Re: RT task scheduling

[Bill Huey (hui) <billh@gnuppy.monkey.org>]

On Fri, Apr 07, 2006 at 12:51:41PM +0200, Ingo Molnar wrote:
> no, i'm discussing precisely the point you raised:

Oh boy.

> >>> You should consider for a moment to allow for the binding of a 
> >>> thread to a CPU to determine the behavior of a SCHED_FIFO class task 
> >>> instead of creating a new run category. [...]
> 
> with the observation that 1) binding is already possible [so your 
> suggestion is apparently knocking on open doors] 2) binding is a 
> separate mechanism (not adequate for all workloads) and it is thus 
> orthogonal to what i'm trying to achieve with the "RT overload" stuff.  
> Really simple and straightforward observations i think.

This is going to take some time to get the terminology right. It's
late now and I'll have to continue this tomorrow.

First thing's first, SCHED_FIFO_GLOBAL for what you want in the main
line is the same thing as SCHED_FIFO in -rt, right ?

bill