Alan Cox wrote: > On Iau, 2006-05-11 at 23:39 +0900, Tejun Heo wrote: >> I'm not sure what you mean by synchronization. But, freezing is indeed >> for shutting the chip up. If the port is frozen, EH doesn't touch >> active qcs nor tries to issue new command. The port is thawed/used only >> after the port & device are put into known good state via successful >> reset. Is this what you're talking about? > > Sort of. I wanted to point out that this can occur > > IRQ is issued on chip > > ->freeze > returns > > ->intr finally arrives and is serviced I see. As far as EH is concerned, this should be okay. Once qc ownership is transferred, irq handler cannot access the qc (ata_qc_from_tag() prevents it) and has no way other than handling it as a spurious interrupt or HSM violation. But, it sounds like it could result in screaming IRQ for controllers without IRQ pending bit. Could it? If so, maybe unconditional irq clearing in ata_interrupt() should be resurrected to handle such conditions. -- tejun - : send the line "unsubscribe linux-ide" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
