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/
vinifm_ has quit [Quit: Ex-Chat]
Maqs has quit [Ping timeout: 255 seconds]
Eigen has quit [Remote host closed the connection]
Eigen has joined #arm-netbook
Eigen_ has joined #arm-netbook
Eigen has quit [Ping timeout: 252 seconds]
freakazoid0223 has quit [Quit: Leaving]
Eigen has joined #arm-netbook
<bsdfox_> grr softwinner sata not working with 3.4 on revc?
Eigen_ has quit [Ping timeout: 252 seconds]
popolon has quit [Quit: Quitte]
bfree_ has joined #arm-netbook
bfree has quit [Ping timeout: 252 seconds]
Eigen_ has joined #arm-netbook
Eigen has quit [Read error: Operation timed out]
Eigen has joined #arm-netbook
stefanro has joined #arm-netbook
Eigen_ has quit [Ping timeout: 260 seconds]
stefanro1 has quit [Ping timeout: 245 seconds]
nollapiste has quit [Quit: Sto andando via]
Eigen_ has joined #arm-netbook
voronaam1 has quit [Quit: Leaving.]
Eigen has quit [Read error: Operation timed out]
Eigen has joined #arm-netbook
Eigen_ has quit [Ping timeout: 248 seconds]
Eigen_ has joined #arm-netbook
Eigen__ has joined #arm-netbook
Eigen has quit [Ping timeout: 252 seconds]
Eigen_ has quit [Ping timeout: 245 seconds]
cnxsoft has joined #arm-netbook
Eigen has joined #arm-netbook
Eigen__ has quit [Read error: Operation timed out]
Eigen has quit [Remote host closed the connection]
dyoung-away is now known as dyoung
penguin42 has quit [Quit: Leaving.]
Eigen has joined #arm-netbook
Eigen_ has joined #arm-netbook
Eigen has quit [Ping timeout: 252 seconds]
Eigen has joined #arm-netbook
freakazoid0223 has joined #arm-netbook
Eigen_ has quit [Ping timeout: 260 seconds]
cnxsoft has quit [Ping timeout: 276 seconds]
cnxsoft has joined #arm-netbook
Eigen_ has joined #arm-netbook
Eigen has quit [Ping timeout: 260 seconds]
mSquare has joined #arm-netbook
Eigen has joined #arm-netbook
Eigen_ has quit [Ping timeout: 255 seconds]
Eigen_ has joined #arm-netbook
Eigen has quit [Ping timeout: 252 seconds]
Eigen has joined #arm-netbook
Eigen_ has quit [Ping timeout: 256 seconds]
cnxsoft has quit [Quit: cnxsoft]
<libv> so who is selling his cubieboards on ebay.de?
<libv> someone who figured out how to clock up the memory to 480MHz :)
aholler_ has joined #arm-netbook
aholler has quit [Ping timeout: 248 seconds]
freakazoid0223 has quit [Quit: Leaving]
<hramrach> what is the default clock? xD
Eigen_ has joined #arm-netbook
<hramrach> it says it' the default clock here http://linux-sunxi.org/Cubieboard
<hramrach> libv:
Eigen has quit [Ping timeout: 252 seconds]
gimli has joined #arm-netbook
Eigen_ has quit [Remote host closed the connection]
Quarx has joined #arm-netbook
dyoung is now known as dyoung-away
sky770 has joined #arm-netbook
Ershov has joined #arm-netbook
gimli has quit [Ping timeout: 272 seconds]
sky770 has quit []
sky770 has joined #arm-netbook
sky770 has quit []
pcat has quit [Ping timeout: 245 seconds]
rz2k has joined #arm-netbook
cnxsoft has joined #arm-netbook
Mazon has quit [Ping timeout: 255 seconds]
Mazon has joined #arm-netbook
cnxsoft_ has joined #arm-netbook
cnxsoft_ has quit [Client Quit]
cnxsoft has quit [Quit: cnxsoft]
cnxsoft has joined #arm-netbook
rellla has joined #arm-netbook
Maqs has joined #arm-netbook
<rz2k> I had wrong import yesterday, half of tags disappeared, now its ok, these are all patches for u-boot from both hdmidongle patches packs
<rz2k> RaYmAn: ^
shalem has joined #arm-netbook
<cnxsoft> rz2k: Thx. Which configs do you use: imx6_defconfig (linux) and imx6_android_defconfig (android)?
<cnxsoft> I mean for linux-imx
<rz2k> I didnt apply patches to defconfig yet, they are conflicting with each other
<rz2k> probably done by two developers
<rz2k> only the patches that formatted good (from Zhang Xiaodong) slightly change imx6 android defconfig
<rz2k> I beleive we will need to figure our own defconfig
<cnxsoft> When I post about that, I'll mention the repo, IRC and google group, as well as quick ways to get started with u-boot and kernel
<rz2k> and from the kernel-imx-ga I used the android defconfig
<rz2k> contact jesbir too
<rz2k> as he has it up and running
pcat has joined #arm-netbook
<cnxsoft> So it be nice to have defconfig for the sticks even if it does not work just yet
<rz2k> this one was used in kernel-imx6-ga.tar.gz imx6_nand_android_defconfig
<cnxsoft> ok but this is for NAND boot, right?
<rz2k> yeah, we need to turn off the micron nand flash
<rz2k> since we have just 2SD slots
cnxsoft has quit [Ping timeout: 248 seconds]
Avernos has quit [Read error: Connection reset by peer]
aesok has joined #arm-netbook
Avernos has joined #arm-netbook
Avernos has joined #arm-netbook
Avernos has quit [Changing host]
cnxsoft has joined #arm-netbook
Ershov has quit [Read error: Connection reset by peer]
Ershov has joined #arm-netbook
Ershov has quit [Read error: Connection reset by peer]
Ershov has joined #arm-netbook
sspiff has joined #arm-netbook
rellla2 has joined #arm-netbook
rellla has quit [Ping timeout: 255 seconds]
aesok has quit [Read error: Connection reset by peer]
sspiff has quit [Ping timeout: 245 seconds]
hg_5 has joined #arm-netbook
hg_5_ has joined #arm-netbook
sky770 has joined #arm-netbook
sky770 has left #arm-netbook [#arm-netbook]
lkcl has quit [Ping timeout: 245 seconds]
sspiff has joined #arm-netbook
lkcl has joined #arm-netbook
hg_5_ has quit [Ping timeout: 256 seconds]
hg_5 has quit [Ping timeout: 256 seconds]
popolon has joined #arm-netbook
pcat has quit [Ping timeout: 260 seconds]
sspiff has quit [Ping timeout: 245 seconds]
sspiff has joined #arm-netbook
pcat has joined #arm-netbook
hg_5 has joined #arm-netbook
hg_5_ has joined #arm-netbook
aesok has joined #arm-netbook
hg_5_ has quit [Ping timeout: 272 seconds]
hg_5 has quit [Ping timeout: 272 seconds]
tinti has quit [Read error: Operation timed out]
Avernos_ has joined #arm-netbook
Avernos has quit [Ping timeout: 272 seconds]
aesok has quit [Remote host closed the connection]
tinti has joined #arm-netbook
andy101 has joined #arm-netbook
mSquare has left #arm-netbook [#arm-netbook]
<andy101> hello
<andy101> I have a board with an Allwinner A13 processor and I'm currently looking for some guys who have experiences with this type of processor
<andy101> I designed my own board and it is very similar to the MICRO board of olimex (but with 512 MB of RAM instead of 256 MB)
<andy101> the thing is, u-boot is successfully loaded from the SD card
<Turl> well, you found the right channel :)
<andy101> but the kernel always gets stuck at "Starting kernel..."
<Turl> andy101: did you load script.bin?
<Turl> before loading the kernel
<andy101> yes I did
<Turl> try booting with earlyprintk on your kernel cmdline and see where it fails then
<andy101> do I have to specify something for earlyprintk?
<andy101> something like earlyprintk=ttyS0,115200
<andy101> because I already tried using earlyprintk but nothing is printed to the console
<andy101> I also used the memory dump util to look at the memory address of __log_buf but the memory gets either overwritten or the content is lost
<andy101> do you recommend a specific kernel version?
<andy101> I tried 3.0.57 and 3.4.x
rellla2 has quit [Remote host closed the connection]
<mnemoc> earlyprintk goes alone, the ttyS0,... stuff to console=
<mnemoc> earlyprintk will just use the memory address and expect u-boot to have configured the pins
<Turl> mripard: ping
gsilvis has quit [Read error: Connection reset by peer]
gsilvis has joined #arm-netbook
<mnemoc> andy101: also note that our kernel is not compatible with the stock u-boot from allwinner. we need the bootloader to pass the proper machine id and mem info
<mnemoc> andy101: if booting from nand, you have to replace the stock nanda's linux/u-boot.bin with one built from the lichee-dev branch of u-boot-sunxi
vinifm has joined #arm-netbook
<andy101> i checked out u-boot from https://github.com/hno/uboot-allwinner.git
<andy101> up to now, I only wanted to boot from sd card
<andy101> do I have to use another u-boot version?
<mripard> Turl: pong
<Turl> mripard: hi
<Turl> mripard: I was working on v2 of the clock series, but I'm hitting some kind of regression
aesok has joined #arm-netbook
<Turl> or maybe not a regression but something caused by splitting the clocks
<Turl> mripard: have you seen this? http://sprunge.us/Thcg
<mripard> Turl: nop, never seen it :S
KoH has quit [Remote host closed the connection]
<Turl> mripard: hm, nevermind
<Turl> mripard: I was missing the parent name when registering the clock :S
<Turl> mripard: clocksource doesn't like 0hz I guess :)
<Turl> dw-apb-uart 1c28000.uart: no clock-frequency property set
<Turl> dw-apb-uart: probe of 1c28000.uart failed with error -22
<Turl> hm, no clocks= support
<mripard> Ah yes, about the UART, they are wired to the apb1
<mripard> could you add it to your patchset as well?
<mripard> if that doesn't take too much effort of course
<Turl> sure
<Turl> what's apb1's parent?
<mripard> 2s
<mripard> let me check
<Turl> hm, it's a mux
<Turl> mux+div
<mripard> yep, looks like it
<Turl> 3.9 merge window closed already, didn't it?
<mripard> yes...
<mripard> and waited a bit too long to send my pull request
<mripard> it seems
<mripard> I was waiting for linus to tell me if he was going to merge a patch
<mripard> and waited a bit too much...
<mripard> I'll try to get this merged in 3.9 anyway, since it's only dt patches
<mripard> but I'm not sure
<Turl> linus picked all of the code, didn't he?
<mripard> yes, but he also merged some of the dt patches
<mripard> so I couldn't merge not directly related patches, because he merged too much
<mripard> and at one point, he couldn't apply my patches anymore
<mripard> and that's when the waiting occured :)
<Turl> http://git.kernel.org/?p=linux/kernel/git/linusw/linux-pinctrl.git;a=search;h=refs/heads/devel;s=Maxime+Ripard;st=author
<Turl> so it's missing the gpio and led dt bitsa
<Turl> bits*
<mripard> the leds are part of the pull request that I delayed too much
<mripard> the patch that caused problem is the one about gpio-controller
maethorechannen has joined #arm-netbook
<Turl> I'll add apb1 then and aim the series for 3.10
<Turl> is there any other clock you need mripard?
<hramrach> btw is that expected that earlyprintk never goes away
<hramrach> on serial console I have every message twice
<hramrach> once with <> prefix once with [] prefix
<hramrach> the early messages appear only once with <> so I would guess the <> messages are earlyprintk
datagutt has joined #arm-netbook
<mripard> Turl: I don't think so
<mripard> the wemac looks connected to AHB, so it will be fine I guess
<Turl> any clock needed for DMA or MMC?
<mripard> probably, but nobody started working on it I'm afraid
<mripard> so don't bother about that :)
<mripard> we'll add it when the time comes
<Turl> ok then
<Turl> stefanro had started work on wemac some time ago
<Turl> I think I got a bounce from his email the other day though :/
<mripard> yes, I continued his work recentrly
<mripard> but it still needs some work
shalem has quit [Quit: Leaving]
<Turl> mripard: did you get a chance to look at the watchdog driver?
<mripard> Turl: nop, sorry :(
<mripard> Turl: I have a loooong trip on Friday, I'll probably do it during it
<Turl> mripard: great then :)
<Turl> hm, my bitbucket got all polluted with branches. I really need to start using tags
CaCtus491 has joined #arm-netbook
<mripard> Turl: haha, you too :)
<mripard> fortunately, 3.9 will be released pretty soon, I'll be able to drop most of them :)
<mnemoc> damn time... it passes too fast :<
<Turl> mripard: mine is the same code over and over, I push it on named branches in case I need to diff with an older version
<Turl> I really should be using tags for that
<mripard> ah, yes
<Turl> mnemoc: yeah, and too slow when you want it to be faster :<
<mripard> I used to do that as well, but it generated waaay too much branches
<mripard> and thinking about it, that wasn't really my workflow
<mripard> because I commit on top the previous branch
<mripard> test
<mripard> and then rebase, when everything's working fine
<mnemoc> quilt rules :p
<mripard> quilt? together with git?
<mripard> never had to use it
<mripard> rebase -i is just awesome for that :)
<mnemoc> guilt implements quilt over git
<mnemoc> so each feature is separated but together at the same time
<mnemoc> but rebase -i is clearly the killer feature of git
<mnemoc> for mercurial people altering history is so taboo
<Turl> mripard: yeah, I do the same
<Turl> and after I finish a feature I push remote branch:branch-N
<Turl> to keep a pseudo history across patch evolution where I can go to if I need to revert to an older version of something
<mripard> well, when I need to do that, I just grabbed the patches I sent :)
<Turl> yeah, but I don't mail every one of those branches I make
<Turl> say, for example today
<Turl> I implemented Gregory's suggestions
<Turl> and then I pushed a branch, but I won't mail it until later when I implement apb1
SouL_ has joined #arm-netbook
SouL_ has left #arm-netbook [#arm-netbook]
cnxsoft has quit [Remote host closed the connection]
mur_ has joined #arm-netbook
KoH_ has joined #arm-netbook
Marex has quit [Ping timeout: 264 seconds]
j^ has quit [Ping timeout: 264 seconds]
j^ has joined #arm-netbook
Marex has joined #arm-netbook
Avernos_ is now known as Avernos
uwe_ has quit [Ping timeout: 264 seconds]
uwe_ has joined #arm-netbook
hg_5 has joined #arm-netbook
merbanan has quit [Ping timeout: 260 seconds]
muts has quit [Quit: Computer has gone to sleep.]
muts has joined #arm-netbook
freakazoid0223 has joined #arm-netbook
SouL_ has joined #arm-netbook
SouL_ has left #arm-netbook [#arm-netbook]
hg_5 has quit [Ping timeout: 252 seconds]
penguin42 has joined #arm-netbook
pcat has quit [Read error: Connection reset by peer]
bfree_ has quit [Ping timeout: 252 seconds]
bfree_ has joined #arm-netbook
rektide has joined #arm-netbook
XenGi_ is now known as XenGi
ganbold_ has quit [Remote host closed the connection]
drachensun has quit [Quit: Leaving]
<Turl> converted branches to tags, much cleaner now :)
andy101 has quit [Quit: irc2go]
drachensun has joined #arm-netbook
<mnemoc> problem of git is that you can't go back to other VCSs... they feel unusably crippled
<traeak> interesting
<traeak> have to check out git again to see if there's some way to accomodate our special needs
<mnemoc> sub-repos still stink :|
<jochensp> mnemoc: you can use git with most other VCSs
<traeak> exactly
<Turl> mnemoc: I was forced to use svn last year
<Turl> it hurt :p
<traeak> we tried to use svn for a while and had to kill it
<mnemoc> jochensp: I mean the frustration of trying to do things on those other crippled VCSs
<Turl> hg is supposedly powerful
<Turl> haven't tried it yet
<traeak> sub repos
<traeak> no one ever did a good job of replicating CVS modules
<traeak> we rely way too heavily on those
<traeak> having atomic commits and actual file/directory renames i wish we had but we're small enough we still deal
<Turl> there's 'repo' for when you need to manage many git repos
<Turl> (the one used for android)
<Turl> its workflow is optimized for gerrit the code review tool
gimli has joined #arm-netbook
andy455 has joined #arm-netbook
<andy455> hi guys... one more question about sunxi. i'm getting an exception at the very beginning of the boot process (data abort)
<andy455> can i figure out where this exceptions occurs with the printed registers?
<jochensp> Turl: git-svn ;)
<Turl> andy455: can you paste the error?
<Turl> jochensp: can you push to svn servers with that?
<RaYmAn> Turl: yes
<jochensp> Turl: sure
<mnemoc> Turl: but `repo` only supports one level
<andy455> Turl: sure
<Turl> mnemoc: one level of what?
<Turl> jochensp: I'll keep it in mind for the next time I'm forced to use svn then :)
<mnemoc> Turl: iirc you can't have a working tree inside another working tree
<Turl> mnemoc: yeah but most of the time you can live without it :)
<Turl> just requires a little planning
<Turl> mnemoc: and maybe a file copy rule for a global makefile or two
<mnemoc> Turl: so it's just like having a bunch of git clone one next to the other
<Turl> yeah except you can sync them all in one shot
<mnemoc> just like with a 3 lines shell script
<mnemoc> otoh, the idea of the manifest is cool
<Turl> andy455: weird error :/
<Turl> andy455: does it happen all the time?
<andy455> almost
<andy455> sometimes it happens nothing
<traeak> at one point we were pasting together different repos. Had to do with keeping ownership separate.
<traeak> talk about pita :(
<mnemoc> can you enable earlyprintk?
<andy455> mnemoc: I enabled earlyprintk but there is no output
<Turl> andy455: what are you using to boot, nand or mmc?
<andy455> mmc
<Turl> andy455: are you using SPL for your specific device?
<Turl> the error really smells like bad/misconfigured ram to me
<traeak> nowadays we spin 5 or 6 different products out of the same repo using CVS modules. we keep everything very flat though libmath/testmath would be shared, libovx/testovx/mainOVX for more end stuff. oh well enough about repos.
<mnemoc> andy455: what defconfig did you use? what device?
<andy455> Turl: yes I'm using SPL
<andy455> mnemoc: I used a13om_defconfig
pcat has joined #arm-netbook
<andy455> mnemoc: so it is for the olimex micro board but mine has 512MB of ram instead of 256MB could that cause this error?
<andy455> Turl: configures the linux kernel the ram or is it all taken from u-boot?
<mnemoc> try using a normal a13_defconfig with mali disabled and the debug uart set to 1
<mnemoc> andy455: you need to use a u-boot-spl for your exact device model because it does the dram initialization
merbanan has joined #arm-netbook
<andy455> mnemoc: is it necessary to enable "Early printk" to accept the earlyprintk command line parameter
<andy455> ?
<mnemoc> yes, and DEBUG_LL
<andy455> ok
hg_5 has joined #arm-netbook
abesis-dev has joined #arm-netbook
Quarx has quit []
<andy455> hm... same exception again
pcat has quit [Ping timeout: 252 seconds]
XenGi is now known as XenGi_
doneill has quit [Ping timeout: 256 seconds]
doneill has joined #arm-netbook
<rm> there seems to be something wrong with wemac performance :/
<mnemoc> first check you got 100FD
<rm> yes I have
<rm> it's fast on LAN
<rm> but downloads from the internet much slower than other hosts behind the same router O.o
<rm> and looks like there are some "dropped" packets in ifconfig
<rm> the count increases by a bit
XenGi_ is now known as XenGi
<Turl> rm: mtu or one of those things?
L84Supper2 has joined #arm-netbook
L84Supper2 has quit [Changing host]
L84Supper2 has joined #arm-netbook
XenGi is now known as XenGi_
L84Supper2 has quit [Client Quit]
<rm> doesn't seem to be
<andy455> does anyone know what's function is called at the very beginning of starting the kernel?
<rm> sent to the mailing list
<RaYmAn> andy455: if you ask what you actually need instead, it might be possible to answer :P
<Turl> andy455: usually it's the decompressor
<RaYmAn> but basically, very early is in uhm, arch/arm/boot/compressed/head.S or something like that.
<Turl> RaYmAn: he's having 'data abort' early on boot
<RaYmAn> ah
<RaYmAn> backlog reading is overrated :P
* ssvb has destroyed an almost perfectly working gles accelerated kwin by doing an unnecessary upgrade :(
<Turl> rm: can you try reverting all the patches on wemac and compare performance?
<rm> define "all" :S
Holo_ has quit [Read error: Operation timed out]
Holo_ has joined #arm-netbook
<mnemoc> Turl: the related change is the replacement of allwinner's obsesive abuse of usleep() with wemac_delay() calls
<mnemoc> it also causes some people to initially sync at 10HD
<bsdfox_> ssvb, is Kwin performance significantly better than x11?
<mnemoc> but no decent fix has been proposed yet
<bsdfox_> err nm
<bsdfox_> I'm thinking of something different
<bsdfox_> Kdrive
<Turl> rm: git checkout 906086575a142b104ff0 drivers/net/sun4i/
<Turl> mnemoc: yeah I've hit the 10HD issue a couple of times
voronaam has joined #arm-netbook
<ssvb> bsdfox_: kwin is just a fancy eye candy stuff, like wobbly windows, etc.
<ssvb> bsdfox_: having it accelerated just means minimizing extra overhead compared to plain x11 stuff
<andy455> now, I recompiled the kernel with another compiler and the data_abort exception is gone. but I don't get any other output. so, i would like to add some "reboot statements" to the code to see if the kernel is executing the parts. has anybody a clue how I can do that?
<Turl> you can arm the watchdog and then busyloop
<jelly-home> ssvb: I didn't know kde compositing could use GLES as engine at all!
<jelly-home> ssvb: candy stuff is less useful than never having to see window redraws again
abesis has joined #arm-netbook
<andy455> Turl: thanks
maethorechannen has quit [Remote host closed the connection]
freakazoid0223 has quit [Quit: Leaving]
eFfeM has joined #arm-netbook
pcat has joined #arm-netbook
freakazoid0223 has joined #arm-netbook
pcat has quit [Ping timeout: 245 seconds]
<rm> Turl, so rolling back to my "net: sun4i-wemac: fix multicast support" kind of helps
<rm> (that's one commit later than you suggested, and before all the tweaks)
<rm> ah no, that's already including one ("net: sun4i-wemac: Attempting minor performance enhancement")
<rm> point is, the internet is now fast, but the LAN is still ~45 mbit
<Turl> rm: are you running on perf governor?
<rm> ondemand, tried performance briefly (before any git rollbacks), didn't help
<Turl> try now
<rm> 45
<Turl> top 3 look sane
<rm> yes, I am looking at it
<Turl> mine is just stylistic cleanup
<Turl> so it must be something between lundman's and hno's commit
<rm> will try at "net: sun4i-wemac: Based on dm9000 driver, remove delays and improve nic." now
<rm> with PHY_POWER changed to 1
<rm> inexplicably he set it to 0 which broke everything
<mnemoc> everuone wanted lundman fixes... and once merged everyone hated them :<
<mnemoc> it was very disturbing
<mnemoc> the only bit that remains is udelay() -> wemac_msleep()
<rm> 82.4K/s
<rm> yepppp
<mnemoc> libv: nice :)
<techn_> finger ;)
<mnemoc> well deserved
<Turl> so wemac_msleep is just a wrapper for msleep right?
<Turl> because in_suspend is 0
<Turl> so we went udelay->msleep?
<mnemoc> i don't remember what was inside wemac_msleep
<Turl> if(suspended) mdelay else msleep
<mnemoc> *g*
<mnemoc> so the diff is the number?
popolon has quit [Quit: Quitte]
<mnemoc> rm: just to be sure, can you revert only the udelay/wemac_msleep change over HEAD?
<rm> mnemoc, it's one huge spanking clusterfuck of a commit
<Turl> so 1msec is 1000usec right?
<rm> am I expected to manually peck in parts of it
<rm> or tell me how/where at least :)
<Turl> we're waiting like 10x?
<mnemoc> Turl: ouch
<mnemoc> and undo the udelay/wemac_msleep changes
<andy455> Turl: do you know if u-boot overwrites the dram?
<Turl> well, yeah, it loads the kernel there :)
<andy455> because if i fill some memory locations with zeros and reset the board (reset button) the content is changed
<andy455> i mean before the kernel is loaded
<mnemoc> andy455: u-boot initializes the dram
<Turl> well uboot might use some ram for itself too I guess
<mnemoc> yes, that too. the spl loads uboot on dram
<andy455> do you know where uboot is located ?
<rm> just always do the read twice
<rm> droppind the conditional
<rm> ng
<mnemoc> rm: that chunk was refactored later
dyoung-away is now known as dyoung
<rm> ah right
<mnemoc> i just want you to confirm me the problem is in udelay/wemac_msleep and not in my race condition fix :)
<rm> I will try it now
pcat has joined #arm-netbook
<hramrach> hmm, I have no ethernet problems but I am not exactly stressing it
<hramrach> I don't think I have any data dource or sink on A10 over 20 mbit
<rm> mnemoc, nope :/
<rm> I did this to HEAD http://dpaste.com/hold/924284/
<rm> doesn't help
<rm> er
<rm> not to head but to 3.0.57-r1
<rm> (shouldn't matter)
vinifm has quit [Quit: Ex-Chat]
<Turl> [ 3] 0.0-10.0 sec 23.5 MBytes 19.6 Mbits/sec
<Turl> that's what I get on my current ancient kernel hm
<Turl> err.. O.O
<Turl> [ 3] 0.0-10.1 sec 108 MBytes 90.2 Mbits/sec
<Turl> [ 3] 0.0-10.0 sec 108 MBytes 90.0 Mbits/sec
<Turl> after setting network to half duplex
<rm> mnemoc, actually
<rm> the "LAN" performance improved from 45 to 70 Mbit/sec
<rm> but from the internet it plummeted again to 500 kbit
<hramrach> heh, is ethernet supposed to work like that?
<mnemoc> uh
<rm> I am trying again with the complete rollback now, just to recheck
<rm> yes, a rollback to Jul 27 2012 solves it
<mnemoc> and chage wise, what does that date mean?
<rm> on my "net: sun4i-wemac: fix multicast support"
<mnemoc> lovely
<mnemoc> btw, what's on your wemac_power ?
abesis-dev has quit [Ping timeout: 252 seconds]
<rm> do you have a wemac device plugged in?
<mnemoc> unfortunatelly no
<rm> would be nice if anyone would try reproducing
datagutt has quit [Quit: kthxbai]
<rm> I don't know what's in wemac_power
<rm> my script.bin is from the latest cubieboard hwpack
<rm> let me decode it
<mnemoc> people has reported that removing that line from .fex solves things
<rm> there is no such line
<mnemoc> ok
simosx has joined #arm-netbook
simosx has quit [Changing host]
simosx has joined #arm-netbook
aholler_ is now known as aholler
<hramrach> yes, a wemac_power = (empty value) is shipped on cubieboard nand image
<hramrach> and reportedly removing that line improved things greatly on some versions of the driver
<mnemoc> plat-sunxi finally backported from 3.4 to 3.0
<SPG> master 6178e4e rhombus allwinner_a10/orders/.mdwn * http://git.hands.com/?p=rhombus.git;a=commitdiff;h=6178e4e
abesis2 has joined #arm-netbook
abesis has quit [Ping timeout: 245 seconds]
<Turl> mnemoc: :)
<mnemoc> i had a ton of 3.4 commits pending because of that
<mnemoc> now I can merge them on 3.0 too
<SPG> master cca6927 rhombus allwinner_a10/orders/venator.mdwn * http://git.hands.com/?p=rhombus.git;a=commitdiff;h=cca6927
hp__ has quit [Read error: Connection reset by peer]
uMinded has joined #arm-netbook
<uMinded> Hey, I have linux-sunxi-3.4.26 via "git checkout origin/sunxi-3.4" but their are not really any source files in /arch/arm/march_sun5i/ and their is no complete Kconfig. If their a patch I need?
dyoung is now known as dyoung-away
aesok has quit [Remote host closed the connection]
<mnemoc> and it's 3.4.29, not .26
<mnemoc> so something is wrong on your working dir
<mnemoc> wrong remote? forgot to pull? wrongly created local branch? ...
<mnemoc> I personally didn't know `git checkout origin/sunxi-3.4` was legal
andy455 has quit [Quit: irc2go]
<techn_> mnemoc: yes you can checkout remote branches.. but you get warnings that you are not in working tree.. if I remember right
<techn_> uMinded: try git fetch
<rm> mnemoc, reverting your rx race fix helps
<rm> although the "LAN" performance is now lowest ever
<rm> just 38 mbit
<mnemoc> :'(
<rm> also
<rm> it's difficult for me to hack a driver
ln2 has joined #arm-netbook
<rm> without understanding WTF all of these
<rm> reg_val = readl(db->emac_vbase + EMAC_INT_CTL_REG);
<rm> actually do
<Turl> reads the value on interrupt control register
<Turl> and stores it on reg_val
<Turl> it's like reg_val = *register
<ln2> Any news on the eoma-68 cards?
<mnemoc> rm: tried enable wemac debugging to see why the fix is harmful for you?
<ln2> Curious about any more up to date developments. =)
<mnemoc> ln2: we only know what's on that page and on the mailing list
<ln2> I see. Thank you. =)
<mnemoc> lkcl posts news as soon as he has something to tell
<ln2> That is more what I was getting at. I saw him in the channel not long ago but unfortunately couldn't stay.
<ln2> I'm curious about any developments with the AMD board specifically.
<mnemoc> nothing has been told about anyone working on making one
ln2 has quit [Read error: Connection reset by peer]
ln2 has joined #arm-netbook
freakazoid0223 has quit [Quit: Leaving]
ipwn has joined #arm-netbook
ln2 has quit [Quit: Yaaic - Yet another Android IRC client - http://www.yaaic.org]
hp__ has joined #arm-netbook
tinti has quit [Quit: Leaving]
focus_it has joined #arm-netbook
<libv> hrmpf. this whole X/mali stuff is also quite fragile.
<libv> heh, seems to be reverting to software rendering
<libv> is it known for AIGLX to default to software rendering?
<libv> libEGL warning: DRI2: failed to open Mali DRI2 (search paths /usr/lib/arm-linux-gnueabihf/dri)
rz2k has quit []
<mnemoc> are there mk802+ a10s with 512? or all are 1GB?
uMinded has quit [Quit: Page closed]
<techn_> libv: yes X is sw rendered
gimli has quit [Ping timeout: 276 seconds]
<libv> ok, so some mesa libs were still in the way
<techn_> ssvb's mali x11 driver does sw rendering better than ARMs
<libv> i want to get the test application ported to X as well, since hramrach lacked the mental capacity for that
<libv> the mali-libs test application
<techn_> purpose of mali x11 driver is currently to only provide gles acceleration
<libv> anyway, i am one step further now, XextCreateExtension is not found
<libv> seems like there is some dependency issue with this binary
<techn_> use libUMP which is compiled by you
<techn_> it will workaround those unresolved dependencies
<libv> i wouldn't know which other libUMP would be in this system
<libv> apart from the one i freshly compiled and installed
<libv> from exactly that url
<techn_> ldd?
<libv> finds the right one
<libv> installed just 12minutes ago
<ssvb> libv: there is a broken binary blob libUMP.so in the mali-libs repository
<techn_> could it be problem with that test app dependencies then :/
<libv> aha
<libv> yes, 12 minutes ago is a bit recent indeed, it was longer ago that i built and installed libUMP.so manually
<libv> ssvb: bingo
<libv> thanks guys
<libv> seems like libUMP.so should be removed from this thing
<techn_> that seems to add X11_LIBS
<libv> and the Makefile of libump could use some love too
<libv> techn_: the manually built libUMP.so seems to work
<libv> so... we need to get rid of the mali-libs libUMP.so for r3p0 x11
pcat has quit [Ping timeout: 240 seconds]
<techn_> yes.. it has workaround which tackles those unresolved dependencies.. libMALI.so has declaration of couple symbols but no definiton or link to needed libraries
<techn_> the real problem is that libMALI.so
<libv> well, that's not something that can be replaced easily, now is it ;)
<techn_> libv: not yet? :)
<techn_> someone proposed elfedit to remove those symbols.. but never tried it
<techn_> dunno what downsides there are
<libv> the manually built libUMP seems to work
<libv> good enough
<libv> i just need to update mali-libs and the howto
<ssvb> techn_: the DRI2SwapBuffers is really needed, and I think it is provided by https://github.com/robclark/libdri2
<libv> also, it might pay to have mali-libs warn about existing swrast libraries installed
<techn_> ssvb: yeah.. that downside :/
<ssvb> techn_: well, apparently some systems may also have libdri2 preinstalled, as explained in http://linux-sunxi.org/Mali400
<ssvb> but that's still also a workaround
<ssvb> rz2k knows better, but he is not around at the moment
<libv> mnemoc: can i push the r3p0/armhf/x11 removal patch directly to the mali-libs tree?
<ssvb> techn_: I think the AIGLX messages are about GLX, which is not supported for GLES, so it's just a noise
vinifm has joined #arm-netbook
<libv> ah, just did so
<ssvb> thanks
<libv> well, i too hate running into issues that others already ran into
eFfeM has quit [Quit: Leaving.]
<ssvb> techn_: now I need to add support for ARGB cursors and also proper page flipping
<ssvb> techn_: page flipping is apparently really need for gnome-shell, because it seems to partially render something immediately after flipping and then wait for a while before rendering the rest, it looks really ugly
<ssvb> techn_: if all the drawing is not really done offscreen before flipping
<ssvb> libv: the performance of xf86-video-mali really suffers from a lot of memory copies, but at least it can be used as a reference to check if the stuff really works more or less correctly
<libv> i am mainly just going to adjust the test application included in mali-libs
<ssvb> yeah, having a test application that just works would be really great
<libv> after that i will be adding mali binary shader loading to my lima tree, so i can run connors freshly compiled shaders on q3a
<libv> then we will have a fully free q3a \0/
<ssvb> btw, so far I have been using this as a simple X11 example - http://www.agnld.uni-potsdam.de/~bergner/egl-example.cpp
<libv> there is some quick X11 GLX port howto i pasted to hramrach many many months ago
<ssvb> about q3a, that's really great, but could you try using openarena or something so that it does not require proprietary files?
<libv> the timedemo is the timedemo :)
<mnemoc> libv: isn't r3p0/armhf/x11 what everyone still uses?
<libv> mnemoc: yes, but the libUMP.so in there gets in the way
<libv> one should use the manually compiled one
<libv> if you use the howto correctly, you will automatically override the mali-libs libUMP.so
<mnemoc> ah, ok
<libv> but if you step slightly out of line, then you run into issues which are non-trivial
<mnemoc> then it's presence is a bug :)
<libv> so imho it is/was best to remove that libUMP.so, which was done just now
freakazoid0223 has joined #arm-netbook
<libv> yeah :)
<libv> it is rather amazing how little these trees do get hacked, even though people know of all sorts of issues with them
<mnemoc> fix it :)
<libv> it's been fixed
<libv> same for the wiki, the things i ran into were well known to some, but noone bothered to fix up the wiki
<libv> also, andoma, why has your issue not been added to the wiki?
<andoma> yeah i know
<andoma> i suck
<andoma> :)
<andoma> i'll fix it right away
<andoma> ok fixed now
<libv> thanks :)
ipwn has quit []
<vinifm> is there any problem if I change struct spi_transfer { ?
<vinifm> in include/linux/spi/spi.h
ipwn has joined #arm-netbook
hg_5 has quit [Read error: Connection reset by peer]
mur_ has quit [Quit: User abort with 5 Ctrl-C's]
Maqs has quit [Ping timeout: 240 seconds]
drachenphone has joined #arm-netbook
<br-> its really quite telling just how little bandwidth you can get away with when booting from usb
* br- in the middle of an ssd install, desktop feels perfectly normal, despite booted from usb2 caddy seemingly constrained to 15mb/sec
focus_it has quit [Quit: Leaving]
raoulh has quit [Remote host closed the connection]
raoulh has joined #arm-netbook