mnemoc changed the topic of #arm-netbook to: EOMA: Embedded Open Modular Architecture - Don't ask to ask. Just ask! - http://elinux.org/Embedded_Open_Modular_Architecture/EOMA-68 - ML arm-netbook@lists.phcomp.co.uk - Logs http://ibot.rikers.org/%23arm-netbook or http://irclog.whitequark.org/arm-netbook/ - http://rhombus-tech.net/
ZaEarl has quit [Ping timeout: 240 seconds]
ZaEarl has joined #arm-netbook
Turl||| is now known as T|
T| is now known as ||||||||||||||||
CaCtus492 has joined #arm-netbook
|||||||||||||||| is now known as Turl_
CaCtus491 has quit [Ping timeout: 244 seconds]
Almamuerta has joined #arm-netbook
Turl_ has quit [Ping timeout: 245 seconds]
<lundman> what's the good word lads
<ManoftheSea> encephlogram
<lundman> elucidate
<lundman> aha, advanced MRI/eeg technique
mad_away has joined #arm-netbook
madmalkav has quit [Ping timeout: 255 seconds]
<ManoftheSea> it's a good word
<lundman> agreed
P4R4N01D has quit [Quit: Konversation terminated!]
<lukestuts> hello lundman!
<lukestuts> Your posts have been massively helpful to my Mele A2000 efforts
<lukestuts> Many thanks
<furan> what would it take to get cedar to work with jb?
<lundman> i made posts?
<lundman> but cool
<rm> which post
<lukestuts> I presume this is you! http://lundman.net/wiki/index.php/MeLe_A2000
<lukestuts> I would not have got started without that
Almamuerta has quit [Ping timeout: 245 seconds]
Quarx has joined #arm-netbook
TheLarch has quit [Quit: puff of smoke]
TheLarch has joined #arm-netbook
TheLarch has quit [Changing host]
TheLarch has joined #arm-netbook
TheLarch has quit [Remote host closed the connection]
TheLarch has joined #arm-netbook
TheLarch has quit [Changing host]
TheLarch has joined #arm-netbook
nibb_ has quit [Read error: Connection reset by peer]
nibb_ has joined #arm-netbook
lukestuts has quit [Ping timeout: 245 seconds]
rellla has joined #arm-netbook
ssspiff has joined #arm-netbook
ssspiff has quit [Changing host]
ssspiff has joined #arm-netbook
sspiff has quit [Ping timeout: 246 seconds]
ssspiff has quit [Ping timeout: 272 seconds]
<rz2k> if anyone will play with ly-f1 - touchscreen works ok with X if you disable multitouch define in ft5x_ts.c
<rz2k> sadly, I cant confirm it myself
<lundman> that it is
<mnemoc> rz2k: wiki? :)
lukestuts has joined #arm-netbook
<rz2k> done
Vayun has joined #arm-netbook
<mnemoc> rz2k: thanks :)
nibb__ has joined #arm-netbook
nibb_ has quit [Read error: Connection reset by peer]
sspiff has joined #arm-netbook
orly_owl has quit [Ping timeout: 265 seconds]
orly_owl has joined #arm-netbook
lukestuts has quit [Quit: Page closed]
<CIA-122> rhombus-tech: Sebastijan master * reafaacf3fedd /allwinner_a10/orders/sv7874.mdwn:
<CIA-122> rhombus-tech: Sebastijan master * r76c774c3a831 /allwinner_a10/orders/sv7874.mdwn:
<CIA-122> rhombus-tech: Joshua master * r03443798ca71 /allwinner_a10/orders/joshuait.html:
<mnemoc> wth, early access to script.bin is only "working" because -1 is true :<
<mnemoc> what a bs api
Almamuetya has joined #arm-netbook
<mnemoc> hno: any idea why, at machine .reserve() time, in 3.0 we can read 0xc3000000 without problems but in 3.4 it just gives 0s?
<RaYmAn> not that I have any clue what 0xc3000000 is, but perhaps difference in mmu setup?
<mnemoc> 0xc3000000 is PHYS_OFFSET + 0x3000000, where script.bin lives
<RaYmAn> is it being accesses through appropiate ioremaps or similar?
<mnemoc> .reserve() happens before that i think
<RaYmAn> before mmu is enabled?
<mnemoc> yes
<mnemoc> guts blame ARM_PATCH_PHYS_VIRT... but still far out of scope for me :<
<RaYmAn> I'd imagine ioremap should take care of mmu vs no mmu? so it's probably saner to do it?
<mnemoc> I'm kind of overwhelmed by new info at the moment... but shouldn't 0xc3000000 vs. bootloaders 0x43000000 mean it's already remapped?
<RaYmAn> hm
<mnemoc> accessing 0x43000000 makes it hung :<
<mnemoc> hang*
<RaYmAn> try something like IO_ADDRESS(0x43000000)
<mnemoc> #define IO_ADDRESS(x) ((x) + 0xf0000000)
<mnemoc> i.e. 0xc3000000
<RaYmAn> right, so that won't make any difference (but, still - you shouldn't hardcode the mapped address :S)
<mnemoc> :)
<mnemoc> arch/arm/mach-sun4i/include/mach/hardware.h:#define __io_address(n) __io(IO_ADDRESS(n))
<mnemoc> what is __io?
<RaYmAn> not sure tbh
<RaYmAn> I guess a compiler flag?
<sspiff> is there a way to run an android build for an Allwinner A10 device from SD instead of flashing it onto the device?
<RaYmAn> there should be if you build up a bootable sd and adjust android init scripts to correctly mount partitions on sd
<RaYmAn> (is there an easy-just-write-image-to-sdcard-solution? Not as far as I know, but it's possible)
<phh> changing init.*rc in the rootfs (usually an initramfs) should be enough
<RaYmAn> you also probably need to unpack the boot.img into seperate components that the sd-booting u-boot can read
<RaYmAn> iirc that one lacks boota(ndroid) command
<mnemoc> __io_address(0x43000000) is NULL :<
<RaYmAn> I'd try ioremap
<RaYmAn> just in case
<mnemoc> ioremap of what? 0x3000000 or 0x43000000?
<RaYmAn> 0x43
<mnemoc> ok
<RaYmAn> and then use ioread32 on the result value
<mnemoc> 0xc3000000 works fine at drivers-time
<mnemoc> the problem is at machine_desc.reserve()
<mnemoc> where 0xc3000000 work in 3.0 but not in 3.4, where it gives just zeros
<RaYmAn> does seem weird
<RaYmAn> but using 0xc3.... is wrong in any case :P Even if it did work (yeah, I know - doesn't really help you at all)
xenoxaos has quit [Quit: ZNC - http://znc.in]
<hno> mnemoc, isn't phys_init before MMU setup?
xenoxaos has joined #arm-netbook
<hno> physical SDRAM base addr is 0x40000000.
* mnemoc overwhelmed
<WarheadsSE> mornin
<mnemoc> moin WarheadsSE
<hno> mnemoc, no, .reserve is after MMU is enabled, but might be before full pagetable has been created.
<hno> SDRAM + 0x3000000 is out of reach until full pagetable is configured.
<mnemoc> so we would need to read the relevant values in another hook earlier than MMU i enabled?
<mnemoc> that explains why 0x43000000 hangs
<hno> There was code to add the script.bin region to the pagetable before I think.
<mnemoc> you mean the stuff in arch/arm/kernel/head.S?
<hno> yes
<mnemoc> it's still there...
<hno> add r0, r4, #((0x3000000 + KERNEL_START) & 0xff000000) >> 18
<hno> ?
<WarheadsSE> Reads about right to me
<mnemoc> it comes from 3.0
<mnemoc> maybe wrong part of head.S?
<mnemoc> I mean, place
rz2k has quit [Read error: Connection reset by peer]
<RaYmAn> I probably asked this before - does regular framebuffer work? E.g. boot into minimal linux -> simple app that reads/writes to framebuffer?
<WarheadsSE> Hmm.. I havent tried it in a while
rz2k has joined #arm-netbook
Almamuerta has joined #arm-netbook
<WarheadsSE> Yeah, fired up w/ fbdev driver
<WarheadsSE> Can't say I have a monitor attached atm though :P
<RaYmAn> lol
<RaYmAn> I more meant non-x11 fbdev stuff, but i guess it's the same
rellla_ has joined #arm-netbook
rellla_ has quit [Client Quit]
* mnemoc can't make any sense of __create_page_tables :<
rellla has quit [Ping timeout: 255 seconds]
<rm> RaYmAn, no
<rm> it doesn't work
<rm> or rather... fb console doesn't work
<rm> which is different from what you asked, I suppose
CaCtus492 has quit [Remote host closed the connection]
CaCtus491 has joined #arm-netbook
<sspiff> are all out/target/product/mk802/*.img's included in the zip when a build is finished>
<sspiff> what I really want to know, is the recovery also flashed when I flash a zip? And what kind of recovery is it?
gimli has joined #arm-netbook
<WarheadsSE> lundman: I think I might use your bits in making a script for Arch installation
croys has quit [Quit: croys]
<lundman> sure
<rm> soo...
<rm> where's the Linux kernel on nand*???
<rm> cant find it in the filesystem on nanda
<rm> is it written raw on nandb/nandc?
<mnemoc> there is a linux file iirc
<mnemoc> in nanda
Almamuerta has quit [Quit: Page closed]
<rm> there's a "linux" directory
ZaEarl has quit [Ping timeout: 240 seconds]
<rm> in it I see linux.bmp, linux.ini and u-boot.bin (<- 245KB)
<mnemoc> ah, ok :<
<mnemoc> and linux.ini doesn't hint the location?
<rm> linux.ini has: img_name = C:\linux\u-boot.bin
<mnemoc> meh
<mnemoc> so magic goes in uboot's env
<rm> http://elinux.org/Hack_A10_devices this suggests it can point to bImage
<mnemoc> and there is a bImage file in nanda?
<rm> not on mine
<mnemoc> here nandb is called "env", hno said that was u-boot's env
<mnemoc> with fw_env
Almamuetya has quit [Ping timeout: 264 seconds]
<rm> nandb and nandc dont seem to mount
<mnemoc> it's not mountable, it's raw
<rm> is bImage = uImage? :S
<mnemoc> you need to point /etc/fw_env.config to nandb
<mnemoc> iirc the unwrap uImage to get bImage
<mnemoc> they
<rm> I don't get the part about pointing /etc/fw_env.config to nandb
<rm> what is /etc/fw_env.config, where is it
<rm> in android?
<mnemoc> fw_setenv and fw_printenv are part of uboot tools
<mnemoc> to read and write the uboot env from userspace
<mnemoc> (I've never done it myself)
<mnemoc> or well.... use the serial console :|
<rm> nandc begins with ANDROID!
<mnemoc> dmesg | grep 'disk name = env'
<rm> grep not found :D
<mnemoc> busydbox grep ?
<rm> ah ok
<mnemoc> but i doubt you can do anything useful from android
<mnemoc> dump the env nand to a file and do it from linux
<rm> i can view it
<rm> w/o newlines
<mnemoc> afaik you can't just hexedit it, because some checksums
<rm> i cant see any reference to the kernel
<rm> hm
<mnemoc> no fatload?
<rm> no
<rm> sorry im typing on a tiniest bluetooth keyboard
<mnemoc> hipboi's uboot repo has a branch for booting from nand
<mnemoc> maybe you can play with that one
<rm> there is: read 40007800 boot
<rm> and this number repeats several times
<rm> i have a feeling the kernel is written raw at some offset :/
<mnemoc> feels like
<mnemoc> +so
<rm> maybe i can just edit linux.ini
<rm> but hos
<mnemoc> ignore uboot and try linux directly
<rm> but how to get bImage from uImage?
<mnemoc> scripts/build_sun4i.sh: ${OBJCOPY} -R .note.gnu.build-id -S -O binary vmlinux output/bImage
<rm> make bImage?
<mnemoc> nope, manually
<mnemoc> [ 0.000000] sunxi: script: base: 0xc3000000
<mnemoc> [ 0.000000] sunxi: script: version: 0.0.0 count: 0
<mnemoc> [ 0.180000] sunxi: script: base: 0xc3000000
<mnemoc> [ 0.180000] sunxi: script: version: 0.1.2 count: 76
<mnemoc> I give up. no script.bin for .reserve(). sh*t happens
<rm> ahh, finally the proper keyboard back
<rm> wtf I don't even have 2 keyboards
QingPei has joined #arm-netbook
<rm> and it just occured to me that with this kind of setup I need to *not* have anything bootable on the SD card
<mnemoc> i suppose it's better to find out why you can't boot from SD
<mnemoc> also, at least for me, SD is faster than nand
<mnemoc> in A10
<rz2k> Last login: Wed Dec 31 18:00:24 CST 1969 on ttyS0
<rz2k> mmm
<mnemoc> wow, got the real epoch
<mnemoc> I assumed it always used 2010-01-01 00:00
<rm> so it now "kind of boots"
<rm> but all symptomps point to "unable to mount root fs"
<rm> which one of mmcX_para is the sd slot?
<rm> and how do I access nanda of a bricked device? :D
<rm> okay I know, adb, right
<rm> or maybe not
<rm> since they talk about a booted android
<rm> I know I can reflash it from LiveSuit
<rm> but then do all the android dance :S
<rm> just to also replace the script bin
<rm> dammit
<rz2k> god damn it, all this time I was poking that Renesas SoC, I've used wrong git branch without tons of fixes. who the hell decied to point HEAD to clear source code drop from authors :/
<rm> don't you hate it when you find your own unanswered mailing list post
<rm> in google
<rz2k> lol
<rm> when searching about your problem
<specing> Haha
QingPei has quit [Remote host closed the connection]
<mnemoc> rm: booting from uSD :)
<rm> mwahaha~
<mnemoc> rm: booting from fel?
<rm> how/what is this
<mnemoc> fel is a low level usb protocol
<mnemoc> there is a `fel` tool in sunxi-tools
<mnemoc> you upload runnable stuff and then call it
<mnemoc> it's what livesuit uses
<mnemoc> no one has made fel-based installer yet, but it's your oportunity to become famous :)
<rm> is FEL the kind of device it shows up as when plugged in with the reset button pushed?
<mnemoc> yes
gimli has quit [Ping timeout: 244 seconds]
<mnemoc> you can read an address, write an address, and jump to an address
<mnemoc> very basic, but all you need :)
hipboi_ has joined #arm-netbook
<mnemoc> wb hipboi_
hipboi__ has quit [Ping timeout: 255 seconds]
<mnemoc> narf
<orly_owl> chinese internet
<orly_owl> #1
gimli has joined #arm-netbook
gimli_ has joined #arm-netbook
gimli has quit [Ping timeout: 272 seconds]
<rm> so
<rm> what's interesting is this bad MK802 has dram_clk = 433
<rm> whereas all others I have seen have =360
<rm> hm wait, this comes from the new firmware
<rm> so perhaps irrelevant
<rm> aaaand yesss it booted
<rm> 1) had to use a different SD card (4GB class 10, previous one was 8GB class 4)
<rm> 2) enters a reboot loop with the displaylink adapter connected - power issue?
<Mazon> fwiw, I had an raspberry pi that didn't want to boot - changed microsdhc-to-sd *adapter* - not the actual microsdhc card - and then it worked, go figure
<rm> but this still does not solve my SD booting issue
<rm> I tried 3 different SD cards, none boot from SD directly
<rm> this time I only succeeded due to putting a kernel with hardcoded root= to nanda
<rm> hahaahhaha
<rm> [ 0.000000] Total Detected Memory: 16MB with 1 banks
<rm> [ 0.000000] 64 MB reserved for MALI
<mad_away> XD
<rm> but then:
mad_away is now known as MaDMaLKaV
<rm> [ 0.000000] Memory: 640MB 16776528MB = 16777168MB total
<rm> [ 0.000000] Memory: 511960k/511960k available, 143404k reserved, 0K highmem
<MaDMaLKaV> some crazy numbers there
<specing> 16 terabytes RAM?
<specing> must have been sent from the future
Quarx has quit []
Vayun has quit [Ping timeout: 264 seconds]
<mnemoc> arch/arm/mach-sun4i/core.c: pr_info("Total Detected Memory: %uMB with %d banks\n", size, banks);
<mnemoc> uh, no. it's not that one
<mnemoc> rm: the hack that removes 64MB for mali isn't prepared for 16MB boards :)
<rm> I find it strange that I have full 512MB here
<rm> this kernel is compiled with Mali
<rm> or... maybe not
<rm> (not strange)
<mnemoc> IGNORE_ATAG_MEM enabled or disabled?
<rm> because this is supposed to be an 1GB board after all
<rm> # CONFIG_SUNXI_IGNORE_ATAG_MEM is not set
<mnemoc> then blame uboot
<rm> what u-boot
<rm> no uboot here!
<rm> the kernel is loaded directly by I dunno what
<mnemoc> then enable CONFIG_SUNXI_IGNORE_ATAG_MEM
<rm> ok
<rm> so where will it get the RAM info then?
schwatter has joined #arm-netbook
<mnemoc> allwinner's hack, a magic address on dram
<mnemoc> some boards don't get any atag_mem by default, so we can fallback to the hack
<mnemoc> but you do receive an atag_mem... crappy but it's there
<mnemoc> so https://github.com/amery/linux-allwinner/issues/60 is not only about using allwinner's hack when there is no atag, but also to fallback when the atag_mem is.... unlikely... whatever that means :<
<mnemoc> I guess we can assume no board will have less than 256? or 512?
<rm> I'd say 256
<rm> [ 2.580000] [cedar dev]: install start!!!
<rm> [ 2.590000] [cedar dev]: install end!!!
<rm> should also have had: zomg it worked!!!!
<rm> :D
<mnemoc> :)
<mnemoc> got your full GB?
<rm> mnemoc, yes I got a ping by mail, I suppose github sends those whenever people are mentioned
<rm> mnemoc, nah, just was amused by a lot of !!! in kernel messages
<rm> and the generally specific style of them :)
<mnemoc> :D
<rm> btw it doesn't seem stable
<rm> I suppose the reservation areas are all wrong
<mnemoc> all reservations are in the first 512MB
<mnemoc> fixed
<rm> apt-get install memtester => lockup
<mnemoc> :(
<rm> nope, a reboot
<rm> and also that reboot loop with the DisplayLink
<mnemoc> :(
<rm> mnemoc, I have Mali enabled and cedar and g2d and whatnot
<rm> yet Linux thinks it's got full 512 MB all to itself
<mnemoc> aha! have highmem enabled?
<rm> so it's either a problem with that, or a power issue
<rm> mnemoc, yes
<mnemoc> :(
<mnemoc> highmem should give you the 1GB-64MB :(
<rm> but I didn't recompile with
<rm> IGNORE_ATAG yet
<mnemoc> ah, ok
<rm> so I am running on that bozo 16MB RAM kernel
<mnemoc> :)
<mnemoc> i wouldn't expect that to be stable
Almamuerta has joined #arm-netbook
<mnemoc> rm: you still happy with your fix to https://github.com/amery/linux-allwinner/issues/59?
<rm> yes
<mnemoc> ok, i'll merge it then
<mnemoc> rm: can you give me proper author info? :)
<mnemoc> rm: don't worry, found it on the ML
<rm> the ignore_atag kernel does not seem to boot :/
<rm> or maybe it botched a write to nanda
<rm> due to the broken kernel not being stable
<rm> but that's a small probability
<rm> I checked md5sums
<rm> theoretically they could've been checked from file cache
<mnemoc> if ignore_atag doesn't work, i can only think of mem=448@0x40000000 mem=512@0x60000000
j1nx_ has joined #arm-netbook
<rm> okay, building with that
<hno> mnemoc, have you figured out the script.bin early access problem?
schwatter has quit [Ping timeout: 245 seconds]
hackandfab has quit [Quit: Leaving]
markvandenborre has quit [Ping timeout: 252 seconds]
ZaEarl has joined #arm-netbook
markvandenborre has joined #arm-netbook
nibb__ has quit [Read error: Connection reset by peer]
nibb__ has joined #arm-netbook
<rm> argh, it trashed the FS while it was unstable
<rm> and I blamed all sorts of unrelated things
<rm> recompiled 2 times more
<rm> due to it not booting
<Turl> rm nanda is extremely fragile
<rm> not on nanda
<rm> on the SD
<Turl> even if you unmount it cleanly the bootloader might not like it :/
<rm> I sync couple of times before and after unmount :)
<Turl> you shouldn't need to :)
<rm> yes but then I just pull the power
nibb__ has quit [Read error: Connection reset by peer]
<Turl> I copied a known good script.bin once
<Turl> unmounted
nibb__ has joined #arm-netbook
<Turl> issued a reboot
<Turl> and the device was dead
<Turl> nad to livesuite it
<Turl> had*
polvo has joined #arm-netbook
ka6sox-away is now known as ka6sox
MaDMaLKaV has quit [Quit: Leaving]
mpthompson has quit [Quit: Depression is merely anger without enthusiasm]
polvo has quit [Quit: ChatZilla 0.9.88.2 [Firefox 13.0/20120601045813]]
mshroyer has joined #arm-netbook
nibb_ has joined #arm-netbook
nibb__ has quit [Read error: Connection reset by peer]
<CIA-122> rhombus-tech: lkcl master * r035b0119dfbb /allwinner_a10/orders/ (31 files in 2 dirs): order processing cleanup
<CIA-122> rhombus-tech: Joshua master * r33bd6a8145ff /allwinner_a10/orders/joshuait.mdwn:
CaCtus491 has quit [Remote host closed the connection]
CaCtus491 has joined #arm-netbook
<CIA-122> rhombus-tech: Joshua master * r97ba91cd4790 /allwinner_a10/orders/not_valid/joshuait.mdwn: Order Information
<CIA-122> rhombus-tech: Joshua master * r21d59e075509 /allwinner_a10/orders/joshuait.mdwn:
<CIA-122> rhombus-tech: Joshua master * r519058ed92ae /allwinner_a10/orders/not_valid/joshuait.mdwn:
<CIA-122> rhombus-tech: Joshua master * r5eb7436bf4f6 /allwinner_a10/orders/not_valid/joshuait.mdwn:
<CIA-122> rhombus-tech: Joshua master * r8110cffd9318 /allwinner_a10/orders/joshuait.mdwn:
<CIA-122> rhombus-tech: Joshua master * r8283f4d7c842 /allwinner_a10/orders/joshuait.mdwn:
j1nx_ has quit [Quit: Few women admit their age. Few men act theirs.]
<CIA-122> rhombus-tech: Joshua master * r91f39b2b60ab /allwinner_a10/orders/joshuait.mdwn:
gimli_ has quit [Quit: Verlassend]
Kraln- has joined #arm-netbook
mshroyer has left #arm-netbook ["ERC Version 5.3 (IRC client for Emacs)"]
Kraln has quit [Ping timeout: 260 seconds]
Vayun has joined #arm-netbook
<mnemoc> hno: no, I gave up. it's way over my current level :(
<RaYmAn> psh
<RaYmAn> excuses! :P
<mnemoc> :)
<mnemoc> I commented out the parts that rely on that capability and moved on
<RaYmAn> I guess that works too :P
<mnemoc> at least it allows me to start cleaning up the gpio code
<RaYmAn> yeah
<RaYmAn> s/cleaning up/rewriting with just register setup used/
<RaYmAn> :P
<RaYmAn> I suspect it'll have very little to do with the current drivers :)
<RaYmAn> I hope anyways :P
<mnemoc> i'm refactoring but keeping old api too
<RaYmAn> :/
<mnemoc> baby steps ;-)
<RaYmAn> yeah, I know
<RaYmAn> can't you just keep the wrapper in seperate files? :>
<mnemoc> that's the idea, yes
<RaYmAn> okay, np then :)
<mnemoc> plat/compat.h :p
<mnemoc> going to sleep, cu tomorrow