Re: Please revert 5adc55da4a7758021bcc374904b0f8b076508a11 (PCI_MULTITHREAD_PROBE)

[Greg KH <greg@kroah.com>]

On Tue, May 08, 2007 at 01:58:10PM -0700, David Miller wrote:
> From: Greg KH <greg@kroah.com>
> Date: Tue, 8 May 2007 07:07:53 -0700
> 
> > Only if it ends up working properly.  The commit you reference above
> > (which removed the PCI_MULTITHREAD_PROBE option), is fine, pci
> > multi-threaded probing is still broken as it is a model that PCI drivers
> > are not yet ready to handle properly yet.
> 
> FWIW I would like to see this working properly at some point.
> 
> It seems to me that the issues are two-fold:
> 
> 1) A proper dependency system is necessary
> 
> 2) Proper mutual exclusion for shared system resources/registers/etc.
>    that are poked at in an ad-hoc unlocked manner currently
> 
> Is that basically what it boils down to?

Yes, that's about it.

Number 1 seemed to cause the most crashes, I don't think number 2 ever
caused any problems, but it might have, there were too many weird oopses
to be able to rule that out.

But in the end, I don't think that PCI really will benefit from this
speed wise, but I think the kernel overall will benefit if we can
document those dependencies somehow.

thanks,

greg k-h