There is no reason not to do that, and it does eliminate some issues. The main objection is that SMPS tend to generate a fair bit of RF noise and so are less suitable for general bench testing.
Other point is that linear PSUs tend to use transformer tappings to reduce heat dissipation. Unless you have a variable SMPS (and they are less common) then you have to put up with one input voltage to your linear reg. For a 30V 5A max output, that would mean dissipating 125W when operating at 5V, 5A. Which would call for a fairly beefy series pass.
Also, use a SMPS with a decent safety approval such as TUV. Many of the no-brand ones are substandard in this respect, and you definitely don't want a short to mains on the test bench.