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/
Guest66946 has joined #arm-netbook
ZaEarl has quit [Read error: No route to host]
ZaEarl has joined #arm-netbook
DEAT_ is now known as DEAT
<xxiao_> it's indeed strange there is not a good android channel
Guest66946 is now known as xenoxaos
<WarheadsSE> mnemoc: here's a new one: X: malloc.c:2810: mremap_chunk: Assertion `((size + offset) & (_rtld_global_ro._dl_pagesize-1)) == 0' failed.
<DonkeyHotei> ok, so i managed to compile mali
<DonkeyHotei> X11 crashes with no message
<DonkeyHotei> last line in Xorg.0.log is:
<DonkeyHotei> [ 17.858] (II) Loading /usr/lib/xorg/modules/drivers/mali_drv.so
<WarheadsSE> just aborts?
<DonkeyHotei> the X process is still in ps
<DonkeyHotei> infinite loop of some sort?
<DonkeyHotei> oh wait
<DonkeyHotei> 705 tty8 S<s+ 0:00 /usr/bin/X :0 -br -once -config /etc/X11/xorg.conf.failsafe -logfile /var/log/Xorg.failsafe.log
<DonkeyHotei> so idk
<DonkeyHotei> simply exits with code 1, no msg
<DonkeyHotei> WarheadsSE: any different for you?
<WarheadsSE> ^
<DonkeyHotei> hmm, seems ld can't load the .so
<DonkeyHotei> wtf
<WarheadsSE> i just get driver errors.
<DonkeyHotei> like what?
<WarheadsSE> X: malloc.c:2810: mremap_chunk: Assertion `((size + offset) & (_rtld_global_ro._dl_pagesize-1)) == 0' failed.
lerc has joined #arm-netbook
<DonkeyHotei> WarheadsSE: where do you see that?
<WarheadsSE> right about the time it should be kicking the hdmi on
<WarheadsSE> rather frustrating.
egustafs has joined #arm-netbook
egustafs has quit [Quit: Ex-Chat]
<furan> hrm
<furan> having trouble figuring out what the orrhi instruction does
<CIA-122> rhombus-tech: Jean-Luc_Aufranc master * rd5d1f445a73e /allwinner_a10/nightly_build_images.mdwn:
<CIA-122> rhombus-tech: Jean-Luc_Aufranc master * rff14595ff804 /allwinner_a10/nightly_build_images.mdwn:
gimli_ has joined #arm-netbook
Almamuerta has joined #arm-netbook
Almamuerta has quit [Quit: Page closed]
ssvb has quit [Quit: Leaving]
P4R4N01D has joined #arm-netbook
lundman has joined #arm-netbook
gimli_ has quit [Remote host closed the connection]
<lundman> Hmm anyone remember how you can register new media file types in android, say if I wanted to add MKV support to the OS, what the API calls are
rellla has joined #arm-netbook
rellla has quit [Client Quit]
hipboi has quit [Quit: Leaving]
rellla has joined #arm-netbook
Quarx has joined #arm-netbook
eebrah has joined #arm-netbook
MMlosh has joined #arm-netbook
<CIA-122> rhombus-tech: Dmytro master * r90df2b864630 /allwinner_a10/orders/dmytro.mdwn:
arokux has quit [Remote host closed the connection]
arokux has joined #arm-netbook
madmalkav has joined #arm-netbook
ssvb has joined #arm-netbook
hipboi has joined #arm-netbook
Quarx has quit []
hackandfab has joined #arm-netbook
Quarx has joined #arm-netbook
<CIA-122> rhombus-tech: Daniel master * rd0253c5374d8 /allwinner_a10/nightly_build_images.mdwn:
<CIA-122> rhombus-tech: Daniel master * r1adeb5ce6080 /allwinner_a10/nightly_build_images.mdwn:
simonckenyon has joined #arm-netbook
<CIA-122> rhombus-tech: Daniel master * r41f5321959b4 /allwinner_a10/nightly_build_images.mdwn:
Vayun has quit [Ping timeout: 264 seconds]
Vayun has joined #arm-netbook
<mnemoc> rm: ping
<rm> waht
<mnemoc> rm: hi, beside NO_HZ what other important CONFIG_ values we need to change in defconfig?
<rm> with NO_HZ disabled you also have to choose which HZness do you want (100 would be fine)
<rm> also that person who were having problems should try the same kernel changing only NO_HZ to on/off
<rm> to confirm this is indeed the problem
<rm> if not, then try disabling PREEMPT
<mnemoc> rm: so you don't have one of the devices with this problem?
<rm> no
<rm> I did not try with NO_HZ being on
<rm> and I did not write that part in the wiki
<rm> the wiki is horrible at being able to tell who wrote which parts
<mnemoc> I was sure you were the source of "NO_HZ causes troubles"
CaCtus491 has quit [Ping timeout: 260 seconds]
<mnemoc> we'll... I'll ask them on the ticket to test
CaCtus491 has joined #arm-netbook
<rm> > Someone wrote
<mnemoc> :)
<rm> every wiki in the world has a "view history" link
<rm> with usernames or IPs of users making changes and a way to view each change individually
<rm> but apparently not this one
<mnemoc> bad side effect of federated login :<
<rm> not really
<rm> the login has nothing to do with this
<rm> i.e. it'd have federated logins in the history
<mnemoc> login = authorship
<rm> if it even had one
<rm> e.g. Dokuwiki supports openID via plugins
<rm> and OpenIDs show up in history just fine
<mnemoc> uhm... reference to NO_HZ is gone
<rm> > tickless kernel doesn't work. (If you turn it on, the system will hang randomly.)
<rm> still there for me
<rm> NO_HZ == tickless
<mnemoc> not for `git grep HZ` :p
<mnemoc> 12e7ec6b6d929e66852d1c61a49d5273ff4eb423 by luke added the comment
<mnemoc> seems to be a paste from the ML
<mnemoc> WarheadsSE: so your X crashes because of an assert in mali libraries?
Almamuetya has joined #arm-netbook
<lundman> ij
<lundman> er ok
hipboi_ has joined #arm-netbook
<rz2k> hehe, renesas emma ev2 support is in 3.5 mainline, time to test that on my old tablet.
Mazon has quit [Read error: Connection reset by peer]
Mazon has joined #arm-netbook
<CIA-122> rhombus-tech: Amelin master * r88219f93712d /allwinner_a10/orders/.mdwn:
<Turl> hey all
<WarheadsSE> mnemoc: what huh
<mnemoc> WarheadsSE: you pasted me a random assert() last night, and I was trying to guess what it was
Almamuerta has joined #arm-netbook
<WarheadsSE> Ah
<mnemoc> wb Turl
<WarheadsSE> yeah, it gets to the point where it ought to be firing on the hdmi, I get that.
<WarheadsSE> I'm working on connecting back so I can get you a whole log.
<mnemoc> WarheadsSE: i haven't used x11/mali myself so I don't have any reference to see what may be different on yours :(
<mnemoc> hipboi_ did the kernel side integration we have currently
<WarheadsSE> well then ping hipboi http://pastie.org/4329891 wwoo graphics wtf.
<mnemoc> :)
<WarheadsSE> You can see everything is going to plan, and then.. poof.
<mnemoc> what kernel are you using?
<WarheadsSE> I am not sure where this symbol _rtld_global_ro is coming from
<WarheadsSE> :) you can always look ;)
<mnemoc> I'm sort of alergic to arch :p
<WarheadsSE> I gank right from git.
<WarheadsSE> commit 365a54c0a2117355e6486258436220118425601d
<mnemoc> can you try with sunxi-v3.0.36-r1 tag?
<mnemoc> that is 970e0e2890476069bac2b7300fd2ccd6ca6d6edb
<mnemoc> .38 hasn't been broadly used yet, but people reported success with .36
<mnemoc> just to discard it's something in the upgrade to .38 or not
<WarheadsSE> Yeah, I had success with .36 :p
<WarheadsSE> brb, coffefeeeee
<mnemoc> WarheadsSE: so you confirm that with sunxi-v3.0.36-r1 x11/mali works and using allwinner-v3.0-android-v2's HEAD it doesn't?
<WarheadsSE> I had been using all the way back to 23e5456879db0175f571dec43095c49e181e0b10
<WarheadsSE> that was the move to 36 & adding zram
<mnemoc> please try sunxi-v3.0.36-r1 (970e0e2890476069bac2b7300fd2ccd6ca6d6edb)
* mnemoc should really get a display to start playing with x11 too
<WarheadsSE> r1 is still .36?
<WarheadsSE> (so i know what i need to do with my config)
<mnemoc> yes, sunxi-v3.0.36-r1 is still .36
<mnemoc> after that there are 2 commits (CONFIG_SUNXI_IGNORE_ATAG_MEM) and then the upgrade to .38
<WarheadsSE> K
<WarheadsSE> let me do a single build locally
DEAT_ has joined #arm-netbook
sspiff has quit [Read error: Connection reset by peer]
sspiff has joined #arm-netbook
sspiff has joined #arm-netbook
DEAT has quit [Ping timeout: 264 seconds]
<zoobab_> hi
<zoobab_> just received my A10 tablet
<zoobab_> wondering if there is a way to root it/load another firmware
<orly_owl> what make and model is it
<Turl> zoobab_: A10 tablets are usually prerooted out of the box :)
<orly_owl> that too
<mnemoc> afaik all A10-based tablets come with `su` and `busybox`
<Turl> and none distribute busybox source afaik
<Turl> mnemoc: does the sunxi IR driver on our kernel source work?
<mnemoc> Turl: no idea... but none of the 'm's in defconfig panics on load :p
<Turl> :p
<orly_owl> does mk802 or a1/2000 support CEC?
<WarheadsSE> mnemoc: it's rebuilding now, it will be a while
<WarheadsSE> CEC?
<orly_owl> control tv/other device via hdmi
<WarheadsSE> ah
<orly_owl> mhm
<orly_owl> consumer electronics control
<WarheadsSE> I'm having troubles, period with hdmi at the moment, so :p
<orly_owl> fair enough ^^
<WarheadsSE> Something in 3.0.38 may have broken my hdmi, so, digging around
* mnemoc hopes .36-r1 fails in the same way
<mnemoc> (easier to narrow the problem within .36s)
<thefrog> rz2k: did you resolve your problem with the gcc > 4.4 built binaries ?
<rz2k> nope
<thefrog> there was some old newsgroup posts about arm binaries being optimized wrong on 4.5+ if you were cross-compiling. But i couldn't run it down
<thefrog> oops... bbl
<WarheadsSE> intereting.. we're not having that problem at all.
* mnemoc uses 4.6 without problems
<WarheadsSE> our entire build is 4.7.1
<rz2k> that is about emma ev2
<rm> I need a script.bin for a 1024MB mk802 :/
<rm> or somewhere to peek at the [dram_para] section from fex
<WarheadsSE> de-compile the mk802 fex, and change it over..
<traeak> WarheadsSE: i helped make some gcc 4.7 fixes :-p
<WarheadsSE> good.
<rm> I have both a FEX and a compiled BIN from an MK802 with 512 MB
<rm> however the 1GB one doesn't seem to boot with it
<WarheadsSE> do you have a 1G uboot ?
<WarheadsSE> *might* be inter-related, idk
<hackandfab> Can someone post a working .config for the 3.0.38+ kernel on a meleA1000/2000 ? (at pastebin for example)
<WarheadsSE> 1 sec
Almamuerta has quit [Quit: Page closed]
<WarheadsSE> linux/android ?
* hackandfab linux
<WarheadsSE> having some odd X issues though
<WarheadsSE> otherwise everything else works.
<WarheadsSE> (as far as I can test on the A100)
<hackandfab> WarheadsSE: thx, I'll try in 3...2...1..
<RaYmAn> rm: I can grab you an mk802 1gb bin when i get home
<rm> I wonder if I also need a 1GB-aware uboot
<hackandfab> thx rm (I'll look at it later)
<mnemoc> rm: the kernel doesn't use the dram section from .bin, that's used to compose the boot0/boot1 by allwinner tools
<mnemoc> rm: so you would need to make uboot 1GB-capable and disable IGNORE_ATAG_MEM in .config
<mnemoc> IGNORE_ATAG_MEM is to be used together with stock boot0/boot1 hackery
<mnemoc> so, from nand
<WarheadsSE> mm, so yes, uboot DOES come into play..
eebrah has quit [Quit: ChatZilla 0.9.88.2 [Firefox 14.0.1/20120713134347]]
<rm> just to extract the first 1 MB + 42KB from there
<WarheadsSE> heh
<thefrog> Has someone soldered new DRAM on the mele boards or is this a different 1G device?
<rm> MK802 1GB
<thefrog> rm: ok.. thanks
<mnemoc> rm: poke hno about the 1GB support on uboot. has has been playing with some methods to extract dram initialization info from stock images
<rm> looks like it finally booted
<rm> but no SSH for some reason
<rm> I wonder if it just booted android
<WarheadsSE> :p
<rm> but android shouldn't have detected the USB NIC...
<hackandfab> WarheadsSE I compiled the kernel with your .config and it loads ok until it gets to rtl8192cu, then after retrieving the mac it gets stuck and 120 seconds later kernel dumps thread_exit
<WarheadsSE> hmm
<WarheadsSE> I don't have that problem
<WarheadsSE> Of course, the 8192cu driver was hacked up to do usb poweron/off in this kernel.
<WarheadsSE> IDK if that got reversed when moving to .38
<hackandfab> WarheadsSE: http://pastebin.com/tPuUjQFh the last lines
<hackandfab> just before readAdapterInfo_8192CU() it says power_mgnt(1),usbss_enable(0) so it can be
<hackandfab> can it be put always on somewhere?
<WarheadsSE> yup
<WarheadsSE> there is a setting in the fex
<WarheadsSE> I think on the Axx its on .. 2?
* WarheadsSE checks
<WarheadsSE> yup
<WarheadsSE> usbc2 usb_host_init_state = 1
hipboi__ has joined #arm-netbook
* hackandfab I'll check the fex and then rm .config
<WarheadsSE> although< i don't think I've tried heartily on .38 yet..
<WarheadsSE> been bashing on X
hipboi has quit [Ping timeout: 244 seconds]
hipboi_ has quit [Read error: Connection reset by peer]
ka6sox is now known as zz_ka6sox
<hackandfab> WarheadsSE: changing the FEX solves it :) <6>usbcore: registered new interface driver rtl8192cu
<WarheadsSE> yw :)
rellla has quit [Quit: ChatZilla 0.9.88.2 [Firefox 14.0.1/20120713134347]]
<WarheadsSE> the driver barks a lot @ console/dmesg
<hackandfab> I can't see it's errors yet just after registering the 8192cu >init: Failed to create pty
<mnemoc> nothing in 8192cu has been changed since Qware's source drop
<xenoxaos> failed to create pty...check your /etc/inittab
<xenoxaos> as well as /etc/securetab
<Turl> mnemoc: I updated the driver but got no feedback yet :<
<CIA-122> rhombus-tech: master * r77807f49d458 /allwinner_a10/orders/alexiss.mdwn:
gimli has joined #arm-netbook
<traeak> mnemoc: what did you break now :-p
<WarheadsSE> hah
<hackandfab> mnemoc: The problem with 8192cu was solved changing the FEX to always on
<mnemoc> hackandfab: that doesn't solve it, that lets you use the mainline driver instead
<mnemoc> traeak: I try to keep my breakages isolated in a wip/ branch :<
<mnemoc> i don't see anything in the .38 that could break 8192cu...
<mnemoc> +update
<mnemoc> I would get rid of that driver, but first we need to add proper power management to the usb driver
<mnemoc> enabling/disabling them based on what other drivers are loaded isn't very nice
<WarheadsSE> ^
<WarheadsSE> should just be turn it on/off, then whatever is attached is recognized
<WarheadsSE> then let udev/systemd handle it accordingly
<CIA-122> rhombus-tech: master * r0db3d4339633 /allwinner_a10/orders/alexiss.mdwn:
zz_ka6sox is now known as ka6sox
<mnemoc> btw, 3.4 is updated too, if someone wants to play with that kernel :p
<mnemoc> WarheadsSE: but only after finding why mali/x11 isn't working in 3.0 for you :) .... need a known-to-work base
<rm> my new mk802 doesn't boot anything but android :<
<rm> writing the Rikomagic lubuntu image now
<mnemoc> tried hooking a custom kernel in nanda's uboot?
<RaYmAn> rm: still needing a mk802 1gb script.bin or was it judged irrelevant?
<rm> RaYmAn, sure, upload it somewhere
<rm> in fact
<rm> if you have an mk802 with 1GB up and running, also post your uboot if you can
<RaYmAn> I haven't really done much on it - it's entirely stock android
* rm 's going to play some Angry Birds >___<
<traeak> mnemoc: same branch?
<RaYmAn> rm: do you just want the nanda dump? =P
<mnemoc> traeak: same brnach for what?
<rm> I don't know about nanda
<rm> I don't use it
<rm> I am trying to boot from an SD card
<RaYmAn> 'k, just script.bin then :P
<RaYmAn> rm: skumler.net/mk802_script.bin
* RaYmAn goes back to playing with his Nexus 7
<WarheadsSE> mnemoc: I's built, I just have to install and reboot.
<WarheadsSE> If it doesnt boot right, lord knows
<WarheadsSE> im not near the power, so if it decides to boot nand I am screwed
<mnemoc> :<
<mnemoc> but it should boot fine... the problem is only on x11 start
<WarheadsSE> right
<mnemoc> ssh should be there
<WarheadsSE> .. right
<WarheadsSE> but this thing has a habit of trying to boot nand
<mnemoc> uhm
<WarheadsSE> idk if it just doesnt like my classs 10 or what
<mnemoc> WarheadsSE: you can also replace the uboot in nanda to give you a "safe" linux when booting from nand
<WarheadsSE> I havent gotten the nand to show up right yet ;)
<mnemoc> uh
<WarheadsSE> always useless, no partitions recognized, no mtds, etc
<mnemoc> that's normal
<mnemoc> the driver isn't mtd
<mnemoc> but /dev/nand[a-h] should be there anyway
<WarheadsSE> K, then how exactly would i reference ot?
* WarheadsSE checks
Alamuerta has joined #arm-netbook
<Turl> /dev/block/nand[a-g]
<WarheadsSE> yup, there
<mnemoc> then just mount them ;-)
arokux has quit [Remote host closed the connection]
<rm> now, how come this mk802 has its time set correctly?
<mnemoc> it feels weird to have accepted it's crap-ness as normal
<WarheadsSE> rm: ntp?
<rm> it doesn't have an internal battery
<rm> I haven't set up the wifi network so it has no internet
<WarheadsSE> rtc
<rm> but... no battery?
<mnemoc> short life, but there
<rm> short life as in 30 days in the mail from china?
<mnemoc> ehm, no
<rm> it never ever had internet here
<rm> well
<rm> it shows "December 31 2009, 7:17"
<mnemoc> hail the gods!
<rm> 7 is wrong
<rm> but 17 is correct
<mnemoc> luck?
<rm> so I assume just the timezone is wrong
<rm> maybe :p
<mnemoc> 2009?
<WarheadsSE> luck
<mnemoc> that looks very much like the 17 minutes after the "epoch" on the chip
<WarheadsSE> my mele kept coming up 2010
arokux has joined #arm-netbook
<mnemoc> TZ?
<WarheadsSE> timezone
<mnemoc> from December 31 2009 to 2010 there is only a TZ of difference
<WarheadsSE> maybe
<WarheadsSE> since the default is chicago
<traeak> is the wireless fixed here/
Quarx has quit []
<WarheadsSE> "here" ?
<traeak> i mean with the new work
<traeak> hmm
<WarheadsSE> ah
<traeak> ugh i still need to submit a patch for drm.h
<hackandfab> It seems my "init: Failed to create pty " error is due to ubuntu12 upstart. Is someone using it?
<traeak> wireless still doesn't compile...i'm just doing 3.4.x
<mnemoc> traeak: a patch in a paste site and a proper From: is enough. no need to fork + pull request
<mnemoc> traeak: yes, that's known. you have to disable WHATEVER8159_SW
<mnemoc> traeak: it needs wakelocks, which don't seem to exist in android-3.4 tree
<mnemoc> traeak: but enabling the usbc in script.bin lets you use the mainline driver
<thefrog> Heh. ARMH announced. Nice.
Vayun has quit [Ping timeout: 244 seconds]
<rm> what?
<traeak> ahh okay
<WarheadsSE> thefrog: "armh" ?
<traeak> mnemoc: pretty much means that rtl8192cu driver is likely irrelevant then i gather
<WarheadsSE> If the usb comes on on its own, yeah
<mnemoc> traeak: at least not a priority for me at the moment
<traeak> for now it's not for me either
<mnemoc> :)
<traeak> so ou tackedon 99.98 onto the module directory?
<mnemoc> the lack of usb gadget in 3.4 is more important... but still... time is limited and volunteers too few :|
<RaYmAn> complete lack of usb gadget?
<mnemoc> RaYmAn: integration needs to be reimplemented
<RaYmAn> ah.
<mnemoc> they did hack it inside android.c instead of a separated sunxi.c
<thefrog> WarheadsSE: Arm Holdings anounced great growth.. analysts were expecting it's growth to slow
<RaYmAn> ouch
<rm> blah
<RaYmAn> that sucks
<rm> where's the open source annoucnement that phoronix promised
<thefrog> WarheadsSE: just makes me excited.
<traeak> always take phoronix with a grain of salt
<thefrog> rm: donno, haven't listened to the conference call
<WarheadsSE> yup
<traeak> mnemoc: where's the 99.98.3.4.5+/ coming from?
<mnemoc> o_O
<mnemoc> never saw that
<traeak> thats from a make modules_install
<mnemoc> weird, sun4i_build/output/lib/modules/3.4.5+/
<traeak> okay i guess time to refreshmy config again
<thefrog> ARM can't make any annoucements during the quiet period before their earnings release (usually.. as i understand it)
<mnemoc> s/$/ here/
<traeak> most of the modules go into 3.4.5+ it's just all the modules.* stuff
<mnemoc> weird
<traeak> might be arch, time to play more then
<mnemoc> i get the modules.* in the same dir, parallel to build/, kernel/ and source/ ... as it should be
Alamuerta has quit [Ping timeout: 245 seconds]
<WarheadsSE> mnemoc: ok, wtf, here goes..
<mnemoc> WarheadsSE: -v ?
<WarheadsSE> i mean ill reboot it while not at home
<mnemoc> :)
<mnemoc> go go go go
<WarheadsSE> dialing back to watch the serial.
<WarheadsSE> i *think* i left my laptop on..
<WarheadsSE> seems RDP saying no no
<mnemoc> uh, RDP
<WarheadsSE> rdp through to my W7 box that has the serial attached to it
<WarheadsSE> can't sem to get it to come up
* mnemoc never expected an arch developer using W7 on his laptop
<rm> nnnnope, doesn't boot ._.
<WarheadsSE> so far.. nada
<WarheadsSE> ip changed.. whee
<rm> maybe I need a smaller SD card
<rm> or a different one. tried 8GB and 32GB so far
<WarheadsSE> With so much ahead, you need strong numbers behind you. Top healthcare providers meet the challenge of addressing managed care receivables and self-pay bad debt. Brandywine Medical Management Services Incorporated’s years of experience in patient accounting leaves us in a strong position for handling managed care accounts. In order to contain costs and enhance your cash flow, BMMSI can maintenance
<WarheadsSE> these active accounts within your system after 60 days of billing. Partnering with BMMSI in this regard frees your staff to work current receivables, therefore reducing time to liquidation. Providers partnering with BMMSI have found this method of accounts maintenance very successful.
<WarheadsSE> I would like to set up a time to speak with you regarding a site visit to discuss our services in further detail. I will be following this email with a phone call in the coming days. In the meantime, please review the additional information below. Thank you and have a great day!
<WarheadsSE> oops
<WarheadsSE> sorry
<WarheadsSE> right click hell
<mnemoc> rm: hno's uboot doesn't know to initialize your dram. you'll have to play from nanda
<mnemoc> rm: only 512M is currently supported
<rm> I am trying this one already: http://rikomagic.co.uk/forum/viewtopic.php?f=2&t=90 Linux for MK802
<rm> that it doens't boot is a bad sign I think :/
<mnemoc> rm: phoenixcard image?
<rm> SD
<WarheadsSE> mnemoc: failed at the same spot
<mnemoc> WarheadsSE: good. and what was the commit you said worked previously for you?
<mnemoc> i'm sure the problem is somewhere else. just need to confirm it :)
<mnemoc> 23e5456879db0175f571dec43095c49e181e0b10
<mnemoc> which is only 4 commits behind sunxi-v3.0.36-r1
<WarheadsSE> yeah, one of those commits was your reservation.
<mnemoc> nope
<mnemoc> WarheadsSE: those are after the tag
<mnemoc> WarheadsSE: and were actually removed
<mnemoc> moved to a branch to be precise
<mnemoc> between 23e5456879db0175f571dec43095c49e181e0b10 and sunxi-v3.0.36-r1 there are only Kconfig changes about the fb module deps
<mnemoc> aha! there is a change indeed
<mnemoc> changing the order fb/g2d/ve are reserved
<WarheadsSE> ah..
<mnemoc> WarheadsSE: can you please re-confirm 23e5456879db0175f571dec43095c49e181e0b10 works?
<WarheadsSE> ill have to move back
<mnemoc> pl
<traeak> probably the depmod screwed up, so have to figure out what stupid things happened
<mnemoc> err
<mnemoc> WarheadsSE: ok, no rush
<WarheadsSE> I still have that floating around.
<mnemoc> no no, please rebuild
<mnemoc> to be sure the chosen hash is the only difference
<WarheadsSE> yeah, 1 sec
<WarheadsSE> rebooting it again
<WarheadsSE> it bit it.
<WarheadsSE> came up android
<mnemoc> meh
<WarheadsSE> I warned you :p
<mnemoc> :(
<mnemoc> no ssh on android?
<WarheadsSE> Will have to wait ~ 3.5 hours
<WarheadsSE> its dead stock
<mnemoc> meh
<xenoxaos> can you get any console over serial?
<WarheadsSE> ...
<WarheadsSE> scroll up xenoxaos XP
<mnemoc> he is 3.5h away from the serial console
<xenoxaos> came up android
<xenoxaos> rdp no work
<WarheadsSE> Yes, and android doesnt expose a console over serial
<WarheadsSE> so, im screwed
<xenoxaos> i think sometimes you can by specifying a bootarg
<xenoxaos> i read it somewhere, but dont remember where
<WarheadsSE> still, cant get to it.
<WarheadsSE> mnemoc: I know that that hash _is_ what i just rebooted, it was my package before building the PKGBUILD
<WarheadsSE> Need to hook up a relay ..
thefrog has quit []
gimli has quit [Ping timeout: 246 seconds]
<mnemoc> WarheadsSE: i don't find impossible to believe that changing the order of the fixed reserves can have side effects, but still sounds very unlikely
<mnemoc> WarheadsSE: so before trying to revert it I would prefer if a fresh build of 23e5456879db0175f571dec43095c49e181e0b10 works while nothing else change
<WarheadsSE> Well if it booted into linux I could start a build :p
<mnemoc> no rush, tomorrow is fine too :)
gimli has joined #arm-netbook
von_fritz has joined #arm-netbook
<WarheadsSE> looks like tonight I may spend the time to grab su & an sshd to put into that android install..
<WarheadsSE> so I can get back to it and reboot it when this happens
Vayun has joined #arm-netbook
<mnemoc> good plan
<rm> is there a way to extract script.bin used by android?
<WarheadsSE> I'' just have to find & install APKs
<mnemoc> rm: yes, mount nanda
<mnemoc> and grab it
<mnemoc> WarheadsSE: `su` and `busybox` should already be there
<WarheadsSE> if the busybox doesnt have sshd :p
<von_fritz> WarheadsSE still segfault using modprobe lcd, If you are interested -> http://pastebin.com/yvcDyhjJ
<WarheadsSE> I'll need to hook up to it somehow
<mnemoc> WarheadsSE: I mean, you only need the ssh server
<WarheadsSE> von_fritz: it doesnt like your FEX
<WarheadsSE> I have managed to replicate that with, say, screen0 mode 3, setting 10
<von_fritz> well selfcompiled kernel with fb modul works :)
<WarheadsSE> yeah, thats the one major difference is fb built in.
<WarheadsSE> which version von_fritz
<WarheadsSE> .3?
* hackandfab I got logon on ubuntu 12.04 armel meleA2000 (ubuntu has a bug with pts I'll document later)
<rm> mnemoc, can I do that while in android? :S
<von_fritz> Linux mele 3.0.38+ #2 PREEMPT Sat Jul 21 21:12:12 CEST 2012 armv7l GNU/Linux
<WarheadsSE> K
<WarheadsSE> thats from HEAD?
<von_fritz> yes
<WarheadsSE> My problem could be related then
<WarheadsSE> see that mnemoc ?
<mnemoc> rm: install a terminal app. `su; mkdir /mnt/rm; busybox mount -t vfat /dev/block/nanda /mnt/rm; ...`
<rm> I'm starting to think some mk802s are more special than others
<mnemoc> WarheadsSE: I doubt it's connected to your problem
<mnemoc> hackandfab: can you file a bug and ask gdb for what line BSP_disp_get_screen_width+0xc refers to?
<mnemoc> hackandfab: `modprobe lcd` works fine for me on 3.0-v2's HEAD
<mnemoc> hackandfab: also, please attach your script.bin somehow
<mnemoc> hackandfab: looks very script related
<WarheadsSE> Uhm, I think you mean von_fritz mnemoc
<mnemoc> WarheadsSE: indeed
<mnemoc> von_fritz: all that ----^ was for you
<mnemoc> hackandfab: sorry :)
<mnemoc> if((gdisp.screen[sel].status & LCD_ON) || (gdisp.screen[sel].status & TV_ON) || (gdisp .screen[sel].status & HDMI_ON) || (gdisp.screen[sel].status & VGA_ON))
<mnemoc> must be the line, so I bet it's passing a bad `sel`
<mnemoc> lovely lovely
* mnemoc hates the disp driver
<DonkeyHotei> ok, so now i have mali x11 running
<hackandfab> mnemoc ;)
<DonkeyHotei> but it doesn't create /dev/dri/card0
<WarheadsSE> Arch Linux ARM recommended by sparkfun
<DonkeyHotei> why not?
gimli has quit [Ping timeout: 255 seconds]
<mnemoc> DonkeyHotei: can you paste the LYCHEE bits of your .config somewhere?
<mnemoc> DonkeyHotei: well... are they 'm' or 'y'?
<mnemoc> DonkeyHotei: zgep LYCHEE /proc/config.gz
<mnemoc> zgrep*
<DonkeyHotei> i'll have to boot it up, hang on
<von_fritz> mnemoc same script.bin same kernel with "my .config" modprobe lcd -> http://pastebin.com/w6T3eFar no problem with X in fbdev
<DonkeyHotei> root@linaro-ubuntu-desktop:~# zgrep LYCHEE /proc/config.gz
<DonkeyHotei> CONFIG_LYCHEE_FB_SUN4I=m
<DonkeyHotei> CONFIG_LYCHEE_FB_UMP_SUN4I=y
<DonkeyHotei> CONFIG_LYCHEE_LCD_SUN4I=m
<DonkeyHotei> CONFIG_LYCHEE_HDMI_SUN4I=m
<mnemoc> DonkeyHotei: looks ok... `lsmod` shows the modules loaded? does people normally have that node at all? :p
<WarheadsSE> DonkeyHotei: check CONFIG_FB
<DonkeyHotei> CONFIG_FB=y
<DonkeyHotei> root@linaro-ubuntu-desktop:~# lsmod
<DonkeyHotei> Module Size Used by
<DonkeyHotei> drm 176951 1 mali_drm
<DonkeyHotei> mali 131217 0
<DonkeyHotei> mali_drm 1958 0
<DonkeyHotei> hdmi 15237 0
<DonkeyHotei> lcd 2706 0
<DonkeyHotei> disp 217545 3 hdmi,lcd
<DonkeyHotei> ump 42553 4 mali,disp
<mnemoc> von_fritz: uh... so the panic isn't on `modprobe` time?
<mnemoc> DonkeyHotei: btw, what branch/version?
<DonkeyHotei> root@linaro-ubuntu-desktop:~# uname -a
<DonkeyHotei> Linux linaro-ubuntu-desktop 3.0.38+ #1 PREEMPT Tue Jul 24 04:13:28 EST 2012 armv7l armv7l armv7l GNU/Linux
<WarheadsSE> hmm, then that matches my config ..
* mnemoc doesn't want to dive into the disp driver .... /me cries
<WarheadsSE> von_fritz: did you rebuild UMP & the x11 driver?
<DonkeyHotei> i did not rebuild ump
<DonkeyHotei> not sure where it would go
<mnemoc> von_fritz: please change drivers/video/sun4i/disp/de_bsp/bsp_display.h:70 to if 0 to get debugging info
<WarheadsSE> DonkeyHotei: you wouldnt need to on ubuntu
<WarheadsSE> they do it all for you
<mnemoc> hipboi made the x11/ump drivers for ubuntu
<DonkeyHotei> i had a hell of a time getting mali x11 working at all because the wiki instructions were so misleading
<mnemoc> no idea how that nightly build script deals with that
<DonkeyHotei> badly
<mnemoc> DonkeyHotei: feel free to do them better at http://linux-sunxi.org :)
<DonkeyHotei> i put a note on the rhombus wiki page
<rm> sooooooooo, I got the script.bin from Android
<rm> and you know what
<rm> the A10 runs at 900 Mhz in this system
<mnemoc> that's a governator thing, not script.bin
<rm> I know
<rm> just checked /proc/cpuinfo and it shows 906 bogomips
<mnemoc> iirc they use that "fantasy" thing
<DonkeyHotei> the sunxi wiki does not appear to have an x11 page at all
<mnemoc> DonkeyHotei: it's waiting for you
<WarheadsSE> DonkeyHotei: I had no issues..
<rm> -dram_clk = 360
<rm> +dram_clk = 433
<mnemoc> rm: dram_* are only used by the tool that creates the boot0/boot1 files
<mnemoc> i.e. ignored when booting from uSD
<rm> -mali_clkdiv = 3
<rm> +mali_clkdiv = 4
<rm> how about that
<rm> anyways I'm going to try it
<rm> maybe it doesn't work either
<mnemoc> mali_clkdiv is relevant, yes
<rm> left side is the problematic mk802
<rm> with 1GB
<rm> right side is from a fex/bin working on a 512MB one
<DonkeyHotei> i put a link to the rhombus wiki on the sunxi wiki
<mnemoc> rm: looks like for a different device
* RaYmAn thought csi was for cameras only
<mnemoc> DonkeyHotei: so you took the time to register only to drop a link to the nightly builds in the mali page?
<WarheadsSE> interesting, complerely different chip too..
<mnemoc> DonkeyHotei: I have to admit I expected better content :<
Mazon has quit [Read error: No route to host]
P4R4N01D has quit [Ping timeout: 246 seconds]
<mnemoc> rm: the diff without sections is kind of hard to understand :<
<DonkeyHotei> mnemoc: documentation is not my forte
<DonkeyHotei> i put a note in the nightly builds page
<mnemoc> DonkeyHotei: i fail to understand how a link to the nightly builds page helps to understand how to build the mali libs
<WarheadsSE> heh
<WarheadsSE> I built them fine.. even moved from armv6w/thumb to armv7-a -fPIC
<rm> nah, nothing seems to help
<rm> I think I will need the serial console
<DonkeyHotei> mnemoc: the nightly builds instruction was preventing them from working
<DonkeyHotei> so, i put a note there
<mnemoc> ok
Almamuetya has quit [Ping timeout: 252 seconds]
hp_ has quit [Ping timeout: 276 seconds]
Mazon has joined #arm-netbook
von_fritz has quit [Quit: vonfritz leaves, don't panic]
<CIA-122> rhombus-tech: Alessio master * ra36e1a679668 /allwinner_a10/orders/Alexenergy.mdwn:
<WarheadsSE> mnemoc: nope, every revision i c.
<WarheadsSE> I can go back to 308 i think..
<traeak> 7.62x51? :-p
NZBBoSS has quit []
<mnemoc> WarheadsSE: so it's not a kernel thing, right?
<WarheadsSE> maybe not
<WarheadsSE> although i still cant replicate that other error with modprobing lcd :
<mnemoc> drivers/video/sun4i/disp/de_bsp/bsp_display.h:70 to #if 0 if you want a verbose disp driver
<mnemoc> after finishing the cleaning of script and gpio stuff, next will have to be that cursed disp driver :<
<RaYmAn> mnemoc: is there a list of work items? :P
<Turl> RaYmAn: check the issue tracker, there's crapton of issues filed :P
<mnemoc> RaYmAn: https://github.com/amery/linux-allwinner/issues/48 is about 3.4 porting
specing has quit [Remote host closed the connection]
specing has joined #arm-netbook
specing has quit [Client Quit]
specing has joined #arm-netbook
nibb__ has quit [Read error: Connection reset by peer]
nibb__ has joined #arm-netbook
<WarheadsSE> mnemoc: it might actually be a glibc problem.
<WarheadsSE> I think I recentl bumped from 2.15 to 2.16
simonckenyon has quit [Quit: Nettalk6 - www.ntalk.de]
lerc has quit [Ping timeout: 255 seconds]