This is the repost of the two patches I posted in earlier this month: - [PATCH 1/2] libsas: remove irq save in sas_ata_qc_issue() Received feedback but nothing really changed. I explained that this is not about "local_irqsave() + spin_lock()" *but* "local_irq_save() + spin_unlock()". This seemed to have been overseen twice. Also there were two opinions about the TODO comment: /* TODO: audit callers to ensure they are ready for qc_issue to * unconditionally re-enable interrupts It is unclear to me if this comment should be removed because it makes no sense or if the intention was indeed to audit callers code for a possible "spin_unlock_irq(ap->lock);". - [PATCH 2/2] qla2xxx: remove irq save in qla2x00_poll() Received no feedback. Sebastian