Re: 2.6.17-rc2+ regression -- audio skipping

[Peter Williams]

Mike Galbraith wrote:
> On Mon, 2006-05-22 at 13:30 +1000, Peter Williams wrote:
> > What about the batch tasks?  How do you ensure that they don't get 
> > starved?  Remember they're "batch" tasks not "background" tasks.
>
> Here, batch means background.  To make them batch as in only static
> priority, I'd just do away with the second array.  Batch as background
> makes more sense to me, and since it's my ball and my playground... ;-)

In reality, both batch and background are useful distinct concepts.  I 
think of a batch task as one that needs to be treated fairly (in 
accordance with its nice value) but for which fairness shouldn't be 
broken to give it a boost as you might for an interactive task or media 
streamer.  I.e. doing useful work but not interactive or a media 
streamer and the occasional long latency isn't a disaster.

Background tasks are ones where you don't care if they ever get any cpu 
:-) except as necessary to prevent priority inversion.

In my schedulers I generalize background to "soft cpu rate caps" with a 
cap of zero being the same as background.  I have patches to add both 
soft and hard cpu rate caps to the standard scheduler but I'm sitting on 
them until things settle down a bit.

Anyway, schedulers based on single priority arrays (such as staircase) 
are looking more attractive every day.  :-)

Peter
--
Peter Williams                                   pwil3058@xxxxxxxxxxxxxx

"Learning, n. The kind of ignorance distinguishing the studious."
 -- Ambrose Bierce
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/