Linux Kernel 2.5.1 is Out

xise writes: "The next installment in the 2.5 Linux Kernel beta series, 2.5.1 is avaliable at the usual place Linux Kernel Archives. Remember to use the mirrors. You can read the changelog here."

Perhaps /. is a bad place for this.

[Anonymous Coward]

Not because "slashdot isn't freshmeat," but becaused judging from the outcry from unsophisticated users who updated to the latest STABLE kernel when they probably should have been sticking with vendor supplied kernels, most slashdotters either already know about the releases, or probably shouldn't.

Any newbie who trys to install 2.5.1 is in for a learning experience (especially if they use SCSI).

Re:2.4?

[Anonymous Coward]

Funny, I always thought that's what the STABLE branch was for....

It irritates me that linux developers insist on adding new "features" to "stable" kernels, rather than keeping a running development kernel year round. Things like the vm change early in the 2.4 series, and some HUGE, server breaking kernel changes should not appear in a stable kernel.

Once new features are found, and coded, they should go immediately into the development kernel, and save the stable kernel for bug fixes, driver updates, and security enhancements.

Re:I'm Bored

[Transient0]

I can't help but agree a little. Slashdot is a great arena for hashing out issues related to tech, but what new insights are going to be raised with each new minor version of the Kernel going into Beta. I can see that maybe it is useful to know that the new kernel exists, but those who need that info check the changelogs daily anyways. I don't see how we are supposed to communicate intelligently about 2.5.1 given a changelog and a mirror, is there no real news?

Let's start a revolution: i for one am in favor of not hearing about uneventful kernel updates anymore... i know i can filter out the entire category if i want, but you never know, there might one day be important news about the kernel(grin).

What are the new things they are working on?

[OS24Ever]

One of the key things for 2.4 if I remeber right was SMP support. Are they going to work on improving SMP support beyond the process level in 2.5? What could one list as the 'key bullet points' for 2.5 if talking to a manager type for futures of the Linux kernal?

Re:What are the new things they are working on?

[KidSock]

99% of the linux boxes out there run only 3 concurent running tasks, so the scheduler has to remain optimized for this situation

Interesting. But could this measurement be simplified to the point of being off base? A large percentage of these machines are webservers sitting idle so what do they care about scheduler optimizations? Same thoughts for single process number cruching ray tracing server farms. Shouldn't the focus be optimizing tasks that will benifit from being optimized? I know we have a few boxes that just run Java apps. I bet they would benifit from a new scheduler if the machine were a 4 way. So what are the bulk of these 99 out of 100 machines doing? They're not desktops.

Also, what prevents Kernel developers from optimizing the scheduler for a Kernel development workstation :-P

Re:2.4? (Offtopic)

[FlowerPotAdmin]

Yes, there is a point to Google caching a copy of itself -- they don't have to code in a special case for their own site.
Value(Code clarity) >= Value(Memory for one cached page)

NTFS r/w

[omega9]

I noticed mention of an upgrade to NTFS in the changelogs. I realize it can be argued as a non issue, but is there any real effort to stablize NTFS read/write? At work we're locked in to using W2k domain controllers, and have W2k in a few other places as well. Samba bridges the gap through the network, but in some cases directly mounting an NTFS partition would prove extremely useful. Or is this a non issue?

So when are...

[doorbot.com]

XFS and JFS supposed to be merged into the kernel? I saw a post a while back on Slashdot that claimed Linus wanted IBM/SGI/etc to wait for 2.5. Well 2.5 is here...

So the 64000 Euro question is... when are we getting ACL support? I've heard the IBM solution was good, but required a lot of kernel patches -- but that's what development kernels are all about!