[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: realtime kernel for Debian

Grammostola Rosea wrote:
Jim wrote:
Grammostola Rosea,

I want also to direct your attention to the kernel, as it has the
possibility to be more supportive of those specific needs, by having
low latency and real-time extensions patched and enabled. The debian
folks (especially "waldi" aka Bastien Blank will say some or all of
these are less stable than they could be -- perhaps googling around or
asking him when he's not so busy will drum up some details.)

Mmh this is interesting, cause there is an realtime kernel available in the ubuntu hardy repo, but not in Debian yet. Would be nice if there was one which users could install. But I'm not an rt-kernel expert at all, so maybe I should forward this to some other people...

But I think it's good to have some discussion about a realtime kernel for Debian on the Debian-dev list...
Looking forward to your opinions on this.

Do you really need real time kernel?
Debian is a technical driven project, but reading the previous two quotes,
"real time" is used as marketing thing.

Standard Linux kernel is good for most real time uses (but with the
right hardware, i.e. common and not obsolete hardware, which could
have not so good maintained drivers, so not modified for lower latencies).
IIRC music industry helped the developer of normal kernel, to have
a good nearly-realtime properties.

With latencytop a lot of latencies was discovered and corrected.

IMHO, for most users I think guarantee realtime is not really need
(but for marketing).
Do you have cases where current kernel are not good? Could you describe it,
so that we can try to correct them?

Anyway current Debian default is CONFIG_HZ_250, an additional CONFIG_HZ_1000
kernel (for desktop user) would be nice, and it would remove most of the
concern of our user.
Note: this is not realtime, but often people confuse the meaning of realtime,
and giving a real realtime kernel could IMO cause more trouble to people.


Reply to: