Re: [uml-user] Could UML run via seccomp traps instead of ptrace?

Richard Weinberger <[email protected]> Mon, 22 May 2017 10:27:00 +0200
Newsgroups gmane.linux.uml.user
Message-ID <CAFLxGvyj8iDw2c=j+x7RGp5e99Ms0X0tBZAZ4KOa5WoBywXU0g@mail.gmail.com>
Dan,

On Mon, May 22, 2017 at 12:39 AM, Dan Kaminsky <[email protected]> wrote:
> The thinking is we'd add another SECCOMP_RET type, such that UML didn't need
> to depend on ptrace.  Ptrace isn't the fastest mechanism, and beyond that,
> if we can avoid it it's a lot easier to syscall firewall UML as a whole (and
> freeze/restore it with CRIU).

Beside of PTRACE_SYSEMU UML also needs ptrace() for ton of other things,
i.e. register read/restore or copy_from/to_user().
So, we won't get rid of ptrace() completely.

I'm not sure whether switching from PTRACE_SYSEMU to seccomp will speed things
up. Do you have some numbers?

-- 
Thanks,
//richard

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot