I have always liked the 'look' of the Stonewall Attack and the concept seems simple enough, but for some reason black never rolled over and died like the examples shown in the book and after a bunch of losses, I gave it up.
Black has three good ways of defending. One is the fianchetto of his King's B and I remember seeing a game won by Euwe as black that left an impression on me that it was a good method of refuting the Stonewall. When black fianchettoes, he reduces, but does not eliminate, white's attacking potential. White's B on d3 does not have any prospects, but what white can do is develop his dark squared B with b2-b3 and Ba3 and he may be able to play, after ...c7-c5, dxc5 which makes it difficult for black to regain his P. White also has the square d4 for his N. What happens when black fianchettoes his B is that white is compelled to change his strategy and when I have met the Stonewall with black, I have found most of my opponents were unaware of this and continued with business as usual.
Black also has a couple of other ways of equalizing. One is with ...Bg4 which is often the choice of engines. The other is to play the rather surprising ...Nc6. Even though it blocks the c-Pawn after 1.d4 d5 2.e3 Nf6 3.Bd3 Nc6 Reuben Fine went so far as to give the move a “!” The move ...Nc6 was also a favorite of Chigorin way back when.
The main thing for white to remember that if black fianchettoes his KB, plays ...Bg4 or ...Nc6, white must change his strategy from the typical plan of straightway attacking black's king.
The following game featuring 3...Nc6 is very complicated; I didn't realize how complicated until I started analyzing it with Houdini, Stockfish and Komodo! It ended in a draw, but the next day Schlechter realized that he had missed a win at move 40. After white's 48.Qf7 the game was agreed drawn; probably both players assumed black, in the face of mate, had to take a perpetual check. GM Andy Soltis gives this game with brief notes in his book The Stonewall Attack and he, like Schlechter, makes no comment on the final position. Of course Soltis was writing an opening book, not a game collection. But, the engines were showing black still had a significant advantage in the final position and considerable analysis with all three engines seems to confirm that black, had he continued, most likely would have won!