I'd like to get some comments on the approach on how to move error handling out from the transport processing thread. Just like I/O submission all EH callers actually are from process context and can block. But given how long EH can take I'd rather not push it into that context as it might end up blocking other LUs from making progress. It seems like having a dedicated EH context for each LU, just like we do in the Linux SCSI midlayer would be appronpinquate here. What makes our life even easier is that there is only one TMF implemented in the target core right now, the LUN reset. Given the nature of a LUN reset there is no real reason to ever process multiple of those at the same time. Can we simply fail a second LUN RESET while the first one is in progress, or do we have to complete them all once the reset is done? -- To unsubscribe from this list: send the line "unsubscribe target-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
