ChanServ changed the topic of #linux-rockchip to: Rockchip development discussion | Wiki at http://linux-rockchip.info | Logs at http://irclog.whitequark.org/linux-rockchip
<naobsd> I don't need whole bleeding edge things
<naobsd> I just want to try new things for rockchip
<naobsd> btw I totally forgot about gmac on rk3288 ;)
<srao_> I think linux-next will have way more bleeding edge code, so it's good in that you will have to pull in less stuff while a platform is still under development upstream (like RK) but it's bad in that you will run into other unrelated bugs
<srao_> probably merging in Heiko's github tree (or just using it as a base) would be easiest
<naobsd> oh I forgot to check what is arm core power source on my box...
<naobsd> guess REG3@act8846
<naobsd> (possibly pwm?)
<naobsd> now trying cpufreq things...
<naobsd> it works, very nice
<naobsd> it survives openssl speed -multi 4 :)
naobsd has quit [Quit: Page closed]
hipboi has quit [Ping timeout: 240 seconds]
hipboi has joined #linux-rockchip
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
GriefNorth has joined #linux-rockchip
Astralix has joined #linux-rockchip
Astralix1 has quit [Ping timeout: 250 seconds]
hipboi has quit [Read error: Connection reset by peer]
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
naobsd has joined #linux-rockchip
ganbold_ has quit [Remote host closed the connection]
ganbold_ has joined #linux-rockchip
hipboi 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
<naobsd> ah, time goes fast on mainline...
<naobsd> +17sec within 1min
<naobsd> ^at 1.6GHz fixed
<naobsd> ah, I did "sleep 60", it may not be real 60sec
<naobsd> ah, sleep 60 doesn't wait 60sec ;)
<naobsd> at 1.6GHz, time goes +17s within 43(60-17)s
<naobsd> at 312MHz, +2.5s within (60-2.5)s
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
<naobsd> "echo jiffies > current_clocksource" is workaround...
<naobsd> (not good)
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
FightingCat has joined #linux-rockchip
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
naobsd has quit [Quit: Page closed]
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
AstralixNB has joined #linux-rockchip
AstralixNB has quit [Remote host closed the connection]
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
naobsd has joined #linux-rockchip
<naobsd> oh, I noticed someone post patches to support Amlogic AML8726-M series SoC...
<naobsd> if my memory correct, their ffmpeg-based code is open
<naobsd> and binary blob for NAND :)
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
AstralixNB has joined #linux-rockchip
VargaD has quit [*.net *.split]
hipboi has quit [Ping timeout: 246 seconds]
hipboi has joined #linux-rockchip
ferric has quit [Ping timeout: 272 seconds]
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
VargaD has joined #linux-rockchip
ferric has joined #linux-rockchip
ferric has quit [Ping timeout: 260 seconds]
cyrozap has quit [Ping timeout: 240 seconds]
cyrozap has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
hramrach_ has quit [Remote host closed the connection]
hramrach_ has joined #linux-rockchip
<naobsd> oh
<naobsd> emmc is always mmc1 and mmcblk1... I thought it will be mmc0
<tomeu> I'm wondering why the thermal series for the rk3288 isn't using the dt bindings for thermal zones
libv has quit [Ping timeout: 250 seconds]
<tomeu> mmind00: do you happen to know?
<mmind00> nope I don't know, but I think it was pointed out in every patch iteration that it _should_ use the thermal zones
<mmind00> tomeu: you could reply to the most recent version and point to thermal zones again ;-)
<tomeu> hmm, for some reason I haven't noticed
<tomeu> but yeah, will do that
<mmind00> at least I know that I asked about this in the first version, and remember the same thing popping up in a later one ...
<tomeu> I saw people pointing out to trip points, guess it should use all of it
<mmind00> yeah, the current binding in v5 seems to implement it's own trip point handling ... which looks very wrong when we have the generic thermal zones
RayFlower has quit [Quit: RayFlower]
<naobsd> oh I cannot set cpufreq higher than 1.2GHz on rk3288... maybe regulator issue?
hipboi has quit [Ping timeout: 260 seconds]
hipboi has joined #linux-rockchip
bengal has joined #linux-rockchip
hipboi has quit [Ping timeout: 245 seconds]
cyrozap has quit [Ping timeout: 245 seconds]
cyrozap has joined #linux-rockchip
hipboi has joined #linux-rockchip
RayFlower has joined #linux-rockchip
FightingCat has quit [Ping timeout: 260 seconds]
bengal has quit [Ping timeout: 245 seconds]
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
nighty^ has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
cyrozap has quit [Ping timeout: 255 seconds]
naobsd has quit [Quit: Page closed]
cyrozap has joined #linux-rockchip
hipboi has quit [Read error: Connection reset by peer]
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
hipboi has joined #linux-rockchip
libv has joined #linux-rockchip
hipboi has quit [Ping timeout: 245 seconds]
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
hipboi has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
GriefNorth has quit [Ping timeout: 258 seconds]
FightingCat has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
<rperier> "Add basic drm support" (see ML) ... mhhh... interesting
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
hipboi has quit [Ping timeout: 272 seconds]
c0d3z3r0 has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
hipboi 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]
naobsd has joined #linux-rockchip
RayFlower has joined #linux-rockchip
<naobsd> I thought there was a HDMI part...
<rperier> karlp: what works today on rk3066 with mainline (I don't know what is your board, but I would like to buy a marsboard probably)
<naobsd> I should try rk3066 too...
ferric has joined #linux-rockchip
<naobsd> there is axp209 driver in mainline, nice...
<rperier> yes, it would be nice to have good support on the rk3188 and rk3066 (by a good support I mean a support in the kernel for 90% of the hw components and a stable support)
<rperier> (at least, this is my goal in few months when I will have development experience in the kernel and more boards)
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
ferric has quit [Ping timeout: 260 seconds]
<karlp> sorry, I didn't get as much time last night as I hoped. It seems to probe the emac and the phy, but it's failing to setup the phy interrupt pin, because I don't know how to figure it out from the factory running kernel
<karlp> I've tried one or two values that seemed plausible, but I don't know where to find it in 3.0.x code, to try more options from the different android kernels.
<karlp> I'm somewhat tempted to script up something that just generates all the combinations for me, and flashes them all, there's not _that_ many gpio pins....
<karlp> be kinda nice if I could just flash a new dtb, instead of having to reappend the kernel and dtb each time.
<karlp> I have a minix neo x5 mini, which is a generic STB style device, and a ug802 clone which is a generic stick device,
<karlp> I'm only really working on the x5 right now, as I've put a serial console on it, and it has an ethernet port, the tv stick is a bit more of a hassle to work with.
<naobsd> ah
<naobsd> I guess
<naobsd> vmac in 3.0.36 can run polling mode
<karlp> that would certainly explain the complete absence of anything I could find that even hinted at setting up an interrupt for the phy :)
<rperier> karlp: is there another way to get the kernel image without ethernet from these devices ? (emmc or something) because in this case you might use kexec instead of flash each time
<karlp> what do you mean?
<karlp> "get the kernel without ethernet" ?
<AstralixNB> are you talking about 3288 or 3188?
<karlp> it's connected via usb and serial, I just power cycle it with the button pressed and use upgrade_tool to flash a new rkcrc(kernel+dtb)
<karlp> this is 3066
<rperier> I mean if you find a way to transfer your kernel image to your device (transfer it as a binary file), you could reboot on your new kernel on the fly using kexec
<AstralixNB> For 3288 there should be a working uboot that just needs the eth device activated.
<AstralixNB> For 31xx and 30xx the uboot is incomplete... I am at it already.
<rperier> as I don't know your devices, I prefer ask :)
<AstralixNB> But I am waiting for my SD-Cards to not always kill my nand flash
<karlp> rperier: yeah, well, either ethernet or wifi would need to work for that, and it looked like ethernet was much closr to working :)
<rperier> AstralixNB: about 3066
<rperier> I see, it sucks
<AstralixNB> rperier, got it.
<karlp> wifi in this device is an sdio connected variant of the rtl8188eus, so I'll leave that off for a bit longer I reckon....
<rperier> AstralixNB: until uboot works on 3188 I use tftp + kexec from user space :)
<AstralixNB> rperier, is this version on github?
<AstralixNB> I have only uboot that either lack of thos or that function... Nothing complete.
c0d3z3r0 is now known as c0
c0 is now known as c0d3z3r0
<naobsd> uboot can load/run kernel/ramdisk from nand/emmc/sd, that's all, and it's enough for rockchip ;)
<AstralixNB> naobsd, which uboot?
<AstralixNB> I do have uboot for 3188 that can load from emmc but not NAND and has no eth.
<naobsd> I only confirmed NAND/SD on _RR_
<AstralixNB> Does this work for 3188? or 3288 only?
<naobsd> and I think it doesn't boot on HTC-T010-V2 board (RK3188)
<naobsd> I cannot try eMMC & RK3188
<AstralixNB> 99% of 3188 have NAND, no eMMC
<naobsd> AstralixNB: that's for Radxa Rock, RK3188.
<AstralixNB> But RK wrote a special version supporting emmc for Toshiba and Lenovo
<naobsd> probably that code supports eMMC
<naobsd> I had to modify a bit for SD boot
<naobsd> (see git history)
<AstralixNB> If it works on radxa, it must support RK-NAND
<naobsd> I said I confirmed NAND(and SD)
<AstralixNB> How can you read the rk-nand without this proprietary driver from them? Or do you use thwe usual open source FTL drivers?
<rperier> naobsd: it's enough except when you build your kernel all the time and you need a fast way to load it on the fly to your device :)
<naobsd> I should try "fastboot boot" with u-boot
<AstralixNB> with the current loaders from RK you can load the loader directly to DRAM and execute them without having to flash them.
<naobsd> AstralixNB: sounds good, how can I do it?
<AstralixNB> in the linux upgrade tool 1.21 you can use the erase flash command to get the MASKROM loader active.
<naobsd> ah
<AstralixNB> Then you can use rb or db command to load and execute any loader
<naobsd> I know, sorry
<naobsd> I misread "load kernel directly" :(
<naobsd> I don't change loader so often...
<AstralixNB> People often tell, that the linux tool is not working, but they miss the information, that for some commands the system has to be in MASK-ROM loader and for other commands a 2nd stage Loader has to be active.
<naobsd> what I want is kernel loader which supports usb/ether/sdmmc/etc
<karlp> well, kexec froma kernel that supports it.... ? ;)
<AstralixNB> But if uboot runs, you should be able to use eth to load a kernel via tftp.
<naobsd> uboot runs,
<AstralixNB> If you do not have an eth, you could modify a fastboot command to your needs?
<naobsd> it loads kernel/ramdisk from NAND/EMMC/SD
<karlp> if uboot runs, and if ubott has ethernet support ... ;)
<naobsd> it's RK uboot.
<AstralixNB> Ok, I think I really switch to uboot with my primary hobby tasks...
<AstralixNB> As soon as my ordered SD-Cards arrive...
<naobsd> I didn't see code so much
<naobsd> but I think EMMC/SD support is not implemented as standard interface in u-boot
<AstralixNB> I cannot promise if i start at uboot or redboot, But it will be a loader for RK3xxx
<AstralixNB> Of course it is... With all devices I have!
<naobsd> it just read raw sector, it's enough for RK!
<AstralixNB> Even the old VoIP Telephone I have has an uboot that supports MMC/SD
<rperier> I agree that it's enough for normal user, but not for kernel developers/contributors. Copy your kernel image to emmc/sd each time is like flash the kernel to NAND , it's a waste of time, imho
<rperier> users*
<rperier> but that's great to have uboot running
<naobsd> I think dwc mmc is supported in u-boot upstream
<AstralixNB> sure it should be
<AstralixNB> but dw_mmc must have been support for a long time. I booted my commercial things based on iMX from eMMC and SD
<naobsd> so port rk core to upstream may be better
<AstralixNB> sure
<naobsd> but probably no arc emac and no dwc usb
<AstralixNB> why not?
<naobsd> I'm not sure
<naobsd> later one is common, it may be supported
<AstralixNB> I'll check if it is more olikely to have that in uboot or redboot. I have closer contacts to redboot. and uboot was a big chaos for many years.
<AstralixNB> -o
<naobsd> no dwc* under host/ (yet?)
<AstralixNB> The USB host will not save the day. Only if you can add a USB-ETH adapter that must be supported too, you can develop on ETH less systems.
<naobsd> but at least there is patch, good
<naobsd> usb storage :)
<AstralixNB> but for what reason? Where is the difference between swapping sticks that swapping cards?
<karlp> is redboot still in active developemnt? I thought redboot was mostly all fractured and localised per device and not used much anymore?
<naobsd> everyone makes root file system on sd?
<AstralixNB> Ouch, sorry. Not reboot, but barebox
<karlp> right, ok :)
<naobsd> and no need to swap card/sticks, just put new file on file system on them
<AstralixNB> naobsd, my primary target was to completely eliminate the RK NAND driver and use an existing open source solution to drive the NAND.
<naobsd> AstralixNB: I see, I really have no idea about NAND, so I prefer other media
<naobsd> btw is there any RK3288 product which uses NAND? all I can see is eMMC...
<AstralixNB> The background is, that we often encounter speeds of 20...40MB/s accessing an SD card of an Android Tablet, but only 2..5MB/s accessing the internal NAND flash, while at least 1 Core is at 100% load.
<AstralixNB> Yes, most 3288 deices use eMMC
<AstralixNB> I only saw a 3288 tablet that has the footprint for NAND and eMMC but all known devices where equipped with eMMC, the NAND was not placed.
<AstralixNB> Ok, I need to do a bit of work again... see you all later
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
ferric has joined #linux-rockchip
ferric has quit [Ping timeout: 272 seconds]
nighty^ has quit [Remote host closed the connection]
<naobsd> oh, firefly is released
<naobsd> should I ask about special offer for developers? ;)
<naobsd> I hope they stay here as like as hipboi...
bengal has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
ferric has joined #linux-rockchip
hramrach_ has quit [Ping timeout: 264 seconds]
hramrach_ has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
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
nighty^ has quit [Quit: Disappears in a puff of smoke]
nighty-_ has joined #linux-rockchip
AstralixNB has quit [Ping timeout: 250 seconds]
BorgCuba has joined #linux-rockchip
<BorgCuba> naobsd, does u-boot work for you?
<BorgCuba> have you tried rkflashtool in conjunction with u-boot?
<BorgCuba> so I assume that the exception vectors are located at 0x0 for ARM?
<naobsd> BorgCuba: I think I said it works on RR
<naobsd> I'm not sure I used rkflashtool with u-boot, but I cannot find any reason not to work
<naobsd> and... is there any relation between u-boot, rkflashtool, and exception vectors...
FightingCat has quit [Remote host closed the connection]
<BorgCuba> for me there is, when I used rkflashtool (doesnt matter which command) it rebooted
<BorgCuba> so I inspected the first 8 dwords at 0x0 which were at least not of the type ldr pc,=handler_addr
<BorgCuba> I think I tried your u-boot code, but I build it myself with the following result: 1) hangs at CPLL, GPLL setup 2) when skipping the pll setup it gave me an unhandled access violation
<BorgCuba> anyway, since the unhandled access violation message comes from u-boot at least this exception works...
<naobsd> where is "first 8 dwords at 0x0" ?
<BorgCuba> yes
<naobsd> what I said is I don't understand...
<naobsd> I guess
<naobsd> if you used rkflashtool with u-boot, you can see panic message from u-boot
<naobsd> around 0x0
<BorgCuba> no, it reboots
<naobsd> anyway
<naobsd> u-boot doesn't work
<naobsd> then, I cannot expect any reasonable behavior
<BorgCuba> at least from me it doesnt work as expected
<naobsd> regardsless of rkflashtool
<naobsd> or
<naobsd> upgrade_tool works with your non-working u-boot???
<BorgCuba> what do you mean by that?
<naobsd> is there any relation between rkflashtool and u-boot?
<naobsd> well
<naobsd> is there any relation between rkflashtool and "u-boot problem"?
<BorgCuba> well, the rockchip u-boot has "rockusb" code inside
<naobsd> I know
<BorgCuba> and I think its the default usb interface
<naobsd> what's "default usb interface"?
<naobsd> u-boot tries to load kernel/boot/recovery from NAND/SD
<naobsd> if load failed, u-boot goes "rockusb" mode
<BorgCuba> be default I mean, if you build u-boot as-is (no config changes) it will enumerate as a "rockusb" device
<naobsd> or if you press a key to go rockusb mode, it goes rockusb mode too
<BorgCuba> yes, exactly
<BorgCuba> currently I am unsure where the problem lies, I put printfs everywhere in the usb and interrupt code but I never see any of these messages
<naobsd> well
<naobsd> u-boot doesn't work
<BorgCuba> okay :-)
<naobsd> but it's not caused by rkflashtool
<naobsd> is there any relation between rkflashtool and "u-boot problem"?
<naobsd> I think rkflashtool with u-boot will work when u-boot works fine
<BorgCuba> yes, exactly
<naobsd> may I forget about rkflashtool when talking about u-boot?
<BorgCuba> naobsd, is there a version of the proprietary loader that support sdram upload und execute?
<BorgCuba> naobsd, yes you may
<naobsd> "fastboot boot" may work (I never tried yet)
<naobsd> ah, it's not proprietary loader, probably.
<BorgCuba> but by default the proprietary loader enumerates as a "rockusb" device
<BorgCuba> can you switch that or is there a special "fastboot" loader?
<naobsd> then, no proprietary loader can load code to sdram(or I just don't know about it)
<naobsd> u-boot supports fastboot mode
<BorgCuba> I am just curious since the commands are defined in the rkflashtool source (although commented out)
<BorgCuba> naobsd, could you probably try it out?
<BorgCuba> that would be great
<naobsd> fastboot?
<BorgCuba> yes, or rockusb
<naobsd> what to try with rockusb?
<BorgCuba> I could give you a slightly modified rkflashtool version
<BorgCuba> upload to sdram
<naobsd> I think you can try with rockusb...
<BorgCuba> I tried with proprietary rockusb and it does not work
<naobsd> yes
<BorgCuba> probably you could try using your version of u-boot?
<naobsd> ah
<BorgCuba> here is a modified version of rkflashtool: http://pastebin.com/QJNM7Zja
<naobsd> you can read u-boot code
<BorgCuba> I added 'n' switch
<naobsd> that command is not implemented
<BorgCuba> it is
<BorgCuba> FW_SDRAMRead10();
<BorgCuba> defined in board/rockchip/common/common/rockusb/rockusb.c
<BorgCuba> or better FW_SDRAMWrite10
<BorgCuba> and there is a CONFIG that allows for booting bare code or a kernel
<naobsd> oh sorry, then I misunderstood
<BorgCuba> I mean this feature would be really nice to have if you want to try out some kernels without having to write some to flash or sdcard every time
<naobsd> strange, this rockusb.c is totally different to one in my memory...
<naobsd> we can do anything with u-boot, it is source code
<BorgCuba> yes, we can modify it an compile it
<naobsd> n command just write
<naobsd> no execute
<BorgCuba> no, not yet
<BorgCuba> but I thought maybe you could do something like upload, download and diff
<BorgCuba> to see if this works
<BorgCuba> gcc rkflashtool.c -lusb-1.0 -o rkft
<naobsd> ah I remembered
<naobsd> when I saw rockusb code
<naobsd> I wanted to find out mask rom command
<BorgCuba> I hat a look at the mask rom code
<naobsd> (human readable code please)
<BorgCuba> human readable maskrom code?
<BorgCuba> btw, I think the bootorder from rk3066 doesnt allow to boot from sdcard
<BorgCuba> this is the md5 of my rk3188-t bootrom: 2595920f6b91cd4e0de0ce08926602bc
<BorgCuba> 16k
<naobsd> ah
<naobsd> totally remembered
<naobsd> what I saw is rockusb code in rk3288 u-boot
<naobsd> 0x1[789] is missing ;)
<naobsd> so I need to try rkflashtool with rk3188 u-boot...
<naobsd> (or implement these commands)
<naobsd> preparing rk3188 board...
<BorgCuba> btw, this function table is located at offset 0x2718 in my maskrom code
<BorgCuba> and is followed by the usb device descriptor at offset 0x277c
<BorgCuba> usb request handler is at 0x198C
<BorgCuba> naobsd, for maskrom usb commands: I think they are similar to the rk27xx rom commands
<BorgCuba> btw, back in the day I reversed the rk27xx usb rom commands and the fw encoding
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
<BorgCuba> yes, use your rk3188 board, boot into u-boot and try out if rkflashtool works (use 'p' as a start, just to see if it works)
<BorgCuba> and if it works try n, then m and check if you can read back what you wrote to dd-sdram
<naobsd> strange, 'n' ignores size?
<naobsd> anyway, it seems working
<BorgCuba> cool!
<BorgCuba> yes, you can remove the comment from "size -= sizeRead"
<BorgCuba> dont know why I put it there
<naobsd> oh...
<BorgCuba> so you wrote to sdram (m), read back from the same address (n) and compared it to what you send in the first step?
<naobsd> and rk loader 2.10 seems not work
<naobsd> I send /dev/zero
<BorgCuba> for writing to ram?
<naobsd> try another data...
<BorgCuba> okay
<naobsd> oh well
<BorgCuba> what happened?
<BorgCuba> btw, here is some code to test executing from ram: http://pastebin.com/ia2Q4z80
<BorgCuba> you need to implement the execute command in rkflashtool before
<BorgCuba> copy http://pastebin.com/ia2Q4z80 to a file named dotty.S and run the commands from the comment at the top
mrueg has quit [Quit: No Ping reply in 180 seconds.]
<naobsd> hmm
<BorgCuba> what happened?
<naobsd> send always 16KiB (whole buf)
<naobsd> otherwise, it hangs
<BorgCuba> okay
<naobsd> well
<naobsd> it may be wrong
<BorgCuba> whats wrong?
<naobsd> anyway, send_cmd/send_buf should use same size
<BorgCuba> yes, you are probably right. I couldnt test it
<naobsd> RKFT_MEM_INCR works too
<BorgCuba> this would only happen if the file isnt a multiple of RKFT_BLOCKSIZE (which is a quite common file size)
<naobsd> 16KiB was intended for NAND access unit
<BorgCuba> yes
<naobsd> and image file is always padded to 16KiB * n
mrueg has joined #linux-rockchip
<naobsd> I don't know about block size for sdram access
<naobsd> but we can read u-boot code
<BorgCuba> so you say after all it is possible to upload data to sdram using rkflashtool und u-boot?
<BorgCuba> let me see
<naobsd> ... and proprietary loader (RK loader 2.10) works too
<naobsd> you can try yourself now
<BorgCuba> there is this line: ReadBulkEndpoint(gCBW.Len, DataBuf);
<BorgCuba> I will
<BorgCuba> could you paste your 'n' code
<naobsd> if ((sizeRead = read(0, buf, RKFT_MEM_INCR)) <= 0) {
<naobsd> send_cmd(RKFT_CMD_WRITESDRAM, offset-0x60000000, RKFT_MEM_INCR);
<naobsd> send_buf(RKFT_MEM_INCR);
<naobsd> well
<naobsd> use same size for send_cmd/buf, that's all
<naobsd> e.g. you can read more at outside of loop ;)
<naobsd> gCBW.Len <- probably any size
<naobsd> ah, probably there is upper limit
<naobsd> but 1 may be ok
<BorgCuba> endpoint size?
<naobsd> xfer size
<naobsd> bulk xfer
<naobsd> uint16 Len;
<naobsd> must be less than 65536
<naobsd> ah UsbBootLinux is commented in u-boot
<naobsd> I hope RK loader enables exec command
<naobsd> probably it needs 2 address, 1 for code, 1 for parameter
<naobsd> there is a screen to load kernel to ram in RKAndroidTool.exe
<naobsd> I think that doesn't work
<BorgCuba> well, I guess it is possible to write a small kernel loader
<BorgCuba> anyway, the data I read back is different from what I send
<naobsd> which address you used
<naobsd> I tried 0x60000000
<BorgCuba> I tried ./rkft n 0x62000000 512 < /dev/zero
<BorgCuba> could you probably past the whole 'n' case here: http://pastebin.com/
<BorgCuba> or maybe just the whole rkflashtool.c
<naobsd> please understand it's 4AM...
<BorgCuba> okay
<BorgCuba> 0x60000000 works for me too
<BorgCuba> naobsd, thanks for helping me out
<naobsd> don't touch running code...
<naobsd> I don't know how much area is mapped in RK loader
<naobsd> but it loads some MB kernel and ramdisk
<naobsd> sleepy
<BorgCuba> thanks again
<BorgCuba> I'll leave now as well
<BorgCuba> I'll try out the "execute" command tomorrow
<BorgCuba> bye
BorgCuba has quit [Quit: leaving]
<naobsd> we can read source, it's very nice...
<naobsd> and I'm too lazy ;)
<naobsd> I tried much more things w/o source...
<naobsd> ^when I wrote rkflash
<naobsd> zzz
nighty-_ has quit [Quit: Disappears in a puff of smoke]
bengal has quit [Ping timeout: 255 seconds]
bengal has joined #linux-rockchip
Bludot 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
ChiaSmurf has quit [Ping timeout: 246 seconds]
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
RayFlower_ has joined #linux-rockchip
RayFlower has quit [Ping timeout: 240 seconds]
RayFlower_ is now known as RayFlower
johnsu01 has quit [Ping timeout: 240 seconds]
johnsu01 has joined #linux-rockchip
johnsu01 has quit [Changing host]
johnsu01 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
bengal has quit [Quit: Leaving]
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
akaizen_ has joined #linux-rockchip
akaizen__ has joined #linux-rockchip
akaizen has quit [Ping timeout: 272 seconds]
RayFlower has quit [Quit: RayFlower]
akaizen_ has quit [Ping timeout: 255 seconds]
<karlp> mmind00: dianders: whos tree has this in it at the moment? "ARM: Rockchip: add cpuclk handling - clock-tree part"
<karlp> I can't find it in github/mmind or git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip.git which is what I presume I need to try out the frequency scaling stressing with i2c that max was after
<karlp> I think that's where mmind's/dianders proposed patches go, but I don't think they're there yet.
<karlp> naobsd: did you say you'd added usb to the dts files for radxarock? have you got them anywhere?
<naobsd> mmind's dts
<naobsd> v3.18-next/dts
<naobsd> I'll push my tree soon...
<karlp> thanks, found it.
<karlp> well, that easily detected them, but it's not linking up some driver somwhere: http://pastebay.net/1495836
<karlp> those vids/pids should have been supported
<karlp> I can see the devices get power, is it save to presume I don't need any vbus gpio, or vbus regulator control? or is it likely that it was just setup by the bootloader, and I'll need to work them out eventually?
Bludot has quit [Quit: Connection closed for inactivity]