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
von_fritz has quit [Quit: vonfritz leaves, don't panic]
rz2k has quit []
Turl has quit [Quit: >.<]
egbert has quit [Disconnected by services]
egbert_ has joined #linux-sunxi
Turl has joined #linux-sunxi
Turl has quit [Quit: >.<]
Turl has joined #linux-sunxi
torqu3e has quit [Ping timeout: 245 seconds]
torqu3e has joined #linux-sunxi
ZaEarl has quit [Ping timeout: 264 seconds]
hipboi has joined #linux-sunxi
wingrime has joined #linux-sunxi
<wingrime> elbrus with delay slots have "branch on cache miss"
torqu3e has quit [Ping timeout: 245 seconds]
e-ndy has quit [Ping timeout: 245 seconds]
e-ndy has joined #linux-sunxi
torqu3e has joined #linux-sunxi
torqu3e has quit [Ping timeout: 245 seconds]
torqu3e has joined #linux-sunxi
shineworld has joined #linux-sunxi
torqu3e has quit [Ping timeout: 245 seconds]
rellla has joined #linux-sunxi
carlo_ is now known as n01
torqu3e has joined #linux-sunxi
<oliv3r> reading backlog here (and maybe it comes up later) but a) what is gadget/sunxi.c and b) what is android specific about it? and c) is it hard to remove the android clutter?
torqu3e has quit [Ping timeout: 245 seconds]
<wingrime> oliv3r: "gadget" this is USB SLAVE device
<wingrime> when you connect a10 to PC
<oliv3r> ah, so the a10 device behaves LIKE a gadget TO the PC
<oliv3r> got it
<oliv3r> non-trivial :)\
<oliv3r> well, I guess i'll continue working on the Fex_Guide wiki page *yawn*
<oliv3r> p.s. script.bin speaks of a dac for the tv-out encoder. I looked at some screenshot of the melee, but it seemed that the TV out port is directly connected to the A10, so the dac is in the A10, but i can't find the pinout for the dac
torqu3e has joined #linux-sunxi
<oliv3r> though I do see 'tv out power supply on R14, and TV-out gnd on N12
hipboi has quit [Read error: Connection reset by peer]
rz2k has joined #linux-sunxi
hipboi has joined #linux-sunxi
<oliv3r> hipboi: cubieboard.org has a few bugs :p slider image 1 doesn't show up (could be my proxy, so possibly ignore that) but because of the missing image the [[1] [2] [3] etc buttons get moved over the text, if there's no image, the box should be stretched to some min-default height?
<oliv3r> anyway, you also say 'TWO flavors, black and white, and transparant, so that should probably read "three flavors, black, white and transparant" :)
<oliv3r> ah, wait, i get it, it is black+white or transparant, so it should read "black and white or transparant" i guess :)
<oliv3r> i do have to say i like hte case :)
torqu3e has quit [Ping timeout: 245 seconds]
<hipboi> oliv3r: i am considering change the style of cubieboard.org
<hipboi> how about this one: http://urlmini.tk/testversion1/
<oliv3r> I liked the old page, but the new page looks nice too, i like the use of black, red and the font, makes it look '80's style' maybe a a bit 'flashy' it bounces up; but it's good
<hipboi> :0
<oliv3r> hipboi: could you help me translate a little? http://sprunge.us/decd there is a section 'SDMMC pins mapping'. I could ifugre out 1 and 2 with google translate, but 3 'gpio detection must be configured sdc_det detect pin-pin port' just doesn't make sense :)
<oliv3r> well the image below the banner, it reminds me of a 80's IBM computer add; very cool
torqu3e has joined #linux-sunxi
<hipboi> oliv3r: 3 is: if you use gpio as detecting method, you must config sdc_det
<oliv3r> oh, it's simpler then I thought :) thanks :)
hansg has joined #linux-sunxi
<oliv3r> hipboi: also, you don't happen to know (from memory) wether bus_width, or sdc_bwid is used by mmc driver? The comment says its sdc_bwid, the example/uage is bus_width (which doesn't match the rest of the parameters)
<hipboi> i guess you need to look into the code
<oliv3r> i'll do that once i've finished Fex_Guide :)
<Turl> hipboi: do you know if batteries on sunxi devices are LiFe or Li-ion?
<oliv3r> hipboi: appearantly, all those parameters are in our sunxi-3.0 not even used, only: sdc_use_wp, sdc_detmode and sdc_used are fetched with script_parser_fetch. The bus_width is set to some default 'somewhere'. Can't find it anywhere in the source anywa y:)
<oliv3r> Turl: http://www.linux-sunxi.org/BORA has a screenshot of a battery in a tablet
<oliv3r> doesn't say which typoe specifically though
<oliv3r> mnemoc: if mux '0' is used, what does that mean? I know 1 means general IO, and 2-5 is one of the muxes, but 0?
rellla2 has joined #linux-sunxi
<oliv3r> 2-7*
<Turl> 0 and 1 are gpio
<Turl> one is in, the other one is out
<oliv3r> ah! i'll document that right away
rellla has quit [Ping timeout: 276 seconds]
<mnemoc> oliv3r: 0 input, 1 output
<Turl> what mnemoc said :)
<Turl> awesome, mike is back from vacations :)
<Turl> hopefully we can get a clock review/merge soonish
<wingrime> Turl: LiFe not used in handled devices
<wingrime> Turl: at-least I have newer saw this
<wingrime> ususly now it Li-Ion
<wingrime> 6 years ago it was NiCd or NiMH
torindel_ is now known as torindel
<mnemoc> NiMH still important.
<wingrime> mnemoc: only Li-Ion-Polymer can be placed to tablet becose it can be "flat"
<wingrime> NiCd or NiMH only in cylidrical form
<mnemoc> yes, sure. for flat devices they are obviusly no-go
<mnemoc> but NiCd is dead also in the cylidrical form world
<wingrime> mnemoc: I still can't understand why udc broken on a13
* mnemoc didn't know it was
<wingrime> mnemoc: I found one problem than send patch for it
<wingrime> mnemoc: but it still not enough
<oliv3r> documented!
<wingrime> [90366.768044] hub 1-0:1.0: unable to enumerate USB device on port 6
<wingrime> PC still not recognise
<wingrime> or maybe it can be in module
<oliv3r> so what if GPIO is I and O
<wingrime> I should try build-in gadgets
<mnemoc> oliv3r: doomed
<mnemoc> oliv3r: it's input, or output, or feature-specific behavior
<oliv3r> mnemoc: heh; I guess you have a choice of an Input, an Output or one of the pre-dfined functions then; those can be bi-directional
<oliv3r> that :)
<mnemoc> oliv3r: what do you need to do?
plan_b has joined #linux-sunxi
<wingrime> Turl: can you test USB-gadget on your a13 devboard
<wingrime> ?
<Turl> wingrime: I don't own any A13 device
<wingrime> so than
<wingrime> give me some a10 dmesg
<wingrime> working
<wingrime> for comparea
<wingrime> with gadget
<Turl> I don't have any of my sunxi devices here to test either :(
<Turl> I know gadget works though, because I use adb on android
<wingrime> Turl: it may be intrsing becose gadget implement ethernet emulation
<wingrime> or audio tuneling
<gzamboni> yesterday i just tested the new gpio driver, its working fine.
<gzamboni> mnemoc i just finished assembling my cubieboard ext board
<wingrime> diff drivers/media/video/sun4i/sun4i_cedar.c drivers/media/video/sun5i/sun5i_cedar.c | less
<wingrime> who can make unifucation patch
<gzamboni> im having some issues with the audio amp. but everything else seems to be working
<wingrime> this not have much differences
<oliv3r> mnemoc: documenting is all :)
<wingrime> hramrach: are you here ?
<gzamboni> when i will have some tim i will document about the new gpio driver on wiki.
<oliv3r> 'yuq' (the mtd guy) did some explaining of certain parameters on the NFC wiki
<Turl> wingrime: I have some pending changes to test and merge for cedar
<Turl> I may unify them after it
<wingrime> Turl: you saw differences, you can easily do this
rz2k has quit []
plan_b has quit [Quit: plan_b]
<Turl> yeah, it's not too many differences
<wingrime> some regs removed and downclocked
hansg has quit [Remote host closed the connection]
<wingrime> Turl: but may be a10 and 13 have same cedar,
<wingrime> other things are marketing
<wingrime> thay can simply downclock module
<wingrime> as you see for that
<wingrime> becose R&D new module costs money
<wingrime> clocks depends only on IC tech-process and operation time
<wingrime> if you simplify module you can make in run on better speed than before
<oliv3r> wingrime: A13 is an castrated A10; so it's definatly the same cedar :)
<wingrime> oliv3: it may be same cristal in other package
<wingrime> I think if we commented out code in a13 and change freq multiper as a10 we will make it decode same things that a10
torqu3e has quit [Read error: Connection reset by peer]
<wingrime> a13 have advantages:
<wingrime> I can solder anything on it
shineworld has quit [Remote host closed the connection]
<oliv3r> a13 is hand-solderable; yeah deff. advantage
<mnemoc> and not only in the A13 case, allwinner-powered PCBs are much simpler than any other
<mnemoc> they need far less passive thingies than the others
<wingrime> mnemoc: this is becose thay can swap pins as thay want on trace
<wingrime> mnemoc: if board makes with cpu in same time
<wingrime> *soc
<oliv3r> i forgot to read back from yesterday, but anybody ordered one of those alibaba A20 products yet? :)
<Turl> oliv3r: I don't think so
<wingrime> [99717.468019] usb 1-6: new high-speed USB device number 54 using ehci_hcd
<wingrime> [99717.612797] usb 1-6: New USB device strings: Mfr=1, Product=2, SerialNumber=0
<wingrime> [99717.612793] usb 1-6: New USB device found, idVendor=1d6b, idProduct=0102
<wingrime> [99717.612800] usb 1-6: Product: Ethernet Gadget
<wingrime> [99717.612803] usb 1-6: Manufacturer: Linux 3.0.62+ with sw_usb_udc
<wingrime> [99718.208567] cdc_eem 1-6:1.0: usb0: register 'cdc_eem' at usb-0000:00:1a.7-6, CDC EEM Device, f2:7f:61:43:6b:1b
<wingrime> [99718.208748] usbcore: registered new interface driver cdc_eem
<wingrime> it begin to work
<wingrime> some how
ganbold__ has joined #linux-sunxi
<wingrime> I bring on ethernet over usb
<mnemoc> wingrime: why do we wrap the #define gadget_is_softwinner_otg ? no one else does it
<mnemoc> the world seems to be pretty happy with the straight strcmp
<wingrime> mnemoc: see whole file
<mnemoc> i'm looking at 3.4
<wingrime> see 3.0
<wingrime> and comare with 3.0
<wingrime> look like not many things required to work
<mnemoc> ok, 3.0 wraps it, 3.4 doesn't
<wingrime> I make other change and make gadget work on 3.0
<wingrime> I try other modules
<wingrime> for bring it to work you have manualy load module for function
<wingrime> for example modprobe g_audio
<wingrime> mnemoc: this not requied android without any change
<wingrime> mnemoc: AW only make patches that make usb mass storage name changabe over fex
<wingrime> mnemoc: and simular
<oliv3r> heh, i just saved the usb mass storage etc on the fex guide :D
hansg has joined #linux-sunxi
<mnemoc> wingrime: in which case, can you submit the missing glue for 3.4?
<wingrime> and I need time to donwnload 3.4 to new tree
<wingrime> wait, it still have problems
<wingrime> and rebuild iy
<wingrime> and test 3.4 kernel
<wingrime> mnemoc: not all gadget drivers works ...
<wingrime> but g_ether and g_audio works
<mnemoc> cp'ing .git is faster than recloning ;-)
<mnemoc> partial gadget support is infinitely better than no gadget support
<wingrime> mnemoc: I have dirty git log
<hramrach> hipboi: the new design looks like it has more design in it than the old so that's good
<hramrach> and it works, too :)
<hramrach> but the about section is sitill empty :s
<oliv3r> wasn't a local clone best (hardlinks?) git clone /home/user/test.git/ /home/user/test2.git/ :)
<hramrach> no
<hramrach> it breaks settings and stuff and not even sure it hardlinks
<mnemoc> techn_: what did you use as base for the hdmi-cec patch?
<wingrime> mnemoc: you forgot add some glue to gadget_chips.h
<wingrime> see END of usb_gadget_controller_number()
<wingrime> on 3.0 and 3.4
<mnemoc> wingrime: I know. added the #define there as only as a reminder
<mnemoc> if you look at the commit, it's your commit. nothing else
<wingrime> actualy my commit only added SUN5I together with SUN4I
<mnemoc> true ;-)
<mnemoc> the stage is rebased, the remind will get squashed into your 3.4 patch
<wingrime> usb need big rewrite
<mnemoc> s/remind/reminder/
<mnemoc> yes, fully agree.
<mnemoc> but after we can deprecate 3.0
<hramrach> working gadgets \o/
<mnemoc> wingrime: that's probably why hansg paniced and chose to improve disp instead of usb (which is his speciality)
<wingrime> hramrach: http://paste.debian.net/242528/
<wingrime> mnemoc: I wan't see some clk driver from mainstream
<wingrime> for make usb work
<mnemoc> wingrime: Turl is mainlining clk
<wingrime> it constain many duplicating clock functions
<mnemoc> Turl: can the clock cleanup be applied on 3.4 to allow wingrime to refactor the usb driver?
<Turl> mnemoc: you mean mainline's? it's incomplete compared to AW's still
<wingrime> Turl: usb has many duplicating clock_enable/disable so we shuld do with this some-thing
<Turl> not sure what's the state of CCF on 3.4 either
<Turl> wingrime: link?
<wingrime> ?
<Turl> link to duplicated clock enable? :)
shineworld has joined #linux-sunxi
<wingrime> whait
<wingrime> wait more...
<Turl> yeah, it pokes at the gates directly
<Turl> the mainline one lets you gate each of them by clk_enable/disable
<Turl> for now i'd keep it like it is, add ifdef for sun3/4/5i
<Turl> when unifying
<Turl> the reg address is the same
<wingrime> turl: It maybe not all
<wingrime> I saw more somewhere
<wingrime> wait
<Turl> yeah, no need to link me more
<Turl> I know what they are now, I've implemented them on ccf
<Turl> wingrime: http://sprunge.us/UgRc
<Turl> look there on AHB gates if you want to know what the shifts mean
<mnemoc> Turl: can 3.4's be refactored to be useful for wingrime and for your mainlining work?
<Turl> wingrime: that last one looks okay
<Turl> mnemoc: it doesn't use the common clock framework
<Turl> mnemoc: I could implement the gates on 3.4 one, but it would be as ugly as poking them directly on the usb driver
<Turl> I think disp also has its fair share of poking registers
<wingrime> Turl: There more one think I wonder
<wingrime> Turl: Why IRQ controller not clear pending after irq exit
<wingrime> Turl: why every driver must clean pending flag
<Turl> no idea, I don't know much about IRQs
<Turl> maybe mripard can answer that one
<mnemoc> Turl: if it reduces code duplication and cleans the other drivers in a usefulway, it sounds like a good thing
<wingrime> and it oblivios that i2c touchscreen driver must NOT configure IRQ controller directly
<Turl> mnemoc: well, the thing is, if you unify the usb stuff, you'd have register poking on one place only
<wingrime> mnemoc: can we drop sun3i code from 3.4
<Turl> but yeah, it might help to clean them
<wingrime> I mean usb
<mnemoc> wingrime: is it too different from sun4i's?
<wingrime> impsible to test
<Turl> talk with the sysfw guy
<wingrime> I think it broken anyway
<wingrime> sysfw must help with testing
<mnemoc> ignore sun3i for now, and lets unify sun[4567]i-usb in 3.4
<wingrime> otherwice: support this will be difficult
<mnemoc> then the sun3i guy can make the needed changes
<mnemoc> but we need to get him onboard
<hramrach> what about sun[45] nand?
<hramrach> well, maybe we get proper mtd driver now
<hramrach> or pm, that's identical in [45]
<wingrime> If we get good dma nand it may be good speedup
<mnemoc> unifying the not-mtd drivers will also help making the sunxi-mtd
<mnemoc> as we don't have documentation, that code *is* our documentation
<mnemoc> only unifying those drivers will tell us the differences
<wingrime> mnemoc: a13 have good register description
<hramrach> oh, nad is only sunxi and sun3i now
<hramrach> nand
<mnemoc> brb. lunch
<mnemoc> wingrime: but let's fix gadget first :)
<mnemoc> pleeeease
<mnemoc> just to be able to tell, "yes, 3.4 is at-least as good as 3.0 in every sense"
<mnemoc> and so deprecate it
<mnemoc> brb
<wingrime> what is still missing ?
<hramrach> if ethernet works it's as good as 3.0
<hramrach> most gadgets were broken to start with
<Turl> does adb work?
<Turl> that should be the test to move to 3.4 :p
<hramrach> how do you use adb?
<wingrime> Turl: I use ssh and wifi
<Turl> hramrach: get adb from platform-tools part of android sdk
<Turl> boot android
<Turl> connect otg to pc and run 'adb shell'
<hramrach> is there a debian package? or wherer do you get taht thing?
<hramrach> oh, android-tools-adb
<wingrime> mnemoc: I still want rtl8188 to merge
<wingrime> mnemoc: I need it for easy-debug using ssh
<hramrach> uloaded to Debian 14 aug 2012
<hramrach> no wonder nobody could tell you where to get it
<hramrach> it was nowhere
<wingrime> hramrach: send patch for enable PM_CONFIG and PM_RUNTIME
<wingrime> in config
<wingrime> expect WAKELOCKS
<wingrime> first boot.....
<ganbold__> so is stage/sunxi-3.4 is current branch where most recent codes go in?
<hramrach> currently it's stage/sunxi-3.0 and stage/sunxi-3.4
<mnemoc> wingrime: what was the commit hash of the driver you wanted?
<wingrime> wait
<ganbold__> ok
<mnemoc> ganbold__: but those are rebased and for testing, not to be used blindly
<wingrime> c2aa8037a9cc831c83ea45151ae2a13df18292e5 1 parent bfd3feb
<wingrime> mnemoc: I have not see anything on screen on 3.4 only usb have power
<mnemoc> wingrime: booting from nand?
<wingrime> nope
<wingrime> mmc
<mnemoc> weird
<mnemoc> serial output shows linux booted?
<wingrime> mnemoc: I not have UART conected
<mnemoc> then connect it ;-)
<wingrime> dissassebly
<mnemoc> ouch
<wingrime> this is tablet not devboard
torqu3e has joined #linux-sunxi
<wingrime> oh funny
<wingrime> CONFIG_WAKELOCK=y
<wingrime> silently
<wingrime> sets
torqu3e has quit [Ping timeout: 245 seconds]
<Turl> hramrach: you did the nand patches right?
vicenteH has joined #linux-sunxi
<Turl> hramrach: it broke my partitions http://sprunge.us/TLVW
<Turl> I suppose it's an off by one somewhere
<wingrime> Turl: why I can't remove CONFIG_WAKELOCK=y using menuconfig ?
<Turl> maybe something else selects it?
<wingrime> in 3.4
<Turl> use / and search for wakelock
<Turl> it will say 'selected by ...'
<wingrime> I find quck fix
<wingrime> I just set "default n" in Kconfig
<wingrime> I hopes it builds
<shineworld> Turl, are you using 3.4 with android ?
<wingrime> oh finaly))
<Turl> shineworld: no, 3.0
<shineworld> ok, I'm just seek a your past message with 3.4 magic world inner ;)
<shineworld> *word
<shineworld> I'm using too the 3.0.42 (from android tv distro)
<wingrime> turl: there some oops in dmesg
<wingrime> related with lcd
ganbold___ has joined #linux-sunxi
<shineworld> do you know if there is some script to put android to uSD instead of NAND ?
ganbold__ has quit [Ping timeout: 264 seconds]
<shineworld> using tools/pack-cm.sh android img is for nand...
torqu3e has joined #linux-sunxi
<Turl> automated? I don'tthink so
<Turl> but it shouldn't be hard to do manually
<Turl> wingrime: that's because the disp thing is called "lcd" and there's an "lcd" module from some backlight thing or leds or sth like that
<shineworld> manually .... uhm... I will try to do but I don't know if will works ... feedback between 30 mins about my tries
<Turl> you just need to make partitions for uboot, system, data, cache
<Turl> copy system to the system partition
<Turl> then separate boot image into kernel + ramdisk
<Turl> edit ramdisk to refer to your partitions instead of nand ones
<Turl> then configure uboot to load kernel+ramdisk
hipboi has quit [Quit: Leaving]
<shineworld> :)
<Turl> fuu github went down :<
<shineworld> I'm pushing without problems ....
<Turl> yeah it's back
<mnemoc> :(
<mnemoc> i need to find the time to setup the new server.... been paying it since oct. :|
<mnemoc> and get a mirror of the git repos there
<mnemoc> and after moving the wiki and git mirror, a mailman
<mnemoc> G ate 4 of hramrach's patches
<Turl> mnemoc: at least it's not like this 142 patches I just got on lakml :P
<Turl> I think some are missing too
torqu3e has quit [Quit: torqu3e]
torqu3e has joined #linux-sunxi
<mnemoc> :)
<wingrime> mnemoc: I add support, gadget_chips.g and enable gadget in config, now will try
<oliv3r> nearly done with fex guide, some i dotting tomorrow and finishing the pmu; then it's done :)
<wingrime> on 3.4
<mnemoc> wingrime: solved the lcd issue?
<wingrime> it begin work, I don't know why, I just rebild with other config
<wingrime> something strange with default config
<mnemoc> wingrime: 8188eu on stage/sunxi-3.0. 3.4 will need some more love (changes in wake_lock)
<wingrime> mnemoc: thanks
<wingrime> mnemoc: I turned off WAKELOCKS on 3.4 and 8188 works good
<wingrime> mnemoc: but agree It need something to do
<mnemoc> btw, I commited a sanitized variant of aw's commit. not pristine
<mnemoc> so it will conflict yours
<wingrime> I can do git revert
hansg has quit [Quit: Leaving]
torqu3e has quit [Read error: Connection reset by peer]
torqu3e has joined #linux-sunxi
rellla2 has quit [Quit: Nettalk6 - www.ntalk.de]
ganbold___ has quit [Remote host closed the connection]
torqu3e has quit [Ping timeout: 245 seconds]
torqu3e has joined #linux-sunxi
<mnemoc> A10S
<mnemoc> do those have uart pads?
torqu3e has quit [Read error: Connection reset by peer]
torqu3e has joined #linux-sunxi
<wingrime> [117071.400021] usb 1-5: new high-speed USB device number 69 using ehci_hcd
<wingrime> [117071.544731] usb 1-5: New USB device found, idVendor=0525, idProduct=a4a7
<wingrime> [117071.544735] usb 1-5: New USB device strings: Mfr=1, Product=2, SerialNumber=0
<wingrime> [117071.544738] usb 1-5: Product: Gadget Serial v2.4
<wingrime> [117071.544741] usb 1-5: Manufacturer: Linux 3.4.29+ with sw_usb_udc
<wingrime> [117072.312149] cdc_acm 1-5:2.0: This device cannot do calls on its own. It is not a modem.
<wingrime> [117072.312267] cdc_acm 1-5:2.0: ttyACM0: USB ACM device
<rm> sub-$35 on Aliexpress
<wingrime> [117072.328829] usbcore: registered new interface driver cdc_acm
<wingrime> [117072.328832] cdc_acm: USB Abstract Control Model driver for USB modems and ISDN adapters
<wingrime> mnemoc: 3.4 some 3.4 gadgets bedin to work
<mnemoc> rm: thanks
<mnemoc> wingrime: \o/
Dreadlish has joined #linux-sunxi
<Dreadlish> hi there :)
<shineworld> hi
mdfe has joined #linux-sunxi
focus has quit [Remote host closed the connection]
focus has joined #linux-sunxi
<wingrime> hramrach: I send patch for 3.4 kernel you can test
<techn_> mnemoc: evening
<wingrime> mnemoc: I done it quck so fogot fix spacing
<wingrime> and patch-check
<wingrime> can you fix it?
<wingrime> It just initial fix, like prove-of-concept
<wingrime> it not for 3.0
<wingrime> techn_: you can also help with gadget
<techn_> wingrime: sure :)
<wingrime> just add patch and enable otg in config
<wingrime> with driver modules ofcouse
<wingrime> some work some not
mdfe has quit [Ping timeout: 245 seconds]
vicenteH has quit [Read error: Operation timed out]
rz2k has joined #linux-sunxi
mdfe has joined #linux-sunxi
<mnemoc> techn_: hi
<mnemoc> wingrime: btw, your fix to suspend looks evil
<wingrime> what patch?
<wingrime> name
<mnemoc> sun5i:usb: fix usb resume on a13 devices
<wingrime> I can explain
<wingrime> on susped modules can suspend / resume on undefied order
<wingrime> some order triggers error
<wingrime> ochi,echi on suspend disable clocks and write to is is_clock_disabled
<wingrime> *is_clk_open
<wingrime> hcd0 (host) do same
<wingrime> if clocks aleady suspended hcd0 skip suspend , and on resume if clocks already enabled skip resume
<wingrime> It works right If suspend order (echi/ochi)->hcd and resume hcd->(echi/ochi)
<wingrime> but when order is opposite
<mnemoc> but sounds like the proper fix is not to remove this code but to make it check for the current state....
<mnemoc> maybe refactor it out to shared function
<wingrime> mnemoc: clk_open have this check
<wingrime> it will not enable clock again
<wingrime> simply suspend do whole work . resume think that we not sleep and not make it all
<wingrime> and leave suspend=1 flag
<wingrime> thats all becose echi ochi also change clk_is_open on suspend resume
<wingrime> when hdc0 resume after echi/ochi clk_is_open =1 so suspend flag leaved in "1" state
<wingrime> good fix: rewite 7 same clock open functions in single one and count open/close event
<wingrime> if numer = 0 do close otherwice number--
<wingrime> mnemoc: you understand ?
<mnemoc> 1m, boss "bothering"
<wingrime> this checks simply make worse
n01 has quit [Ping timeout: 264 seconds]
<mnemoc> uhm
<wingrime> I not realy know why a10 have right load order and a13 not
<techn_> mnemoc: [15:02:43] <mnemoc> techn_: what did you use as base for the hdmi-cec patch?
<mnemoc> techn_: already pushed
<hramrach> wingrime: what should I configure for gadget?
<mnemoc> wingrime: the sun5i code in our main branches is much older than the sun4i code
<oliv3r> do we have source for the ssd253x-ts touchscreen driver?
<wingrime> oliv3r: nope, but I make zet6221
<oliv3r> crap, my tablet has a ssd253x :S
<techn_> oliv3r: you can check lichee branches
<wingrime> oliver: you can write it
<oliv3r> maybe in one of the source dumps we had are those drivers, i'll git grep
<oliv3r> no skills :(
<wingrime> hrameach: wait
<wingrime> hrameach: I found where in makeconfig
<oliv3r> does git grep search all branches? or just current?
<techn_> oliv3r: you could use git foreach
<wingrime> hramrach: <*> USB Gadget Support --> <*> USB Peripheral Controller (SoftWinner SUN5I USB Peripheral Controller) -
<wingrime> then
<wingrime> <*> SUN5I USB2.0 Dual Role Controller support ---> USB0 Controller support (otg support) -
<wingrime> that must me build in kernel
<oliv3r> ah, i wrongfully assumed git grep would search through all branches
<wingrime> and don't forget buid gadget as modules and load some one for test
<wingrime> *I mean g_serial etc
<hramrach> I don't have the softwinner controller in usb gadget support
<libv> Q3A with open source generated shaders!
<Dreadlish> wingrime: hm, you are making driver for zet6221?
<hramrach> nice :)
<oliv3r> libv friggin' awesome :) you, are my hero!
<libv> cwabbott is
<libv> he did all the insane compiler work for it
<wingrime> libv: you awesome
<Dreadlish> nice.
n01 has joined #linux-sunxi
<wingrime> libv: I still want xf86-sunxi-video based on ES GL
<oliv3r> cwabbott your awesome too!
ZaEarl has joined #linux-sunxi
<wingrime> Dreadlish: I write testing request in maillist
<libv> oliv3r: our mali vertex shader is really quite insane
<wingrime> Dreadlish: No reports still
<libv> it's really rather amazing what cwabbott did there
<oliv3r> i thought you said though, that the mali was a really realy good design
<wingrime> Dreadlish: you can see https://github.com/wingrime/zet6221-ts-drv
<libv> oliv3r: a fantastic design
<libv> oliv3r: but insane at the same time
<mnemoc> libv: congrats!
<libv> oliv3r: insane in the right places
<Dreadlish> wingrime: oh, i would
<libv> oliv3r: which is why it is so fast, and why we are able to match binary performance that easily
<oliv3r> how's the gallium driver coming along
<Dreadlish> wingrime: im looking for it for loong time ;d
<wingrime> Dreadlish: you can test
<libv> oliv3r: shutup.
<libv> :p
<wingrime> Dreadlish: : you have to add it to Makefile / Kconfig in drivers/input/touchscreen/ with files
<Dreadlish> i know
<Dreadlish> i was doing it for some cpufrq governors and realtek drivers
<hramrach> wingrime: CONFIG_USB_SW_SUN4I_USB0_HOST_ONLY=y
<hramrach> the other option is USB0_NULL
<wingrime> heramrach: no
<hramrach> and neither shows a xcontroller in gadget submenu
<oliv3r> libv I was honestly innocently wondering!
<wingrime> hramrch: you have to USB Gadget Support --> <*> USB Peripheral Controller (SoftWinner SUN5I USB Peripheral Controller)
<wingrime> hramrch: I too looking for it for a long time
<wingrime> hramrch: It need be NON MODULE
<wingrime> hramrch: <*>
<wingrime> hramrch: <M> USB Peripheral Controller <- it need be NON MODULE
<wingrime> in USB Gadget Support
<wingrime> hramrach: than you will se sunxi controller
_whitelogger has joined #linux-sunxi
<wingrime> Dreadlish: I not sure that mapping code same this other firmware/connection
<Dreadlish> wingrime: we'll see ;d
<wingrime> Dreadish: stil one-touch (not problem to add)
<wingrime> hramrach: finaly found ?
<Dreadlish> i would be happy if it would work at all
<Dreadlish> i will*
<Dreadlish> eh, my english still sucks
mdfe has quit [Remote host closed the connection]
<hramrach> yes, thanks
<hramrach> btw you get build error with gadgets built as module
<hramrach> it useless configuration for sunxi anyway but shhould probably not happen
eebrah has joined #linux-sunxi
<wingrime> hramrach: some gadgets falis to build
<wingrime> hramrach: I tesed g_serial as module and it works
<hramrach> enabled all so will see which fail
<wingrime> acm_ms
<hramrach> it fails in header, though
<hramrach> so not exactly fault of the gadget code it would seem
<wingrime> hramrach: wait
<wingrime> hramrach: you forgot mnemoc patch
<wingrime> also this
<wingrime> already in stage
<wingrime> trul: sun5i:usb: fix usb resume on a13 devices is look good for you ?
<wingrime> techn: what are you doing with display ? /dev/disp will be outdated ?
<hramrach> eventually
<hramrach> it warns on every ioctl
<hramrach> ehternet does not build http://paste.debian.net/242631/
<wingrime> hramrch: I noticed it
<wingrime> hramrch: It need separate commit
<wingrime> hramrch: you can help me if add
<wingrime> hramrch: fix
<hramrach> extra arguments are easy to fix ;-)
<wingrime> 23f9885b2b0ae628bb6ac21587aca35a433b5913
<wingrime> revert this
<mnemoc> uhm.... that was supposed to fix things...
<mnemoc> wingrime: are you still happy with your gadget integration patch for 3.4 or should I wait for a v2?
n01_ has joined #linux-sunxi
<wingrime> mnemoc: not happy
<hramrach> mnemoc: yes, I did a manual revert of that
<wingrime> mnemoc: it fail on rmmod than insmod
<hramrach> the added arguments are superfluous unless you amend the function to accept them
Epsylon3 has joined #linux-sunxi
<wingrime> mnemoc: and need testing
<mnemoc> ok
<mnemoc> hramrach: wtf :<
<wingrime> mnemoc: this patch for others dev to help me with testings etc...
<mnemoc> hramrach: I'll revert it on stage
<mnemoc> wingrime: ok, I'll apply it on stage then. and squash with the next
<hramrach> maybe 3.0 has them but 3.4 does not
<wingrime> mnemoc: 3.0 have other patch
<hramrach> all gadgets compile with the revert
<mnemoc> wingrime: hramrach: pushed
<hramrach> thanks
<wingrime> mnemoc: 3.0 need now only comment out "sw_udc_enable(udc)" as in 3.4 patch
<hramrach> will A10 lock up when you connect the OTG to the other USB port?
<wingrime> hramrach: you need load gadget first
<wingrime> hramrach: modprobe / inmod
<hramrach> still compiling ..
<mnemoc> wingrime: paste me the git-format-patch please
<hramrach> just was thinking that it's technically posible to connect to itself and that it might not like it
<mnemoc> hramrach: slave mode needs an extra pin
<hramrach> so it won't work unless I select the device only mode on otg?
<hramrach> or won't work at all without special cable?
torqu3e has quit [Ping timeout: 245 seconds]
<mnemoc> i think it won't work at all, at all
<mnemoc> special cable, on special port, configured as slave or otg
<mnemoc> afaik
<mnemoc> usb1 and usb2 only have d+/d-
<mnemoc> the id and vbus thingies are only for usb0
<hramrach> you can connect usb0 to usb1
<hramrach> which would be the same as connecting usb0 to another random device
calris has quit [Ping timeout: 264 seconds]
<Epsylon3> usb id is otg protocol
<wingrime> otg is simply additional wire
<Epsylon3> id to gnd or id to vcc
<Epsylon3> or id not connected
<wingrime> that solve who will host and slave
<Epsylon3> phones without real otg use the pin to detect a charger
mdfe has joined #linux-sunxi
calris has joined #linux-sunxi
<wingrime> Epsylon3: some chargrs simply connecd D+ and D-
<wingrime> together
<wingrime> for loopback
<Epsylon3> or a resistor wall
<Epsylon3> to define a specific voltage on id
<Epsylon3> seen on the defy
<hramrach> the endless possibilities how to crew with a standard intrface
<hramrach> screw
<Epsylon3> about that... the cubieboard has analog i/o ?
<Epsylon3> i mean to drive a small rbg led... with or without transistor
<Epsylon3> to have all colors, you need to set some levels or R/G/B
<Epsylon3> of
<wingrime> Epsylon3
<wingrime> use PWM
<Epsylon3> but you need 3 of them
<Epsylon3> its not really important but ;)
<Epsylon3> the rgb notification led is handled on android
<Epsylon3> you can tune colors for different kind of apps
<mnemoc> hramrach: ah, ok. I didn't understand what you mean initially
<Epsylon3> i need a baseboard :p
<wingrime> hramrach: you understand why sun5i:usb: fix usb resume on a13 devices look that way ?
<Epsylon3> with a lot of holes
<Dreadlish> wingrime: drivers/input/touchscreen/zet6221.c:47:29: fatal error: mach/sys_config.h: Nie ma takiego pliku ani katalogu
<Dreadlish> wingrime: aka no such file or directory
<mnemoc> change mach to plat
<Dreadlish> ok
<wingrime> mnemoc: when it happends move to plat?
<mnemoc> months ago
<wingrime> mnemoc: I still used old slice/....
<wingrime> mnemoc: I should tese it on new 3.4
<Dreadlish> wingrime: and one typo ;d
<mnemoc> wingrime: 3.0 and 3.4 have sys_config.h in plat/
<wingrime> Dreadlish: you can send me pool
<hramrach> wingrime: that check relies on devices suspending/resuming in particular order and/or the clock for ehci and otg being differnt
<wingrime> Dreadlish: that I mean whan talk that driver is ~alpha~
<hramrach> which is probably not the ase on your board
<Dreadlish> yup, i understand ;d
<mnemoc> wingrime: can you update your tree? :) the 3.0 udc patch doesn't apply either :p
<hramrach> it happens to work here on sun4i, though
<wingrime> mnemoc: You can quck find this like and coment out , qucker than I send you new version of same .....
<mnemoc> wingrime: yes, I already did
<mnemoc> wingrime: but still would like you to update your branch :)
<wingrime> mnemoc: I will contine dev on 3.4
<wingrime> as base
<mnemoc> both stage branches should be the same
<mnemoc> sunxi-wise
torqu3e has joined #linux-sunxi
Dave77 has joined #linux-sunxi
<Dreadlish> wingrime: it works :)
<wingrime> mnemoc: less problem for us to maintrain single 3.4
<wingrime> Dreadlish: how it works?
<wingrime> Dreadlish: you on debian ?
<wingrime> Dreadlish: you or android ?
<mnemoc> wingrime: sure, but we can't just trash 3.0
<mnemoc> wingrime: and these things that can be fixed in both trees should be fixed in both
<wingrime> mnemoc: "only bug fix"
<Dreadlish> wingrime: i am on debian
<wingrime> Dreadlish: what model
<mnemoc> wingrime: yes. more serious refactoring can be left for 3.4-only
<wingrime> Dreadlish: what tablet ?
<Dreadlish> wingrime: q88b
<mnemoc> wingrime: getting sun6i/sun7i in too
<Dreadlish> also known as a13-b
<wingrime> DreadLish: same as mine
<Dreadlish> ;)
<wingrime> DreadLish: you can help me wirh android testing
<Dreadlish> okay
<Dreadlish> i can flash andro kernel with that ;)
<wingrime> Dreadlish: I have pen release timer that make useable touch on debian
<Dreadlish> mhm
<wingrime> Dreadlish: but android handle event differ and not need such thing
<Dreadlish> nice to know
<wingrime> Dreadlish: slight move make ts send many pen release event
<Dreadlish> ouch
<techn_> Turl: tried cables opposite.. http://linux-sunxi.org/File:JTAG_MicroSD_UART-front.jpeg , now it works :(
<wingrime> Dreadlish: what desktop are using? xfce , others ?
<Dreadlish> wingrime: now? its on dwm ;D
<wingrime> Dreadlish: some time ago I used awesome on desktop
<Dreadlish> but i can put gnome-shell on it
<Turl> techn_: :P
<Dreadlish> (what im doing now)
<hramrach> soo
<hramrach> I have tunl0 and usb0 on A10
<wingrime> perhaps it only make resonable with ts
<hramrach> when I connect the OTG to anoutehr PC I get usb0 there as well
<hramrach> not usb1 when I connect to another port on A10
<mnemoc> wingrime: did you really compile rtl8188eu in 3.4? it heavily relies in an struct which doesn't exist anymore.... struct beacon_parameters
<hramrach> and when I ping one of the A10 tunl0 or usb0 from pc
<wingrime> mnemoc: I cheated ; I set WAKELOCKS=n in config
<hramrach> the *repiles* only appear on tunl0
<hramrach> but not on pc
Dave77 has quit []
<wingrime> mnemoc: I don't know why I can't disable wakelock in menuconfig, so I simply set /power/Kconfig "default n" on WAKELOCK
<mnemoc> unfortunatelly I need to fix it properly :p
<mnemoc> hate those rtl drivers
<hramrach> hmm usb0 shows incoming packets
<hramrach> so the gadget almost works
<hramrach> ther replies exist, just are never sent
<wingrime> *almost*
<wingrime> hramrach: set route /ip
<wingrime> to both
<wingrime> sides
eebrah has quit [Read error: No route to host]
paulk-desktop has joined #linux-sunxi
<oliv3r> mnemoc: do we have @linux-sunxi.org addresses or aliases? :)
<mnemoc> i don't have an MX there yet. but the dns hosting (pointhq.com) provides aliases. want one?
<wingrime> mnemoc: better, I think rename linux-sunxi google group to linux-sunxi-users
<wingrime> or leave same name but add linux-sunxi-dev
<oliv3r> oliver@linux-sunxi.org -> oliver@schinagl.nl please
<wingrime> for patches and dev
<mnemoc> wingrime: that's part of the plan
<Turl> wingrime: why -users? it's a patch ML :P
<oliv3r> mnemoc: wanna try to request an official datasheet for the SDD233x
<mnemoc> oliv3r: test
<oliv3r> and rather have an linux-sunxi.org alias, then my personal email
<mnemoc> Turl: he means setting up a "real" mailing list and leaving the current googlegroups based thing for user support
<mnemoc> oliv3r: sure, totally agree
<wingrime> how to make "Signed-off-by" automaticly ?
<mnemoc> git commit -s
<oliv3r> mnemoc: works
<oliv3r> ty
<mnemoc> oliv3r: my pleasure
<wingrime> thanks
<Epsylon3> - {"sndhdmi", AUDIO_NAME_HDMI},
<Epsylon3> + {"sun4isndhdmi", AUDIO_NAME_HDMI},
<oliv3r> i do have the datasheet allready btw; it was on some random UK site
<oliv3r> just wanna see if I can request it 'officially'
<Epsylon3> was the change between 3.0.52 and 3.0.62 ?
<wingrime> git log "commit1" "commnit2"
<wingrime> or git diff "commit1" "commnit2"
<Epsylon3> lol
<mnemoc> wingrime: you forgot .. between both
<Epsylon3> between those, there is a year of commit
n01_ has quit [Quit: leaving]
<Epsylon3> i try fix the audio hal
<mnemoc> Epsylon3: git log origin/reference-3.0..origin/sunxi -- filename
<mnemoc> that will narrow it to sunxi changes
<oliv3r> looks like i'll have to talk to wingrime a lot
<Epsylon3> ive no idea where that alsa stuff is
<Epsylon3> but that should be a common problem no ?
<Epsylon3> i was on the opentv rel2
<wingrime> Epsylon3: alsa works with mplayer and mpd
<Epsylon3> and upgraded the kernel
<wingrime> Epsylon3: but need config...
<Epsylon3> 96
<mnemoc> Epsylon3: this channel is not cubieboard-centric, don't assume anyone knows what you mean by "opentv rel2"
<Epsylon3> 20:19:26 | Epsylon3: i think something has changed in hdmi audio between 52 and 62
<Epsylon3> 20:19:32 •››› Join: n01_
<Epsylon3> 20:20:17 | mnemoc: Epsylon3: for code related discussions please prefer #linux-sunxi
<Epsylon3> lol
<Epsylon3> love you
<Epsylon3> next channel ?
<hramrach> hmm, enabling tunl0 caused the packets to get lost due to routing
<mnemoc> Epsylon3: we can't know what vendors put in the images they share
<wingrime> gadget is awesome for pen-devices on sunhi
<mnemoc> Epsylon3: only hipboi knows how he made that "opentv" thing
<oliv3r> wasn't OpenTV some propriatary tech in use on set-top boxes for TV's?
<oliv3r> it was 'opentv' because everybody could use the binary blob :p
<mnemoc> no idea
<oliv3r> pretty sure
<hramrach> I can ping over the gadget
<oliv3r> wikipedia opentv should tell ya; i worked with it years ago for $work; it's a crappy stack
<mnemoc> oliv3r: I think hipboi means something different
<wingrime> mnemoc: how looks for you : [PATCH] sunxi:i2c: Don't print initial register state on boot
<mnemoc> like android for tv dongles
<Epsylon3> not sure its related
<mnemoc> wingrime: fine. but as it's not critical, I'm just giving time for people to scream
<Epsylon3> in fact there is not even a readme
<Epsylon3> its a tarball without .git refs
<Epsylon3> of a whole 5GB repo
<Epsylon3> ^^
von_fritz has joined #linux-sunxi
<mnemoc> Epsylon3: it's probably built with allwinner's GPL violating SDK
<Epsylon3> ive cyanogenmod.com kernel code stuff
<Epsylon3> with the wiki ref
<Epsylon3> :)
<mnemoc> can't comment on vendor released images/tarballs
<Epsylon3> no
<Epsylon3> its a source tarball
<Epsylon3> i dont use the image
<Epsylon3> check my whois
<mnemoc> Turl and RaYmAn are the CM fans
<Epsylon3> im "tpruvot" on github
<wingrime> lets call Quarx
<mnemoc> Epsylon3: anyhow, the point is we focus in a community driven repository
<Epsylon3> Quarx is on the defy team with me
<Epsylon3> and ?
<mnemoc> Epsylon3: we can't possibly keep track of all source tarballs in the wild
<Epsylon3> in fact we can
<Epsylon3> i made it for the kernel first
<hramrach> seems disconnecting usb cable while gadget running causes panic
<mnemoc> Epsylon3: we have allwinner sources in github.com/amery/linux-allwinner lichee/* branches
<Epsylon3> ive reintegrated the git history
<mnemoc> Epsylon3: if you find things there that are work importing, please tell us the hash or submit a patch
<Epsylon3> (on the kernel)
<Epsylon3> to check the diff
<Dreadlish> hm
<Dreadlish> usb driver is sort of crazy
<Epsylon3> the 3 last commits is the only diff
<mnemoc> Epsylon3: we have complete sources, with history, of every soc. but lack of time to review it all
<Epsylon3> from sunxi
shineworld has quit [Remote host closed the connection]
<mnemoc> Epsylon3: with commits up to march 2nd.
<Epsylon3> and where ?
<Epsylon3> because there is no readme
<mnemoc> github.com/amery/linux-allwinner, lichee/* branches
<wingrime> mnemoc: I saw CM have LG git too
<Epsylon3> LG stuff is arcee
<Epsylon3> often
<Epsylon3> Ricardo Ceiqueira
<wingrime> LG git with history)))
<hramrach> hmm, no
<oliv3r> Andrew Tarabaras wrote a driver for the SSD232x driver in january against both 3.0 and 3.4 tree's
<hramrach> removing and reinserting the g_ether causes the driver stop working
<mnemoc> Epsylon3: we are working in integrating, cleaning and improving for all sun?i platforms. if you find something sunxi-3.0 and sunxi-3.4 are missing, search for it in the lichee branches, and help us to bring it in
<hramrach> which includes the kernel panic on reinserting the cable
<wingrime> hramrach: seems API changes and we report some things to udc-core
<mnemoc> Epsylon3: since we got allwinner -dev branches there is no need to try to rebased a .git-less dump
<mnemoc> s/rebased/rebase/
<Epsylon3> i like to see the diff made by "officially" published roms
<Epsylon3> on cubieboard mainpage
<Epsylon3> could be interresting... and it is
<wingrime> mnemoc: git-less rom can be eventaly constain interesting vedor stuff
<mnemoc> he hardy has time to care about anything beside manufacturing... he even seem to have gave up on selling directly
<mnemoc> wingrime: of course. but in this case the vendor is tom
n01_ has joined #linux-sunxi
<mnemoc> wingrime: and afaik he hasn't added anything
<Epsylon3> well i will continue to find the hdmi audio diff
<Epsylon3> search
<mnemoc> wingrime: and easy trick is to copy over a .git and find the commit which causes the smallest diff
<Epsylon3> yep
<Epsylon3> not so easy but yep
<Epsylon3> i do it often with vendor opensource kernels
<mnemoc> we have the advantage of having a complete lichee/a10-dev
<wingrime> Epsylon: bisect ?
<mnemoc> so you can script around a `git diff | wc -l`
n01_ has quit [Quit: leaving]
<Epsylon3> git log <file_changed>
<mnemoc> manually, sure
<Epsylon3> to see if the diff is inside or not
<Epsylon3> at different points in time
<Epsylon3> you finally find the closest point
<Epsylon3> its often tagged
<wingrime> Epsylon3 : Is easy for CM if we drop nand[abcde] to nand[12345] ?
<Epsylon3> the merge points
<mnemoc> Epsylon3: yes, I pushed the tags I got too
<Epsylon3> wingrime: yes
<Epsylon3> but ...
<wingrime> mnemoc ^
<Epsylon3> why not mmcblk ?
<mnemoc> Epsylon3: nand is not mmc
<Epsylon3> damn mtd
<mnemoc> it's not mtd either (yet)
<Epsylon3> there is direct io ?
<Epsylon3> i mean direct addresses ?
<mnemoc> the nand driver implements everything
<Epsylon3> ive seen that on old phones
<Epsylon3> froyo era
<Epsylon3> 2.6.29
<Epsylon3> but yep, the numbers are more compatible
mdfe has quit [Ping timeout: 248 seconds]
<Epsylon3> like mmc parts
<wingrime> Epsylon3: I wonder why sunxi make nand[abcd]
<Epsylon3> china history
<hramrach> you get ep already claimed when reinserting g_ether
<Epsylon3> ive seen first MIDs with that
<Epsylon3> SmartQ5
<hramrach> so hte terdown does not work on unregistering
* mnemoc has one of those
<Epsylon3> i have it somewhere...
<Epsylon3> i just moved, all are in boxes
<Epsylon3> ^^
<Epsylon3> was maybe nand0 nand1
<Epsylon3> dont remember exactly
<wingrime> Epsylon3: you save cedar libs from some china ip ?
<Epsylon3> huh ?
<oliv3r> mnemoc: do you have sunxi-archive from january?
<mnemoc> wingrime: I have a copy of those. tons of blobs for android
<Epsylon3> some are built by the system
<mnemoc> oliv3r: eh?
<Epsylon3> but... like all socs ive seen, the codec part is not opensourced
<Epsylon3> most of them
<oliv3r> Andrew submitted a patch on the 13th of january, he was told to come back with a V2, he said on 22'nd he would in a week, but can't see if it got merged
<oliv3r> mnemoc: if not, i'll take his patch, clean it up and re-submit it
<mnemoc> oliv3r: I have no idea what you are talking about
<oliv3r> [PATCH] Add support for ssd2533 touchscreen controller used on some A10 Ta
<oliv3r> blets
<mnemoc> oliv3r: found the thread. 3 replies
<Epsylon3> sunxi was merged in 3.8 right ?
<Epsylon3> (real) mainline
<wingrime> Epsuylon: base-parts
<oliv3r> Epsylon3: 'part 1' :)
<wingrime> Epsuylon: I hope It boots to UART
<Epsylon3> oh right, ive seen next is the gpio mux stuff
<mnemoc> oliv3r: need me to merge it?
<Dreadlish> uart should be set in uboot config iirc
<Dreadlish> console=/dev/ttyS0 or somewhat
<Epsylon3> yea, uboot before kernel
<mnemoc> oliv3r: for some unknown reason I didn't [star] it...
<Epsylon3> really useful :)
<Epsylon3> that change dev life from phones
<oliv3r> mnemoc: i'll clean it up as requested tomorrow
<mnemoc> oliv3r: cool. thank you
<oliv3r> just making sure that i didn't miss anything
<mnemoc> oliv3r: i had missed it entirely
<Epsylon3> i skip the 3.4 branch for the moment... i will just follow the changes... like hdmi one ;)
<Epsylon3> new leds and gpiolib etc
<mnemoc> Epsylon3: those are also in sunxi-3.0
<Epsylon3> i know, i experience them
<Epsylon3> so we maybe have some outputs event
<Epsylon3> oups
<Epsylon3> my dmesg
<wingrime> If i have docs for zet6221 I could make better support
<oliv3r> did you google it? :)
<wingrime> yep
<oliv3r> i noticed that the sdd2533x can do 10 fingers!
<oliv3r> and we have docs for that :)
<Dreadlish> docs are rather not available
<oliv3r> did you try requyest it?
<Dreadlish> only contact on zeitec site is to sales
<wingrime> I do same , no any answer
<wingrime> so I use IDA and mind
<Dreadlish> they said it has 5point multitouch
<Dreadlish> so ;)
<wingrime> marketing
<wingrime> ts IC - are simple MCU
<Dreadlish> hm
<Dreadlish> they could hide sth
<Dreadlish> and unhide it in ic of other number ;d
<hramrach> well, the gadget thing is leaking endpoints
<hramrach> when g_ether is first loaded it gets ep1-ep2
<hramrach> next time they are already claimed
<wingrime> andorid don't need gadgets as modules\
<wingrime> so AW make it simpler
<hramrach> or just did not test
<Epsylon3> f_ether is old/deprecated if i remember
<Epsylon3> try rndis which do the same and more if you have
<hramrach> g_ether uses rndis
<Epsylon3> but this part was rewrote by google
<Epsylon3> and co.
<Epsylon3> cat /sys/class/android_usb/f_rndis/ethaddr
<Epsylon3> 02:00:47:48:49:40
<hramrach> I don't have android support
<Epsylon3> a kernel is not android
<oliv3r> wingrime: done, wrote a request for the datasheet, don't expect any piracle
<hramrach> # CONFIG_ANDROID is not set
<wingrime> oliv3r:thanks a lot, I want miracle)))
<Epsylon3> oh nice loop
<Epsylon3> [ 3424.920000] WRN:L2357(/repos/cubieboard-tv-sdk/kernel/allwinner/common/drivers/usb/sun4i_usb/udc/sw_udc.c):ERR : dev->driver=0xc0829514, dev->gadget.speed=0
<Epsylon3> until usb unplug/replug
<Epsylon3> when i switch to rndis
<mnemoc> Epsylon3: using sunxi kernel or that opentv thingy?
<Epsylon3> for that i can say sunxi
<Epsylon3> because the only diff is related to the wifi we dont have and hdmi
<Epsylon3> and some new config, not related to usb
<Epsylon3> leds, gpio, ext fs
<Epsylon3> ext4 security
<Epsylon3> the stuff made in sunxi defconfigs
<Epsylon3> damn, i dont know alsa, where can i find these keys
<Epsylon3> mixer channels etc
<Dreadlish> alsamixer etc.
<Epsylon3> stuff like ""Playback PAMUTE SWITCH"
<Epsylon3> Quarx ones are fully different
<Epsylon3> mine/crane-common ones : http://epsy.ath.cx:888/paste/?934
<Epsylon3> nope, allwinner/common/hardware/audio
<hramrach> the gadget stuff is way to convoluted
<hramrach> maybe it's cool but requires lots of lxr reading to figure out how it fits together
<techn_> anyone interested of a13 nand flash log?
<Epsylon3> im sure you can use android usb layer to have working gadgets
<Epsylon3> even on a normal linux
<Epsylon3> that was cleaned in ics, before all vendors were doing it differently
<Epsylon3> (kernel layer)
<Epsylon3> then its just sysfs stuff
<Epsylon3> and udev rules
<hramrach> how does that fix the driver not allowing gadget unload?
<wingrime> Epsylon3: you can help with 3.4 kernel on android
<mnemoc> wingrime: hey! we still haven't deprecated 3.0 :)
<Epsylon3> new stable line is 3.8.1
<Epsylon3> :p
<wingrime> hramrach: spoiler: 3.0 kenrnel unload gadgets right way
<mnemoc> Epsylon3: our next stop is mainline, not 3.8
<Epsylon3> cf 3.8.x
<Epsylon3> 3.8.3 now
<mnemoc> i mean, "master"
<Epsylon3> oh right, longterm:3.4.36
<Epsylon3> didnt seen that last time
<Epsylon3> about that... 3.0.69 is out :p
<Epsylon3> nice number
<mnemoc> :)
<wingrime> mnemoc: we must test 3.4 on andorid before deprecate 3.0
<mnemoc> wingrime: fair point
<Epsylon3> Turl is working on it
<Epsylon3> he have both a tablet and a cubieboard
<Epsylon3> which devices contains the sun6i/sun7i ?
<mnemoc> allwinner A31 is sun6i
<mnemoc> there are a bunch out there
<hramrach> meaning useless
<hramrach> 7i not available
<Epsylon3> its the A20 ?
<mnemoc> A20 is sun7i, not on sale yet
<wingrime> at this price a31 are notintresting at all
<mnemoc> A31s is sun6i too, also not on sale yet. very phone oriented
<Epsylon3> k, A31 or A13 ?
<Epsylon3> their numbering is weird :p
<mnemoc> A13 is sun5i, A31 is sun6i
<wingrime> I have 13 tablet
<wingrime> a13 - costrated a10
<hramrach> 10s is more interesting than 13
<Epsylon3> and which ones are dual cores ?
<mnemoc> Epsylon3: the sun?i thing is chronological. the A?? naming is marketing
<hramrach> hdmi is kind of useful
<wingrime> a31 - 4-core
<wingrime> cortex a-7
<mnemoc> Epsylon3: the A20 is dual A7
<Turl> mnemoc: still I cannot make sense of why they did A31 first :p
<hramrach> new market
<Turl> usually one does single-dual-quad core
<Epsylon3> and the S4 has 8
<wingrime> modern=expensive
<mnemoc> Turl: the A20 is just a drop-in solution for A10 lovers
<Epsylon3> i was waiting the omap5
<Epsylon3> but it never came...
<mnemoc> Turl: but castrated to not harm the sales of their A31
<Epsylon3> maybe the next moto phone...
<Epsylon3> in 2 months
<Epsylon3> ^^
<Epsylon3> at google io
<mnemoc> Epsylon3: going?
<Epsylon3> no, i dont even know where it is
<Epsylon3> but i will follow online
<Epsylon3> like last years
<Epsylon3> all is on youtube or broacast tvs
<wingrime> Epsylon3: you will not bye expensive thing when cheaper slily slower exists
<mnemoc> Epsylon3: and been frustrated when watching the give aways ;-)
<Epsylon3> erghh
<Epsylon3> yea... last years 3 devices
<Epsylon3> year
<Epsylon3> Nexus Q
<Epsylon3> never seen here in france/suisse
<Epsylon3> the TV media box
<mnemoc> did they sell the Q anywhere?
<Epsylon3> never seen online neither
<Epsylon3> for europe at least
<hramrach> did TI give up on arm?
<Epsylon3> thats weird, there was demos on youtube about omap5
<hramrach> maybe just nobody built device on it?
<mnemoc> from before they slaugthered the omap division
<Epsylon3> multi hdmi movies on a 3D desktop
<hramrach> ah
<Epsylon3> multi fullHD oops
<hramrach> well, chinese turn out better at making usable chips, anyway
<Epsylon3> chips, yea
<wingrime> I like TI SoCs and any ICs
<hramrach> or they just make so many that you find one that is usable
<Epsylon3> socs are fine, but some other chips...
<wingrime> msp430
<wingrime> nice architecture
<hramrach> wrt TI SoCs
<hramrach> when boards based on thema were pretty much the only available
<hramrach> I was waiting for them to make an usable one
<hramrach> and now I am here :p
<hramrach> well, there were those groadcom and marwell SoCs you could use for mini-server
<hramrach> but not desktop
<Epsylon3> i bought the cubieboard for the dock ability
<hramrach> and hte only mini-server based on that that I found was guru plug and was unusable due to heat problems
<Epsylon3> that means you can make a board for a specific usage, and just plug a cubieboard as heard
<Epsylon3> heart*
<hramrach> yes, pins on bottom
<Epsylon3> and unplug it for maintenance
<Epsylon3> yep
<Epsylon3> its a very good idea
<hramrach> but makes putting the board in small spaces more challenging
<Epsylon3> too bad the 2mm space
<Epsylon3> you can t use the normal dev boards
<Epsylon3> with holes
<hramrach> yes but then not pluggable
<Epsylon3> maybe one side was possible
<Epsylon3> its hard to do an adapter
<Epsylon3> you need to make pcbs and get connectors
<Epsylon3> which double the price
<Epsylon3> for almost nothing
<hramrach> adapter for what?
<Epsylon3> to have 0.1" double DIN headers
<Epsylon3> to plug it fast on a prototype board
wingrime has quit [Ping timeout: 256 seconds]
<Epsylon3> like sure electronics ones
<hramrach> that's because half of the world uses inches and another metric units
<hramrach> can't fit everywhere
<Epsylon3> no, all pcs connectors are on DIN
<Epsylon3> ATA cables too
<Epsylon3> PATA ;)
<hramrach> so what
<hramrach> btw some PC boards use smaller distance headers for some connectors to save space already
<hramrach> or maybe to just prevent connecting the wrong cables
<Epsylon3> on laptops yep
<Epsylon3> ;)
<hramrach> and ITX boards
<Epsylon3> i cant find their hole boards with all connectors
<Epsylon3> audio etc
<Epsylon3> sdcard also
<Epsylon3> indeed this sensor use smal pins ;)
<Turl> techn_: what's on your env?
<hramrach> it's not like you will be connecting PATA cables to cubieboard ;-)
<Turl> techn_: I suppose you want to use 'boot' instead of 'kernel'
<techn_> nand read 50000000 kernel;bootm 50000000 => nand read 50000000 boot;bootm 50000000 ??
<Epsylon3> first on photo, i was thinking... its a good idea to recycle our old PATA cables
<Turl> yeah techn_
<Epsylon3> ive plenty of new ones in motherboard boxes ;)
<techn_> I'll try.. thanks :)
<hramrach> only the old staright ones are useful
<techn_> Turl: "Wrong Image Format for bootm command" :)
<Epsylon3> 3$ sensor
<hramrach> the new ones do som voodoo with interleaving ground between signals or what
<hramrach> so you could design gpio connector for pata cable but would need to decide that upfront
<Epsylon3> its just to connect to another board
<Epsylon3> more easily
<Epsylon3> without buying special pcbs
<hramrach> also the master/slave wire is not connected to all headers on some overly clever cables
<Turl> techn_: what did you put on the partition?
bfree_ has joined #linux-sunxi
<techn_> Turl: cwm10
<Epsylon3> (and you can cut wires as you want them)
bfree has quit [Ping timeout: 252 seconds]
<Turl> I suppose bootm is not the command to boot android boot images :P
<Epsylon3> like 40pins to 12 + 10
<Epsylon3> etc
<Turl> Epsylon3: can you printenv your android uboot env?
<Epsylon3> yep
<Turl> also, koush is already at v10? :P
<Epsylon3> boota so
<Turl> techn_: ^
<Epsylon3> i used bootm alors on openwrt
<Epsylon3> orion arm
<Turl> bootm is for uImage
<Epsylon3> hmm /dev/block/ is only on android ?
<Epsylon3> i see a lot of /dev/nand references
<Epsylon3> and here /dev/mmcblk0p4
<mnemoc> /dev/block/ is android-style, yes
<Epsylon3> right
<Epsylon3> ls /dev/block/
<Epsylon3> 1:0 1:1 1:10
<Epsylon3> ...
<Epsylon3> 3 lines of that ;)
<Epsylon3> 8:97
<Epsylon3> on mepis
<techn_> Unknown command 'boota' - try 'help'
<Epsylon3> did you try help ?
<Epsylon3> :p
<techn_> :D
<Epsylon3> its how i understood the ftpd to save my router
<Epsylon3> tftp
<Epsylon3> you have more help then if i remember with help <command>
<hramrach> the problem here is that the command is not available
<hramrach> not compiled in or not in this version
<Epsylon3> boota is maybe (a)ndroid
<Epsylon3> yes it is
<Epsylon3> base - print or set address offset
<Epsylon3> boot - boot default, i.e., run 'bootcmd'
<Epsylon3> boota - boota - boot android bootimg from memory
<Epsylon3> bootd - boot default, i.e., run 'bootcmd'
<Epsylon3> bootm - boot application image from memory
<Epsylon3> nboot - boot from NAND device
<Epsylon3> go addr [arg ...]
<Epsylon3> - start application at address 'addr'
<Epsylon3> well... im out
Epsylon3 is now known as E3-zzZ
Turl has quit [Ping timeout: 264 seconds]
Turl has joined #linux-sunxi
von_fritz has quit [Quit: vonfritz leaves, don't panic]
paulk-desktop has quit [Quit: Ex-Chat]
chujalt has joined #linux-sunxi
torqu3e has quit [Read error: Connection reset by peer]
chujalt has quit [Quit: Saliendo]
torqu3e has joined #linux-sunxi
<bsdfox> anyone worked with the vivante gc2000 gpu? I'm wondering if there is better linux support for it than mali. If so http://dx.com/p/gk802-quad-core-android-4-0-google-tv-player-w-air-mouse-wi-fi-bluetooth-1gb-ram-8gb-rom-177884?utm_source=dx&utm_medium=cdn&utm_campaign=nhomeleft2-177884 could be interesting