David, The more I look at the patch, the more it's not done correctly. For example: - The rcv function jumps to the msg_type, and thereby jumps over the transport designator. This should be a msg_type under the generic SCSI Transport (aka SCSI_NL_TRANSPORT). We shouldn't steal msg_types from the transports. There was a layering in the design, granted it wasn't used much yet, that was subverted by this. - The message needs to be it's own message type - not munged into the FC event message. - The send function, should not be a function under the fc transport, but rather a generic function exported by the scsi netlink module. - Your recently added scsi_host_put() was very wrong. You never had a corresponding get(). I'm testing a patch which corrects the above, and adds the following: - adds generic support for any transport to receive messages, and send to pid - the shost vendor message is a now a message type of the generic scsi transport - adds checks for vendor_id to the reception of vendor message - adds event propogation to the scsi hosts too - updates shost for the netlink rcv function, event function, vendorid - Keeps the basic semantics you already had for entry into the shost. I'll post it when testing is complete. -- james s -- 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
