See Silverfox's
alkaline battery shootout thread and examine the curves for D cells.
Yes, D alkalines can deliver some decent current. But Siverfox's 3A discharge curve shows that alkalines are very poor at holding their voltage under high current drain. A D alkaline will quickly sag below 1V after around 10 min of run time.
Do you know what your LED's forward voltage is at 1A drive or 3A drive? If the forward voltage is high, then you typically need more input voltage from your battery stack to get the driver to run in full regulation. It's a circuit thing.
Direct drive is certainly simple and with 3xD alkalines, you have very little risk of overdriving your P7. The downside is that the output is unregulated and the light will start out bright, sag very quickly, and continually head downhill as the cells drain.
If you use a 7135-based linear regulator at 2.8A output, I would use a 4D Mag, not a 3D, if you plan to use only D alkaline cells. The 3D will work, but the run time in full regulation will be short (probably less than 5 min). The 4D probably can hold full regulation for closer to 20-30 min. Alternatively, you could get a 1050mA 7135 driver and a 3D Mag probably can run in full regulation for close to 100 min.
The problem with your DX buck driver is that the driver requires a certain minimum input voltage to run in full regulation (meaning it delivers the constant 925mA that it is design to provide). At 925mA, and Seoul P7 LED could have a forward voltage of about 3.25V (based on the P7 datasheet). That's why it helps to know what the actual forward voltage of your LED is at the driver's rated drive current of 925mA. That DX driver typically needs about 0.6V of extra voltage above the LED's forward voltage to reach full regulation. The lower your LED forward voltage is, the lower the required input voltage to reach full regulation. Based on the datasheet, you may need at least 3.85V input from your battery stack, or close to 1.3V per cell. Of course, a real flashlight system has parasitic resistances in the circuit arising from things like cell-to-cell contacts, spring contacts, and switch contacts. So you might need more like 4.0V input, or around 1.35V per cell.
So one possibility to explain why your battery stack is delivering only 0.6A is that you are not running in full regulation and the combination of battery sag and system resistance results in what you get.
Another possibility is that maybe you do have a very low forward voltage P7. Let's say it is 2.9V at 925mA drive current from the driver. If the driver is 85% efficient as a guess, then based on power delivered to the driver = power output to the LED,
Vbatt*Ibatt*driver efficiency = Vf*If
Let's assume Vbatt is about 1.35V (I picked off an early time voltage from Silverfox's 1A draw curve for D cells since I assume that you made your measurement right at the beginning, and not an hour after turning on the light). Vf is 2.9V. If is 0.925A.
Thus,
(3*1.35V)*Ibatt*0.85 = 2.9V*0.925A
Ibatt=(2.9V*0.925A)/(3*1.35V*0.85) = 0.78A
So we are in the ballpark. However, it's tough get to a calculated 0.6A, which is what you measured. I'd have to increase the battery voltage to 1.4V per cell, drop the LED forward voltage to 2.8V, and increase the driver efficiency to 90%. And I still get an Ibatt of 0.66A. Thus, to get to 0.6A, I think that I'd have to assume some unreasonable set of parameters, meaning most likely the reason your battery current draw is so low is probably because of 3xD alkalines aren't delivering enough voltage to reach full regulation. Your light is essentially running in direct drive and delivering 600mA of drive current to the LED, instead of the full 925mA.
So to use that DX driver effectively, I would suggest driving it with 4xD alkalines.
Otherwise, I'd switch to a 7135-based driver that delivers 1050mA drive current (such a driver will have four 7135 chips on-board, each delivering 350mA nominal) and use that in your 3D Mag mod. I think that higher drive currents are not practical if you plan to use only alkaline cells.