Re: [PATCH 07/13] libata-hp: implement unload-unplug

[Jeff Garzik <jgarzik@pobox.com>]

Tejun Heo wrote:
> Jeff Garzik wrote:
>> Tejun Heo wrote:
>>> Implement unload unplug - driver unloading / PCI removal via hot
>>> unplug path.  This is done by ata_port_detach() which requests detach
>>> of all devices, schedules EH and wait for it to complete.  EH path is
>>> slightly modified to handle this (e.g. force zero eh_info during
>>> unloading).  With this patch, EH and unloading are properly
>>> synchronized and unloading should be safe under any circumstances.
>>>
>>> Signed-off-by: Tejun Heo <htejun@gmail.com>
>>
>> How does this behave for controller hotplug?  For that case, there is 
>> nothing to EH...
> 
> I've only tested with module unloading but when EH is invoked for 
> unplugging, autopsy and report are skipped and recover is called with 
> zero ehi but with all DFLAG_DETACH set.  This amounts to unconditional 
> detachment of all devices without any other operation (well, of course, 
> other than being synchronized to other EH invocations).  So, I guess it 
> should work okay.  Can you tell me how to test such condition?  I guess 
> I can yank out my sil3132 and unload sata_sil24.  Would that be enough?

ISTR there is some sysfs trigger for PCI eject, that will allow you test 
this.

	Jeff