[lkml]   [2009]   [Sep]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [RFC] [PATCH 2/2] ACPI: introduce ACPI ALS device driver
On Wed, 2009-09-23 at 16:11 +0800, Jonathan Cameron wrote:
> >> Are there any plausible race conditions associated with doing the allocation this way?
> >
> > well, theoretically, yes.
> > simultaneous add calls may happen in hotplug case, but I don't think an
> > ACPI ALS device supports hotplug.
> That would make that point irrelevant!
> Thinking further on this. What happens on a box that for some reason repeatedly inserts and
> removes this module?
I don't see a problem in this case.

> This is the sort of problem idr's are meant to overcome. They will cost you a bit
> in terms of complexity though.
If we use generic names, say als0, als1, ..., it's worth using idr in
the ALS class driver.
But here, als_id is just used to fix the duplicate device name problem,
in a native ALS driver. IMO, it's overkill to implement the idr stuff.


 \ /
  Last update: 2009-09-24 04:07    [from the cache]
©2003-2011 Jasper Spaans