[plt-scheme] thread safety of structure field mutator procedures

Section 10.1 of the reference manual [1] gives some information about
the thread safety of things like set! and vector-set!

I would like to know whether similar guarantees can be given for
structure field mutator procedures.

I am guessing that structure field mutator procedures are not
necessarily atomic, but that they boil down to a vector-set! or
something similar that is effectively atomic.

Basically I am looking for assurance that it is ok to call the accessor
procedure for a structure field while another thread may be calling the
mutator procedure.

Thanks.

David

[1] <http://docs.plt-scheme.org/reference/threads.html>