Hello, again. This is part of patchset series described in [T]. This is the third take of new-EH-framework patchset. This patchset contains 11 patches. Changes from the last take[L] are * ata_eh_schedule_port() is replaced with ata_port_schedule_eh(), ata_port_abort() and ata_port_freeze(). This is cleaner and makes adding link level abort easier. * abort now uses ATA_QCFLAG_FAILED to tag a qc as failed and simply call ata_qc_complete(). This simplifies abortion. * ATA_EH_PENDING is set when EH is requested and cleared right before beginning EH. After completing EH, if the flag is still set, EH is repeated. This way no EH request is lost. Later EH implementation also includes a mechanism to avoid repeating EH actions unnecessarily in such cases. * ATA_DFLAG_FAILED removed. It's replaced with ehi->dev implemented in the EH patchset. * Port can be in FROZEN state after EH. Previously, EH framework repeated EH automatically in such cases. Now whether to leave the port frozen or not is the EH's call. (Note that user can request another shot at recovery or detection with warm plug request) * ops->thaw() added. Reset or postreset methods now don't automatically thaw the port. This change is necessary for later PM support where multiple resets are performed while frozen. It also cleans code up a bit. * Timedout qcs are now handled by ata_scsi_error() after EH kicks in. ata_scsi_timed_out() path is used only for old EH. To transfer a timedout qc's ownership to EH, it's necessary to freeze the port. Previously, the transfer happened on timeout detection in ata_scsi_timed_out() forcing other active commands to get aborted. With this change, the transfer is done after all other commands are drained. * ata_exec_internal() EH fixed/improved. As commands can be aborted during PIO, PIO task needs to be flushed whether it timed out or not. * Error messages from SCSI qc completion path is suppressed if new EH is in use. New EH itself reports better. This patchset is against upstream (acc696d93dcf993dec123d69d599979e1456ffec) + [1] prep-for-new-EH patchset -- tejun [T] http://article.gmane.org/gmane.linux.ide/9957 [L] http://article.gmane.org/gmane.linux.ide/9524 [1] http://article.gmane.org/gmane.linux.ide/9959 - : 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
