To prevent people from modernizing the loop, add a comment that
fixupImmediateBr can append to ImmBranches.

Revoke the change in https://github.com/llvm/llvm-project/pull/146198

https://github.com/llvm/llvm-project/pull/146198 changes
```
    for (unsigned i = 0, e = ImmBranches.size(); i != e; ++i)
      BRChange |= fixupImmediateBr(ImmBranches[i]);
```
to
```
    for (ImmBranch &Br : ImmBranches)
      BRChange |= fixupImmediateBr(Br);
```
Unfortunately, they are not NFC and cause the buildbot error. e.g.,
https://lab.llvm.org/buildbot/#/builders/24/builds/9943
https://lab.llvm.org/buildbot/#/builders/169/builds/12570
Use make_early_inc_range to fix the issue

The map variable `BlockJumpTableRefCount` was added in commit
f5f28d5b0ce76af8f6944774aa73bad9e328b020 to track whether a basic block
was the target of any jump table entries. This was used in the function
`fixupBTI` to insert and remove BTIs after jump tables had been
modified.

Commit 3b742242a53ed0c2a2e1b6bb2352cace43c22030 removed `fixupBTI` on
the grounds that the work was now being done elsewhere. That left
`BlockJumpTableRefCount` still being created, but now nothing is using
it. So we can garbage-collect that variable and all the code that
populates it.

The MS bit-field packing ABI depends on the storage size of the type of
being placed in the bit-field. This PR addresses a number of cases in
llvm where the storage type has lead to suboptimal packing.

This reverts commit e48916f615e0ad2b994b2b785d4fe1b8a98bc322.

MinNoSplitDisp was first introduced in D16890 to handle cases where the
ConstantIslands pass fails to converge in the presence of big basic
blocks. However, the computation of the variable seems to be wrong as it
currently computes the offset immediately following UserBB. In other
words, it represents the distance from the beginning of the function to
the end of UserBB. The distance from the beginning of the function does
not seem to be a good indicator of how big the basic block is unless the
basic block is close to the beginning of the function. I think
MinNoSplitDisp should compute the distance between UserOffset and the
end of UserBB instead.

Identified with misc-include-cleaner.