1 |
On Thu, 2002-01-24 at 17:37, Nathan Myers wrote: |
2 |
> I notice Gentoo is using RML's pre-emptible kernel instead of |
3 |
> Andrew Morton's low-latency kernel, despite both Linus's and |
4 |
> Alan's recommendation of the latter. Maybe it would be worth |
5 |
> posting a position paper (or comment) about why. If not, I'm |
6 |
> curious, anyway. (ISTM that the low-latency patch should be much |
7 |
> less hazardous, and ought to be substantially more efficient on |
8 |
> single-CPU systems anyway.) |
9 |
|
10 |
We're probably going to go back to low-latency, although the two can be |
11 |
complementary sometimes. |
12 |
|
13 |
For a server, we'll probably do: |
14 |
XFS CVS kernel |
15 |
+IO read latency fixes from Andrew Morton |
16 |
+low-latency |
17 |
+rmap (these include the above-mentioned IO read latency fixes) |
18 |
+O(1)? |
19 |
|
20 |
Desktop will probably be similar but we may include a few more things. |
21 |
|
22 |
-- |
23 |
Daniel Robbins <drobbins@g.o> |
24 |
Chief Architect/President http://www.gentoo.org |
25 |
Gentoo Technologies, Inc. |