leowt changed the topic of #linux-rockchip to: Rockchip development discussion | http://linux-rockchip.info | http://irclog.whitequark.org/linux-rockchip
tonikasch has joined #linux-rockchip
tonikasch has quit [Quit: Bye!]
cnxsoft has joined #linux-rockchip
FreezingCold has quit [Ping timeout: 255 seconds]
rz2k has quit [Read error: Connection reset by peer]
FreezingCold has joined #linux-rockchip
hramrach has quit [Write error: Broken pipe]
hramrach has joined #linux-rockchip
FreezingCold has quit [Ping timeout: 258 seconds]
FreezingCold has joined #linux-rockchip
_hipboi_ has joined #linux-rockchip
hipboi_ has quit [Ping timeout: 240 seconds]
AstralixNB has joined #linux-rockchip
AstralixNB has left #linux-rockchip [#linux-rockchip]
<mnemoc> :o
cnxsoft has quit [Ping timeout: 265 seconds]
FreezingCold has quit [Ping timeout: 240 seconds]
AstralixNB has joined #linux-rockchip
cnxsoft has joined #linux-rockchip
<cnxsoft> The original Chinese article just says Intel may invest in Rockchip, not directly buy it.
<cnxsoft> That is if Google translate did the right job
_massi_ has joined #linux-rockchip
viric has quit [Ping timeout: 255 seconds]
<AstralixNB> Woha... cnxsoft, can you post me the link again?
viric has joined #linux-rockchip
<AstralixNB> cnxsoft, thanks!
<hramrach> That's what my translator says too
<hramrach> also see the page URL ;-)
<hramrach> anyone got Linux working on a random RK tablet?
cnxsoft has quit [Ping timeout: 265 seconds]
* hramrach looking up flashing tools
nighty^ has joined #linux-rockchip
naobsd has joined #linux-rockchip
<naobsd> definition of "working" must be provided... ;)
<naobsd> in general, most Linux userland can work with Android kernel
<AstralixNB> In mainline kernel lcdc is still missing so you won't see anything on the tablet
<naobsd> stock Android kernel may be enough for some case ;)
naobsd has quit [Ping timeout: 240 seconds]
<cosm> hramrach: I've tried that yesterday
<cosm> on this particular tablet most stuff seems to work out of the box, except video output
<cosm> with an 802iv board file
<cosm> and I suspect it doesn't get correctly set up even if enabled in the kernel
<cosm> HDMI might be working, but I don't have a mini HDMI adapter yet
<cosm> and I didn't have time to take a closer look at what's wrong
<cosm> regarding naobsd's comments: an Android kernel might be enough to boot a GNU/Linux userspace, but usually some feature will be missing or some android-specific stuff might be enabled which makes it less convenient
<cosm> so I'd start with a kernel that you can be built from source
<cosm> I've asked the manufacturer for their source, but I'm not very hopeful
<AstralixNB> You could check the already existing distros like picuntu. These should work on the MK80x sticks and clones.
<AstralixNB> All of the sources are available free. So you just need to concentrate on adding the board of your tablet and the HDMI/LCDC conversion chip.
<cosm> AstralixNB: no, the GPLv2 license means that the manufacturer has to provide the source code required to build the kernel binary they're distributing
<AstralixNB> Ok, you can insist on getting it and wait for it... See you in a hundred years...
<cosm> I know how these fly-by-night operations work
<cosm> which is why I've worked around it
<cosm> I'm just saying it's not accurate to say "All of the sources are available free."
<AstralixNB> depends
<AstralixNB> I am pretty sure, that the sources of your tablet are either not available, not complete, not matching, outadted or any combination of these options.
<AstralixNB> But the kernel and linux of these MK80x sticks provided by alok and others should be available completely.
<AstralixNB> I know that as I know the guys who spent night over night to get it working and they will insist on having all sources open and available.
naobsd has joined #linux-rockchip
<AstralixNB> But for tablets... I have seen 50+ kernels with according sources and it costs me about 5min each to validate it in a disassembler.... only very very few where matching.
cnxsoft has joined #linux-rockchip
<naobsd> Astralix: cosm is talking about tablet, not mk* stick
<AstralixNB> yes but there is so much to do if you start from plain tablet kernel up to working linux
<naobsd> Astralix: cosm is the man who can understand things like source code
<AstralixNB> If you didn't do that once or twice, you'll likely fail
<AstralixNB> ok
<AstralixNB> I just proposed a different approach that results in less work and probably better result
<naobsd> he already have working kernel source for his stick. I'm not speaking such a thing
<AstralixNB> So where is the problem?
<naobsd> I just say he is "developer", not the man who ask dumb questions
<AstralixNB> I hop I didn't give any dumb answer
<naobsd> he talked about tablet and it's source
<AstralixNB> And I told him that it will not work out
<cosm> I think there's some confusion around, maybe I wasn't clear
<AstralixNB> Chances to get real sources are less than 10%
<naobsd> <cosm> I've asked the manufacturer for their source, but I'm not very hopeful
<naobsd> ''not very hopeful'
<naobsd> there are many people "if I ask maker about source code, I must get it soon"
<AstralixNB> May be this will change some day. If customs in here will feed their computers with large blacklists of chinese products that violate GPL
<naobsd> we are not
<naobsd> such a people
<naobsd> (I'm talking about rockchip and/or similar chinese soc world)
<cosm> I had some limited time to play with it, so I've tried a generic kernel/board config
<cosm> most stuff is working, the major issues being the internal display
<AstralixNB> I would never request a current and working snapshot of a GPL code and expect it in a minute. But I hav requests of GPL code running for over 1.5 years now. Have not seen any line of code.
<cosm> when I'll get the time I'll look at fixing it myself, but in the offshot that the vendor will provide the source code, I've also sent them an email
<AstralixNB> cosm: What I found out is, that many times almost every thing is correct, but then one or two things are simply missing
<AstralixNB> In some cases, the single driver needed for the LCD was just deleted before putting the GPL code on the servers.
<AstralixNB> You could see it as it was still mentioned in Kconfig and Makefile but the .c and .h files where not there
<AstralixNB> with one tablet you just needed to exchange the not installed photo flash light GPIO with the lcd-enable GPIO and it worked
naobsd has quit [Ping timeout: 240 seconds]
<AstralixNB> but cosm, If you can tell me the device, I can check if I have something working for you or a prossible matching driver for the LCDC interface.
<cosm> but I'm fairly sure they get them from an ODM
<cosm> I have some board pics, wait a minute
naobsd has joined #linux-rockchip
<cosm> it feels reasonably well built for a cheap tablet, by the way
<naobsd> is ssd2828 sit left-above from ap6330??
<cosm> naobsd: yes
<cosm> the one glaring exception being the uSD socket which is designed to be surface mounted on a board, not in a cutout
<cosm> and I may have accidentally popped it open
<naobsd> kernel/drivers/video/rockchip/transmitter/ssd2828.c hmm...
<cosm> naobsd: yeah, there's that driver, which I've enabled
<cosm> but maybe it needs something else set up in the board file
<naobsd> spi and gpio?
<cosm> I didn't have time to look at how it's supposed to work
<AstralixNB> As I said before. Often it is more a GPIO problem than a missing driver. Check if the EN or RESET GPIOs for display and SSD are set correctly. Check if the chip has I2C interface and if, if it is recognized at the right bus with the right address. If it doesn't work out, send me the original binary kernel image and I check some of the details via disassembler
<cosm> naobsd: yeah, it looks like you're right
<cosm> for some reason I thought it's i2c
<AstralixNB> ok, spi, not I2C, so it needs a CS GPIO line
<AstralixNB> oh my goodness... They use a GPIO based bitbanging driver... I ever wondered why they have several SPI interfaces in the SOC but never use them... So it must be true that they are buggy
<AstralixNB> However full datasheet of the SSD is available here: http://www.datasheet-pdf.com/datasheet/SOLOMON/774632/SSD2828.pdf.html
<cosm> haha, I've seen shorter MCU datasheets
<AstralixNB> Long times ago, yes.. the latest I actually read is over 1900 pages
<naobsd> pipo u8? I can find some pages with "Q7901"
<cosm> naobsd: yeah, it looks like the same hardware
<cosm> nice!
<AstralixNB> A bonus of Pipo is that they are so kind in placing serial port test pins.
<naobsd> very minor difference around frontcam ;)
<AstralixNB> Yes, the pipo pictures show a board revision 1.2 while yours is a 1.7
<naobsd> RK3188_P7901_V1.2 2013-05-07,
<AstralixNB> :)
npcomp has joined #linux-rockchip
<naobsd> anyway, both should be equivalent... no info/source, probably ;)
<AstralixNB> cosm, in case everything fails, disassembler helps to get the right GPIOs
<naobsd> "struct ssd2828_t *ssd2828 = NULL;" global? ;)
<naobsd> I sometimes like this kind of cheap thing from china
<cosm> it looks like there's going to be a bit of guesswork involved in setting up that ssd2828_t
<naobsd> e.g. rockchip don't change bootloader USB protocol, firmware format,
<cosm> naobsd, yes, in a strange way they're sort of hacker friendly
<naobsd> I can hack ASUS MeMO Pad 8 very easily ;)
Tsvetan has quit [Remote host closed the connection]
<naobsd> it seems ASUS guys tried to hide some things, they don't make physical key to enter RK bootloader mode, they changed kernel reboot func to go fastboot mode instead of RK bootloader mode
<naobsd> but probably they forgot to change "reboot-bootloader" in fastboot, it goes RK bootloader mode
<cosm> haha
<cosm> this one uses the "esc" key to enter the RK bootloader
<naobsd> yep, most cheap products have it, sort of hacker friendly ;)
<naobsd> I really wonder why people try to "rooting" on RK. NAND can be modified freely :)
<cosm> maybe they use the production hardware for firmware development to reduce cost
<naobsd> they use development hardware for production
<cosm> :)
<naobsd> making new board needs $$$
<naobsd> so we can access serial console :)
<cosm> but that source linked in the original post doesn't seem to actually set up ssd2828
<cosm> any idea how they got the GPIO dump here: http://www.freaktab.com/attachment.php?attachmentid=3700&d=1382971982
<cosm> oh nevermind, it was linked a few comments above
<AstralixNB> yes it is made by a module that you can install that was made by some people hacking rk for long time...
<naobsd> is CONFIG_DEBUG_FS enabled?
<AstralixNB> check the version, some of these modules work on either 3066 or 3188 but not both.
<naobsd> on stock rom
<AstralixNB> Normally not, but the gpio module works without as the gpio names are in even without config fs I guess
<cosm> naobsd: it seems to be
<naobsd> I guessed it's gpiolib_dbg_show(), but format is different a little...
<naobsd> it should say in/out hi/lo
<cosm> cool, I know what I'm doing after work
<naobsd> but if CONFIG_DEBUG_FS is enabled, there should be /sys/kernel/debug/gpio
<cosm> naobsd: it seems to be empty
<naobsd> did you mount?
<hramrach> and I do not have micro HDMI so cannot connect to external screen
<hramrach> I have sort of universal adaptor which has many connectors on it so you can figure out which one you have by trying all but it
<cosm> naobsd: yes
<hramrach> is so bulky it always gets in the way of other nearby connectors
<naobsd> cosm: then, let's try GPIO_DUMP ;)
<cosm> naobsd: yes, when I get home
<naobsd> I didn't know that module can print label
<hramrach> how do you take apart these tablets?
<hramrach> it has no screws or anything
<cosm> hramrach: mine had no screws, I just took a plastic spudger to it :)
<naobsd> hacker friendly :)
<naobsd> need more $$$ for screws!
<cosm> watch out for any wires going to the back panel (e.g. for a speaker) and for things that might get stuck in the cutouts (e.g. a micro SD card)
<hramrach> it has a serial number printed on the plastic :o
<hramrach> but no button descriptions ..
<cosm> hramrach: one other thing to be careful about: the battery
<cosm> it's a soft pouch, make sure you don't damage it
<hramrach> hmm, could be tricky when the case is not made to open easily
<hramrach> but might need to see the chips on the board eventually
<cosm> the case should open fairly easily
<AstralixNB> strong fingernails or an old credit-card is needed and almost all of them open up easy
<naobsd> btw, I have interest to CONFIG_RK_USB_UART
<naobsd> I guess D+/- can be used as TX/RX
<hramrach> I have an a13 tablet which opens really easily with just fingernails but this one has more tight fitting case
<naobsd> but I have no idea about detail. clock? voltage? :(
<cosm> naobsd: I would have thought that it configures the USB OTG port as a usb-serial device
<AstralixNB> enter into the slit in the middle of a long side without speakers or plugs, then try to open into direction of the small side with the plugs. A second old card helps to keep the slit open
<hramrach> ok, need to find some suitable plastic parts for that
<AstralixNB> For this I always have a collection of old ensurance-, credit- or rebate-cards :)
<naobsd> cosm: then it needs driver on host. I guessed it's simple function, just bypass UART signal...
<cosm> usb-serial is supported out of the box on any modern OS
<naobsd> I sometimes use not-old-card, and sometimes say ahhhhhhh
<hramrach> did not need them so did not collect any
<naobsd> cosm: there are drivers too
<AstralixNB> Yes, I am not aware of any special hardware on any of the USB ports. So I guess it is simply a USB-CDC made by rk implementation.
<naobsd> ftdi, profilic, etc etc
<naobsd> ah prolific
<naobsd> hmm
<hramrach> but my current card is so worn that using it for opening a tablet can't possibly make any difference
GriefNorth has joined #linux-rockchip
<AstralixNB> rk wrote many drivers on their own even they implemented an IP that already has drivers in the kernel
<naobsd> enabling that config is easy, I'll try and see what happen...
<naobsd> I have RK2926 tablet and I still cannot find UART pads ;)
<AstralixNB> 2926? Ok... Do you have some detailed photos of it?
<naobsd> I guess I told this several month ago ;)
<naobsd> I want to get rk302x device for my collection...
GriefNorth has quit [Remote host closed the connection]
<cosm> naobsd: at least you have nicely labelled voltage rail pads
<naobsd> cosm: no TX/RX, unfriendly ;)
<cosm> I have a faulty rk3066 tablet which I'm fairly sure has some power supply issues
RayFlower has quit [Ping timeout: 240 seconds]
<cosm> but it'a real PITA to figure out what's going on
<cosm> it was UART pads, but it doesn't get to print anything :)
<naobsd> RK2926 pin assign info may be enough there are pins I can touch directly
<naobsd> (and if I have good probe)
<naobsd> my soldering skill is not enough for that pitch
<cosm> holy shit, I've just realised that's a hand solderable Cortex-A9 SoC
<cosm> too bad they don't publish datasheets
<naobsd> you have very nice hand ;)
<AstralixNB> you need to know the tricks, it works with a 5mm width tool tip.
<naobsd> http://www.rock-chips.com/a/en/products/RK30_Series/2013/0731/371.html "pin-to-pin rk2926" you can get hand solderable dual-core A9 too
GriefNorth has joined #linux-rockchip
<AstralixNB> But you need to buy at least 1k of them otherwise you have to pay multiple times the price and you still not get any datasheets
<AstralixNB> not even sure if 1k is enough...
<AstralixNB> probably 10k or more
<naobsd> buy few cheap tablets, and unsoldering? ;)
<AstralixNB> Just use the tablets PCB and add on your own application on a second PCB you connect via I2C or USB. So you get a display for free :)
<naobsd> and battery (it may explode easily) :)
RayFlower has joined #linux-rockchip
<AstralixNB> LiIon is critical to handle but easy to charge. Just keep the values. You should not leave it alone connected to your lab supply
ganbold_ has joined #linux-rockchip
<AstralixNB> There is absolutely no pinout of the 2928 or 2926 available... So you just can try with a scope or a low voltage serial probe
<AstralixNB> naobsd: I would search the serial signals around the corner where the WiFi module is.
<cosm> naobsd: you don't need such a steady hand
<AstralixNB> In one tablet I found the pads hidden as a non placed resistor.
<cosm> the surface tension of solder pulls components in place
<naobsd> Astralix: I see, worth to try
<naobsd> I hope I can find TX, and I really hope RX is near by TX... ;)
<AstralixNB> I even searched my own archives of datasheets... but unfortunately no 292x
<AstralixNB> hard to tell as the 3xxx are all in BGA and there they are almost near by
<AstralixNB> no idea how they mapped it at the QFP packages
<AstralixNB> simple way to try is to us a low power USB dongle and unsolder battery (+). Then put the rx to a pin and manually connect battery. Within a second the loader shows up. If not, disconnect battery, try next pin.
<AstralixNB> I prefer these USB-UART dongles as they reliably work even with very low voltage pins: http://shop.in-circuit.de/products/Home/Schnittstellenmodule/27/USB-A-UART-Bridge
<naobsd> it's for tx(output)
<AstralixNB> no, if you have a cheap 5V dongle you might get problems if the SOC just sends with 3.3V
<naobsd> sorry
<naobsd> "... loader show up" is for tx
<AstralixNB> But I not only work with RK SOC I even develop on low power cortex M series and there serial port may have 1.2V or 0.8V
<naobsd> what I worried is input
<naobsd> ping: ganbold
<AstralixNB> put in 1K resistor between TX of dongle and the pin to test, so you'll not kill anything
viric has quit [Ping timeout: 276 seconds]
<ganbold_> naobsd: pong
<naobsd> ganbold: FreeBSD kernel.img can boot from SD
<ganbold_> naobsd: cool :_
<ganbold_> :)
<naobsd> (of course I'm sure there is no OS restriction in bootloader)
<naobsd> and I found armv6 snapshot on ftp. I got BBB one and make USB root file system
<naobsd> userland works, of course :)
<naobsd> then
<naobsd> can you add axe(4)? ;)
viric has joined #linux-rockchip
<naobsd> I don't have cross build env yet ;)
<ganbold_> I see
<naobsd> I'm thinking to use RR as build machine
<naobsd> I hope I'm not so crazy
<naobsd> lol
<ganbold_> I will make it for you tomorrow afternoon
<naobsd> did you try urtwn(4)?
<naobsd> I can see it
<ganbold_> urtwn works :)
<naobsd> very nice
<ganbold_> are you sure you have axe(4)? I have similar one, but that is udav
<naobsd> then I can use urtwn(4) instead of axe(4)
<naobsd> very sure, I like collecting USB wired/wireless nic
<naobsd> I also have axen(4)
<ganbold_> ok, I will add both and let you know tomorrow afternoon
<ganbold_> I think urtwn is already there
<ganbold_> so you can use it
<naobsd> I guess I can get source via urtwn
<naobsd> yes
<naobsd> I almost forgot about FreeBSD, but I can remember :)
<ganbold_> I'm sure you can remember, it should be easy for you
<naobsd> or probably I can put source on Linux machine...it should support ufs
<naobsd> I really wonder why I don't have any *BSD machine in my home
<ganbold_> nfs :)
<ganbold_> maybe you should :)
<naobsd> nfs needs network :)
<ganbold_> right
<naobsd> there were only *BSD machines til few years ago
<ganbold_> if we can get arc vmac driver that would be great
<naobsd> and now I have only Ubuntu machine and too many cheap Android crap^Wdevices... ;)
<naobsd> yes, I really have interest about it
<ganbold_> yeah, if you can have cross build env, then I guess everything should be ready to start working on it :)
<naobsd> ioctl[SIOCS80211, op=16, arg_len=0]: Invalid argument
<naobsd> mmm
<naobsd> which wired/wireless drivers are included?
<naobsd> I guess no other driver... trying urtwn
<naobsd> IEEE80211_IOC_ROAMING ??? :(
<naobsd> ah
<naobsd> ifconfig wlan create wlandev urtwn0...?
<ganbold_> did you try by looking at http://www.freebsd.org/doc/handbook/network-wireless.html ?
<ganbold_> quick start
<naobsd> yes, I have to refresh my brain...
cnxsoft has quit [Quit: cnxsoft]
cnxsoft has joined #linux-rockchip
cnxsoft has quit [Client Quit]
<naobsd> it seems I have to go multiuser for wpa_supplicant
<naobsd> now everything working
npcomp has quit [Quit: leaving]
<naobsd> I just use USB disk to copy src ;)
AstralixNB has quit [Ping timeout: 252 seconds]
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
<nighty^> ganbold, ping
<nighty^> ganbold, how is the raxda support now in FreeBSD ?
rz2k has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
_massi_ has quit [Remote host closed the connection]
nighty^ has quit [Remote host closed the connection]
nighty^ has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
nighty^ has quit [Quit: Disappears in a puff of smoke]
nighty^ has joined #linux-rockchip
bgal has joined #linux-rockchip
npcomp has joined #linux-rockchip
npcomp has quit [Quit: leaving]
GriefNorth has quit [Ping timeout: 240 seconds]
billy55 has joined #linux-rockchip
billy55 has quit [Client Quit]
nighty^ has quit [Remote host closed the connection]
cosm has quit [Quit: Leaving]
tonikasch has joined #linux-rockchip
tonikasch has quit [Quit: Bye!]
FreezingCold has joined #linux-rockchip
FreezingCold has quit [Remote host closed the connection]
FreezingCold has joined #linux-rockchip
bgal has quit [Ping timeout: 240 seconds]
cosm has joined #linux-rockchip
cosm_ has joined #linux-rockchip
cosm has quit [Ping timeout: 265 seconds]
cosm_ is now known as cosm
<cosm> so based on naobsd's find I thought I'd use the pipo u8 firmware for a free kit kat upgrade
<cosm> and guess what, it worked
<cosm> but it claims to run a 3.4.0 kernel...?
<cosm> oh, nevermind
<cosm> someone got creative again
<cosm> uname says it's still 3.0.36+
<ferric> moo.
<ferric> haha @ someone getting creative.
<cosm> I got the GPIO mapping for the pipo u8 tablet: http://pastebin.com/wFbnVwfr
ojn has quit [Ping timeout: 265 seconds]
_whitelogger_ has joined #linux-rockchip
ojn_ is now known as ojn
naobsd has quit [Quit: Page closed]
architekt_ has joined #linux-rockchip
architekt has quit [Ping timeout: 265 seconds]
honx_ has quit [Ping timeout: 265 seconds]
mac- has quit [Ping timeout: 265 seconds]
_whitelogger has quit [Ping timeout: 265 seconds]
mmind00_ has joined #linux-rockchip
mmind00 has quit [Write error: Broken pipe]
ganbold has quit [Excess Flood]
ganbold has joined #linux-rockchip
ganbold_ has quit [Remote host closed the connection]
FreezingCold has quit [Ping timeout: 240 seconds]
Astralix has quit [Remote host closed the connection]
FreezingCold has joined #linux-rockchip
<ferric> cosm: awesome... how?