<rz2k>
allwinner guys user mail at reuuimllatech.com
<rz2k>
which is free...
<rz2k>
s/user/use in gpl'd files/
<rz2k>
lets register that domain and redirect it to linux-sunxi :p
<ZaEarl>
mnemoc, is that an a31 kernel source?
<Turl>
rz2k: go read reuuimlla backwards and upside down
<rz2k>
HMMM...
* rz2k
feels dumb
<rz2k>
interesting way of providing email
<rz2k>
ZaEarl: since it has mach-sun6i fully populated...
<rz2k>
and powerVR kernel module is also there
<Turl>
and sun7i is A20 :)
<ZaEarl>
where'd it come from? does it work?
<rz2k>
git clone and test if it compiles
<rz2k>
I have no a20/31 devices :(
<ZaEarl>
I'm cloning, but it's going to take quite a while
<Turl>
ZaEarl: do you own a20/a31 devices?
<ZaEarl>
yes a31 tablet
<ZaEarl>
haven't had luck finding any source for it yet
<Turl>
it's kind of your lucky day then :)
<ZaEarl>
guess I'll be busy this weekend
<Turl>
night all
rz2k has quit []
mnemoc has quit [Excess Flood]
mnemoc has joined #linux-sunxi
mnemoc has quit [Excess Flood]
mnemoc has joined #linux-sunxi
<ZaEarl>
Is sun6ismp_fiber_defconfig a good config?
<uMinded>
This is a slow chat so I went upstairs, sorry.
<uMinded>
I want to use the 64bit counter already present in march-sun5i/clock/aw_clocksrc.c but it uses a different reload style than timers0-4
<uMinded>
I want to compile without hi resolution timer support and then use the 64bit counter in my kernel module. Im just not sure how to access its functions from userspace for testing
mnemoc has quit [Excess Flood]
mnemoc has joined #linux-sunxi
<Skaag>
mnemoc: so I finally got to test the newly compiled kernel on my A13_Olinuxino, and it works great
<theOzzieRat>
i was just going to ask if you could tell me if I had submitted my 2 spi patches correctly?
<mnemoc>
theOzzieRat: yes, submit is correct. can't tell about the changes themselves yet
<theOzzieRat>
cool. that all i was after atm.
<theOzzieRat>
actually I do have another question, in spi_sunxi.c there is a function aw_spi_set_waitclk_cnt that is not called from anywhere. For some spi transfers it needs to be called so that the slave has time to load data. Any idea what the best way to go about changing stuff to fix this? I was thinking and field in struct spi_transfer (in spi.h) was most likely the best way.
<techn_>
mnemoc: \ o/ :D
<mnemoc>
techn_: i might make another git push -f today, but it's a good start anyway
<techn_>
hmm.. why they are using unstable kernel as base? :/
<libv>
mnemoc: i just fixed up r3p2 xf86-video-mali with hansg and rz2ks changes
<libv>
i need this for the odroid
<libv>
can i just go and push this into sunxi, as it is generic
<libv>
basically, i have redone the xserver 1.13 port
<mnemoc>
libv: as long as you don't remove the old version.... :)
<libv>
k
<libv>
i of course would not remove existing code
<libv>
instead of copying from the arm site, i noticed that we had a clone of this (a non-unified clone at that) as xf86-video-mali, with the r3p2 branch included
<mnemoc>
techn_: I suppose we should focus in merging the 4 in sunxi-3.4
<mnemoc>
libv: oh, I didn't know
<mnemoc>
techn_: will push the newer "lichee" A10, A13 and A10s kernels (constructed upon the qware base we used as starting point) to my github today as well
<techn_>
mnemoc: nice.. cherry-pickking time :)
<libv>
un-be-liev-able
<libv>
r3p2 shifts the GET_API_VERSION up two ioctls
<libv>
any chance of hipboi getting us r3p2 binaries?
<libv>
ok, r3p2 libump is also unchanged compared to our tree.
<mnemoc>
ssvb: fyi, sun6i (A31) and sun7i (A20) have g2d drives in the source drop
<mnemoc>
drivers*
<ssvb>
mnemoc: good to know that the G2D hardware is still there
rellla has joined #linux-sunxi
<ssvb>
mnemoc: where can one find this source drop?
<mnemoc>
ssvb: my github
<mnemoc>
i might do another `git push -f` today, but it's already sanitized/cleaned, so it can be reviewed easily
<libv>
haha
<libv>
r3p2 has the same x11 bug as our r3p0 had
<mnemoc>
:)
<libv>
the one for which i provided a binary fix
<ssvb>
mnemoc: did allwinner officially release the sources?
<mnemoc>
no
<mnemoc>
hopefully we can get a GPL confirmation now that the code is public
<mnemoc>
that's why it's (at least for now) kept on my personal github
* ssvb
facepalm
<mnemoc>
for now it counts as... liberated gpl code
hansg has joined #linux-sunxi
gzamboni has quit [Remote host closed the connection]
<ssvb>
libv: would it make sense to submit a bugreport to ARM about this r3p0 and r3p2 issue?
<ssvb>
but surely this long "arm -> samsung -> hardkernel -> end user" chain is absurd and not very encouraging
ZaEarl has quit [Ping timeout: 244 seconds]
Liberator has joined #linux-sunxi
<libv>
ssvb: nah
<libv>
ssvb: wait a few months, and lima should be useful
<mnemoc>
:)
ZaEarl has joined #linux-sunxi
<ssvb>
mnemoc: looks like a newer revision of G2D (in sun6i) is supposed to have premultiplied alpha support, hope the same applies to sun7i
<mnemoc>
i suppose they use the same ip in both
<ssvb>
ok, then I need to do some experiments with G2D in sun4i and try the same bits
<ssvb>
*if* premultiplied alpha is actually supported in sun4i but is just not documented, then this is a great news
<ssvb>
but I believe that just a newer revision of g2d ip is more likely
<mnemoc>
i was thinking in a20 vs a31, not a10 :)
<ssvb>
ah, ok
<ZaEarl>
$ ./gen_check_code
<ZaEarl>
can't open source file
<ZaEarl>
how do I run gen_check_code?
<bfree>
am I misinterpretting things or is there a u-boot+script.bin config which works to some degree across all (or most) a10 boards (a10s or a13 obviously a bonus)? and if so which one is it?
<ssvb>
mnemoc: anyway, "a fixed version of g2d is here to stay in newer hardware" is surely more encouraging than "g2d is lacking important features and is going away" that some people thought before :)
* ssvb
wants some a20 based hardware
* mnemoc
too
<Skaag>
mnemoc: stage/sunxi-3.0
<Skaag>
which is weird because I was certain I switched to another branch you told me about
<mnemoc>
so you confirm stage/sunxi-3.0 works ok? need to merge those changes into sunxi-3.0 (and the same for 3.4)
zub has joined #linux-sunxi
ganbold__ has joined #linux-sunxi
<zub>
ls
<zub>
wrong window, sorry :)
ganbold_ has quit [Remote host closed the connection]
Liberator has quit [Quit: Lost terminal]
<zub>
has anybody tried OE for building images for the allwinner?