Hi Christoph, On 02.04.2015 10:29, Christoph Hellwig wrote: > [..] > For that case writing a SCSI target would be more useful I think. One > protocol that is really missing is SAS, and I think a SAS target using > libsas should be fairly doable, especially for chips like the mvsas > family that is heavily used in targets. I think, youre right, that SAS is another useful case. But for learning AoE seems more interesting for me. Additionally, the SAS part is the device-part, AoE ist the network-part. So if Ming Li want to learn the network-part, SAS is the wrong thing. > >> And also hope it could be useful. > I think a mainline ATAoE target would be very useful, but I doubt it > should be merged into LIO, there just isn't a whole lot of overlap. Ok, that is a point i dont know. Imho LIO is a framework for exporting devices over net in different ways (iscsi, fiber, aoe, udp ;-) ), so I could imagine, that AoE is a relativly easy part to implement: @Ming Lin: If you need a tester for Aoe, ich would do it. I used Aoe in different ways and would be glad to see support in LIO. It would solve different problems for me. In general, I follow the discusion for AoE in LIO for several years (2011?) and the arguments against it werent convincing to me. The only case I could is the uselessness for the data storage system of Datera, the company of Nicholas. But for Datera this could be another unique selling point, after coraid is no longer on the market. Just my 2ct Oilver -- 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
