I went to your link, and made a quick scan for "free fall" If it is defined there, I missed it. For me "free fall" in a gravity field is without any detectable force inside a box that is "free falling", until it hits the gravity source. How, other than that, do you define "free fall" ?
Free fall is the natural motion [inertial motion] of an object, described by a geodesic path, thru the gravitational field. This includes light following a natural path described by the null geodesic. Most everything in GR is about the natural path. Orbits are geodesic paths. This is a derivation of the natural path for Einstein orbits including the prediction for Mercury's natural precession. This derivation is from the Schwarzschild geometry. Newton's Law of Gravity predicts there is no natural precession and the rate of radial oscillation, w^2_r, and the rate of angular velocity, w^2_phi, is M/r^3. No natural precession. This is what Einstein predicts for w^2_r and w^2_phi including the weak field calculation for the natural precession.
Derive the natural precession rate of Einstein orbits. All Einstein orbits naturally precess.
Start with the Schwarzschild metric, in geometric units, setting theta at 0.
dTau^2 = (1-2M/r)dt^2 - dr^2/(1-2M/r) - r^2(dphi)^2
Substituting constants of geodesic motion E/m and L/m for dt and dphi
dt = [(E/m)/(1-2M/r)]dTau
dphi = [(L/m)/r^2]dTau
The solution relates squared values for radial motion (dr/dTau)^2, energy per unit mass (E/m)^2, and the effective potential per unit mass
(V/m)^2 = (1-2M/r)(1+[(L/m)^2/r^2]).
(dr/dTau)^2 = +/- (E/m)^2 - (1-2M/r)(1+[(L/m)^2/r^2])
Taking some license for the weak field and multiplying through by 1/2 after multiplying out the squared effective potential
1/2(dr/dTau)^2 = 1/2(E/m)^2 - [1/2 - M/r + (L/m)^2/2r^2 - M(L/m)^2/r^3]
setting (V/m)^2 = U/m
U/m = 1/2 - M/r + (L/m)^2/2r^2 - M(L/m)^2/r^3
1st derivative
d(U/m)/dr = M/r^2 - (L/m)^2/r^3 + 3M(L/m)^2/r^4
2nd derivative d'2(U/m)/dr'2 = rate of radial oscillation = w^2_r
w^2_r = M(r-6M)/r^3(r-3M)
Without writing down details the rate of angular velocity becomes
w^2_phi ~ (dphi/dTau)^2 = M/r^2(r-3M)
The difference.
w^2_phi - w^2_r = 6M^2/r^3(r-3M)
We can find a factor * M/r^3 which closely approximates
6M^2/r^3(r-3M)
That factor is 6M/r
(6M/r)(M/r^3) = 6M^2/r^4
The last step is further weak field approximation
(6M/r)^1/2 ~ 1/2(6M/r) = 3M/r
So a very close approximation for the rate of orbital precession, in the weak field, is 3M/r. You can plug in numbers and get an answer that matches observation.
3M_Sun = 4431m
r_mean Mercury = 5.8x10^10 meters
415.1539069 times Mercury orbits the Sun in 100 Earth years
360 degrees per year
3600 arcseconds per degree
etc...