Subject: Re: locking out device attach/detach?
To: Bill Studenmund <email@example.com>
From: Iain Hibbert <firstname.lastname@example.org>
Date: 12/07/2005 10:25:29
On Tue, 6 Dec 2005, Bill Studenmund wrote:
> On Mon, Dec 05, 2005 at 10:06:10PM +0000, Iain Hibbert wrote:
> > How should I lock out device attach/detach routines while for example
> > walking a device list that I've added to in said routines in order to
> > protect against the list being modified while I use it? (only short scans)
> Today, you don't have to. We still have one kernel biglock, so your access
> will be single-threaded.
> The right way will be some form of spinlock.
Ah, right - thanks. This means then that the kernel threads will never be
involuntarily put to sleep? although they can be interrupted I guess?
(and by soft interrupts?)
Looking at what I have written, nothing interrupt driven touches that list
so I guess it should be ok for now..
I saw some mention here recently about kernel biglock, out of interest,
would that be
a) being removed as we speak
b) undesireable but nobody working on it
c) just what we have
d) staying, we love it.
e) controversial, please dont mention it