The patent is publicly available. There's a link right in the summary. The application was confidential, but it was confidential for less time (7 months) than patent applications are by default (18 months) before it was granted. Damages for patent infringement can include time during which the application was pending, but only if the infringer has notice of the application. Since the application was confidential, that doesn't apply here.

It's pretty much a patent on software. Software for controlling a vehicle, but software nonetheless. Claim 15, for example:

This claim format is known as a Beauregard claim. Such claims have long been used as one of several ways of claiming software-implemented inventions, though their long-term viability is somewhat suspect (hence all the "tangible non-transitory" hedging language). The other claims are to a method (also basically software) and to a vehicle running the software on a control module (since Google didn't invent the car, that's basically another software claim).

What Google did was make a request for confidentiality under 35 U.S.C. 122(b)(2)(B). Normally patents are kept confidential for 18 months after filing, whereupon they are published by the Patent Office. A 122(b)(2)(B) request keeps the application confidential so long as the applicant doesn't file in another country or make a Patent Cooperation Treaty filing that requires publication at 18 months. Essentially it's confidentiality at the expense of not being able to go international. In this case, it wouldn't have made a difference because the patent issued about 7 months after filing. Even without the 122(b)(2)(B) request the application would not have been published before the patent issued.

Apparently you don't understand how summary judgment works. Summary judgment will not be granted if there is a genuine dispute as to a material fact (i.e. a fact that could affect the outcome of the case). The purpose of summary judgment is to resolve issues of law that would not be given to the jury, or to save the time and effort of a trial if the facts are undisputed. Whether the claims are patent eligible under 101 is an issue of law that would not be submitted to the jury in any case.

The Monsanto cases typically involved farmers who intentionally harvested and planted GMO seeds, not accidental pollination. The most famous such case occurred in Canada, by the way, not the US.

Apple can now argue that the patent has already made it through two reviews by patent examiners and that further evaluation is unnecessary.

That's not the standard under which reexamination is made. Reexamination requires a "substantial new question of patentability." As long as a substantial new question is raised, it doesn't matter how many times it's been examined in the past.

Further, a reissue is not a reexamination. Examination is only made with regard to the amendments, not to the patent as a whole.

Why reissue it?

Apple filed the reissue application in order to amend the claims. It appears that some small, slightly narrowing tweaks were made to add references to "beacons." The reissue does not affect the term of the patent. Check out 35 U.S.C. 251 for more details. Notably, if anyone did anything that would infringe the reissued claims but didn't infringe the original claims, then they can continue doing so. This can even apply to taking substantial preparation towards something that would infringe. In other words, a reissue is not retroactive, per 35 U.S.C. 252. This is a protection against broadening reissue. In this case, though, it looks like the amendments narrowed the claims rather than broaden them.

You do realize that that patent cites as prior art numerous patents describing linked lists and other data structures, right? The patent doesn't claim a plain, CS101 linked list. The idea is a linked list with two pointers per node such that the pointers describe two different sequences (e.g. in a list of customer records, one set of pointers traverses the list according to last name and the other according to first name).