Re: [PATCH libata-dev-2.6:upstream 02/02] libata: add host_set locking to ata_scsi_error()

[Jeff Garzik <jgarzik@pobox.com>]

Tejun Heo wrote:
> 02_libata_add-locking-to-ata_scsi_error.patch
> 
> 	SCSI EH can start before ata_qc_complete is completely
> 	complete.  so, latter part of ata_qc_complete can run
> 	side-by-side with ->eng_timeout(), interfering EH.
> 
>         This patch makes ata_scsi_error() to grab and release host_set
> 	lock before invoking ->eng_timeout().
> 
> 	Note: host_failed decrementing and eh_cmd_q banging are moved
> 	above ->eng_timeout() invocation such that they're done while
> 	holding the lock.

I think it's better to follow a path that sets a "EH owns this port now" 
flag inside spinlock,

	spin_lock_irqsave()
	ap->in_eh = 1;
	spin_unlock_irqrestore()

and then check that flag in ata_qc_complete()

	if (qc->ap->in_eh)
		return; /* do nothing else with command */

so that for the case of a late interrupt, the interrupt handler should 
acknowledge the interrupt, but not actually touch the ata_queued_cmd state.

	Jeff