On 2018-05-24 16:20:33 [+0800], Jason Yan wrote: > NO, this: > > /* TODO: audit callers to ensure they are ready for qc_issue to > * unconditionally re-enable interrupts > */ > local_irq_save(flags); > spin_unlock(ap->lock); indeed, I have no idea how I could have overseen this. At least as of today, the interrupts are never "re-enabled" in this function. I *assumed* that the intention was to audit the code for this spin_unlock_irq(ap->lock); change instead. But if this is or was never intended than I could indeed remove the TODO comment. Sebastian