> This is one of the strangest bugs I've ever seen.
> 
> If there is a suspended tail process, new tail processes don't see lines
> added to the file:

[example deleted]

I've finally had time to look at this.  It seems that tail uses mmap, but 
never calls munmap.  When you execute the second tail, it grabs the same
mmap'd region, which has not been updated with the new data.

Adding appropriate munmap()'s seems to fix the problem.  This was done in 
4.4lite.  I'll be checking in a patch tonight, and then merge with the 
4.4lite version soon.

	--jtc