If I follow your question correctly(and what was posted to the GRASS maillist), then I guess you have:
- A point vector of settlements (multiple settlements)
- A second point vector of the archeology site (single point)
- A raster map where each pixel represents the time of walking thru that cell
In this case I think that r.cost (manual) and a simple r.mapcalc expression will do what you want.
First run r.cost with the walking time raster as input, use the settlements vector as start_points and the archeology site as stop_point, and call the output "accum_cost".
Then, assuming the walking time raster was in seconds, run:
r.mapcalc "within_1hr = if(accum_cost<=3600, 1, null())"
This should create a new raster such that all those areas within 1 hr walking will have value 1, and all other areas are null.
The added value of r.walk, vs the simpler r.cost, is the option to combine both an elevation layer and a friction layer. So if in your model, walking from the settlements to the site involves both climbing hills and i.e. cutting thru a jungle, then you could combine the costs of both with r.walk. You can even set the speed parameters (you've probably already seen the r.walk manual for this.)