<meetpatty>
Has anyone had success in getting the cursor to show in Virtual Box? After adding logging to hildon desktop I can see that with a pointing device of either PS/2 Mouse or USB Tablet that if Mouse Integration is disabled then the device is reported as mode Relative and wm_set_cursor_visibility is called with visible = 1 yet no cursor is displayed. This is using the latest chimaera image,
<meetpatty>
20250601-chimaera.
<meetpatty>
Further to this, if I change the call to wm_set_cursor_visibility to always show the cursor is visible if Mouse Integration is enabled but not visible if Mouse Integration is disabled.
meetpatty has quit [Ping timeout: 252 seconds]
meetpatty has joined #maemo-leste
System_Error has quit [Ping timeout: 264 seconds]
System_Error has joined #maemo-leste
joerg has quit [Ping timeout: 244 seconds]
joerg has joined #maemo-leste
vectis_ has joined #maemo-leste
_fab has joined #maemo-leste
System_Error has quit [Remote host closed the connection]
<freemangordon>
meetpatty: you should disable mouse integration, however, no idea why cursor is not visible. I am using virtualbox and have no issues
LeePen has quit [Remote host closed the connection]
LeePen has joined #maemo-leste
meetpatty has joined #maemo-leste
xinomilo has joined #maemo-leste
RedM has quit [Ping timeout: 244 seconds]
RedW has joined #maemo-leste
leste has joined #maemo-leste
<dsc_>
freemangordon: alright
<dsc_>
thx
Anasko has quit [Quit: Quit]
Anasko has joined #maemo-leste
pere has quit [Ping timeout: 260 seconds]
meetpatty has quit [Ping timeout: 276 seconds]
System_Error has joined #maemo-leste
pere has joined #maemo-leste
mkfx has joined #maemo-leste
Blikje has quit [Remote host closed the connection]
xinomilo has quit [Quit: Leaving]
arno11 has joined #maemo-leste
kiva has joined #maemo-leste
<kiva>
dsc_: if remember right hw-rotation/zoom is outside of Mali GPU, it is AllWinner feature, read some documents about it, but I did not find enough low level info about it that can make driver. There was also mentioned two layer feature with alpha channel, would be usefull if make game with sprites in old school style. Or one layer has Hildon up-bar
<kiva>
and another has window it self and zoom/rotate it.
<dsc_>
i dont really understand the concept of hw rotate tbh ;p
<dsc_>
we have a vertex and fragment shader, you can rotate them both with uniforms, this is standard shader stuff I believe... in this case the vertex shader should be rotated to update viewmodel matrix (?)
<dsc_>
but embedded GPU drivers is not my area of expertise
<kiva>
and I only know low level of 80s Yamaha GPUs (yes they have 2d acceleration)...but because AllWinner has rotate is after Mali GPU part..I think viewport in landscape should be 1440x720 and then AllWinner hw-rotate should rotate it to 720x1440 screen.
<kiva>
I think that hw-rotate just read memory in different order to get pixels in right places to screen. Or I am designer I would do it that way.
<kiva>
And if it is made in chip, it does not have any speed penalty.
<sicelo>
although in some (weak?) SoCs, there is still some penalty :p
<sicelo>
N900 & vrfb :'(
<kiva>
DSP of N900 could do rotate on fly ;)
_fab has quit [Quit: _fab]
<kiva>
But back to PP...I think modern Linux do not have _good_ drivers for Mali400 or AllWinner. I think the are more like generic just work drivers.
ceene has quit [Ping timeout: 252 seconds]
arno11 has quit [Quit: arno11]
<kiva>
But more bad open source driver is in latest Debian for PowerVR SGX535..you cannot stream video in any size..ok I did not test under 120p.
<Wizzup>
uhh
<Wizzup>
the lima driver is pretty darn good
<Wizzup>
maybe you're talking about the 2D unit in some of the allwinner socs
<sicelo>
and what do you say is in Debian?
<Wizzup>
lima is in mesa
<kiva>
wizzup and sicelo: fine..maybe I have some bias :) And it not just in Debian..and original reason is lack of documentation from hardware maker.
<kiva>
but linux fastly drops hw-acceleration if it does not support latest apis.
<sicelo>
i was asking what powervr driver you say is in debian?
<kiva>
these days none...old days they had some closed source with that extra DVD.
_fab has joined #maemo-leste
<kiva>
reason was that Intel GMA500=PowerVR SGX 535 (Inside Nokia RX-72 and RX-75), GMA3600=PowerVR SGX545, Intel Atom Z24xx=CPU+PowerVR SGX540
_fab has quit [Client Quit]
<kiva>
btw..Meego Netbook edition had video acceleration driver for full screen video playing for those Nokias.
<kiva>
sorry about flooding about this...back to Leste...and that have good PowerVR drivers!
mkfx has left #maemo-leste [Error from remote client]
<kiva>
at least with Droid4.
<uvos__>
powervr may mostly work now
<uvos__>
but the propriatary powervr drivers we use have been the source of countless headaches
<uvos__>
lima is def mutch less hassle
<Wizzup>
yup
<kiva>
And lima driver will be good enough for me if Allwinner rotates its view to landscape without lot penalty. I have already streamed tv-news at 360p with FireFox. So thank you for good job.
<dsc_>
kiva: I watched a 15min youtube video @ 480p in Jib on PP and lost 8% battery
<kiva>
dsc_: If Jib have .MOV (Quicktime Video) support in it, with embedded streaming feature that .MOV can have. Then I would tell web-page that will work even with N900.
<kiva>
even with 2.5G connection!
<kiva>
eh...2.5G might need 240x136px quality :)
apac has joined #maemo-leste
apac has quit [Read error: Connection reset by peer]
<kiva>
ah..good old days with Netscape and Quicktime videos ;)
<kiva>
with dial-up modems.
apac has joined #maemo-leste
<kiva>
btw..None of Chromium based brownsers does not support .MOV, so if Jib do that someday then it would be better than any Chromium based browser :)
sfa has quit [Ping timeout: 248 seconds]
sfa has joined #maemo-leste
panzeroceania has quit [Ping timeout: 268 seconds]
LeePen has quit [Remote host closed the connection]
pabs3 has quit [Ping timeout: 272 seconds]
panzeroceania has joined #maemo-leste
hexnewbie has quit [Ping timeout: 272 seconds]
LeePen has joined #maemo-leste
hexnewbie has joined #maemo-leste
apac has quit [Read error: Connection reset by peer]
pabs3 has joined #maemo-leste
kiva has quit [Quit: Client closed]
<freemangordon>
rotation is not GPU related
<freemangordon>
on d4 for example, it is the TILER module in the SoC
<freemangordon>
n900 (omap3) uses VRFB
<freemangordon>
perhaps allwinner has similar HW module
<freemangordon>
but there issue is with damn modesetting driver
<freemangordon>
which simply do SW rotation
<freemangordon>
and nobody seem to care to fix it, given the WL hype
xs has quit [Quit: Cheers!]
xs has joined #maemo-leste
Smatkovi has joined #maemo-leste
<inky>
> i was asking what powervr driver you say is in debian?
<inky>
i have read that there's an effort to create open powervr drivers, i think pmos podcast also mentioned it.
<freemangordon>
hmm, EGL_KHR_image_pixmap is supported
<freemangordon>
what the?
<freemangordon>
you should not see "fallback" mexxages on PP
<freemangordon>
*messages
<dsc_>
are you refering to hildon-desktop?
<freemangordon>
yes
<dsc_>
ok, I think this also depends on the state of my enviromnent (variables)
<freemangordon>
no, why is that?
<dsc_>
I dont think I had these errors after, ehm, doing things
<freemangordon>
hmm... what things? maybe SW rendering?
<dsc_>
was doing hildon-desktop debugging this weekend, and I pasted this error but it could have been related to my debug/dev setup
<freemangordon>
ah, ok
<freemangordon>
have to run now, but will ping you when I have some spare time to confirm there is really no issue with TFP on PP, ok?
<freemangordon>
will check on my allwinner as well
<dsc_>
sure
<dsc_>
freemangordon: are you interested in checking for 'texture_from_pixmap' ?
<dsc_>
because I can check myself, I can debug hildon/clutter/cogl these days from my IDE
<dsc_>
GLX_EXT_texture_from_pixmap ?
<freemangordon>
GLX?
<freemangordon>
no EGL_KHR_image_pixmap
<freemangordon>
:)
<freemangordon>
but, you already confirmed it is there
<freemangordon>
so, because you anyway provided logs from a session where it is not used, I want to make sure it was because you 'broke' it somehow, and not because clutter does not work properly
<dsc_>
im awaiting it now, but still wondering if e.g. osso_abook_aggregator_get_default can be faster
<dsc_>
idk if 5000ms is exclusive to arno11 or everyone has such boot times on n900 for conversations
<dsc_>
such boot times = >30sec
<freemangordon>
sorry, have to run, ttyl
<dsc_>
ok
nmdw has joined #maemo-leste
<kiva>
inky: If I remember right, those PowerVR drivers for Debian was PowerVR SGX535-545 on closed source drivers DVD about ten years ago. And they was under brand Intel.
<kiva>
..or was those on Ubuntu DVDs..second thought I am not sure.
<kiva>
and binarys was for x86.
<kiva>
x86=i386 (and real minimum was i686)
pere has joined #maemo-leste
nmdw has quit [Quit: Leaving]
kiva has quit [Quit: Client closed]
<Wizzup>
I am planning to migrate maedevu today or tomorrow, fyi
<sicelo>
migrate to?
<Wizzup>
a server that I don't have to pay 80 euro a month for :)
<Wizzup>
but with faster internet and storage
<sicelo>
nice
pinedsc has joined #maemo-leste
arno11 has joined #maemo-leste
<arno11>
dsc_: @boot time, it can't be exclusive to me. it is probably even worse on most n900's...
<dsc_>
arno11: can I push another version to -devel and you report back the .timings file ?
<dsc_>
i included more logging statements
<arno11>
sure
<dsc_>
arno11: its in -devel
<arno11>
can't do it now, but in 1h, it will be ok
<dsc_>
thx
leste has quit [Remote host closed the connection]
leste has joined #maemo-leste
pinedsc has left #maemo-leste [#maemo-leste]
pinedsc has joined #maemo-leste
_fab has joined #maemo-leste
pinedsc has quit [Read error: Connection reset by peer]
pinedsc has joined #maemo-leste
<arno11>
dsc_: i suppose .60 is OK for the test ?
<dsc_>
arno11: yes
ac_laptop has joined #maemo-leste
<arno11>
ok
arno11 has quit [Quit: arno11]
pinedsc has left #maemo-leste [#maemo-leste]
pinedsc has joined #maemo-leste
apac has joined #maemo-leste
pinedsc has left #maemo-leste [#maemo-leste]
pinedsc has joined #maemo-leste
pinedsc has left #maemo-leste [#maemo-leste]
pinedsc has joined #maemo-leste
arno11 has joined #maemo-leste
pinedsc has quit [Read error: Connection reset by peer]
<arno11>
maybe related to overall qt slowness on n900 but afaik, conversations has been always very slow on boot
_fab has quit [Quit: _fab]
<Wizzup>
parallel boot is a very active time in the device
<Wizzup>
I don't think that's a good way to measure performance
<Wizzup>
the I/O load is high, as well as the cpu load
ac_laptop has quit [Read error: Connection reset by peer]
<arno11>
high or rather very high :P
ac_laptop has joined #maemo-leste
<arno11>
anyway, i think it is not a major issue for n900 (conversations on boot)
ac_laptop has quit [Ping timeout: 252 seconds]
<dsc_>
ok, thanks for testing
<arno11>
extreme qt5 slowness (by default) in daedalus is far more problematic imo
<arno11>
dsc_: np
<arno11>
@qt5, not a blocker but a real pita for new users
<dsc_>
arno11: I have a n900 somewhere...
<arno11>
oh really ?
<dsc_>
courtesy of wizzup ;p
<dsc_>
ill install leste and strace a qt5 app
<arno11>
:D
<arno11>
really cool
<arno11>
ty
<dsc_>
oh, it doesnt charge. oh well ;p
<arno11>
?
<dsc_>
always have trouble with charging d4/n900, maybe the adapter, maybe the cables
<dsc_>
tried multiple
<arno11>
ok
<arno11>
could be faulty usb connector too
<sicelo>
same issue on D4, so cables
ac_laptop has joined #maemo-leste
<arno11>
ok
kiva has joined #maemo-leste
<kiva>
dsc_: if N900 does not charge, try this with original Nokia charger: put cable only 1-2 for seconds, then put it again for longer time. Somehow it works sometimes and I dont know reason for that.