Atmel had two families of FPGAs, the AT40K and the
AT6000. They are both
 tiny and limited by today's standards, but teh AT6000 (if I remember
 correctly) did have its configuration bits fully documented and for a
 while I considered using it in my projects for this reason. 
I assume said chips are very hard to find now. That's the other problem I
find withFPGAs and their ilk -- they're discontinued and replaced very
freqeuntly. And then you have to updata your tools for the new device and
hope it really does behave identically. At least with plain old TTL I can
still get a 14 pin chip that behaves in much the same way as the
original 7400, albeit faster and with less power consumption
  There was a project to figure out the configuration
bits for Virtex
 chips which was part of a larger project to build a PDP-10 with FPGAs. I
 don't think either was finished:
 
http://neil.franklin.ch/Projects/VirtexTools/
 Here is a series of papers about more recent efforts to generate
 configuration bits without using the vendor tools:
 Roman Lysecky, Frank Vahid, and Sheldon X.-D. Tan. 2004. Dynamic FPGA
 routing for just-in-time FPGA compilation. In Proceedings of the 41st
 annual Design Automation Conference (DAC '04). ACM, New York, NY, USA,
 954-959.
 
http://dl.acm.org/citation.cfm?id=996819&preflayout=flat
 Etienne Bergeron, Marc Feeley, and Jean Pierre David. Toward on-chip JIT
 synthesis on Xilinx VirtexII-Pro FPGAs. In International IEEE Northeast
 Workshop on Circuits and Systems (NEWCAS'07), pages 642-645, August 2007
 
http://www.iro.umontreal.ca/~feeley/papers/BergeronFeeleyDavidNEWCAS07.p
 df
 Etienne Bergeron, Marc Feeley, and Jean-Pierre David. Hardware JIT
 Compilation for Off-the-Shelf Dynamically Reconfigurable FPGAs. In
 International Conference on Compiler Construction (CC'08), volume 4959
 of Lecture Notes in Computer Science, pages 178-192, March 2008
 
http://www.iro.umontreal.ca/~feeley/papers/BergeronFeeleyDavidCC08.pdf
 Etienne Bergeron, Louis-David Perron, Marc Feeley, and Jean Pierre
 David. Logarithmic-Time FPGA bitstream analysis: A step towards JIT
 hardware compilation. ACM Transactions on Reconfigurable Technology and
 Systems, Vol 4, No 2, Article 12 (May 2011), 27 pages
 
http://dl.acm.org/citation.cfm?id=1968502.1968503&preflayout=flat 
Am I theonly person to feel it's ridiculous to have reserach papers
re-discovering information that is already known ot some people (those
who work for the FPGA manufactuer). I feel that research should solve
genuine problems that nobody knows the answer too, not artificial ones to
whci hthe anser is known but 'we won't tell you'.
-tony