On 20.11.2010, at 20:52, ronnie sahlberg wrote: > Resending since the mail might have been lost... > > List, > > > Please find attached a patch that add an initial iSCSI client library > and support to kvm-qemu. > This allows to use iSCSI devices without making them visible to the > host (and pollute the page cahce on the host) > and allows kvm-qemu to access the devices directly. Please resend with an in-line patch to qemu-devel@xxxxxxxxxxx KVM is really the wrong mailing list for this, as it only deals with KVM specifics and this falls into generic qemu code. Also, I'd recommend CC'ing kwolf@xxxxxxxxxx, as he is the block layer maintainer and this is a block driver. If you want thorough review of the iSCSI parts, I'd also recommend to CC hare@xxxxxxx and nab@xxxxxxxxxxxxxxxx > > > This works with iscsi resources, both disks and cdrom, make available > through TGTD. > > See this as experimental at this stage. > This is a very early implementation but id does works with TGTD. > Significant effort is required to add additional features to ensure > full interoperability with other initiators. > > > The syntax to describe an iscsi lun is : > iscsi://<host>[:port]/<target-name>/<lun> > > Example : > -drive file=iscsi://127.0.0.1:3260/iqn.ronnie.test/1 > or > -cdrom iscsi://127.0.0.1:3260/iqn.ronnie.test/2 > > > Future work needed : > - document this in the manpage > - implement target initiated NOP > - implement uni- and bi-directional chap > - implement a lot of missing iscsi functionality > ... > - maybe get it hooked into bus=scsi just like it today can hook into > /dev/sg* devices and send CDBs directly to the LUN, > that part of the code could be enhanced to also provide straight > passthrough SCSI to these iscsi-luns too. The whole passthrough thing should be a property of the block layer, yes. Multiple backends potentially have the chance of sending direct scsi commands somewhere and should be able to decide on that themselves. Alex -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
