On Fri, 2019-07-12 at 12:06 -0300, Jason Gunthorpe wrote: > I wonder if the CRB in the nvs actually needs to be save/restored in > AMD's implementation? > > Sounds like the answer is to have TPM CRB somehow detect NVS and > acquire the resource from it, while somehow sequencing suspend/resume > properly. Would be cool if we would be able to get someone from AMD to give some details just to ensure that we are doing the right thing. /Jarkko
