There could happen a deadlock when two threads change for example
message info flags, where one threads holds message info's summary lock
and is waiting for the info's property lock, which the other thread
hold's info's property lock and is waiting for the summary lock.
It's safer to hold summary lock before acquiring the message info's
lock when anything will be changed in the message info, which didn't
happen here.

This had been reported downstream at:
https://bugzilla.redhat.com/show_bug.cgi?id=1649373