michaelkuc6/blog_os
1
0
Fork 0
mirror of https://github.com/phil-opp/blog_os.git synced 2025-12-16 22:37:49 +00:00
Code Issues Projects Releases Wiki Activity

Describe unwrap function

Browse Source
This commit is contained in:
Philipp Oppermann
2015-12-07 15:24:23 +01:00
parent 57fbd2efc5
commit fac58a003a
1 changed files with 6 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
posts/DRAFT-paging.md
View File

@@ -664,8 +664,7 @@ pub fn identity_map<A>(&mut self,
``` ```
### Unmapping Pages ### Unmapping Pages
To unmap a page, we set the corresponding P1 entry to unused:
TODO
```rust ```rust
fn unmap<A>(&mut self, page: Page, allocator: &mut A) fn unmap<A>(&mut self, page: Page, allocator: &mut A)
@@ -680,13 +679,15 @@ fn unmap<A>(&mut self, page: Page, allocator: &mut A)
.expect("mapping code does not support huge pages"); .expect("mapping code does not support huge pages");
let frame = p1[page.p1_index()].pointed_frame().unwrap(); let frame = p1[page.p1_index()].pointed_frame().unwrap();
p1[page.p1_index()].set_unused(); p1[page.p1_index()].set_unused();
unsafe { tlb::flush(page.start_address()) };
// TODO free p(1,2,3) table if empty // TODO free p(1,2,3) table if empty
allocator.deallocate_frame(frame); allocator.deallocate_frame(frame);
} }
``` ```
TODO The assertion ensures that the page is mapped. Thus the corresponding P1 table and frame must exist for a standard 4KiB page. We set the entry to unused and free the associated frame in the supplied frame allocator.
Huge pages…
We can also free the P1, P2, or even P3 table when the last entry is freed. But checking the whole table on every `unmap` would be very expensive. For now, we leave the `TODO` in place until we find a good solution. I'm open for suggestions :).
_Spoiler_: There is an ugly bug in this function, which we will find in the next section.
## Testing it ## Testing it