Linked by David Adams on Mon 19th Mar 2012 17:04 UTC, submitted by diegocg
Linux Linux 3.3 has been released. The changes include the merge of kernel code from the Android project. There is also support for a new architecture (TI C6X), much improved balancing and the ability to restripe between different RAID profiles in Btrfs, and several network improvements: a virtual switch implementation (Open vSwitch) designed for virtualization scenarios, a faster and more scalable alternative to the "bonding" driver, a configurable limit to the transmission queue of the network devices to fight bufferbloat, a network priority control group and per-cgroup TCP buffer limits. There are also many small features and new drivers and fixes are also available. Here's the full changelog.
Order by: Score:
Better writeup:
by kragil on Mon 19th Mar 2012 17:33 UTC
kragil
Member since:
2006-01-04

http://www.h-online.com/open/features/What-s-new-in-Linux-3-3-14668...

The delay fixes for slow storage devices is what I look forward to the most.

Reply Score: 5

Android merge
by fran on Mon 19th Mar 2012 21:01 UTC
fran
Member since:
2010-08-06

Regarding the merge of kernel code from the Android project.
What precisely does this do?
Is a part of the Android code not java and such.
Or is it just real low level stuff.

Edited 2012-03-19 21:02 UTC

Reply Score: 3

RE: Android merge
by TemporalBeing on Mon 19th Mar 2012 21:59 UTC in reply to "Android merge"
TemporalBeing Member since:
2007-08-22

Regarding the merge of kernel code from the Android project.
What precisely does this do?
Is a part of the Android code not java and such.
Or is it just real low level stuff.


Android/Google/OHSA had a number of changes in the Linux Kernel portion per timers and battery efficiency. They were initially rejected due to very big differences with the mainline kernel. So this is the merging of that code reworked to be more compatible with the mainline kernel - including drivers, etc.

At least, that's what I gather based on previous information about the incompatibilities of the patches to Android vs. the mainline kernel.

Reply Score: 9

RE[2]: Android merge
by Lennie on Tue 20th Mar 2012 15:15 UTC in reply to "RE: Android merge"
Lennie Member since:
2007-09-22

Actually, this is most if not all custom kernel code in Android except for the "portion per timers and battery efficiency" (Wavelocks):

"Immature

As anticipated, a number of Android specific drivers have now entered the staging area, including some drivers previously merged into the staging area, but later removed as because they were not being maintained (1, 2, 3, 4, 5 and others). These changes mean that the kernel should now contain everything required to boot an Android userland, but is still missing wake lock code or equivalent [1], which is required to achieve decent battery life.

John Stultz has put together some background information on the status of integration of Android-specific changes in this LWN.net article [2], in which he reports on a meeting of the Android mainlining interest group. [3]"

[1] http://thread.gmane.org/gmane.linux.kernel/1249726/focus%3D1256...

[2] http://lwn.net/Articles/484196/

[3] http://www.h-online.com/news/item/Android-drivers-to-be-included-in...

Progress is/seems slow, but still moving.

Reply Score: 5

RE: Android merge
by Elv13 on Tue 20th Mar 2012 02:36 UTC in reply to "Android merge"
Elv13 Member since:
2006-06-12

Android AOSP/LiNARO are getting bored of porting an heavy patch set to keep the Android fork of Linux up to date. So instead of another massive code drop that have no chance of getting merged, they go bit by bit to make the merge easier for everybody. This is a good thing if this get merged because Android drivers will start to work fine on regular Linux, a win for everybody (graphic drivers excluded).

Reply Score: 4

vSwitch?
by Kebabbert on Tue 20th Mar 2012 12:57 UTC
Kebabbert
Member since:
2007-07-27

Cool, I did not know that Linux also copied the Crossbow concept from OpenSolaris
http://www.ibm.com/developerworks/linux/library/l-virtual-networkin...

Soon Linux has copied everything that Solaris invented.

Reply Score: 0

RE: vSwitch?
by Laurence on Tue 20th Mar 2012 17:14 UTC in reply to "vSwitch?"
Laurence Member since:
2007-03-26

Cool, I did not know that Linux also copied the Crossbow concept from OpenSolaris
http://www.ibm.com/developerworks/linux/library/l-virtual-networkin.....

Soon Linux has copied everything that Solaris invented.


OpenSolaris equally owes a large debt of gratitude to Linux too. GRUB and Gnome are direct ports from GNU/Linux and essential components to OpenSolaris. This isn't even just Sun/Oracle copying the concepts of those GNU projects - it's direct code rips.

All software borrows from each other. It's impossible to write 100% original concepts 100% of the time.

Edited 2012-03-20 17:15 UTC

Reply Score: 3

RE[2]: vSwitch?
by Kebabbert on Wed 21st Mar 2012 10:24 UTC in reply to "RE: vSwitch?"
Kebabbert Member since:
2007-07-27

OpenSolaris equally owes a large debt of gratitude to Linux too. GRUB and Gnome are direct ports from GNU/Linux and essential components to OpenSolaris. This isn't even just Sun/Oracle copying the concepts of those GNU projects - it's direct code rips.

The difference is that Solaris is doing new and innovative development. I mean, GRUB is nothing new nor innovative. Boot loaders have existed earlier. GNOME is nothing new nor innovative.

But ZFS is new and innovative. Crossbow too. Zones too. DTrace too. etc

So where is the new innovations in Linux? GRUB? Gnome? Nowhere. It is just polished versions of old software, and nothing new and unique.

Thus, Linux copies stuff from Solaris but no new innovative tech comes from Linux. Everything is just copies. Heck, even the entire Linux is just a Unix copy. Everything in Linux is a copy.

Reply Score: 2

RE[3]: vSwitch?
by Laurence on Wed 21st Mar 2012 11:54 UTC in reply to "RE[2]: vSwitch?"
Laurence Member since:
2007-03-26


The difference is that Solaris is doing new and innovative development. I mean, GRUB is nothing new nor innovative. Boot loaders have existed earlier. GNOME is nothing new nor innovative.

But ZFS is new and innovative. Crossbow too. Zones too. DTrace too. etc

So where is the new innovations in Linux? GRUB? Gnome? Nowhere. It is just polished versions of old software, and nothing new and unique.

Thus, Linux copies stuff from Solaris but no new innovative tech comes from Linux. Everything is just copies. Heck, even the entire Linux is just a Unix copy. Everything in Linux is a copy.

Well yeah, Sun had pioneered a number of fantastic technologies. But they were the exception rather than the norm. Most software houses and most OSs are re-implementations + improvements. Be that Linux, OS X, Windows, FreeBSD or whatever.

Sadly Oracle have pretty much killed any innovation Sun had, so this argument is rapidly becoming moot.

[edit]
I've just read your other post re giving credit vs denying inspiration and that explains your stance a lot clearer than you had here. Taking that post into account, I 100% agree with you ;)

Edited 2012-03-21 11:58 UTC

Reply Score: 2

RE[4]: vSwitch?
by Kebabbert on Wed 21st Mar 2012 12:43 UTC in reply to "RE[3]: vSwitch?"
Kebabbert Member since:
2007-07-27

I've just read your other post re giving credit vs denying inspiration and that explains your stance a lot clearer than you had here. Taking that post into account, I 100% agree with you ;)

Ok, Laurence, you seem to be a normal sane guy. I am glad there are Linux people (if you are that?) that gives credit where it is due.

I myself copy, but I always give credit. Everybody copies, but you should always give credit and not try to claim it as your own idea.

Reply Score: 3

RE[5]: vSwitch?
by Laurence on Wed 21st Mar 2012 13:50 UTC in reply to "RE[4]: vSwitch?"
Laurence Member since:
2007-03-26


Ok, Laurence, you seem to be a normal sane guy. I am glad there are Linux people (if you are that?) that gives credit where it is due.

I myself copy, but I always give credit. Everybody copies, but you should always give credit and not try to claim it as your own idea.

hehehe thank you though I wouldn't really call myself a "Linux person". I'm actually a big fan of Solaris as well as Linux. I remember trying Zones out for the 1st time and being blown away by how powerful they were and yet how simple to create. Those systems have since been migrated to FreeBSD after the Oracle/Sun takeover, however I still run a number of old Solaris 8 boxes on ageing SPARC hardware.

I think if I had to pick a favourite, I'd probably side with FreeBSD - which has only become even more epic since they've ported ZFS.

So all in all, I wouldn't say I have any particular allegiance to Linux over any other *nix OS.

Reply Score: 2

RE: vSwitch?
by alexz on Wed 21st Mar 2012 04:30 UTC in reply to "vSwitch?"
alexz Member since:
2012-02-25

Cool, I did not know that Linux also copied the Crossbow concept from OpenSolaris
http://www.ibm.com/developerworks/linux/library/l-virtual-networkin...

Soon Linux has copied everything that Solaris invented.


I can feel a bit of frustration in your post. It also sounds a bit like what software patent proponents try to protect "OMG we did it first, you shouldn't be allowed to do it too!"

How would any operating progress if everybody had to be 100% original? Every OS needs to introduce ideas, if they do not, they die. But eventually, the most popular particularities have to become mainstream and be adapted to other OSes to benefit most of us. You can always rejoice on the fact that licenses were not compatible, linux dev had to re-implement it, they couldn't steal code!

Reply Score: 2

RE[2]: vSwitch?
by Kebabbert on Wed 21st Mar 2012 10:43 UTC in reply to "RE: vSwitch?"
Kebabbert Member since:
2007-07-27

I can feel a bit of frustration in your post. It also sounds a bit like what software patent proponents try to protect "OMG we did it first, you shouldn't be allowed to do it too!"

How would any operating progress if everybody had to be 100% original? Every OS needs to introduce ideas, if they do not, they die. But eventually, the most popular particularities have to become mainstream and be adapted to other OSes to benefit most of us. You can always rejoice on the fact that licenses were not compatible, linux dev had to re-implement it, they couldn't steal code!

I dont have any problems with porting or stealing ideas. Not at all.

The problem I have, is when the stealer does not give credit to the original innovative hard work. Instead they pretend they are doing it all themselves from scratch and thinking deep and hard, but for a knowledgeable person it is apparent they are making a copy. It only looks ugly, when they try to hide that. Assume you have an unique idea and you tell it to a person, and that person steals your idea and claims it as his own. Would you have a problem with that? Wouldn't it be better if that person gave you cred and said it was your idea?



To exemplify, VMware copied DTrace and called it vProbes, and did gave cred to Solaris. I love that. I really hope VMware can improve DTrace and they can help each other:
http://x86vmm.blogspot.se/2007/09/presenting-vprobes.html

"...In the "credit where it's due" department: we owe an enormous debt in our thinking about this problem to our colleagues at Sun. I've never hidden my admiration for DTrace...

VMware never tried to hide the fact they copied from Solaris, they dont try to steal and grab all credit themselves.



Nor did FreeBSD or Mac OS X hide the fact they have ported DTrace. They do not claim it as their own original work.




On the other hand, Linux people copied DTrace, and what happened? Linux people tried to get all cred themselves, and pretend they did new and innovative work from scratch. That sucks badly, dont you agree?
https://blogs.oracle.com/ahl/entry/dtrace_knockoffs

"...Amusingly, in an apparent attempt to salvage their self-respect, the SystemTap team later renounced their inspiration. Despite frequent mentions of DTrace in their early meetings and email, it turns out, DTrace didn't actually inspire them much at all:

Log message:
Removed refs to dtrace, to which we were giving undue credit in terms of "inspiration."..."




Now Linux people have copied ZFS, and calls it BTRFS. Even though Chris Mason says he looks at ZFS for things to copy, Linux fans insists that BTRFS is not a copy, it is a new and original work. In a few years, all Linux people will try to hide the fact that BTRFS is a ZFS wannabe.
-No no no, BTRFS is new and innovative. It is original work. It is not a ZFS wannabe.




But IBM is not better, they are also copying DTrace and calling it AIX ProbeVue. But IBM never gives credit. Among others IBM also copied Solaris Zones, and calls it AIX WPAR. And calls WPAR new and innovative. Ugly, yes?

Solaris copied LPAR from IBM AIX, and Solaris calls it LDOMs, but this is no secret.


I have problems with people not givin credit to the originator doing all his hard work. In science you always cite other scientists work. If you tried to steal others work, it would look very ugly. But not among Linux people, apparently.

I wish Linux people did new cool stuff that everybody wanted and copied or ported. But that has not happened. Everybody copies/ports from Solaris today.

Reply Score: 2