Building the graph takes time where C is the number of clauses. Each clause adds exactly two edges to the graph. SCC computation is for both Kosaraju and Tarjan. Here V = (n variables, each has two nodes) and E = (each clause makes two edges).
Checking satisfiability and constructing the assignment both take time. Total: , which is linear in the input size. That is why -SAT is tractable while general SAT is not. The restriction to two literals per clause makes all the difference.
Space complexity is for the data structures used.