Mark Haverkamp wrote: > > The recent dv changes to scsi_transport_spi.c have fixed the problems > that I had with my U160 drives being set to async/narrow but there are > still problems with my U320 drive causing indefinite dv on my system. > mptscsih: ioc0: attempting task abort! (sc=f734bc80) > command: cdb[0]=0x12: 12 00 00 00 60 00 > mptbase: Initiating ioc0 recovery > jejb - The problem is the task abort occurs as a result of the factor 0x8 max speed inquiry test timing out to the end device. The task abort results in a host reset. The host reset results in domain validation running again. The thread running the host reset returns control back to the device that failed the max speed inquiry test and exits dv; meanwhile the worker thread is still running and requesting domain validation on all devices, to include the one that failed. When get back to the failed device, the previous dv run had already completed, and we are running dv all over again the same device, which will timeout the inquiry test. I had already provided a patch that will fix this problem on the mailing list. The idea is we don't issue domain validation again if were in host reset and when some device is in the middle of domain validation. Instead we send previous nego parameters out to devices that have been already configured. This needs to be fixed one way or another. If mid layer didn't issue host reset when command timed out, such as a target reset, we wouldn't encounter this. Eric Moore - To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
