, which doc was that? The docs I have read say that SPI0-3 all have CS0 and CS1. And the spi_sunxi.c is different again.
<Turl> user135: you need a device tree like _zatab or _mk802ii ones
<Turl> the example repo got outdated
<Turl> you'll be better off cloning the zatab one and modifying it to fit
<user135> ok
<Turl> theOzzieRat: google 'a10 datasheet', first result
<theOzzieRat> I think that overview is wrong in section 15.1. In the same datasheet under the multiplex functions all SPI ports have CS0 and CS1. The overview also doesn't mention SPI2 or SPI3
<rz2k> allwinner datasheets have errors in DDR3 setup
<rz2k> what do you want them in SPI :p
<theOzzieRat> Just trying to confirm it before I submit a patch to fix spi_sunxi.c
<theOzzieRat> rz2k, talking about SPIx_CSy
<user135> okay, Turl, is there any specific changes I need to make to ZaTab for a13_mid besides the defconfig?
rz2k has quit []
<Turl> user135: well, I don't know what your mid has, you might need to adjust the sensor definitions, partition layout, recovery keymapping, etc
<user135> ok
<user135> well I'm only making a recovery image, but I guess I won't know the keymappings until I test
user135 has quit [Quit: Leaving]
ZaEarl has quit [Ping timeout: 246 seconds]
user135 has joined #linux-sunxi
<user135> Turl, I've made some progress I think, the screen now lights up
hipboi has quit [Ping timeout: 252 seconds]
hipboi has joined #linux-sunxi
user135 has quit [Quit: Leaving]
ZaEarl has joined #linux-sunxi
anunnaki has quit [Remote host closed the connection]
anunnaki has joined #linux-sunxi
rellla has joined #linux-sunxi
anunnaki has quit [Remote host closed the connection]
Guest77256 is now known as fredy
fredy is now known as Guest30019
Guest30019 is now known as fredy
anunnakiii has joined #linux-sunxi
rellla has quit [Remote host closed the connection]
anunnaki has joined #linux-sunxi
user135 has joined #linux-sunxi
paulk-desktop has joined #linux-sunxi
<paulk-desktop> hi there
<libv> i will be deprecating mali-libs and libump today.
<paulk-desktop> any way to get u-boot prompt on sun5i tablets.
<paulk-desktop> ?
<libv> just had someone mail the lima ml where he first explained how he installed mali-libs :(
<libv> paulk-desktop: uart.
<paulk-desktop> libv, where are the connectors?
<libv> paulk-desktop: that's the million dollar question.
<libv> paulk-desktop: every device is different.
<libv> paulk-desktop: if it is not documented on our wiki, then you are on your own.
<libv> notice the 's' at "A13_Tablets"
<paulk-desktop> libv, I'd like to understand a bit better all this:
<paulk-desktop> basically, A13 is a SoC made by Allwinner
<libv> *sigh*
<paulk-desktop> but the tablets are all unbranded?
<libv> paulk-desktop: SoC != device
<paulk-desktop> libv, but A13 is the SoC, not the device?
<libv> yes
<paulk-desktop> so no misunderstanding there
<paulk-desktop> A13B/Q88 are the tablets
<paulk-desktop> libv, so I guess each tablet has a separate board file in the kernel?
<libv> now, if you understand that, what is the difficulty in understanding that there are multiple devices out there, from various manufacturers, some with branding, some rebranded, some without branding
<libv> yes
<libv> with wildly different board layouts as well
<paulk-desktop> libv, then I'm interested in knowing which tablets have public board files and defconfigs :)
<libv> paulk-desktop: linux-sunxi.org
<libv> paulk-desktop: if you had gone through FirstSteps, you would know
<paulk-desktop> should I assume the ones from http://linux-sunxi.org/Category:A13_Tablets are supported by the linux-sunxi kernel on github?
<paulk-desktop> ah sorry then
<libv> should be, yes
<libv> unless marked differently in the respective wiki-page, but beware, wikis are never 100% correct and up to date
<paulk-desktop> libv, if I buy a Q88 tablets on ebay, should I assume that all the Q88 tablets I can find are the same to this one?
<libv> no
<paulk-desktop> or are there differences even between tablets of the same model
<libv> paulk-desktop: we are not the makers of this hw
<paulk-desktop> I see
<libv> paulk-desktop: board/device makers do crazy things
<libv> there is a high probability though
<libv> but no assurance.
<paulk-desktop> so a Q88 tablet I buy may, in facts, not be suitable for the kernel you release?
<libv> what are you trying to achieve?
<paulk-desktop> libv, I'm a Replicant (fully free Android distribution) developer and my problem is that we release prebuilt images, so if the hardware changes between tablets, it's not suitable for replicant
<paulk-desktop> basically, we won't tell our users: rebuilt a kernel with your particular tablet config
<paulk-desktop> rebuild*
<libv> paulk-desktop: kernel is not the issue
<libv> which you would've known if you had run through FirstSteps
<libv> paulk-desktop: uboot and script.bin are device specific
<paulk-desktop> ah then this is an highly uncommon process then :)
<libv> really?
<user135> not really, I wrote a script to do it
<libv> i expect DT to be a similar but more widespread system for the same
<paulk-desktop> libv, other android devices come with one kernel per board
<paulk-desktop> best case scenario, one defconfig per board
<libv> paulk-desktop: DT is devicetree
<paulk-desktop> libv, yes but nobody uses DT for now
<paulk-desktop> it's just beginning
<paulk-desktop> current Android devices are still released with 1 kernel and 1 defconfig per board
<paulk-desktop> (which is the usual case for embedded ARM devices)
<libv> paulk-desktop: well, we have 1 defconfig for sun4i and 1 for sun5i
<paulk-desktop> that's pretty good
<libv> and then we have the uboot and script.bin differences
<paulk-desktop> libv, are the u-boot sources for all tablets released?
<paulk-desktop> it's GPLvsomething IIRC
<libv> it needs device specific info
<libv> to set up dram
<libv> we have a small tool to find out the dram settings for us
<paulk-desktop> yeah but that should have been released by the vendors
<libv> paulk-desktop: welcome to the real world.
<paulk-desktop> :)
<paulk-desktop> that's why I prefer to work with mainstream manufacturers
<libv> sure.
<paulk-desktop> at least they release what they have to release
<libv> it took long enough to convince them to do so
<libv> and they are not immediate
<paulk-desktop> yeah
<libv> what do you concretely want to do?
<paulk-desktop> libv, I would be interested in buying such a tab and helping to port replicant to it
<libv> and how does that relate to your direct uboot question you started out with?
<paulk-desktop> because I like the SoC and u-boot support
<paulk-desktop> libv, that was to help user135, who has such a tab
<paulk-desktop> libv, but I'm interested in buying one too, so I'm asking general questions as well
<user135> indeed, but I don't have GPU drivers!
<user135> :)
<libv> paulk-desktop: best chance is indeed finding a tablet that is listed in our wiki
<user135> I think that's what I'm a needing anyways.
<paulk-desktop> libv, so basically, u-boot and script.bin are already in the tab when I buy it?
<user135> yes
<paulk-desktop> but in binary for obviously
<paulk-desktop> so if I want to rebuild, I have to figure it out from binary/from the device
<user135> well, you'll need to compile uboot
<libv> paulk-desktop: you should also verify sunxi-boards for the related .fex file, which can then be turned into script.bin
<user135> and script.bin is in binary on nanda
<libv> paulk-desktop: somewhat/yes.
<paulk-desktop> libv, alright
<libv> paulk-desktop: the dram parameters for that we have a special tool
<paulk-desktop> libv, good
<paulk-desktop> so it's doable?
<libv> and those parameters go into uboot
<libv> the script.bin can be read out from the fs directly, and then converted to a .fex
<paulk-desktop> libv, is u-boot specific too or is the specific part only script.bin ?
<libv> both are specific
<libv> but we can get the uboot specific data separately
<paulk-desktop> so I won't be able to release u-boot for all sun5i boards, got it
<paulk-desktop> libv, just curiosity, how come Linux can figure out which touchscreen to use (for instance) ? cmdline?
<libv> no idea
<paulk-desktop> ok but it works?
<libv> i am a graphics driver guy, haven't bothered with touchscreen at all
<paulk-desktop> nice
<paulk-desktop> libv, mhh I think I remember seeing you around in #lima
<libv> paulk-desktop: yes
<libv> paulk-desktop: given that i started the lima thing, that's no surprise
<paulk-desktop> libv, I see :)
<user135> so say if I had a a13_mid tablet, and I had compiled u-boot, and the screen turned on but it was blank ... what could be the problem?
<paulk-desktop> I was searching for the FOSDEM presentation to figure out your name
<user135> and would there be anyway of diagnosing the problem? :/
<paulk-desktop> libv, I have to say, it's really really good to have people like you doing this
<paulk-desktop> so thanks a lot
<paulk-desktop> (as a Galaxy S2 user)
<paulk-desktop> anyway I guess I'll get a Q88 tab and see
<libv> it is a very important problem that needs a solution
<paulk-desktop> libv, one told me that you almost got rid of the non-free shader compiler
<libv> paulk-desktop: cwabbott has thrown me some shaders that he compiled with his own compiler
<libv> these shaders should enable a fully free q3a
<libv> but the road to a full, performant, and correct compiler is still long
<paulk-desktop> libv, but that quake 3 arena video -- was it with the non-free compiler or not?
<libv> non-free
<paulk-desktop> ok
<libv> but that demo only requires 2 vertex shaders, and 4 fragment shaders
<libv> and we could hand-assemble the fragment shaders
<paulk-desktop> that's pretty amazing anyway
<libv> matching performance with the binary driver is the amazing one
<libv> and this with only some general optimizations
<libv> on mali, we can have a free software driver which matches the binary
<libv> which is unique
<paulk-desktop> given that we're talking about graphics, that's truly great
<libv> user135: describe in more detail what you have done to get this display to light up
<user135> okay, it's a bit complex, but I'll explain
<user135> compile u-boot sunxi
<user135> for a13_mid
<user135> (my tab is a13b)
<libv> user135: did you run through Firststeps?
<user135> pretty much
<libv> apart from?
<user135> once I did that ...
<libv> what rootfs are you using?
<user135> replicant's recovery
<user135> so I got the replicant sources
<user135> and the common files
<user135> then I used zatab's device configs, and modified them
<user135> to use the a13_nuclear_defconfig --> paulk pointed me in the right direction here
<user135> and added uboot support
<libv> user135: so the backlight comes on?
<user135> yes
<libv> user135: have you tried adb?
<libv> i do hope that replicant enables adb per default
<user135> that's what I was thinking ...
<user135> :)
<paulk-desktop> definitely
<libv> otherwise there is just no chance to even try to debug android
<user135> let me check
<paulk-desktop> serial would be better for that
<paulk-desktop> libv, is there dmesg on serial too?
<libv> depends
<user135> no adb :(
<libv> never used serial with adb
<libv> err, s/adb/android/
<libv> user135: in theory, the backlight should not turn on until the lcd is initialized
<user135> how do I kill the adb daemon paulk, I forget
<libv> another option is to edit one of the init scripts to give you wireless out of the box
<user135> I get list of attached devices: with a blank line
<libv> android is a pain though, i worked with android-x86 half a year ago...
<libv> no usb-device, so i needed to get eth initialized at boot and then have adb enabled on that
<libv> could be that the adb problem is on the host side
<libv> paulk-desktop should be able to help you there
<paulk-desktop> user135, well killall adb should do it
<paulk-desktop> then sudo adb start-server
<user135> still no luck.
<user135> nm
<user135> libv, is it possible to get serial over wireless? Or logs?
<libv> paulk-desktop: erm, why are you not mentioning things like lsusb, udev?
<paulk-desktop> user135, do you have an UART usb thing?
<paulk-desktop> libv, lsusb in case serial is via usb
<paulk-desktop> which happens
<libv> ?
<paulk-desktop> like on gta02
<libv> paulk-desktop: are you really a "android developer"?
<paulk-desktop> libv, who said that?
<libv> user135: let me dig out the link
<user135> thanks
<paulk-desktop> libv, why this question?
<libv> paulk-desktop: his device seems to have booted
<paulk-desktop> libv, right but serial would greatly help to see what's going on
<libv> paulk-desktop: there is a chance that his device has adb enabled
<libv> paulk-desktop: but that his host just is not fully set up for adb
<libv> to whatever the device is telling the host that it is now
<paulk-desktop> libv, lsusb would let us know anyway
<user135> I don't know what it means ^^
<libv> user135: right, no device there
<paulk-desktop> libv, so serial seems relevant to me
<libv> serial is severely non-trivial
<libv> user135: not sure where Turl keeps his stuff
<libv> user135: but he is also doing android for sunxi
<libv> user135: maybe turl has something prebuilt you can use
<user135> cool
<libv> user135: another option is to try to get like a linaro rootfs to boot on your device first
<libv> user135: and then exchange the kernel, and the rootfs with what paulk-desktop provides
<libv> this would verify the first part of firststeps
<libv> i think you should also be able to just boot a linaro rootfs on an android enabled kernel, but i haven't tried this myself
<user135> I shall try
<user135> libv, which one should I go for https://wiki.linaro.org/Platform/DevPlatform/Rootfs ?
<libv> user135: work through the full firststeps
<libv> that includes that u-boot stuff you got from hno his separate wiki
<paulk-desktop> user135, maybe installing a regular android would be a good start indeed
<libv> a regular linux is what i suggested
<libv> and only then step things up to android
<libv> user135: make sure that you properly replace mmcblkx everywhere in that howto
<libv> i wrote that howto, and i always get bitten by that
<libv> when your sdcard doesn't boot, then check whether /dev/mmcblkx exists
<libv> if it does, then you probably dd'd to memory instead of to the actual sd card :)
<libv> paulk-desktop: that looks like http://linux-sunxi.org/A13B indeed
<libv> paulk-desktop: which is the same that user135 has
<paulk-desktop> good :)
<libv> and which is probably the most common a13 out there
<paulk-desktop> too bad there is no GPS, I would have used that to do OSM mapping
<user135> why not get an a10?
<user135> just a suggestion.
<user135> then we could get two devices working. :)
<libv> :)
<paulk-desktop> I'll look
<libv> paulk-desktop: i think there must be some devices with gps
<user135> The A10 I found has better support anyway compared to the A13B
<libv> paulk-desktop: be sure to add a separate page to our wiki when you have something useful
<paulk-desktop> ok
<paulk-desktop> maybe with USB host it could work too
<user135> libv, all the rootfs links are broken on linaro
<libv> user135: ok
<libv> heh, seems like noone is using this howto
<libv> grmbl-
anunnakiii has quit [Remote host closed the connection]
<user135> I'm sorry! :)
<libv> ah, the rootfs links on linaros own pages, you mean?
<libv> i have never known those to work :)
<libv> it's been like that for 5 months :)
<libv> but now my direct link broke as well
<user135> indeed
<paulk-desktop> user135, A10 tablets seem to be more expensive and to come in more vendor-specific derivatives
<paulk-desktop> I guess I'll stick with A13
<user135> ok
<user135> sup to you! :D
<libv> heh
<libv> the linaro assholes deleted precise
<libv> great for continuity.
<user135> hmm, what to do?
anunnakiii has joined #linux-sunxi
<libv> i'll fix up the pages, just give a minute while i complain on the linaro wiki
<user135> lol k
<libv> no. f them.
<libv> they should not delete known working rootfs
<libv> apt-get distupgrade should just work and get the new version as well
<libv> so f them for simply deleting old files, and therefor breaking everything tryint to link to them
<user135> awkwardness really
<libv> it's as if they do not want users.
<libv> because many users tend to give up instead of fix things or complain
<user135> I'm in the complainer category I think ;)
<user135> Well, at the moment anyway.
<libv> which is already rare today
<libv> i am in the "wtf, why do i have to go fix _that_ shit as well, when i already have a bazillion of much more important things on my todo list"
<libv> + category :)
<user135> I've asked at #linaro
<libv> user135: reload
<user135> k
<libv> there is now a rant, and then a working link
<user135> lol
<user135> cool, thanks
<user135> which one is most likely to work with an android kernel?
<libv> no idea, just try the latest
<user135> k
<libv> i do not think there is any issue with using a proper linux on an android kernel
<libv> but i do not know for sure
<mnemoc> beside android's paranoic network "feature", it's pretty normal
<libv> mnemoc: ah, good morning
<mnemoc> good morning :)
<libv> mnemoc: is it ok if i deprecate the mali-libs and libump repos, as i described before, today?
<libv> it will probably break the bsp, but we should try to get that fixed today as well
<mnemoc> the bsp uses submodules, so it's bound to certain hashes
<mnemoc> nothing against the deprecation of those modules
<mnemoc> but getting the bsp improved is always a good thing
arete74 has quit [Ping timeout: 276 seconds]
arete74 has joined #linux-sunxi
<paulk-desktop> libv, do you know if there is a first stage bootloader before uboot?
<paulk-desktop> libv, and an android kernel with gnu/linux userspace works, with small differences for Xorg (see the SHR project)
hramrach has quit [Remote host closed the connection]
<user135> libv, I cannot extract any images from here: https://snapshots.linaro.org/quantal/images/alip
anunnakiii has quit [Quit: Leaving]
<user135> tar: binary/dev/*: Cannot mknod: Operation not permitted
<user135> nm, had to use sudo to extract
bfree__ has quit [Ping timeout: 252 seconds]
bfree__ has joined #linux-sunxi
user135 has quit [Ping timeout: 260 seconds]
user135 has joined #linux-sunxi
rz2k has joined #linux-sunxi
<libv> hrm... i forgot the .gitignores on the mess that is libump...
<libv> not sure if it pays to go fix that one...
ZaEarl has quit [Ping timeout: 246 seconds]
<libv> hrm, so no-one has bothered to write up wiki things on how to set up usb-otg gadget things like networking
* libv has had this working since december
<libv> due to hansg and others' usb fixes
ganbold_ has quit [Remote host closed the connection]
<paulk-desktop> is there any uboot tree that can boot from NAND?
<paulk-desktop> libv, ^^^
<Turl> yes, the lichee* branches if I recall correctly
user135 has quit [Ping timeout: 252 seconds]
uminded has joined #linux-sunxi
<uminded> Every kernel I build and ones that I load that others have working give the error: "Unable to handle kernel NULL pointer dereference at virtual address 00000018" is this a hardware issue??
user135 has joined #linux-sunxi
hramrach has joined #linux-sunxi
servili007 has joined #linux-sunxi
<mnemoc> uminded: might be a script.bin issue
<mnemoc> identify the driver and see if it's something your device actually has. if not, disable it in script.bin
<uminded> mnemoc: how do I recompiler the script.bin back to text??
<uminded> decompile*
<hramrach> bin2fex
<uminded> mnemoc: theirs a whole bunch of things in here I don't have like hdmi and tv out, and the PPL6 freq is wrong
<uminded> the one provided must have been all possible options and not board specific
<hramrach> I don't think the script.bin frequencies are used
anunnaki has quit [Quit: Changing server]
<uminded> I got it working now
<uminded> on the wiki under Fex_Guide it says "Configuration and GPIO. Doc" do you know which one it refers to?
<uminded> wow the wiki really needs an overhaul, you need to pick through page links to find something simple
<mnemoc> go ahead :)
<user135> Does anyone know how to enable usbserial on an a13_mid?
<user135> zio, I think it's called
<mnemoc> the ftdi driver is called something zio
<mnemoc> but it's kind of weird you want usbserial on a tablet...
<mnemoc> anyhow, recompile your kernel with the stuff you need enabled
<user135> how do I go about enabling it?
<mnemoc> eh? enabling the right CONFIG_ in .config before compiling it?
<mnemoc> never compiled a kernel before?
<user135> hmm, usb serial is enabled already.
<user135> thanks anyways.
<mnemoc> nice
<mnemoc> they might use it for 3g modems
<techn_> user135: have you tried framebuffer console?
<user135> now trying it
<user135> gotta build the kernel again
<user135> and make a boot.cmd
<mnemoc> btw, to connect to the serial console of a device you need the usb driver in your computer, not on the device
<mnemoc> anyhow, not following the discussion, and time to sleep
<mnemoc> good night
<user135> night
rz2k has quit []
user135 has quit [Quit: Leaving]
lous_ has joined #linux-sunxi
paulk-desktop has quit [Quit: Ex-Chat]
lous_ has left #linux-sunxi [#linux-sunxi]
user135 has joined #linux-sunxi
<user135> Hi, does anyone know why I'm getting these errors whilst enabling the framebuffer:
<user135> error: implicit declaration of function 'acquire_console_sem'
<user135> error: implicit declaration of function 'release_console_sem'
<user135> ?
<user135> in kernel/power
<user135> consoleearlysuspend.o
<hramrach> maybe because you enabled early suspend
<hramrach> my DVD-rom died :s