<wens>
use the sunxi-next branch from u-boot-sunxi repo, or get it from the official u-boot repos
<wens>
ah
vitorcarvalhoml has joined #linux-sunxi
<alpha_one_x86>
and what about cpu frequency to have minimal power consumption?
steev has quit [Ping timeout: 258 seconds]
Black_Horseman has quit [Remote host closed the connection]
steev has joined #linux-sunxi
<mripard>
alpha_one_x86: not supported yet
<alpha_one_x86>
ok
dlan has quit [Ping timeout: 258 seconds]
naobsd has quit [Quit: Page closed]
dlan has joined #linux-sunxi
dlan has joined #linux-sunxi
<alpha_one_x86>
The power gatting work for GPU when linux kernel modules is disabled?
cosm has quit [Ping timeout: 245 seconds]
cnxsoft1 has quit [Remote host closed the connection]
cnxsoft has joined #linux-sunxi
JohnDoe_71Rus has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org]
Akagi201 has joined #linux-sunxi
naobsd has joined #linux-sunxi
Andy-D has joined #linux-sunxi
cosm has joined #linux-sunxi
naobsd_ has joined #linux-sunxi
naobsd has quit [Ping timeout: 246 seconds]
Quarx has joined #linux-sunxi
selfbg has joined #linux-sunxi
<afaerber>
wens, when you submit, please add my Sob for the mmc DT nodes after I spent time figuring that data out.
<afaerber>
wens, your usb branch has been working well in the meantime, I used a USB-Ethernet adapter to make the system usable
<afaerber>
been looking into the rgx driver, but only clocks and regulators to be seen there, apparently no registers to read revision and stuff like on TI systems
xinj has joined #linux-sunxi
lioka has quit [Ping timeout: 265 seconds]
lioka has joined #linux-sunxi
cnxsoft has quit [Quit: cnxsoft]
<alpha_one_x86>
cubieboard + 3.17 = crashed under load...
afaerber_ has joined #linux-sunxi
<alpha_one_x86>
again...
afaerber has quit [Ping timeout: 265 seconds]
Dodger78 has quit [Read error: Connection reset by peer]
afaerber_ is now known as afaerber
lioka has quit [Changing host]
lioka has joined #linux-sunxi
<alpha_one_x86>
I have do: /usr/bin/ionice -c 3 /usr/bin/nice -n 19 dd if=/dev/sda of=/dev/null -> 624 kB/s why so slow?
<alpha_one_x86>
I have say nothing
<wens>
afaerber: rgx?
<libv>
rogue
<libv>
powervr 6xxx
<afaerber>
yep, the one that's not been leaked "luckily"
<wens>
oh
<libv>
afaerber: some things will be the same
<libv>
from what i can tell from the kernel side, without digging too much
<afaerber>
libv, I guess I could ask you general questions but would get "no comment" ;)
naobsd_ has quit [Quit: Page closed]
leviathanch has quit [Ping timeout: 256 seconds]
<libv>
afaerber: i remember no microcode/userspace details anyway
<afaerber>
like, where in memory does the thing actually sit - is that accessed via /dev/mem from userspace or am I blind?
<libv>
afaerber: check the kernel driver
<afaerber>
well, I did - probably not in the right place
<libv>
its register space is accessible from the kernel
<libv>
it does mmu stuff in the kernel as well
<libv>
it also uploads the microcode from the kernel
<afaerber>
libv, so, taking a step back - freedreno is in drivers/gpu/drm/msm iiuc. where would mali and pvr even go given that they're not specific to exynos/sunxi/etc.? gpu/drm/pvr?
<libv>
they are not drm drivers.
<libv>
afaerber: why does this question come up now?
<afaerber>
so gpu/pvr then? note that I am not talking about the bad downstream drivers but about where proper Open Source drivers should go - should they not be integrated with drm framework somehow?
<libv>
at some point yes
<afaerber>
at least freedreno needed a libdrm which I assume is communicating via ioctls with the drm/msm driver
<libv>
afaerber: so the plan is to create a skeleton drm driver, and only then to figure out where pvr register/memory/etc live?
xinj has quit [Ping timeout: 265 seconds]
<afaerber>
libv, that's what I've recently done for the Epiphany, yes. it's independent of the Mesa etc. side of course, which I know less about
<libv>
i see, parallella
selfbg has quit [Quit: Leaving]
naobsd has joined #linux-sunxi
<afaerber>
libv, btw, what about your sunxi kms driver? would that be usable on the a80 as well? right now there's no graphics to play with - I didn't get very far with U-Boot
<hramrach_>
hmm, livesuit WorksOnWindows(tm)
<hramrach_>
but it is 1.09
<hramrach_>
maybe you really need ancient livesuite for A13
<wens>
afaerber: which patches do you want tagged?
arokux has quit [Remote host closed the connection]
<alpha_one_x86>
my problem is when I use AHCI + fs (samebug with btrfs as ext4)
arokux has joined #linux-sunxi
fredy has quit [Ping timeout: 256 seconds]
<wens>
afaerber: updated
fredy has joined #linux-sunxi
tomcheng76 has quit [Read error: Connection reset by peer]
tomcheng76 has joined #linux-sunxi
jemk has joined #linux-sunxi
HeHoPMaJIeH has quit [Ping timeout: 264 seconds]
ssvb has quit [Quit: Leaving]
gianMOD has quit [Ping timeout: 258 seconds]
gianMOD_ has joined #linux-sunxi
leviathanch has joined #linux-sunxi
trollinger has joined #linux-sunxi
tomboy64 has quit [Remote host closed the connection]
fredy has quit [Ping timeout: 260 seconds]
tomboy64 has joined #linux-sunxi
Netlynx has joined #linux-sunxi
Netlynx has joined #linux-sunxi
fredy has joined #linux-sunxi
konradoo77 has joined #linux-sunxi
<rellla>
jemk: i thought you didn't want to invest time in libvdpau anymore :p ?
deasy has joined #linux-sunxi
afaerber has quit [Quit: Verlassend]
<jemk>
rellla: i mainly didn't want to invest time in output improvements with disp
konradoo77 has quit [Ping timeout: 245 seconds]
<rellla>
jemk: you can't explain, why vaapi-gl uses this many handle locks and unlocks, can you?
afaerber has joined #linux-sunxi
Gerwin_J has quit [Quit: Gerwin_J]
konradoo77 has joined #linux-sunxi
<jemk>
rellla: i guess because they want to be absolutely sure nothing can go wrong, but vdpau documentation says the player has to take care itself not to use the same handle concurrently
<jemk>
"Note, however, that this simply guarantees that internal VDPAU state will not be corrupted by thread usage, and that crashes and deadlocks will not occur. Completely arbitrary thread usage may not generate the results that an application desires. In particular, care must be taken when multiple threads are performing operations on the same VDPAU objects."
<rellla>
so we can miss every handle lock at top and handle unlock at the end of the vdp_functions
<rellla>
only locking the fifo queue is necessary
domidumont has quit [Read error: Connection reset by peer]
<jemk>
there might be places where locking handles is necessary, but i didn't come across any
<rellla>
jemk: when using one queue and a extra thread for presentation it might?
fakesn has quit [Ping timeout: 256 seconds]
<jemk>
i don't think so, the presentation thread should only read and nobody else should touch the VideoSurface while its in the queue (block_until_idle)
<rellla>
read and pop the first element out of the queue.
<jemk>
the queue needs to be locked of course
<rellla>
that's what i tried to implement. i did some exercises this afternoon. and just want to be sure, that it is not ONLY nonsense :p
protoCall7 has joined #linux-sunxi
<jemk>
btw, the handle_acquire from fwolter is nonsense, it locks the same mutex for all handles
quitte has joined #linux-sunxi
trollinger has quit [Quit: No Ping reply in 180 seconds.]
<rellla>
jemk: hey, that's why i recently bought Michael Kerrisk's TLPI :p
protoCall7 has quit [Quit: protoCall7]
trollinger has joined #linux-sunxi
fakesn has joined #linux-sunxi
protoCall7 has joined #linux-sunxi
protoCall7 has quit [Quit: protoCall7]
gianMOD_ has quit []
leviathanch has quit [Ping timeout: 258 seconds]
bonbons has joined #linux-sunxi
konradoo77 has quit [Ping timeout: 245 seconds]
Skaag has quit [Quit: Leaving.]
Zboonet has quit [Quit: Leaving]
fakesn has quit [Ping timeout: 240 seconds]
Zboonet has joined #linux-sunxi
deasy has quit [Quit: Nom d'un quark, c'est Edmonton !]
<diego71_>
Netconsole was designed to be as instantaneous as possible, to
<diego71_>
enable the logging of even the most critical kernel bugs
<diego71_>
So it should have a chance to send a backtrack before hang.
jemk has quit [Quit: leaving]
<alpha_one_x86>
But the syntom seam indicate will not work
<rah>
libv: the particular devices which a GPU is used in is an important factor in people's direction of effort
<rah>
libv: and the SGX is used in many devices which are likely to attract skills
<rah>
libv: such as, much as it pains me to admit, the GTA04
<rah>
"why have you not used your time to help any of the other ARM GPU drivers? Why did you suddenly decide to work on _this_ project, and why did you before not care about the other projects?"