I see a couple of things here.
1. There's nothing wrong with setting baselines for a campaign. You, the DM, pitch the concept to the players, get their buy in and we're all good to go. In my next Primeval Thule campaign, I am going for a very specific feel - very low magic. So, I flat out stated that all classes with at-will casting are off the table. Not that you can't play any sort of caster, one of the players pitched the Witch Hunter class to me and I'm groovy with two of the three subclasses.

There was some back and forth on this, but, I got the buy in from the players and so, we're all on the same page. I would be pretty annoyed if someone then pitched me a Sorcerer at this point.
Once the buy in is there from the players, generally, most problems go away.
2. Where I tend to fall off the train is that I see a lot of problems occurring AFTER character generation. Let me spin a tale. Oooh, gaming story.
Years ago, in a 2e game, I was playing a Necromancer using the Necromancer's Guide (2e). I forget the name of the kit, but, basically a mad scientist type. It was my (clearly) stated goal that I wanted this guy to be a researcher, pushing the boundaries of magic and making some new spells. I spent several levels (and many weeks of play) finding a lair, stocking it with a library and laboratory (all rules pulled from the 2e Complete Wizard) and, around sixth level, I was good to go.
I wrote up a new spell - an improved Unseen Servant (I was going with the idea that my necromancer was using spirits to do his bidding) with a bit longer duration and some minor attacks. Nothing too major. The DM was fine with it. Nothing overpowered. So, using the 2e Wizard's Guide, I calculated my chances of successfully creating a new spell after two or three weeks of in game time. Something like 30% as I recall. If you failed, you burned money and time, but you could try again.
DM: No, that's not right.
Me: I think so. See, here, I do this and this and this, yup, 30%. Can I roll?
DM: No. Creating new spells is something only very powerful wizards should do. You have a 3% chance of success.
Me: But, my entire player concept centers around the idea of making new spells... this is a major change to my character.
DM: Doesn't matter. I don't want a bunch of new spells in the game.
Me: But, you didn't have a problem with this spell five minutes ago.
Back and forth, and the DM stuck to his guns. Completely nerfed my entire concept, after weeks of play.
So, when people on online forums talk about how the DM is always right, and the campaign is 100% under the control of the DM and whatnot, it does tend to fly up my left nostril. Because, by the arguments put forward by [MENTION=6677017]Sword of Spirit[/MENTION] and others, I should have just nodded and smiled and not been the slightest put out by having a DM flush my character down the toilet. I made a character whose entire schtick was creating new spell
s. Not spend level after level trying to make a single third level spell for the entire campaign. Talk about frustrating.
I guess my basic point is, no, the campaign is not always 100% under the ownership of the DM. DM's make mistakes. DM's don't think of everything. None of us do. And putting your campaign world ahead of what the player's want is almost always (presuming good faith on the part of the player) a very bad idea. When your players are enthusiastic about an idea, run with it. If it runs roughshod over your campaign world? So what? Make up new stuff. That's what world building is all about.
My advice to DM's is to never presume that your ideas are better than those of your players.