The Court's Opinion in Brendlin v. California
is here, via SCOTUSBlog. Justice Souter's opinion gets it right; the issue is control over the passenger, and a traffic stop controls the passenger as much as the driver.

  At least on a first read, I take the Court to hold that the seizure begins when the car comes to a stop. The passenger is seized because most car passengers in a traffic stop wouldn't initially feel free to leave at that point. Justice Souter leaves open the possibillity of a different result with different facts; Footnote 6 suggests that there may be a different result for taxi cab and bus passengers. As a I blogged about before, I tend to think this approach misses the initial brief seizure as the car is coming to rest; I think the passenger is seized in the few seconds when the car is slowing down and coming to rest as much as after the car is stopped, so that even the taxi and bus passenger would be seized for a brief interim even if they would feel free to leave after the car is stopped. But this is only a very minor quibble, and not one that will make a difference in 99.8% of the cases.