Add uefi::boot module #1255
Add uefi::boot module #1255
Conversation
nicholasbishop
force-pushed
the
bishop-add-boot-mod
branch
from
d986905 to
172b40d
Compare
uefi/src/boot.rs
Outdated
| ty: AllocateType, | ||
| mem_ty: MemoryType, | ||
| count: usize, | ||
| ) -> Result<PhysicalAddress> { |
There was a problem hiding this comment.
Seems like an API inconsistency with allocate_pool, which returns Result<NonNull<u8>>.
There was a problem hiding this comment.
That's true. This inconsistency is present in our BootServices methods and in the UEFI spec; AllocatePool produces a void* pointer, while AllocatePages produces a EFI_PHYSICAL_ADDRESS. Notably, EFI_PHYSICAL_ADDRESS is always a UINT64 rather than being pointer sized, so there is a real difference in behavior. The docstring notes this in the context of ia32 PAE (I copied it from the BootServices method, the text was originally added in this commit).
That said, I think in almost all cases the first thing a caller is going to do with the EFI_PHYSICAL_ADDRESS is cast it to a pointer, and I can't currently think of any real-world counter example. So, while I don't feel very strongly in either direction, I don't mind converting this to return NonNull<u8> instead of PhysicalAddress.
There was a problem hiding this comment.
Updated. For full consistency, I also changed both free functions to use NonNull<u8> as well.
nicholasbishop
force-pushed
the
bishop-add-boot-mod
branch
from
172b40d to
f6a2dbd
Compare
The initial version contains allocate_pages/allocate_pool and the corresponding `free` functions. Unlike the `BootServices` methods, these functions all use `NonNull<u8>` for consistency.
phip1611
force-pushed
the
bishop-add-boot-mod
branch
from
f6a2dbd to
f1c9f28
Compare
The initial version contains
allocate_pages/allocate_pooland the correspondingfreefunctions.Checklist