Audit problematic `(bits + 7) / 8` pattern more and fix it more nicely

This is a follow-up issue to this discussion: #15879 (comment)

Here's an example of this possibly problematic pattern that is found in many places in the compiler and the std:

            const bytes = try self.readBytesNoEof((@typeInfo(T).Int.bits + 7) / 8);

The type of the bits integer is very relevant here. @typeInfo(T).Int.bits is of type u16. If the bits value could be 65535 (note the integer types with the greatest bit size are u65535/i65535) and the bit size of the bits int type is less than 17, there is a problem.
The fix could be:

-            const bytes = try self.readBytesNoEof((@typeInfo(T).Int.bits + 7) / 8);
+            const bytes = try self.readBytesNoEof(@intCast(u16, (@as(u17, @typeInfo(T).Int.bits) + 7) / 8));

However @jacobly0 discovered a slightly nicer pattern which is std.math.divCeil(bits, 8). I think we might want to do that instead of (bits + 7) / 8 or @intCast(u16, (@as(u17, bits) + 7) / 8). I think it's easier to read.

So the goal is to replace the existing pattern with that and also fix more of these issues because I think there are more.

Here are two commits that fix some of these issues for more reference:

• 4ac6e38
• 3fa310d

A blocker for this might be #16022 if we decide to do @divCeil(bits, 8) everywhere which would be even nicer.