On Fri, 30 Jul 2004 16:11:37 -0400
Lee Revell <rlrevell(a)joe-job.com> wrote:
On Fri, 2004-07-30 at 09:31, Florian Schmidt wrote:
Hi,
can someone here maybe shed some light on the naming conventions of
Ingos volountary preemption patches? What's the difference between
Ix, Jx and Mx [where x is a positiove integer number]?
A change in the letter represents a major change to the patch, and the
numbers represent bugfixes within an iteration. Ingo issues a patch,
people report any problems or latency issues, and based on this
feedback he posts a new one.
For example the -L series moves all hardirq processing to the irqd
thread. I had to hack the kernel source to make the soundcard and RTC
interrupts 'direct'. In the -M series this can all be done via /proc.
There is not really any point in using anything but the latest
version, because the patches are improving every day.
Lee
Ah, ok.. i thought they were different subsets of the functionality.
Sure, makes sense.. I use 2.6.8-rc2-M5 now and it works quite nicely..
The M5 patch against mm1 didn't really work. I got a warning at compile
time and my mouse didn't work upon bootup [ps/2].. Hmm, i just see i get
this warning with M5 applied to 2.6.8-rc2 directly, too:
CC init/main.o
init/main.c: In function `init':
init/main.c:672: warning: implicit declaration of function
`spawn_irq_threads'
Is this serious?
--
Palimm Palimm!
http://affenbande.org/~tapas/