On Thu, Jul 23, 2020 at 09:41:37AM -0700, Dave Hansen wrote: > On 7/23/20 9:25 AM, Sean Christopherson wrote: > > How would people feel about taking the above two patches (02 and 03 in the > > series) through the KVM tree to enable KVM virtualization of CET before the > > kernel itself gains CET support? I.e. add the MSR and feature bits, along > > with the XSAVES context switching. The feature definitons could use "" to > > suppress displaying them in /proc/cpuinfo to avoid falsely advertising CET > > to userspace. > > > > AIUI, there are ABI issues that need to be sorted out, and that is likely > > going to drag on for some time. > > > > Is this a "hell no" sort of idea, or something that would be feasible if we > > can show that there are no negative impacts to the kernel? > > Negative impacts like bloating every task->fpu with XSAVE state that > will never get used? ;) Gah, should have qualified that with "meaningful or measurable negative impacts". E.g. the extra 40 bytes for CET XSAVE state seems like it would be acceptable overhead, but noticeably increasing the latency of XSAVES and/or XRSTORS would not be acceptable. > I thought KVM had its own vcpu->arch.guest_fpu buffers which mirrored > the size and format of task->fpu. Can we have KVM support today without > task->fpu support? I see some XSS munging in the KVM code so I think > this might be *possible*, but I don't see all of the plumbing that would > make it actually work. It'd be possible, but long term I don't think it's a good idea for KVM to diverge from the kernel's FPU support, i.e. fully converting KVM to it's own implementation will likely lead to pain and maintenance problems. Without fully converting KVM to a custom implementation, adding one off support for CET would be a massive hack job.
