I am just a leisure-mathematician but I almost never read any texts which do not contain exercises. Basically what I will do is to select some interesting unit (which could be a book chapter or lecture slides). First read through the exercises (respectively lecture exercises), then quickly read through the content accompanying the exercises to get an overview (no notes taken as this will just slow you down at this stage). Then I will try to do the exercises on paper. When I am happy with the solution I will type it up on LaTeX. Obviously while solving the exercises I will browse the chapter, but in a selective way. Sometimes I'll also browse other sources like Wikipedia or survey papers at this stage.

The advantage of this technique is that you will not be wasting your time for writing extensive notes of things which have already compiled in the textbook but rather stimulate your mind and put you in pseudo-research situation. In addition you create your own content, which helps to build your mathematical taste / personality.

Also a huge mistake you can make is to force yourself linearly through one or two textbooks only and not consider alternative sources. In the end the goal is to solve the exercises, get your own hands dirty and ultimately formulate your own research conjectures etc. Which resources you have used on the way is secondary.