On Jan 4,  3:10am, smb@cs.columbia.edu ("Steven M. Bellovin") wrote:
-- Subject: Re: crash from ohci detach

| On Fri, 4 Jan 2008 02:47:50 +0000 (UTC)
| christos@astron.com (Christos Zoulas) wrote:
| 
|  
| > >
| > >That said, I think the attached patches should be committed
| > >nevertheless.  Comments?
| > 
| > Yes, if we call pmf_register we should also call pmf_deregister.
| > 
| My primary concern is the location of the call.  Since the pmf_register
| calls were at the end of the attach sequence, I put the pmf_deregister
| calls at the beginning of the detach sequence.  But what happens if
| there's a suspend event in the middle of either sequence?

Sounds right to me.
I don't think that we support suspend in the middle of attach or detach...
It is something we'll have to solve later.

christos