Preemptible Linux Kernel: Interviews and Info 238
An anonymous submitter sends: "MontaVista and Robert Love are developing a patch for the Linux kernel to make it fully preemptible. Lots of users are involved, and tests show huge reductions in latency. Robert's kernel patches are here. Finally, an interview with Robert, on preemption and more."
Wow! (Score:1, Interesting)
So will that make Linux a superior audio platform? (Score:3, Interesting)
Finally.. (Score:2, Interesting)
Hmm (Score:3, Interesting)
I'm wondering about this paragraph:
Can anyone give a nice layman's description of what he's talking about here?
I'm not sure... (Score:2, Interesting)
I think that (for now) using this patch on workstations is a pretty good idea. And I think that there should be a better solution for the problem witch should THEN be something along the lines of kernel 3.0
I am not a kernel developer or anything, but I am currently reading up on the source and the mailing lists.
Basically all I am trying to say is: Make it work NOW and solve the real problem later. Just make sure that is WILL be solved... (no microsoft coding ways here
What does that mean? (Score:2, Interesting)
Re:I'm not sure... (Score:3, Interesting)
Re:Windows 2000/NT (Score:4, Interesting)
There is a difference between pre-emptive multitasking and pre-emptible kernel.
Pre-emptive multitasking means that the kernel can interrupt any thread and give control to another thread, so that a thread cannot hog the CPU resources. This is what all modern operating systems do, except Windows 3.1/9x/Me and MacOS (pre- X), though it could be argued that Windows 3.1/9x/Me is not an operating system much less a modern one
Pre-emptible kernel is a different beast. It means that the kernel can interrupt itself (i.e. a thread running in the kernel mode) and give control to another thread running in the kernel mode. This is used in real-time operating systems where you need to have a guaranteed maximum response time (i.e. a thread must not wait longer than a certain amount of time before it gets control). However, this is not all that useful for general-purpose OSes and may even be detrimental to servers, where throughput matter more than response time. So it's good to know that this will be a compile-time option.
When will the Linux HZ be bumped up to 1000? (Score:1, Interesting)
Re:Preempt Patches in Kernel (Score:4, Interesting)
Isn't that like saying that you'd rather fix all buggy applications instead of providing a protected memory environment?
Re:When will the Linux HZ be bumped up to 1000? (Score:1, Interesting)
Re:needed badly (Score:2, Interesting)
There's a bug in recent 2.4 kernels where a multithreaded app dumping core could livelock the system. You might try setting a hard limit on core file size to zero and see if the crashes go away.
You'd want to do this in /etc/profile, of
course.
Re:needed badly (Score:2, Interesting)
Will the linux kernel allow a user process to be killed that is blocked in a kernel call? In my experience, Solaris and Tru64 do not: a user program that is blocked in a kernel call will stay blocked until the kernel call returns, regardless of any action (short of rebooting the machine) that a user can take. I assume that there is some well-thought-out reasoning behind this, but sometimes (e.g. during device driver development) I wish it were somehow a configurable behavior.
I'd think that infinite loops would be too much of a newbie bug.
Lots of times, the most junior person gets stuck writing device drivers. And even experienced programmers can have brain farts.