<Turl>
hno: I don't see the revision on dmesg, maybe it's only on 3.4?
<Turl>
I see brom ver 1100 1100 1623
<hno>
do you have sunxi-dbgreg.ko?
<hno>
or console access to u-boot?
<Turl>
no console access
<Turl>
it's unlikely I have that module too, I like to y/n
Quarx has joined #arm-netbook
<hno>
ls -l /sys/devices/virtual/misc/sunxi-reg/
<Turl>
no such thing hno
* Turl
builds the module
<Turl>
ok hno, it's loaded
<hno>
Ok, then we need to read 0xF1C20D3C
<hno>
you got the directory now?
<Turl>
yes
<hno>
What does it contain? (never done this with out kernel, only AW A13 kernel which is slightly different)
<Turl>
hno: 0x80
<hno>
So rev C then. Which is the rev we applied major clocking changes for.
<hno>
the "SATA fix".
<Turl>
my tablet doesn't have sata :p
* Turl
needs sleep
<hno>
The "SATA fix" changes how the PLL clock generators are used for many I/O blocks, not only SATA, and I can imagine that some clocks temporarily gets upset until reinitialized by the kernel to match the PLL settings.
<Turl>
it's possible
<Turl>
good night hno
arete74 has joined #arm-netbook
<hno>
Good night.
<Turl>
I'll have a look tomorrow at the sata patch
RITRedbeard has quit [Read error: Connection reset by peer]
sspiff has joined #arm-netbook
DEAT_ has quit [Read error: Connection reset by peer]
DEAT_ has joined #arm-netbook
<mnemoc>
hno: so r-pi now? :p
<andoma>
morning
<mnemoc>
hno: what's that header you put on top of the uSD/audio plugs of your a13 olinuxino?
<mnemoc>
hno: jtag?
popolon has joined #arm-netbook
<oliv3r>
mnemoc: it is fixed :D
<oliv3r>
and good morning :)
<oliv3r>
was reading the cnbeta.com post via googletranslate. Its funny to read 'full-Chi' to as their translation for Allwinner and AMLogic as Crystal Morning
<hno>
mnemoc, yes there is an R-Pi in the board collection now.Have not powered it on yet.
<hno>
mnemoc, yes that's a 20 pin JTAG header, wired to the corresponding SD signals on a resistor mat just below. Was a hell to solder.
<mnemoc>
oliv3r: what's fixed?
<mnemoc>
hno: :D
<mnemoc>
morning andoma
jelly has joined #arm-netbook
<hno>
mnemoc, it's also what makes the A13 a very handy dev board for me, in combination with an JTAG SD card.
<hno>
== a card with a boot sector that forces JTAG and then halts.
<mnemoc>
nice
<hno>
doing the same on the cubie with the breakout is a mess.
<mnemoc>
tried to get the mmc where the nand is soldered yet?
<hno>
Not yet.
gsilvis has quit [Read error: Connection reset by peer]
<mnemoc>
hno: what about asking tom for a board without nand?
gsilvis has joined #arm-netbook
<mnemoc>
the spl halt-on-reboot problem is awful
<hno>
Yes. Should do that.
<hno>
but seem to never meet in IRC. Here at different times.
<oliv3r>
mnemoc: the wiki, the wiki!
<mnemoc>
hno: he isn't comming often, a mail could do better
<mnemoc>
oliv3r: :)
<mnemoc>
oliv3r: installing dvipng solved it. there is something buggy in the generation via gs
<hno>
math?
<mnemoc>
yes
<mnemoc>
for the CCM page
<mnemoc>
there are some fancy formulae there
<oliv3r>
It's only very basic math that is used (for now) but i do think \frac{foo}{bar} is far more readable (as output) then foo/bar
<mnemoc>
absolutely
<oliv3r>
so thank you for adding it, even if it is only for that one simple page (for now :p)
<mnemoc>
oliv3r: just be sure to check the theorical stuff with the values in real code
<oliv3r>
the question always remains, who will be lieing :)
<oliv3r>
but will absolutly do
<mnemoc>
code always win
<oliv3r>
but what if the code is buggy! (see sata/pll4 stuff :p)
<mnemoc>
:)
<oliv3r>
i know what you are saying of course, but fact is, you cannot ever be 100% sure that either is 100% correct :(
<mnemoc>
a sad truth. so in the cases of conflict, make it clear in the docu :)
<oliv3r>
aye
<mnemoc>
and <ref />
<oliv3r>
... <ref />? O: )
<oliv3r>
not sure if it is wise to first literally document things, then modify based on code or not. if done so, you have it somewhere in the history at the least
<mnemoc>
good point
<oliv3r>
strange, clock.h only lists half if that of the stuff the doc has
<oliv3r>
even more funny, the code has more info then the doc :D
<andoma>
jo
<oliv3r>
why do you always make me harder then I planned :(
<andoma>
sorry wrong window
<oliv3r>
andoma: sup :p
<andoma>
oliv3r: nothing really. dayjobing :)
<mnemoc>
oliv3r: eh? what i've complicated?
<oliv3r>
making me verify it against the source etc! i was gonna be lazy today :(
<oliv3r>
shall i issue my pull request against that?/
<oliv3r>
and retract the one i have?
<mnemoc>
+1
<oliv3r>
see! make me work hard!
<mnemoc>
:)
<oliv3r>
lets see if i can figure out how to do that in github :p
<oliv3r>
why do you have .md appended to your README files?
<mnemoc>
github's template
<mnemoc>
.md means markdown syntax
<oliv3r>
ah, ok
<oliv3r>
never saw that before is all
<oliv3r>
bah, github being slow :(
<mnemoc>
one usually has something own to push as first commit instead of using github's template
<oliv3r>
only getting 5MiB/s cloning/updating BSP
<mnemoc>
bah, I never get more than 1
<mnemoc>
oliv3r: actually, I'll give you push access to allwinner-tools
<mnemoc>
oliv3r: beware that makes you responsible of it :p
merbanan has joined #arm-netbook
<oliv3r>
i don't want that responsibility (yet) :p
<oliv3r>
i'm too unfamiliar with git, i'll break things I promise
<mnemoc>
:)
<mnemoc>
ok, fork and make the pull request.... for now
<oliv3r>
forked it allready, just waiting for the BSP to finish updating
<oliv3r>
i probably have to create patch/apply patch or something magic; i'll make it happen :)
tzafrir_laptop has quit [Ping timeout: 260 seconds]
cat_x301 has joined #arm-netbook
<oliv3r>
mnemoc: done the pull request, but accidentally closed it (wrong window, do'h ment to close the other one)
<oliv3r>
mnemoc: i assume you can unclose it? i can't i don't think
<mnemoc>
yes, you can unclose it
<oliv3r>
then i will find it
<oliv3r>
'reopen' sounded reasonable enough :p
<oliv3r>
sorry for the noise :)
<mnemoc>
you only ding'ed my phone, both tablets a couple of times ...
<oliv3r>
:D
<oliv3r>
sorry :(
<mnemoc>
don't worry
<mnemoc>
eventually I'll learn and disable those notifications
<mnemoc>
or at least disable them in the tablets
<oliv3r>
But how will we grab your attention :)
<mnemoc>
:)
<oliv3r>
is it common practise (as it by proper kernel devs) to use register numbers in struct/define/variable names? e.g. __CCMU_PLL!_CORE_REG000
<mnemoc>
it is a good practice to give names to things instead of "magic numbers"
<oliv3r>
thought so
rellla has quit [Ping timeout: 255 seconds]
<oliv3r>
i just find ccmu_regs.h (and others) to be globberd full with them
<oliv3r>
won't even mention the fact that it's all plastered in typedefs
<mnemoc>
that's not a good practice
<oliv3r>
i'll admit, i'm looking at an old 'blob' branch, just verifying is all
rellla has joined #arm-netbook
<oliv3r>
is there an easy way to add a tag to unverified values? I now put <unverified> behind such content, but i'm sure there's some wiki magic for this?
<mnemoc>
a `{unverified}` template seems a good idea. .... like wikipedia's citation needed
<lundman>
odd i read that blog just yesterday, for no real reason
merbanan has quit [Ping timeout: 265 seconds]
alcides has joined #arm-netbook
alcides has joined #arm-netbook
alcides has quit [Changing host]
mnemoc has quit [Excess Flood]
mnemoc has joined #arm-netbook
<oliv3r>
wb
<mnemoc>
meh
<oliv3r>
what is the wiki's session timeout? It could be my proxy, but it often complains that it can't save my document due to something in the session. saving it then work okay
<mnemoc>
oliv3r: happens if you take too long to save a change
<mnemoc>
you can keep your session alive in a separated tab
<mnemoc>
re-login there if required
<mnemoc>
after that you'll be able to submit your change without troubles
<oliv3r>
i don't get logged out, it just says (i'll copy paste it later)
<oliv3r>
it hasn't caused a problem yet, i just have to hit save again, just wondering if it really takes me that long to write stuff :)
<hno>
mnemoc, please post your reply to list and we continue there.
<mnemoc>
hno: the reply about g2d would have been a perfect one to continue on the ML :p
<mnemoc>
i really haven't replied anything useful in that thread :|
<oliv3r>
I don't know struct bit wise addressing things (the name:bitsize; thing in a struct) I read up on it very briefly. It's quite new to me. Having said that, https://github.com/linux-sunxi/linux-sunxi/blob/sunxi-3.
<oliv3r>
0/arch/arm/mach-sun4i/include/mach/ccmu_regs.h#L51 has a struct, but without the 'reserved:2' added at that specific line. Unless my math is very wrong, InitFreq:7 should be 7 bits wide, starting at bit16. If VCOGain should start at bit26, what happend to bit 24 and 25? Could this cause bugs (if used at all?)
<oliv3r>
the manual actually conviniently forget the entire PLL1_tuning section :S
merbanan has joined #arm-netbook
tzafrir_laptop has joined #arm-netbook
<mnemoc>
hno: what reply in particular you want me to repost on the ML?
deathz0r- has joined #arm-netbook
DEAT_ has quit [Ping timeout: 252 seconds]
Almamuetya has joined #arm-netbook
<oliv3r>
quite some inconcistencies between manual - code :S
<oliv3r>
quite nasty actually
<mnemoc>
yes :<
<mnemoc>
that's why we can't trust the manual :|
<oliv3r>
i bet
<mnemoc>
but eventually we will be able to trust the wiki :)
<oliv3r>
so we have a manual with actual real bugs (some caused by typo's i'm sure, some caused by forcefully ripping bits out), and code, that could be quite buggy even so
<oliv3r>
eventually yeah lol
<mnemoc>
:p
ZaEarl has quit [Ping timeout: 260 seconds]
<oliv3r>
i mean, if it's AW code, god knows how true that is :)
<mnemoc>
at least we know code "works"
<oliv3r>
do we know af A10's running current code for months without issues/crashes?
<oliv3r>
My tablet doesn't seem particiular stable (not tried my own built kernel of course)
<Turl>
oliv3r: well I've run my mele-server for weeks at times, but I reboot it to install new software :P
<oliv3r>
:p
<oliv3r>
how in gods world do they expect people to work with a datasheet like that :S I feel sorry for the people who actually have to derive information out of it
<alcides>
^ true
<oliv3r>
<smeagul>My precious! I must keep it zzeeecrettz, My precious chip, only I shall know! Prrreeccieeous</smeagul>
<Turl>
mnemoc: patches stuck on moderation? :(
<Turl>
hm nvm, the web has them
<lkcl>
oliv3r: look at TI as a counter-example. how's their "comprehensive documentation" policy working out for them, really? they've *failed* in the mass-volume mobile space, and are currently in advance negotiations with amazon for the sale of their entire Mobile SoC Division.
mysteryname has quit [Ping timeout: 268 seconds]
<Turl>
lkcl: I don't think TI failed
<oliv3r>
lkcl: TI opened their specs entirely? or you being sarcastic?
<Turl>
no idea about specs, but TI has *the best* reference code
<oliv3r>
anyway, OMAP345 are used somewhat reasonable? and it failing in the mobile (e.g. GSM and maybe even tablet) market, I don't think it's their fault really. Who are the big handset players, Apple and Samsung. Sure HTC is of an 'ok' size, but not doing so well atm either. Apple 'makes' their own soc (or would like us belive that), samsung has their own SoC. So who's left to make the omap be amazing? Nokia? HTC?
<oliv3r>
I doubt documentation has much meaning in this case
<Turl>
lkcl: amazon uses TI chips, most motorolas are TIs, nexus devices are TI-based
<oliv3r>
nexus 7 is tegra3 :p
<Turl>
oliv3r: HTC uses qualcomm pretty much exclusively
<Turl>
oliv3r: galaxy nexus and nexus q are omap
<oliv3r>
Turl: really? Didn't know. I do have an HTC though :)
<oliv3r>
but the mobile market is a cut-throat business if you ask me
<oliv3r>
specs or no specs won't matter much, price per unit probably means everything
<oliv3r>
price per mip/watt if you must
<Turl>
personally I dislike HTC :P
<Turl>
their quality is meh :P
<oliv3r>
Turl: it was a free gift :)
<oliv3r>
it's one of the first androids, the HTC hero :)
<Turl>
I won't complain if it's free :P
<Turl>
oliv3r: MSM7201A? :)
<oliv3r>
I even have an ol winmo HTC (Touch cruise) running android
<oliv3r>
Turl: yep
<Turl>
oliv3r: I call with one daily :P but motorola
<oliv3r>
TI should have made those mirror repositories a seperate directory, it's very crowded now with mirrors :S
<oliv3r>
Turl: motorolla msm7201a?
<Turl>
oliv3r: yep
<oliv3r>
ah :)
<Turl>
oliv3r: cliq xt / quench, depending on where you live
<oliv3r>
Eu
<Turl>
so probably quench then :)
<oliv3r>
ti even OSSed their bootloader?
<oliv3r>
(omapboot)
<RaYmAn>
TI uses u-boot + hacked up mini-uboot for SPL afaik
<oliv3r>
ah wait, that's USB downloader and USB 2nd stage
<RaYmAn>
might have been different in older OMAP's though
pawel5870 has quit [Ping timeout: 248 seconds]
<oliv3r>
Ok, done for today!
<oliv3r>
No more energy :p
<oliv3r>
or will power for that matter
<oliv3r>
anybody feel free to continue to work on PLL4 :)
zeug has joined #arm-netbook
mikey_w has quit [Remote host closed the connection]
<hg_5>
i have uhost device, connected it to pc with 2usb, how to get access to its storage?
<hg_5>
linux-sunxi.org/UHost
<hg_5>
what linux distribution is suitable with this device?
<mnemoc>
any that supports arm
<mnemoc>
just be sure to get a serial console
<mnemoc>
or the first steps will be... painful
<techn_>
arm-linux-gnueabihf-ld.bfd: error: /usr/lib/gcc/arm-linux-gnueabihf/4.6/libgcc.a(bpabi.o) uses VFP register arguments, u-boot does not
<techn_>
arm-linux-gnueabihf-ld.bfd: failed to merge target specific data of file /usr/lib/gcc/arm-linux-gnueabihf/4.6/libgcc.a(bpabi.o)
<techn_>
uboot must be hardcoded to armel?
<RaYmAn>
what's the argument for wanting to use armhf for u-boot? I kind of doubt u-boot does a lot of floating point calculations ;)
<techn_>
RaYmAn: yes.. but why not :)
<RaYmAn>
well, only argument to why not is: there's no benefit from doing it. It makes no real difference :P
<mnemoc>
techn_: install the sf compiler
<mnemoc>
it will fallback silently
<mnemoc>
gcc-multilib if building natively
<mnemoc>
gcc-arm-linux-gnueabi in the case of ubuntu x86 host, etc
<techn_>
sf compiler?
<mnemoc>
"soft float" :)
<techn_>
yeah.. I have it installed
<techn_>
but I was wondering why you changed u-boot to use CROSS_COMPILER if it can be build only armel
<mnemoc>
as a "political statement" that it should be hardcoded
<mnemoc>
shouldn't*
<mnemoc>
also, here it does fallback to sf nicely
<mnemoc>
in the worst case `make u-boot CROSS_COMPILE=arm-linux-gnueabi-` is always there
<techn_>
mine doesn't fallback :(
<mnemoc>
uhm
<mnemoc>
as there are several toolchains with different prefixes we can't blindly ignore the provided CROSS_COMPILE= and assume arm-linux-gnueabi- of u-boot
<mnemoc>
sure you have `gcc-4.6-arm-linux-gnueabi` installed?
<techn_>
"gcc-arm-linux-gnueabi is already the newest version."
<mnemoc>
that's virtual... do you have the other?
<mnemoc>
i'm not clear why hno forced soft float
<techn_>
gcc version 4.6.3 (Ubuntu/Linaro 4.6.3-1ubuntu5) => arm-linux-gnueabi-gcc arm-linux-gnueabi-gcc-4.6
<techn_>
gcc version 4.6.3 (Ubuntu/Linaro 4.6.3-1ubuntu5) => arm-linux-gnueabihf-gcc arm-linux-gnueabihf-gcc-4.6
<techn_>
or did you mean "gcc-4.6-arm-linux-gnueabi is already the newest version." :)
* mnemoc
runs a test `make u-boot` on his laptop
<hg_5>
hey, Allwinner A10 1GHz Cortex-A8 is one or dual core?
<mnemoc>
hg_5: all cortex-a8 are single core
<hg_5>
oh ;., thx
<mnemoc>
techn_: and it failed! wtf
<mnemoc>
will test tonight if it boots compiled by hf with the soft-float imposition removed
rsalveti_ has joined #arm-netbook
hg_5 has quit [Read error: Connection reset by peer]
<libv>
pfff, the mental midgets...
<libv>
value &= mask_bit & mask_bit & mask_bit;
<mnemoc>
o.o
<libv>
third file i encountered that in already...
<mnemoc>
in the disp driver?
<libv>
yeah
<mnemoc>
the sadest part is that i don't find it hard to believe :<
<libv>
i am sure that suddenly, for instance, the clock code will start working a bit better
<mnemoc>
libv: if you are doing changes already, please use the stage branch as base
<mnemoc>
either 3.0's or 3.4
ZaEarl has joined #arm-netbook
merbzt has joined #arm-netbook
RITRedbeard has quit [Read error: Connection reset by peer]
RITRedbeard has joined #arm-netbook
rsalveti_ is now known as rsalveti
Brandon15811 has quit [Excess Flood]
Brandon15811 has joined #arm-netbook
<bsdfox_>
I just got my toolchain working but I'm hitting the -msoft-float issue now
<bsdfox_>
where do I disable that?
<mnemoc>
bsdfox_: pass the prefix to a soft-float toolchain as CROSS_COMPILE
<bsdfox_>
I'm not sure that I follow. I've got a hard-float toolchain but I believe I', hitting this issue https://bugs.meego.com/show_bug.cgi?id=13140 due to this error ... uses VFP register arguments, u-boot does not
<bsdfox_>
I'm looking at the patch specified but there aren't any file names
<bsdfox_>
I'm trying to use a10-hwpack-bld.sh right now
<mnemoc>
that error means you are using a hardfloat compiler and u-boot wants a softfloat compiler
<mnemoc>
blindly patching u-boot to build with a hardfloat compiler is a bad idea
<mnemoc>
it might not boot and kill your pets
<bsdfox_>
ok I just built armv7a-hardfloat-linux-gnueabi. perhaps I should give it a shot with armv7a-softfloat-linux-gnueabi?
<mnemoc>
yes
<mnemoc>
hf is fine for everything else, just not for u-boot
eFfeM has joined #arm-netbook
<bsdfox_>
ok
* bsdfox_
thinks sata support is fixed since 10/16 or so
<bsdfox_>
but the builds on sunxi-linux.org aren't working
<mnemoc>
we have a fix to the sata fix in the stage branch
<rzk>
u-boot was working with hf compiler week ago
<rzk>
when hno pushed new commits from u-boot master git that stuff happend
<bsdfox_>
perhaps I am using the wrong sources. I had been using the hwpack and and a1x-media-create
<mnemoc>
rzk: :<
<mnemoc>
rzk: so it was upstream u-boot who imposed sf?
<mnemoc>
bsdfox_: problem is those scripts are unmaintained, and things change
<bsdfox_>
I've got a working userland. should be able to copy kernel/modules over right?
<mnemoc>
yes
gimli has quit [Quit: Verlassend]
eFfeM has quit [Quit: Leaving.]
<rzk>
mnemoc: yes, it is from upstream
rsalveti has quit [Ping timeout: 260 seconds]
rsalveti has joined #arm-netbook
rsalveti has quit [Read error: No route to host]
rsalveti has joined #arm-netbook
hp__ has quit [Read error: Connection reset by peer]
<bsdfox_>
mnemoc, where would I find the stage branch? I just compiled the main one and got it working but no sata drivers (I just compoiled defconfig)
pawel5870 has joined #arm-netbook
mysteryname has joined #arm-netbook
Svet has joined #arm-netbook
Svet has quit [Changing host]
Svet has joined #arm-netbook
Sv has quit [Ping timeout: 276 seconds]
Entropy512_ has joined #arm-netbook
Entropy512_ is now known as Entropy512
valhalla has quit [Ping timeout: 252 seconds]
<rzk>
bsdfox_: git branch -a to view branches, git checkout <name of the branch> to switch
<rzk>
switch to remote/origin/stages/3.0 or something like that