On Fri, Jan 23, 2004 at 02:51:16PM -0500, Alan Stern wrote:
> Off the LKML now...  For those who just tuned in, this continues a 
> thread starting with
> 
> http://marc.theaimsgroup.com/?l=linux-kernel&m=107487719002239&w=2
> 
> 
> On Fri, 23 Jan 2004, Greg KH wrote:
> 
> > That patch that Alan pointed to is wrong anyway, it dies a horrible
> > death, and was only a hack to try something.
> 
> Really?  I didn't see any problems in it.  Do you remember just what was 
> wrong with the patch?

Did you try it?  It dies horribly.  Now that means either that we have
our reference counting all wrong, or the patch isn't sane.  I didn't
take the time to check out either.

> > I think it might be a better idea to just prevent the module unload of
> > the USB host controller until all drivers bound to devices owned by it
> > are also unbound in order to fix these kinds of problems.
> 
> That sounds too extreme.  How is someone supposed to unbind all the
> drivers, if they want to unload the HCD, short of unloading them too?  
> What about the hub driver bound to the virtual root hub? -- there's no way
> to unbind that.

Yeah, icky hub driver :(

> I thought the whole problem was simply that the HCD module was unloading 
> before it had finished releasing its bus.  If that is fixed, shouldn't 
> everything else be okay?

Yes, it should.  But as bug reports show, this isn't fixed yet :)

thanks,

greg k-h


-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to