Re: Flaw in ide_unregister()

[Richard Purdie]

Alan Cox:
> Currently there are four things in the -ac tree that use this feature
>
> 1. ISA PnP, where we don't support hotplug (and anyway the only maker of
> consumer hotpluggable ISA docking stations I know of - IBM - wont
> provide docs on them)
> 2. ide-cs
> 3. delkin (cardbus IDE)
> 4. PCI layer
>
> Of those the PCI one is a common shared function so I put the supporting
> logic in the IDE PCI helper function, The others need different handling
> at the PCMCIA or Cardbus level in order to free up resources and clean
> up.

PCI seems to use __ide_unregister_hwif() directly. Case 2 and 3 seem to need 
the "retry until the device is unregistered" behaviour and case 1 seems to 
be compatible with that.

That suggests ide_unregister_hwif() could call itself back via a work queue 
until the device unregisters successfully. Anything that doesn't want this 
behaviour can use __ide_unregister_hwif() directly as PCI does?

Richard 

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/