cvs commit: src/sys/dev/syscons/apm apm_saver.c src/sys/i386/bios apm.c apm.h

[Poul-Henning Kamp]

In message <20060525.220611.74708877.imp at bsdimp.com>, Warner Losh writes:

>At the very least, we should mandate that timeouts are a non-sleepable
>event.  Sleeping just doesn't work there.  taskqueues, I'm less sure
>of, since short sleeps there work, but do degrade performance.  I like
>this idea.

A couple of years back I did a survey of our timeout callbacks because
I was worried about this issue in the context of SMPng.

The majority are "mostly OK" (there is a lot of memory allocations with
M_WAITOK) and a few are just plain wrong.

It's certainly a fixable problem.

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.