So yeah, if an F-15 blindly stumbles within the effective range of an S400 it will be shot down.
The fingerprint of the S400's radar are detectable from thousands of kilometers away and a nice little icon will be plopped onto the pre-mission overlays before the F-15 takes off.
You don't even have to detect the radar beam, you can detect the data link.
As for casualties, what casualties?
The only casualties a drone launching a glide bomb results in are those on the receiving end.
Now this doesn't mean destroying air defenses is easy. Factor in smaller SAM sites dispersed around th larger ones, as well as enemy air superiority fighter operating in the area and SEAD gets more complicated. SAM sites also sometimes turn their radars off until planes are close and ambush them. But SAM sites alone are not an impervious defense against air power.
2. Non-SEAD RWR is generally limited to "There is light shining on me." Not where it's coming from, let alone providing targeting information.
Passive sensors have advanced significantly.
I am going to admit to moving the bar - but at this point you have MIG-31s in the air looking for SEAD and taking down any AWACS.
It’s worse than that (for the defender). Missiles can loiter and wait, if the emitter is switched off, then strike when it switches back on again.
The other bit of information that might be perhaps missing from some (other!) explanations is that this arises entirely because the ground station has to receive a radar reflection _of the original beam_ from the target. Given that power goes as 1/r^2, this means that in practice range drops off sharply for the ground-based station as 1/r^4. Meanwhile, the aircraft can just _detect_ some sort of blip-over-thermal power spectral density coming from somewhere. Given identical amplifier performance (which, with good designs, should be limited by physics) and you end up with the attacking aircraft "winning" from a back-of-the-envelope point of view.
This ignores the fact that it can be traditionally quite easy to confuse radar by "just" transmitting the "right" signals at the right time. There's a good 1960s video that explains a lot of the basic ideas here: https://www.youtube.com/watch?v=pyFqaaqqph0
You ignore the same problems with air platforms emitting.
Your claim about detecting the data link is absurd.
Even point to point microwave links can be detected off-axis.
Hundreds of watts of point-to-point means dozens of watts of sidelobes.
Dozens of watts in the middle of the desert is easy to detect.
If the AN/MLQ-40 could do it at tens of kilometers 20 years ago I bet datalinks could be detected from space today.
What part are you confused about?
There is no such thing as a perfectly directional antenna. Radiation profiles consist of multiple lobes which generally point towards a single direction on 3 axes, but there are always side lobes, like so: [1]. You can't hide these, therefore they can be detected and targeted.
There's a whole class of missiles that simply look for RF sources and head that way. The only way to avoid such missiles is to stop being an RF source. But if you do that, you're also unable to target the attackers or defend against those missiles. You also have to turn off your data link and lose situational awareness.
The S-400 is the best air defence currently fielded, but it is by no means magical or invulnerable.
https://www.cisco.com/c/dam/en/us/products/collateral/wirele...
You don't think the s400 has any jamming capability and protection of its own? You don't think the system can detect when it's being targeted and locked on?
It can definitely be defeated but not without suffering heavy casualties. That's _why_ it's called anti-access/area denial technology.