lkml.org 
[lkml]   [1998]   [May]   [1]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
Subjectscheduling

Shaleh writes:

> I seem to recall a "QNX real time" patch for the kernel.
> This caused it to assign processes priority in a different
> way than nice does. You could make a process take over or
> a process rarely seen.

With that patch, doesn't an infinite loop stop everything less
important forever? That is only good for real-time and idle eaters.
It would be nice to have the features without the troubles,
so we need a way to mix the schedulers.

Perhaps we could have 3 levels of the "QNX scheduler" and run
the normal timesharing scheduler within each of those. If there
isn't a standard name for it already, I suggest "priority bands".
Then we would have something like this:

20 .. 59 idle eaters
-20 .. 19 normal
-60 .. -21 important

Alternate solution:

If a less segmented system is desirable, then just don't ever run
a process when there is another process which is n levels more
important. With n=40 we get the current system, and with n=1 we
get the QNX scheduler AFAIK. The default n could be 40.


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu

\
 
 \ /
  Last update: 2005-03-22 13:42    [W:0.031 / U:1.084 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site