By: Doug S (foo.delete@this.bar.bar), October 18, 2021 2:35 pm
Room: Moderated Discussions
zArchJon (Anon.delete@this.anon.com) on October 18, 2021 10:35 am wrote:
> Speaking as someone who works a lot on defining new functions on IBM Z and having to deal
> with the complications of live migration, it is not a simple problem to deal with unless
> the application is aware something can change underneath. If every application checked proper
> CPU feature bits before using functions it would be easier, than the ones that do things
> like try an instruction and handle the interrupt if there is an illegal instruction.
I would suppose a lot of applications don't use new instructions directly, but call a library that uses vector instructions, privileged instructions aren't used directly by applications but only via a system call, etc. which would help.
> Speaking as someone who works a lot on defining new functions on IBM Z and having to deal
> with the complications of live migration, it is not a simple problem to deal with unless
> the application is aware something can change underneath. If every application checked proper
> CPU feature bits before using functions it would be easier, than the ones that do things
> like try an instruction and handle the interrupt if there is an illegal instruction.
I would suppose a lot of applications don't use new instructions directly, but call a library that uses vector instructions, privileged instructions aren't used directly by applications but only via a system call, etc. which would help.