sgx: avoid unnecessarily creating a slice

[RalfJung]

Cc @jethrogb -- no idea why this created a slice only to directly convert it back to a raw pointer, but we can avoid this and in fact make the entire function safe. I didn't change the function signature (it's still an unsafe fn) as I know nothing about the surrounding code.

[rustbot]

r? @ChrisDenton

rustbot has assigned @ChrisDenton.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[jethrogb]

no idea why

ptr::slice_from_raw_parts_mut was added in 8b21b07. This code was added in 1e44e2d, which predates that.

[jethrogb]

make the entire function safe

So the Safety documentation section of the trait method seems to imply that the following invariant must be upheld

• T: UserSafe implies that the memory location of T is in user (not enclave) memory.

However, I think that may be incorrect. I think a similar invariant is upheld by the UserRef and User types. Also UserSafe is implemented for regular primitive Rust types (like u8), so it sounds weird that implementing that trait somehow adds this kind of restriction.

[RalfJung]

It is also generally odd to have a safety comment on a trait impl that's not the same as the one in the trait itself -- callers that use a generic T: UserSafe will not know about the specific requirement of this impl.

Anyway I won't be conducting a refactor of that trait here, I just noticed this little improvement inside the slice impl. I will leave follow-up cleanup to people that understand the surrounding context better :)