[PATCH RFC] seccomp: Implement syscall isolation based on memory areas

[Paul Gofman]

On 5/31/20 20:31, Matthew Wilcox wrote:
> If it's the cost of the syscall that's the problem, there are ways
> around that.  We'd still want a personality() call to indicate that
> the syscall handler should look (somewhere) to determine the current
> personality, but that could be issued at the start of execution rather
> than when we switch between Windows & Linux code.

Sure, we can call personality() at start and specify the location to
look at, the only thing is that the location should be thread specific,
that is, based on fs: or gs: or whatever else which would allow us to
have different threads in different "personality" state. If anything
needs to be set up at thread start we can do that also of course.

If there will be any proof of concept solution I will be happy to make a
proof of concept Wine patch using that and do some testing.