> From: David Woodhouse <dw...@infradead.org> > Sent: Tuesday, July 22, 2025 8:04 PM > > On Tue, 2025-07-15 at 20:35 +0800, Yi Liu wrote: > > > > David is talking about the IF and SW flags. And he is correct. Spec has > > below sentence. It means a wait descriptor can have both IF and SW set > > and indeed completion interrupt happens later than status write. Let's > > go on refine the patch. :) > > Are there any refinements you believe are necessary?
No extra change I suppose. I was just waiting for a respond from VT-d arch. I've now got a clear answer. It's ok to generate the interrupt no matter the status write failed or not if IF is set. Sorry for the delayed respond. > https://gitlab.com/dwmw2/qemu/-/pipelines/1941324674 looks like it's > going to be clean, and I think we resolved everything we talked about. Reviewed-by: Yi Liu <yi.l....@intel.com> Thanks, Yi Liu
