Messages in this thread |  | | Date | Fri, 21 Jun 2002 13:07:47 +0200 | From | Kurt Garloff <> | Subject | New SCSI exc handlers: Returning cmnds? |
| |
Hi,
converting the dc395x_trm (Tekram DC3x5 / TRM-S1040) to new exception handling, I wonder what needs to be done with queued commands when eh_abort_handler() or eh_bus_reset_handler() are called.
In the old abort handler, I was feeding the command back with DID_ABORT and call scsi_done() on it, if successful. In the old reset handler, I was feeding all commands (including the one that was passed when _reset was called) that were sitting in the host adapter driver's queue with DID_RESET back to midlayer by calling scsi_done() on them.
How should this handled in the new EH code? Is the new EH code implicitly aborting the commands, so I don't have to do it? Is the ML prepared to have scsi_done() called when doing EH? Or do I have to use some bottom half/tasklet ... type of mechanism?
Do I have to give the cmnds in queueing order, or does the ML take care of the correct ordering?
http://www.andante.org/scsi_error.html is scarily ignorant on that subject.
Thanks for advice! -- Kurt Garloff <garloff@suse.de> Eindhoven, NL GPG key: See mail header, key servers Linux kernel development SuSE Linux AG, Nuernberg, DE SCSI, Security [unhandled content-type:application/pgp-signature] |  |