Hi Michael, Thanks. I checked the applied v2 (60a792b97933) in your vhost branch and it looks good.
One small thing on top of it: vringh_bad() there still passes the raw desc->len, so the reported length reads wrong when the vring byte order differs from the CPU byte order. The October 21 posting had this right; the rest is fine as is. Happy to send a one-line follow-up on top of your branch if you want it; otherwise fine to leave as is. Thanks, Fang
