Re: [PATCH 1/3] libata: implement hotplug by polling

[Tejun Heo <htejun@gmail.com>]

Jeff Garzik wrote:
> NAK, there is no reason why a global poll list is needed.  Simply having 
> a struct workqueue in each schedule-able entity implicitly creates such 
> a list.  But regardless, this is another example of adding 
> cross-controller synchronization, where none is needed.
> 
> If the worker fired per-host_set, then you could use an ata_port dynamic 
> flag to indicate the poll-active ports, for the ata_hp_poll_worker() loop.
> 
> If the worker fires per-port, then no loop or mutex is needed at all. 
> You could simply call the hp_poll hook.

It was basically an optimization to reduce storage usage and spread all 
polling operations at least by polling interval even when there are many 
ports to poll.  Cross-port synchronization aside, it's very simple too. 
  Hmm... I don't think it makes much difference either way though.  I'll 
convert to per-port workqueue + timer.

>> @@ -633,6 +637,10 @@ struct ata_port_operations {
>>      void (*error_handler) (struct ata_port *ap);
>>      void (*post_internal_cmd) (struct ata_queued_cmd *qc);
>>  
>> +    void (*hp_poll_activate) (struct ata_port *ap);
>> +    void (*hp_poll_deactivate) (struct ata_port *ap);
>> +    int (*hp_poll) (struct ata_port *ap);
>> +
>>      irqreturn_t (*irq_handler)(int, void *, struct pt_regs *);
>>      void (*irq_clear) (struct ata_port *);
>>  
> 
> All new hooks require at least a one-two sentence description in the 
> DocBook docs, telling driver writers how to use them.

Will do.

-- 
tejun