Re: [PATCH] x86/svm: Merge hsa and host_vmcb to reduce memory overhead

2020-10-28 Thread Jan Beulich
On 27.10.2020 20:30, Andrew Cooper wrote: > On 27/10/2020 15:24, Jan Beulich wrote: >> On 26.10.2020 14:50, Andrew Cooper wrote: >>> The format of the Host State Area is, and has always been, a VMCB. It is >>> explicitly safe to put the host VMSAVE data in. >> Nit: The PM calls this "Host Save Are

Re: [PATCH] x86/svm: Merge hsa and host_vmcb to reduce memory overhead

2020-10-27 Thread Andrew Cooper
On 27/10/2020 15:24, Jan Beulich wrote: > On 26.10.2020 14:50, Andrew Cooper wrote: >> The format of the Host State Area is, and has always been, a VMCB. It is >> explicitly safe to put the host VMSAVE data in. > Nit: The PM calls this "Host Save Area" or "Host State Save Area" > afaics. > > I rec

Re: [PATCH] x86/svm: Merge hsa and host_vmcb to reduce memory overhead

2020-10-27 Thread Jan Beulich
On 26.10.2020 14:50, Andrew Cooper wrote: > The format of the Host State Area is, and has always been, a VMCB. It is > explicitly safe to put the host VMSAVE data in. Nit: The PM calls this "Host Save Area" or "Host State Save Area" afaics. I recall us discussing this option in the past, and not