Bugzilla

I'll jump ahead to nsPresContext.cpp:
* Its first call is in `nsPresContext::GetInProcessRootContentDocumentPresContext()`. That one's definitely fine as-is since it's in an explicitly "in-process" function.
* Its second call is in `nsPresContext::GetRootPresContext()`, which seems to be a rabbit-hole in and of itself, with [lots of callers that need to be triaged](https://searchfox.org/mozilla-central/search?q=symbol:_ZNK13nsPresContext18GetRootPresContextEv&redirect=false)
* Its third call is in nsPresContext::NotifyInvalidation, which seems to be about handling paints and invalidation.  This one looks fine as-is; I don't imagine paint-invalidations would need or want to cross process boundaries.

Besides those and comment 2, the other GetParentPresContext calls are in these files:
PresShell.cpp
ViewportUtils.cpp
nsLayoutUtils.cpp
nsGfxScrollFrame.cpp

Most of these calls look to be APZ-related, and I'm not familiar enough with our APZ code/invariants to know what to make of those.  (And I'm not sure which [if any] of those might be GeckoView-specific, which could put them out of scope for the first major fission release.)

I'll jump ahead to nsPresContext.cpp:
* Its first call is in `nsPresContext::GetInProcessRootContentDocumentPresContext()`. That one's definitely fine as-is since it's in an explicitly "in-process" function.
* Its second call is in `nsPresContext::GetRootPresContext()`, which seems to be a rabbit-hole in and of itself, with [lots of callers that need to be triaged](https://searchfox.org/mozilla-central/search?q=symbol:_ZNK13nsPresContext18GetRootPresContextEv&redirect=false)
* Its third call is in `nsPresContext::NotifyInvalidation`, which seems to be about handling paints and invalidation.  This one looks fine as-is; I don't imagine paint-invalidations would need or want to cross process boundaries.

Besides those and comment 2, the other GetParentPresContext calls are in these files:
PresShell.cpp
ViewportUtils.cpp
nsLayoutUtils.cpp
nsGfxScrollFrame.cpp

Most of these calls look to be APZ-related, and I'm not familiar enough with our APZ code/invariants to know what to make of those.  (And I'm not sure which [if any] of those might be GeckoView-specific, which could put them out of scope for the first major fission release.)