Active Topics
-
Firefox with Leste (10)
to Maemo 7 / Leste by teroyk - 3 days, 6 hrs ago - more...
 |
|
2014-03-20
, 00:22
|
|
Posts: 10 |
Thanked: 14 times |
Joined on Mar 2014
|
#22
|
This is very cool - looking forward to this hack! 
I'll be glad to test something if you will need more people.
I'll be glad to test something if you will need more people.
 |
|
2014-03-20
, 02:38
|
|
Posts: 634 |
Thanked: 3,266 times |
Joined on May 2010
@ Colombia
|
#23
|
Originally Posted by marmistrz
That's a possibility but I wouldn't worry about that right now. Just get something working first.
Can't our Linux kernel take care of whole power management and our Dalvik app just let be controlled?
Originally Posted by lexik
Spot on.
Correct me if I'm wrong, but yes, as we are not emulating anything. Drivers are still needed inside (and outside) the LXC, because chrooted system has no knowledge about it's HW.
Originally Posted by Android_808
Correct.
AIUI, LXC containers share the same kernel still - see Oracle documentation on LXC
Originally Posted by Android_808
That was always going to be the situation, even with a normal chroot. It's not possible to run multiple kernels without a hypervisor.
In that case your now looking at patching host kernel to enable Android, which I dare say is going to break a few things a long the way.
Originally Posted by Android_808
libhybris shouldn't be needed. The kernel and drivers will be running under Maemo with glibc and will continue to talk to the hardware as normal.
If that's the case I'd start looking at a Sailfish approach by running libhybris, if only for the possibility of detaching MCE and co from N900 specifics to allow porting to other devices.
Originally Posted by Android_808
The big advantage of OS-level virtualisation is that there is little or no overhead. It will only use whatever is normally used by Android.
Problem is how much RAM and performance loss your going to see.
Originally Posted by taixzo
Drivers (kernel modules) are an extension to the kernel hence they will be provided (shared) by the host OS. If there are any userspace drivers they would have to be loaded inside the container.
If the container has the same kernel, does that mean it has to load two sets of drivers for everything? Or are the android drivers only loaded inside the LXC container?
Originally Posted by Alecsandru
I'm not familiar with Android/Jolla patches and without knowing the details I can only speculate but it's possible that they were userspace drivers.
and why the drivers (the camera one and others ) were patched to work in android(jolla) if they share the same kernel , i don°t think they share the same drivers
Originally Posted by Alecsandru
That demo was a long time ago so it would have been based on an older version of Android running a 2.6 kernel, most likely patched.
by the way why alien dalvik won°t run on kernels prior to 3.x.x (jolla , leaving qualcomm s**t aside) , from what I remember it was demoed on n900 , that phone was with a updated kernel? or our kernel was patched?
__________________
DebiaN900 - Native Debian on the N900. Deprecated in favour of Maemo Leste.
Maemo Leste for N950 and N9 (currently broken).
Devuan for N950 and N9.
Mobile devices with mainline Linux support - Help needed with documentation.
"Those who do not understand Unix are condemned to reinvent it, poorly." - Henry Spencer
Maemo Leste for N950 and N9 (currently broken).
Devuan for N950 and N9.
Mobile devices with mainline Linux support - Help needed with documentation.
"Those who do not understand Unix are condemned to reinvent it, poorly." - Henry Spencer
| The Following 3 Users Say Thank You to wicket For This Useful Post: | ||
|
|
2014-03-20
, 05:59
|
|
Posts: 1,225 |
Thanked: 1,905 times |
Joined on Feb 2011
@ Quezon City, Philippines
|
#24
|
Originally Posted by reinob
As on Harmattan, the systems runs okay enough even with Android drivers enabled and loaded - no app uses them, so they're left well alone
Quote from the linked page: "You’ll need a kernel with the Android patches (binder, ashmem, logger etc.)"
[...]
Neither Fremantle will run on Android kernel nor Android will run on Fremantle kernel.
I think that sorta sums it up good enough.
And the same situation in reverse - on my Android device, I can stop Zygote, take control of the framebuffer, and start Xorg with the GNU Userland from a terminal.
Originally Posted by marmistrz
Assuming that Android runs inside Maemo via ParallelDroid, and static PM support isn't available, Maemo's (IMO superior) dynamic power control will reign supreme.
Supposed there is static pm missing, what would happen while trying to run an Android app? //edit: Can't our Linux kernel take care of whole power management and our Dalvik app just let be controlled?
__________________
N9 PR 1.3 Open Mode + kernel-plus for Harmattan
@kenweknot, working on Glacier for Nemo.
N9 PR 1.3 Open Mode + kernel-plus for Harmattan
@kenweknot, working on Glacier for Nemo.
 |
|
2014-03-20
, 06:26
|
|
Posts: 439 |
Thanked: 282 times |
Joined on Oct 2012
|
#25
|
so , let"s start with kernel , what patches and how to compile them without breaking maemo?
__________________
N900 Beast featuring :
8GB OPTFS@ext4(*performance mount options)
overclock: cpu[125 - 1,150(*VDD1)+Dsp overclock
CSSU-T latest version
N900 Beast featuring :8GB OPTFS@ext4(*performance mount options)
overclock: cpu[125 - 1,150(*VDD1)+Dsp overclock
CSSU-T latest version
|
|
2014-03-20
, 06:41
|
|
Posts: 959 |
Thanked: 3,427 times |
Joined on Apr 2012
|
#26
|
Originally Posted by Alecsandru
Can't Maemo run on the Nitdroid kernel? If so, that would be a good place to look.
so , let"s start with kernel , what patches and how to compile them without breaking maemo?
__________________
Saera: A Siri clone for the N900, N9(50) and Jolla
Follow me on Twitter: twitter.com/taixzo
Saera: A Siri clone for the N900, N9(50) and Jolla
Follow me on Twitter: twitter.com/taixzo
|
|
2014-03-20
, 17:22
|
|
Posts: 203 |
Thanked: 445 times |
Joined on Mar 2010
|
#27
|
I'd say help freemangordon, pali, et al. to run maemo on 3.x kernels, then go from there.
Win-win for everyone.
Last edited by foobar; 2014-03-20 at 18:13.
Win-win for everyone.
Last edited by foobar; 2014-03-20 at 18:13.
|
|
2014-03-20
, 18:43
|
|
Posts: 71 |
Thanked: 177 times |
Joined on Aug 2013
|
#28
|
Well, I was trying to compile LXC and when generating configure by running ./autogen.sh I'm getting:
aclocal: configure.ac: 205: macro `AM_COND_IF' not found in library
aclocal: configure.ac: 219: macro `AM_COND_IF' not found in library
aclocal: configure.ac: 234: macro `AM_COND_IF' not found in library
aclocal: configure.ac: 252: macro `AM_COND_IF' not found in library
aclocal: configure.ac: 269: macro `AM_COND_IF' not found in library
aclocal: configure.ac: 304: macro `AM_COND_IF' not found in library
aclocal: configure.ac: 315: macro `AM_COND_IF' not found in library
aclocal: configure.ac: 377: macro `AM_COND_IF' not found in library
That's probably related to old version of autoconf used in scratchox. Will update & report.
<- solved by downloading .tar.gz tarball & running configure directly.
Another error log, when running make:
In file included from lxc_attach.c:37:
utils.h:56: error: static declaration of 'setns' follows non-static declaration
/usr/include/bits/sched.h:91: error: previous declaration of 'setns' was here
Ideas? <- caused by broken version of libc6 from previous projects. Reinstalling whole scratchbox, will let you know soon
#lexik
Last edited by lexik; 2014-03-25 at 12:32.
aclocal: configure.ac: 205: macro `AM_COND_IF' not found in library
aclocal: configure.ac: 219: macro `AM_COND_IF' not found in library
aclocal: configure.ac: 234: macro `AM_COND_IF' not found in library
aclocal: configure.ac: 252: macro `AM_COND_IF' not found in library
aclocal: configure.ac: 269: macro `AM_COND_IF' not found in library
aclocal: configure.ac: 304: macro `AM_COND_IF' not found in library
aclocal: configure.ac: 315: macro `AM_COND_IF' not found in library
aclocal: configure.ac: 377: macro `AM_COND_IF' not found in library
That's probably related to old version of autoconf used in scratchox. Will update & report.
<- solved by downloading .tar.gz tarball & running configure directly.
Another error log, when running make:
In file included from lxc_attach.c:37:
utils.h:56: error: static declaration of 'setns' follows non-static declaration
/usr/include/bits/sched.h:91: error: previous declaration of 'setns' was here
Ideas? <- caused by broken version of libc6 from previous projects. Reinstalling whole scratchbox, will let you know soon
#lexik
Last edited by lexik; 2014-03-25 at 12:32.
|
|
2014-03-20
, 19:35
|
|
Posts: 3,328 |
Thanked: 4,476 times |
Joined on May 2011
@ Poland
|
#29
|
Originally Posted by Hurrian
So what's this whole kernel discussion for? This would mean we don't need those static PM patches... And lexik, don't get scared by any "It won't work" posts. The same was said about Harmattan, that Harmattan apps won't work on Fremantle (hard/soft float)
Assuming that Android runs inside Maemo via ParallelDroid, and static PM support isn't available, Maemo's (IMO superior) dynamic power control will reign supreme.
|
|
2014-03-20
, 19:52
|
|
Posts: 439 |
Thanked: 282 times |
Joined on Oct 2012
|
#30
|
Originally Posted by foobar
that it°s a longshoot , it should be more easy to apply to our kernel instead of the 3.x.x one due to missing drivers in the last one
I'd say help freemangordon, pali, et al. to run maemo on 3.x kernels, then go from there.
Win-win for everyone.
__________________
N900 Beast featuring :
8GB OPTFS@ext4(*performance mount options)
overclock: cpu[125 - 1,150(*VDD1)+Dsp overclock
CSSU-T latest version
N900 Beast featuring :8GB OPTFS@ext4(*performance mount options)
overclock: cpu[125 - 1,150(*VDD1)+Dsp overclock
CSSU-T latest version
 |
| Tags |
| android, chroot, easy debian, nitdroid |
«
Previous Thread
|
Next Thread
»
|
All times are GMT. The time now is 18:37.
8GB OPTFS@ext4(*performance mount options)
overclock: cpu[125 - 1,150(*VDD1)+Dsp overclock
CSSU-T latest version
Last edited by Alecsandru; 2014-03-19 at 22:34.