(ITS#5726) syncrepl must look at runqueue also after successful connection

[rein@OpenLDAP.org]

Full_Name: Rein Tollevik
Version: CVS head
OS: linux, solaris
URL: 
Submission from: (NULL) (84.215.36.97)
Submitted by: rein


The first fix to ITS#5454 in revision 1.392 appear to be wrong under curtain
situations, runqueue manipulation must be done after a paused syncrepl
connection is rescheduled.  Ref:

http://www.openldap.org/lists/openldap-devel/200810/msg00014.html

It looks as if switching from trylock to lock at the start of do_syncrepl()
should be the only fix for ITS#5454, and that the optimization this revision
introduces should be reverted.  I'll commit that shortly.

Rein Tollevik
Basefarm AS