I have seen that signal info before.
I've done some more digging, and I can of course avoid the problem by just looping and using the LOCK_NB flag. I've implemented and tested that with success. The file I was locking was also on NFS, and when I tried it with a local file the problem does not exist either. This is good enough for my purposes, since I'm be using it on a local file in actual usage. My development is just done on an NFS share.
At least some NFS implementations and locks don't mix too well. Several experts consider locks on NFS at least unreliable. Maybe things changed with newer NFS versions, but I would not bet on that.
Alexander
--
Today I will gladly share my knowledge and experience, for there are no sweeter words than "I told you so". ;-)