On Mon, Oct 07, 2019 at 03:04:12PM +0300, Jarkko Sakkinen wrote:
> On Mon, Oct 07, 2019 at 11:10:24AM +0300, Jarkko Sakkinen wrote:
> > Actually, maybe like this:
> >
> > struct sgx_enclave_add_page_desc {
> > __u64 addr;
> > __u64 offset;
> > __u64 secinfo;
> > __u16 mrmask;
> > __u8 reserved[6];
> > };
> >
> > struct sgx_enclave_add_page {
> > __u64 src;
> > __u64 nr_pages;
> > __u64 pages;
> > };
>
> Of course we should remove @addr:
>
> struct sgx_enclave_add_page_desc {
> __u64 offset;
> __u16 mrmask;
> __u8 reserved[6];
> };
>
> struct sgx_enclave_add_page {
> __u64 src;
> __u64 secinfo;
> __u64 nr_pages;
> __u64 pages;
> };
>
> That is something we have forgot to do. We should have started to use
> offset instead of address when we moved to fd based API. Anyway I think
> this kind of API where you give array of descriptors from one source
> would be optimal.
>
> Also, @secinfo is better to be out of the descriptor so that let say
> LSM checks could be done with a single callback.
Famous last words, but hopefully I can get this to you tomorrow, as well
as the vDSO changelog rewrite.
