On Tue, Apr 02, 2013 at 09:27:57AM +1030, Rusty Russell wrote: > "Michael S. Tsirkin" <mst@xxxxxxxxxx> writes: > > Rusty's currently doing some reorgs of -net let's delay > > cleanups there to avoid stepping on each other's toys. > > Let's focus on scsi here. > > E.g. any chance framing assumptions can be fixed in 3.10? > > I am waiting for your removal of the dma-compelete ordering stuff in > vhost-net. > > Cheers, > Rusty. Now, it looks like it's actually a smart datastructure. It allows signalling consumptions from multiple without any locks, with multiple consumers, and just a single kref counter. Nothing simpler than a producer/consumer does this. Yes it can in theory delay some tx completions a bit but normally no one is waiting for them. We can refactor it to save some memory, and cleanup the code, playing with this now. _______________________________________________ Virtualization mailing list Virtualization@xxxxxxxxxxxxxxxxxxxxxxxxxx https://lists.linuxfoundation.org/mailman/listinfo/virtualization
