Re: [patch] "HT scheduler", sched-2.5.63-B3

Aaron Lehmann (aaronl@vitelus.com)
Sat, 8 Mar 2003 10:28:21 -0800


On Fri, Mar 07, 2003 at 08:36:12AM +0100, Ingo Molnar wrote:
> > I was able to reproduce them by selecting text in Mathematica (ugh, not
> > a very helpful example). The skips were shorter and about three times as
> > hard to trigger as on 2.5.63.
>
> okay, just as a data point, could you try to renice the player
> process/thread to -2? Does it make the skipping harder to trigger?
> How about -5, or -10?

Renicing all xmms threads to -2 makes the problem impossible to
trigger in this way. I'm sorry for having such a stupid testcase;
Mathematica is proprietary software, and judging by the way it makes
XMMS skip, it's probably doing something stupid. XMMS also has a "Use
realtime priority" option, which makes it do:

sparam.sched_priority = sched_get_priority_max(SCHED_RR);
sched_setscheduler(0, SCHED_RR, &sparam);

Obviously, this requires root privileges, however, it also prevents
the skipping. I wonder if setting CAP_SYS_NICE on xmms and hacking it
to ignore geteuid() would be safe.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/