fix(hal-x86_64): don't stop pagetable walk on huge tables (##502)

Currently the code for finding or creating the next-level page table
for a page will bail out if the entry pointing to the page table is a
huge page. This is because we don't currently support *mapping* huge
pages. But, if a present page table entry set up by the bootloader
happens to be a huge page, that's fine and we can still twiddle the
mappings on that table.

This fixes the kernel crashing with the latest `bootloader` crate.

Fixes #501

hal-x86_64/src/mm.rs

@@ -298,13 +298,7 @@ impl<R: level::Recursive> PageTable<R> {
             tracing::debug!("entry not present!");
             return None;
         }
-        // XXX(eliza): should we have a different return type to distinguish
-        // between "the entry is huge, so you stop the page table walk" and "the entry is not
-        // present"? we definitely should...
-        if entry.is_huge() {
-            tracing::debug!("page is hudge!");
-            return None;
-        }
+
         let vaddr = R::Next::table_addr(idx.base_addr());
         tracing::trace!(next.addr = ?vaddr, "found next table virtual address");
         // XXX(eliza): this _probably_ could be be a `new_unchecked`...if, after
@@ -328,13 +322,6 @@ impl<R: level::Recursive> PageTable<R> {
             tracing::debug!("entry not present!");
             return None;
         }
-        // XXX(eliza): should we have a different return type to distinguish
-        // between "the entry is huge, so you stop the page table walk" and "the entry is not
-        // present"? we definitely should...
-        if entry.is_huge() {
-            tracing::debug!("page is hudge!");
-            return None;
-        }
 
         // If we are going to mutate the page table, make sure it's writable.
         if !entry.is_writable() {