Implement core::iter::Step trait for for PhysFrame
#292
Conversation
|
blocked by fix linked in #294 |
|
@ncatelli can you implement this on EDIT: we've also fixed the CI issue, so you should be able to rebase and have things work |
ncatelli
force-pushed
the
feature/212/implement-step-trait-for-PhysFrame
branch
from
5e29388 to
9bdd526
Compare
ncatelli
force-pushed
the
feature/212/implement-step-trait-for-PhysFrame
branch
from
9bdd526 to
644cc68
Compare
These changes have been applied per your request. |
|
For now I've used an implementation of |
| #[cfg(feature = "step_trait")] | ||
| impl<S: PageSize> core::iter::Step for PhysFrame<S> { | ||
| fn steps_between(start: &Self, end: &Self) -> Option<usize> { | ||
| PhysAddr::steps_between(&start.start_address, &end.start_address) | ||
| } | ||
|
|
||
| fn forward_checked(start: Self, count: usize) -> Option<Self> { | ||
| PhysAddr::forward_checked(start.start_address, count) | ||
| .and_then(|start_addr| Self::from_start_address(start_addr).ok()) | ||
| } | ||
|
|
||
| fn backward_checked(start: Self, count: usize) -> Option<Self> { | ||
| PhysAddr::backward_checked(start.start_address, count) | ||
| .and_then(|start_addr| Self::from_start_address(start_addr).ok()) | ||
| } | ||
| } | ||
|
|
There was a problem hiding this comment.
These implementations are incorrect. We need to be taking the page size into account (so that we are consistent with the PhysFrame<S>: Add<u64> implementation).
Specifically, steps_between should be:
end.checked_sub(start)?.try_into() and forward_checked should be:
start.checked_add(count.try_into()?)(similar for backward_checked)
|
Oof, that was my mistake, I tried to move it quick to get feedback and didn't think it through. Sorry for the inconvenience. I agree with your above sentiment that it's it should just wait on #293 I'm going to close this out for now and will revisit this after we settle on the implementations for at least the addressing types. |
General
This is an initial naive first pass at implementing
core::iter::StepforPhysFrame. I initially had it align off the start address of a frame and look for overflows of theu64space from the start. But I can imagine it would be worth checking chat the end overflows as well and figured I'd post to get your feedback before deciding.Related Issues
#212