<hramrach_>
building kernel is one thing but building android requires tons of disk space
<mnemoc>
his monster machine can deal with it
<hramrach_>
I wonder if this machine has more disk trays
* mnemoc
has only a 60GB drive
<mnemoc>
(and the external esata/usb raid)
<hglm>
ssvb: slightly off-topic: I ported sunxifb to RPi. It is a little faster than the standard fbdev they use.
<ssvb>
hglm: :)
<mnemoc>
this seriously smells like a fbdev2 driver is needed
<ssvb>
hglm: are you using RPI DMA for blits and fills?
<hglm>
ssvb: No, I haven't touched "DMA-hell" on the RPi. They seem to require busy-waiting, no interrupt.
<ssvb>
hglm: well, this does not seem to be entirely true
<ssvb>
hglm: they have a bit messed up dmaer module, but this should be just rewritten properly
<hglm>
ssvb: I guess so, I looked at dmaer, it could be improved.
<ssvb>
hglm: where does the improvement come from?
<ssvb>
hglm: have you implemented a VFP based overlapped copy function?
<hglm>
ssvb: I ported the cpu backend to ARM simply using memcpy, it helps a lot.
<hglm>
ssvb: memcpy seems to be optimized on the RPi.
<ssvb>
it's not very good actually (unless they improved it since the last time I had a look at it), optimal memcpy needs aligned prefetches
<hglm>
ssvb: OK, I guess the two-pass approach in the CPU back-end works reasonably well with standard memcpy (because some of the arguments are aligned).
_BJfreeman has joined #linux-sunxi
_BJfreeman is now known as BJfreeman
<hglm>
ssvb: I am seeing 70-100% speedup for unaligned screen-to-screen blits using the sunxifb CPU back-end on the RPi compared to standard fbdev, 200% for rightwards overlapped blits (dragging windows to the right).
<ssvb>
hglm: fbdev with or without shadowfb enabled?
<hglm>
ssvb: without shadowfb, with shadowfb the screen is updated only sporadically
<ssvb>
hglm: yeah, shadowfb 'cheating' a lot, but that's how fbdev ddx driver is normally designed to work
<ssvb>
hglm: also RPi is interesting because external L2 cache enabled for the framebuffer, that's making it kinda unique
<hglm>
ssvb: L2 cache enabled? That explains some things. Is that a configuration bug?
<ssvb>
hglm: L2 cache is a part of the GPU, and ARM CPU can observe it as physical memory having L2 cache
<hglm>
ssvb: I understand, the RPi seriously lacks a good real CPU cache (that's why it's slow I think).
eebrah|away is now known as eebrah
<ssvb>
hglm: but in any case, if we want to also have a good RPi support, DMA needs to be taken into use
<ssvb>
hglm: this means a bit of kernel hacking and proper use of IRQ which can signal DMA completion
<hglm>
ssvb: Yeah IRQ would be nice. Not sure if they are using DMA IRQ's in the RPi kernel though (you would assume so though).
ganbold_ has joined #linux-sunxi
<ssvb>
hglm: also about your G2D related patches for sunxifb, could you please pick a few initial patches that you think are ready to be pushed and make a separate branch for them?
<hglm>
ssvb: OK, I'll do that.
Kadosch has joined #linux-sunxi
<ssvb>
hglm: ok, thanks
<hglm>
ssvb: BTW I can upload my RPi port to github, but it does change a lot of identifiers and some configurations, it is not in a state mergeable with sunxifb.
<ssvb>
hglm: yes, this is to be expected, writing clean code needs a bit more work than a quick proof of concept prototype
<ssvb>
hglm: but clean and maintainable support for multiple platforms would be surely a nice addition
<mnemoc>
xorg-video-ssvbfb :p
xlab has quit [Remote host closed the connection]
<hglm>
ssvb: Are you're OK with me uploading the experiment public Rpi repo? Since it's your code mostly.
<ssvb>
hglm: that's a free software, derivative works are encouraged (unless you try to remove the original copyright notices and claim that it's written by you) :)
<ssvb>
hglm: you are also free to add your own copyright notices when adding non-trivial changes
<hglm>
ssvb: Thanks, and I did add my own copyright notices.
Yaku321 has quit [Read error: Operation timed out]
Yaku321 has joined #linux-sunxi
rellla has joined #linux-sunxi
shineworld has quit [Quit: Leaving]
<hno>
<hno> analog.
<hno>
<hno> i2s is a digital audio interface.
<wingrime>
~ssvb: strage. after boot 3.0 suspend not hung with xserver
<wingrime>
ssvb: I should try new 3.4
<wingrime>
ssvb: It will be strange if bug presents only on 3.4
<hno>
wingrime, there is many bugs which is in 3.4 only, but and many that is 3.0 only. But most Allwinner bugs are common.
<wingrime>
huo: simply say , I have some hung on suspend when Xorg running
<wingrime>
huo: using standard framebuffer driver
* hno
is nog guilty for the Reuuimlla ploy
<hno>
s/nog/not/
\\Mr_C\\ has joined #linux-sunxi
<wingrime>
what a Reuuimlla ?
<mnemoc>
allwinner backward
<mnemoc>
their "smart" idea to reduce their legal responsibility
Kadosch has quit [Read error: Connection reset by peer]
piyushverma has joined #linux-sunxi
<mnemoc>
there is s/allwinner/newbie/ too
Kadosch has joined #linux-sunxi
<mnemoc>
but they stupidly transfer the ownership to the first random troll who bought the domain
<wingrime>
mnemoc: register new company )))
<wingrime>
huo: can we have some boot select in uboot?
<Turl>
olimex asking question for an LCD :)
<wingrime>
mnemoc: how much possible that a10s,a13 are same chip with different case(package)?
<mnemoc>
wingrime: very unlikely. a13 is rgb/lcd-only and a10s is hdmi/vga-only
<wingrime>
mnemoc: so maybe we need move soc detection to uboot and simply send to kerenl machine number ?
<wingrime>
and we defenetly get right MACHINE_START macro
<mnemoc>
machine_start only tells the sunNi
<mnemoc>
but I believe soc-detect would help to have a single-sunxi u-boot bin too
<wingrime>
we can make machine number for every soc
<mnemoc>
machine numbers are obsolete
<wingrime>
realy?
<mnemoc>
yes, "DT rules" (tm)
<mnemoc>
no point in registering new machine numbers
<wingrime>
DT are build-time only or is recived from uboot?
<mnemoc>
passed by uboot
* tzafrir
tries following the XMBC page from the wiki
_BJfreeman has quit [Read error: Connection reset by peer]
<oliv3r>
hno: you haven't commented/applied the wemac->emac patch yet; any plans on doing anything with that?
_BJfreeman has joined #linux-sunxi
<oliv3r>
mnemoc: how 'well' does your sramc detection branc work so far? Are we at 99% detection accuracy? for sun[4567]i? (not thinking a13 vs a12 vs a10s
<oliv3r>
e.g. pure sramc based
jemk has quit [Read error: Operation timed out]
ganbold_ has quit [Ping timeout: 252 seconds]
ganbold_ has joined #linux-sunxi
mdp has quit [Excess Flood]
Kadosch has quit [Read error: Connection reset by peer]
<Turl>
also, why do they still ship with that ugly red thingy? :)
<mnemoc>
it's actually easier to control than the touchpad :p
<Turl>
I used it once
<Turl>
it was a pain in the neck
<mnemoc>
specially on the move, or with the hands wet :p
<mnemoc>
there are use cases for both, so they have both
<Turl>
I wouldn't use a computer with my hands wet :p
<mnemoc>
you have a desktop computer with built-in ups
<Turl>
mnemoc: the other issue is screen size
<Turl>
the terminal itself is like 5-6 inches
<mnemoc>
that's font-size relative
<Turl>
so I cannot, say, read a site and work on a terminal simultaneously :p
<Turl>
mnemoc: well, anything smaller than that would mean I either shrunk from 80x24 or I'm using an unreadable font
<Turl>
(or painreadable)
<mnemoc>
Using `terminator` I have two full screen tiled terminal screen. 225x60 total space
<mnemoc>
on each screen. perfect for 4 large terminals on each virtual desktop
<mnemoc>
6 some times
<Turl>
I prefer to tab them on gnome terminal :p
<mnemoc>
ctrl-tab to switch across virtual desktops is faster :p
ganbold_ has quit [Remote host closed the connection]
<mnemoc>
err, <win>-tab
<Turl>
mnemoc: ctrl+pgup/down
<mnemoc>
with tabs you can't see many terminals at the same time
<Turl>
I click most of the times though
<Turl>
mnemoc: I untab them when I need so
<mnemoc>
one virtual-desktop per topic, all terminals right there
<mnemoc>
tiled, full screen. but want smaller pixels :\
<Turl>
mnemoc: chromebook pixel ;)
<mnemoc>
waiting for the exynos octa chromebook pixel
hramrach_ has quit [Remote host closed the connection]
hramrach_ has joined #linux-sunxi
<Turl>
mnemoc: N10 with ubuntu and kbrd? :p
<oliv3r>
wingrime: ping
ganbold_ has joined #linux-sunxi
<oliv3r>
mnemoc: what's the policy on building stuff on linux-sunxi?
<mnemoc>
explain
<oliv3r>
cross compiling stuff
<oliv3r>
kernels, rootfs
<oliv3r>
e.g. used cpu time; used diskspace
<Turl>
oliv3r: what do you mean by policy?
<Turl>
oliv3r: ah, building on the server?
<mnemoc>
the linux-sunxi virtual server is not for building
<oliv3r>
how do you build the dailies?
<mnemoc>
on another virtual server
<oliv3r>
oh ok
hglm has quit [Quit: leaving]
<mnemoc>
different partitions, different ip address, different limits, different packages
<oliv3r>
fair nuff
<oliv3r>
mnemoc: i've submitted the backport patch; just need hramrach's patchset so i can adapt my patchset with his (new) files
<mnemoc>
so yours goes after hramrach's?
<oliv3r>
yeah
<mnemoc>
ok
<oliv3r>
i'll rebase them ontop of his work
<mnemoc>
same for 3.4?
<oliv3r>
if he submits it in reasonable time
<oliv3r>
might aswell yeah
<oliv3r>
3.0 doesn't touch the defconfigs though
<oliv3r>
so we can do that one right now, and do a follow up after hramrach's
user_2 has joined #linux-sunxi
<Turl>
damn it windows
<Turl>
making my doors shut themselves repeatedly with the wind is not funny :)
<user_2>
hi all. im getting crazy. im running linaro and debian on cubieboard, both od distors hase some positive side (linaro run gnome-mplayer with audio and video on hdmi, debian dont) and debian recognize ORG usb, wireless, dongle, but linaro dont. so, how to merge audio/video support inside debian?
<oliv3r>
lol, i was thinking a whole different level of windows
<oliv3r>
user_2: first, you can always try the latest 3.4 kernel that gives you best support
<oliv3r>
user_2: also, give hansg's Fedora 18 a try
<user_2>
debina tun 3.4.29
<user_2>
debian run 3.4.29
<user_2>
fedora? i will do. but im a debinist so i wish to use debian... aniway, now i try fedora, i already downloaded from dl.cubiebaord.org, let me dd on uSD.
<user_2>
oliv3r, do you use fedora? if yes, do you feel good on it? (i mean, youtobe works, audio/video works, etc)
<oliv3r>
i haven't used it that much; i've only tested things with it; but it does work quite well from what I saw
<rellla>
why you want to download these? you don not need it anymore!?
mdfe has quit [Remote host closed the connection]
<user_2>
hello. Im running on cubieboard a debian i got from http://linux-sunxi.org/. It is nece: OTC usb works, wireless dongle works, youtube works with video. THe proble mare te follow: youtibe video has not audio, and gnome-player does not play (no video, no audio at all)
mdfe has joined #linux-sunxi
<user_2>
instead linaro play audio and video but no usb OTC wotks, no wireless dongle works. Ho to add video capabilities to debian?
<Turl>
user_2: check sound settings and make sure hdmi sound output is selected
<user_2>
Turl, i ave to reboot, wait.
<user_2>
Turl, ok, im in.
<user_2>
where is the control? i am using lxde
<user_2>
seems i mess the audio panel
<user_2>
what sound setting program i should isntall?
<hno>
oliv3r, I have been working 15h/day the last week. Brain a bit fried. Which patch are we talking about?
<oliv3r>
hno: oh, sorry, didn't wanna push you :)
<oliv3r>
[PATCH 0/2] Standardize Allwinner EMAC name
<oliv3r>
05/25/2013
<oliv3r>
user_2: did the fedora image not work?
<user_2>
oliv3r, no it dont boot. mow im re-.download from link you gave me.
<user_2>
65% done ,12min to go
<oliv3r>
i think it wasn't a download issue, but you have to run setup for it to actually work
<user_2>
the hell! i forgot!
<hno>
oliv3r, is that an u-boot patch or kernel?
<user_2>
it is 3 days m fighting on cb, now im going to loose loose my mental capabilites
<user_2>
look, i have some uSD her, now i did confusion and.... well, i wil lstart again: comlete donwload, unzip, dd, setup... i need tome. i home you still here along the evening :-)
<oliv3r>
hno: sunxi-current
<hno>
ah, it's an u-boot patch. Thought it was a kernel patch so didn't pay attention.
<hramrach_>
oliv3r: still building the 3.4 configs :s
<hramrach_>
found a few more build errors
<hramrach_>
ugh, the axp152 has *lots* of warnings
<hno>
oliv3r, fails to apply for me for some reason.
<oliv3r>
hno: shall I rebase
<oliv3r>
i did it ontop of the ... 'axp research' we did a few days ago
<oliv3r>
i'll pull and see
<hno>
Please rebase to sunxi-current.
<mnemoc>
do the AXPs came with their i2c id hardwired?
<mnemoc>
come*
<hno>
mnemoc, yes I think so.
<oliv3r>
hno: it was based on a reasonable new sunxi-current; i'll pull to make sure
<hno>
oliv3r, git am is a bit picky...
<oliv3r>
wasn't the AXP version id thing what causes some issues? I think my cubieboard had a newer version
<mnemoc>
hno: based on the patch from hansg that would mean we can "detect" which PMU is used by the id
<mnemoc>
at least in a couple of known cases
BJfreeman has quit [Ping timeout: 276 seconds]
<oliv3r>
resent
<hramrach_>
I would really like a decent notebook case with the innards ripped out and replaced with something based on moderately powerful SoC like a20/rk3066/Exynos
<oliv3r>
hno: did you do change anything with regards to the different axp ID on my cubieboard?
<hno>
oliv3r, yes, i disabled the ID check as that field is actually unknown, not specified at all in the manual
<hno>
only seen vaguely in some headers, and looking at boot1 there seems to be some magic involved, not just a plain compare.
<rellla>
for cross compiling every header is taken from the rootfs. $HOME is part of the path. you can change it to whatever you want. it's adapted to my environment...
<rellla>
tzafrir: $HOME is maybe a misleading name..
<tzafrir>
rellla, confusing, indeed. Read lines 10 and 11 in that file
<rellla>
tzafrir: it has to be cleaned up a little bit ;) - esp. the comments. see it as a d&q hack of me to easily change the base dir without taking care of the rest
<tzafrir>
there's no clear separation there between 'make' and 'make install'
Kadosch has quit [Read error: Connection reset by peer]
Kadosch has joined #linux-sunxi
<tzafrir>
I want a target to run as a user and a separate target to run as root that will install but not download or build
<tzafrir>
and will be simpler to inspect
<rellla>
make -C xbmc install is missing
<rellla>
as i said. it's very improveable ...
_BJfreeman has joined #linux-sunxi
BJfreeman has quit [Ping timeout: 245 seconds]
_BJfreeman has quit [Ping timeout: 245 seconds]
_BJfreeman has joined #linux-sunxi
rellla has quit [Remote host closed the connection]
mdfe has quit [Remote host closed the connection]
<hno>
rz2k, did it work?
<rz2k>
hno: will grab a10s olinuxino nanda tomorrow and then launch it, as far as I see, there are no a10s targets available
Kadosch has quit [Remote host closed the connection]
rz2k has quit []
Undertasker has quit [Read error: Connection reset by peer]
Undertasker has joined #linux-sunxi
<hno>
Oh, there is a whopping 1 funders so far.
<libv>
people seem to be capable of learning
<oliv3r>
lol yeah
<oliv3r>
hno: pushed a little patch, let me know what you think. i would think it would look nice in spl too
<hno>
pushed where?
techn_ has joined #linux-sunxi
<hno>
oliv3r?
nove has quit [Quit: nove]
stekern_ has joined #linux-sunxi
mripard_ has joined #linux-sunxi
Tartarus_ has quit [Quit: Ex-Chat]
Dreadlish has quit [*.net *.split]
stekern has quit [*.net *.split]
techn__ has quit [*.net *.split]
mripard has quit [*.net *.split]
Dreadlish has joined #linux-sunxi
<hno>
oliv3r, ok found it on the ml. Yes it's probably good, but at the same time a bit useless.. the build is always for a specific arch. We should be able to derive the same info from there.
<hno>
Good night.
<lkcl>
night
hramrach_ has joined #linux-sunxi
user_2 has quit [Ping timeout: 260 seconds]
<lkcl>
whew. got a20 doing fel-boot. uploaded u-boot (success), loaded kernel from sd/mmc (success), kernel execute BARF but that's just like with android kernel anyway.
Undertasker has quit [Quit: Leaving.]
sanka has quit [Remote host closed the connection]