On Wed, Jul 25, 2018 at 04:39:28PM +0200, Igor Mammedov wrote:
> Fixed dynamic tables are usually more or less self sufficient
> so it's safer to reload so I'm ok with their reloading.
Well it's just a matter of excercising self-control and
building sane APIs to make sure we do.
E.g. let's say we agree on prefix DQ for dynamic tables.
Now, isn't
Name(DQCS, Package() {...} )
which is constrained not to use any names from the main
table not self-sufficient?
To ensure this contraint, we can add APIs like dynamic_name
that add the "DQ" prefix, and we can verify that a dynamic
table does not use a name without this prefix.
> As for current cpuhp interface, it was designed to be extendable
> and we can extend it for CST. It most likely would be much less
> code to extend than it is in this series.
I do have patches that use cpuhp to signal CST changes.
I don't think cpuhp solves the mess that is dynamic
package generation though.
--
MST
