8 ms·

Linux 3.6 released

177 points by moonboots 14y ago

DiabloD3 14y ago
And I still can't sleep my Powerbook /w Radeon because post KMS radeon doesn't support it yet (no one ported the pre-KMS code from the X driver portion of radeon into the post-KMS radeon kernel driver). OTOH, I can just leave KMS off and have no 2D or 3D acceleration. /rant
mtgx 14y ago
Looking forward to Linux 3.7 with its support for ARM64: http://www.phoronix.com/scan.php?page=news_item&px=MTE5NjQ http://www.phoronix.com/scan.php?page=news_item&px=MTE5N...
- ajross 14y ago
  The code is already out there, all 3.7 integration would represent is a stamp of approval. Consider that there is no hardware for the architecture yet, I don't know why approval would be very interesting: http://git.kernel.org/?p=linux/kernel/git/cmarinas/linux-aarch64.git http://git.kernel.org/?p=linux/kernel/git/cmarinas/linux-aar...
jeltz 14y ago
See the Kernel newbies page for a good summary of the news in 3.6. http://kernelnewbies.org/Linux_3.6 http://kernelnewbies.org/Linux_3.6
Nux 14y ago
Lots of goodies!
- viraptor 14y ago
  From the desktop perspective, I really like the "Suspend to disk and memory at the same time" part. It's great that you don't have to think anymore if you're going to use your laptop in the next 2 days or not. If it runs out of power, it does - nothing gets lost! Adding bufferbloat pieces into the main release is also good news (although that probably won't hit the servers that really need it for many more months...)
  - jlgreco 14y ago
    Do you think this will obsolete the 'hybrid suspend' functionality in pm-utils? I think that works by suspending to ram, waking up in 10 minutes then immediately suspending to disk. Maybe a combination of these two approaches could be useful.
jeffdavis 14y ago
I'd really like to see btrfs start to stabilize and become more mainstream, but everything still seems to say that it's experimental. The main reasons that I want it on my desktop are: * easy backups * checksums * maybe it will become integrated with package manager (I'm on ubuntu) so that I can roll back package updates if they don't work properly
- triplepoint217 14y ago
  I have been running Ubuntu 12.04 with my / as btrfs for a couple of months now, with no trouble (and not even thinking about it truth be told). That said, I left my /home ext4 because I didn't know how much to trust it.
  - buster 14y ago
    Did the same, btrfs for everything but home and will switch to ext4 with 12.10 because it feels much slower then the ext3 i had earlier. But i'll be sure when i upgrade. To be honest, what do i need btrfs for? I don't (and likely wouldn't) use all of its advanced features anyway.. In all the benchmarks ext4 seems to be a reasonable choice in terms of speed and stability.
    - triplepoint217 14y ago
      I am looking forward to when we have fire and forget snapshotting system so that we all have rolling backups. I would also like the send snapshot features they have added recently for keeping multiple drives in sync (I currently use unison, but I think this could be faster/better). Plus copy on write should be pretty cool (cp becomes as fast as mv), though I could see it allowing me to engage in some suboptimal behaviors of copying tons of stuff and counting on CoW to avoid actual duplication.
- rogerbinns 14y ago
  It will be marked as experimental for a long time. There isn't a magical point at which it will suddenly be perfect and bug free. That said kernel 3.2 was a major point of stability. https://btrfs.wiki.kernel.org/index.php/FAQ#Is_btrfs_stable.3F https://btrfs.wiki.kernel.org/index.php/FAQ#Is_btrfs_stable.... It is integrated with the package manager on Ubuntu already. I just upgraded my laptop to Quantal and part of the upgrade process spotted I was using btrfs and created a snapshot. (It was a complete waste of time as I have automatically created hourly, daily, weekly and monthly snapshots.) Doing regular package updates is very slow with btrfs. This is due to fsync() being called several times per file. You'll also want Quantal for an updated btrfs-tools package. For example it lets you do scrubbing. You also need it (and kernel 3.3+) to change raid levels after filesystem creation. The Ubuntu 12.04 installer does create and install to btrfs but didn't let you configure things like raid level so you were stuck with whatever it did. I've been running all my systems on btrfs for several months now on Ubuntu 12.04 (kernel 3.2) on both SSD and HDD including using RAID 0, RAID 1, dmcrypt/LUKS, inplace conversion from ext4 and who knows what else. (Across my server, workstation, laptop and HTPC.) The only problem I have had is when filesystems fill up. I've never lost data but it can be quite frustrating trying to find files to delete (also need to be removed from snapshots), rebalancing etc to get things running smoothly again. The various ways of doing df are mostly a work of fiction. My data is duplicated to offline backups, Dropbox, multiple systems and git/hg repositories so the failure of any system would be annoying but I'd never lose anything. You should make sure you are in that position first, independent of what filesystems are being used.
alexholehouse 14y ago
OK, so here's a question. I've always wanted to spend some time reading through the Linux kernel - I really enjoyed operating systems when I studied it (and reading through MINIX src) but I don't really know where/how to start with Linux. I mean, I guess I can just dive in, and maybe that's the best approach, but is there a strategy anyone would recommend for reading the Linux source in terms of it making sense as a combined unit of code (as opposed to a collection of algorithms, if that makes sense)?
- jey 14y ago
  Don't read the Linux source code unless you have to. Read xv6's source code instead. http://www.google.com/search?q=xv6 http://www.google.com/search?q=xv6
- Jarihd 14y ago
  start off with reading 1) linux kernel development 2) linux device drivers 3) linux kernel module programming guide have - understanding the linux kernel as your reference manual. by now, you should be comfortable to read/understand the kernel source; download linux kernel source and start browsing through the code. simply reading books wont get you anywhere - you need to play around with kernel source inorder to understand the linux kernel behavior and different problems you may come across. write simple kernel modules to get a hang of how you can interact/modify with the kernel. join some opensource project and start fixing bugs you're comfortable with or just play around with your local linux kernel source - make changes; build and deploy and observe what happens. EDIT: if you have no prior knowledge of OS Theory and Fundamentals; then you should start here first - read either of the following books 1) Operating System Concepts by Galvin, Silberschatz OR 2)Modern Operating Systems by Tanenbaum For programming related - system calls and stuff read 1) Advanced Programming in the UNIX Environment - by Richard Stevens
- acomjean 14y ago
  I got sent to redhat in my companies project to go from Solaris/HPUX -> Linux. It was government software and the project got put on hold.... The instructor was excellent. https://www.redhat.com/training/courses/rhd361/course-exam-outline https://www.redhat.com/training/courses/rhd361/course-exam-o... Well thats kinda expensive.... One thing I found a little different was that the OS has its own libraries for everything (string.h etc..) which makes sense if you think about it. If you want to browse the source code on-line with this software website thing called lxr (Linux Cross Reference). Its got a good search tool and linked headers. Clicking on a function name shows you where that function is used. You can install it yourself and I think its much faster.. http://lxr.linux.no/linux+v3.6/ http://lxr.linux.no/linux+v3.6/ http://rhkernel.org/ http://rhkernel.org/ There is linux weekly news too, which is a decent site when I was still in the Linux porting world. Like many Linux sites, seems to lack in style, but makes up in content. http://lwn.net/ http://lwn.net/
DASD 14y ago
TCP Fast Open (client side) is quite interesting. Is anyone aware of testable server examples/implementations yet?
- nine_k 14y ago
  I suppose [some] Google services may be the server side with Fast Open support: http://googlecode.blogspot.com/2012/01/lets-make-tcp-faster.html http://googlecode.blogspot.com/2012/01/lets-make-tcp-faster....

177 points by moonboots 14y ago

And I still can't sleep my Powerbook /w Radeon because post KMS radeon doesn't support it yet (no one ported the pre-KMS code from the X driver portion of radeon into the post-KMS radeon kernel driver). OTOH, I can just leave KMS off and have no 2D or 3D acceleration. /rant

mtgx 14y ago

Looking forward to Linux 3.7 with its support for ARM64: http://www.phoronix.com/scan.php?page=news_item&px=MTE5NjQ http://www.phoronix.com/scan.php?page=news_item&px=MTE5N...

ajross 14y ago
The code is already out there, all 3.7 integration would represent is a stamp of approval. Consider that there is no hardware for the architecture yet, I don't know why approval would be very interesting: http://git.kernel.org/?p=linux/kernel/git/cmarinas/linux-aarch64.git http://git.kernel.org/?p=linux/kernel/git/cmarinas/linux-aar...

jeltz 14y ago

See the Kernel newbies page for a good summary of the news in 3.6. http://kernelnewbies.org/Linux_3.6 http://kernelnewbies.org/Linux_3.6

Nux 14y ago

Lots of goodies!

viraptor 14y ago
From the desktop perspective, I really like the "Suspend to disk and memory at the same time" part. It's great that you don't have to think anymore if you're going to use your laptop in the next 2 days or not. If it runs out of power, it does - nothing gets lost! Adding bufferbloat pieces into the main release is also good news (although that probably won't hit the servers that really need it for many more months...)
- jlgreco 14y ago
  Do you think this will obsolete the 'hybrid suspend' functionality in pm-utils? I think that works by suspending to ram, waking up in 10 minutes then immediately suspending to disk. Maybe a combination of these two approaches could be useful.

jeffdavis 14y ago

I'd really like to see btrfs start to stabilize and become more mainstream, but everything still seems to say that it's experimental. The main reasons that I want it on my desktop are: * easy backups * checksums * maybe it will become integrated with package manager (I'm on ubuntu) so that I can roll back package updates if they don't work properly

triplepoint217 14y ago
I have been running Ubuntu 12.04 with my / as btrfs for a couple of months now, with no trouble (and not even thinking about it truth be told). That said, I left my /home ext4 because I didn't know how much to trust it.
- buster 14y ago
  Did the same, btrfs for everything but home and will switch to ext4 with 12.10 because it feels much slower then the ext3 i had earlier. But i'll be sure when i upgrade. To be honest, what do i need btrfs for? I don't (and likely wouldn't) use all of its advanced features anyway.. In all the benchmarks ext4 seems to be a reasonable choice in terms of speed and stability.
  - triplepoint217 14y ago
    I am looking forward to when we have fire and forget snapshotting system so that we all have rolling backups. I would also like the send snapshot features they have added recently for keeping multiple drives in sync (I currently use unison, but I think this could be faster/better). Plus copy on write should be pretty cool (cp becomes as fast as mv), though I could see it allowing me to engage in some suboptimal behaviors of copying tons of stuff and counting on CoW to avoid actual duplication.
rogerbinns 14y ago
It will be marked as experimental for a long time. There isn't a magical point at which it will suddenly be perfect and bug free. That said kernel 3.2 was a major point of stability. https://btrfs.wiki.kernel.org/index.php/FAQ#Is_btrfs_stable.3F https://btrfs.wiki.kernel.org/index.php/FAQ#Is_btrfs_stable.... It is integrated with the package manager on Ubuntu already. I just upgraded my laptop to Quantal and part of the upgrade process spotted I was using btrfs and created a snapshot. (It was a complete waste of time as I have automatically created hourly, daily, weekly and monthly snapshots.) Doing regular package updates is very slow with btrfs. This is due to fsync() being called several times per file. You'll also want Quantal for an updated btrfs-tools package. For example it lets you do scrubbing. You also need it (and kernel 3.3+) to change raid levels after filesystem creation. The Ubuntu 12.04 installer does create and install to btrfs but didn't let you configure things like raid level so you were stuck with whatever it did. I've been running all my systems on btrfs for several months now on Ubuntu 12.04 (kernel 3.2) on both SSD and HDD including using RAID 0, RAID 1, dmcrypt/LUKS, inplace conversion from ext4 and who knows what else. (Across my server, workstation, laptop and HTPC.) The only problem I have had is when filesystems fill up. I've never lost data but it can be quite frustrating trying to find files to delete (also need to be removed from snapshots), rebalancing etc to get things running smoothly again. The various ways of doing df are mostly a work of fiction. My data is duplicated to offline backups, Dropbox, multiple systems and git/hg repositories so the failure of any system would be annoying but I'd never lose anything. You should make sure you are in that position first, independent of what filesystems are being used.

alexholehouse 14y ago

OK, so here's a question. I've always wanted to spend some time reading through the Linux kernel - I really enjoyed operating systems when I studied it (and reading through MINIX src) but I don't really know where/how to start with Linux. I mean, I guess I can just dive in, and maybe that's the best approach, but is there a strategy anyone would recommend for reading the Linux source in terms of it making sense as a combined unit of code (as opposed to a collection of algorithms, if that makes sense)?

jey 14y ago
Don't read the Linux source code unless you have to. Read xv6's source code instead. http://www.google.com/search?q=xv6 http://www.google.com/search?q=xv6
Jarihd 14y ago
start off with reading 1) linux kernel development 2) linux device drivers 3) linux kernel module programming guide have - understanding the linux kernel as your reference manual. by now, you should be comfortable to read/understand the kernel source; download linux kernel source and start browsing through the code. simply reading books wont get you anywhere - you need to play around with kernel source inorder to understand the linux kernel behavior and different problems you may come across. write simple kernel modules to get a hang of how you can interact/modify with the kernel. join some opensource project and start fixing bugs you're comfortable with or just play around with your local linux kernel source - make changes; build and deploy and observe what happens. EDIT: if you have no prior knowledge of OS Theory and Fundamentals; then you should start here first - read either of the following books 1) Operating System Concepts by Galvin, Silberschatz OR 2)Modern Operating Systems by Tanenbaum For programming related - system calls and stuff read 1) Advanced Programming in the UNIX Environment - by Richard Stevens
acomjean 14y ago
I got sent to redhat in my companies project to go from Solaris/HPUX -> Linux. It was government software and the project got put on hold.... The instructor was excellent. https://www.redhat.com/training/courses/rhd361/course-exam-outline https://www.redhat.com/training/courses/rhd361/course-exam-o... Well thats kinda expensive.... One thing I found a little different was that the OS has its own libraries for everything (string.h etc..) which makes sense if you think about it. If you want to browse the source code on-line with this software website thing called lxr (Linux Cross Reference). Its got a good search tool and linked headers. Clicking on a function name shows you where that function is used. You can install it yourself and I think its much faster.. http://lxr.linux.no/linux+v3.6/ http://lxr.linux.no/linux+v3.6/ http://rhkernel.org/ http://rhkernel.org/ There is linux weekly news too, which is a decent site when I was still in the Linux porting world. Like many Linux sites, seems to lack in style, but makes up in content. http://lwn.net/ http://lwn.net/

DASD 14y ago

TCP Fast Open (client side) is quite interesting. Is anyone aware of testable server examples/implementations yet?

nine_k 14y ago
I suppose [some] Google services may be the server side with Fast Open support: http://googlecode.blogspot.com/2012/01/lets-make-tcp-faster.html http://googlecode.blogspot.com/2012/01/lets-make-tcp-faster....