Christophe de Dinechin

> [AMD APM Vol 2](https://www.amd.com/system/files/TechDocs/24593.pdf) section 15.34.6 Page Table Support says: > > > Note that because guest physical addresses are always translated through the nested page tables, the >...

@larrydewey Responding after being ping'ed on Slack. Sorry for the delay, KVM Forum + waiting for more replies to my email. FWIW, I have only received one reply to my...

I had a direct discussion with @larrydewey, where he shared with me some exchanges with @jpecholt where she was sharing various problems she ran into on SNP systems. The issues...

@tlendacky Thank you for your response. I think you seem to imply that a guest can use PTE / TLB bits that are marked reserved in the host. I was...

The last paragraph of 15.34.6 indeed seems to confirm @tlendacky's interpretation. Lifting my request for change.

I can't say that I really like the approach of building without seccomp by default, and require active changes in all downstream consumers to re-enable it. This is likely to...

Hi @gulamf. I think that you are right. Good catch. I will fix that when I return from KVM Forum.

@ddebroy This is very interesting. I added an issue on the Kata side to make sure we track this. Are you aware of any similar effort for CNI? I'm asking...

Possibly related is that when @sameo pulled a more recent version in [kata containers PR #2332](https://github.com/kata-containers/kata-containers/pull/2332/commits), this seems to have introduced a regression in the systemd case. From a discussion...

> I would intuitively think a contradictory definition like above would simply equate goedel to something like the "never"/! type in rust; a type to which no value belongs. Not...