On Tue, Mar 31, 2020 at 7:58 AM Jarkko Sakkinen <jarkko.sakkinen@xxxxxxxxxxxxxxx> wrote: > > On Mon, Mar 30, 2020 at 05:42:29PM -0400, Nathaniel McCallum wrote: > > On Mon, Mar 30, 2020 at 4:48 PM Jarkko Sakkinen > > <jarkko.sakkinen@xxxxxxxxxxxxxxx> wrote: > > > > > > On Mon, Mar 30, 2020 at 11:08:06AM -0700, Sean Christopherson wrote: > > > > Nathaniel pointed out that __vdso_sgx_enter_enclave() is tantalizingly > > > > close to being callable from C (with caveats and a cooperative enclave). > > > > The missing pieces are preserving %rbx and taking @leaf as a standard > > > > parameter. > > > > > > > > v2: > > > > - Rebase to Jarkko's latest master, commit 402fb35a477a, "docs: ...") > > > > - Add CFI directive for RBX. [Cedric] > > > > > > I'm sorry for throwing stick's constantly but I think having a real > > > ELF loader is for better. This statement seems like you are juxtaposing having __vdso_sgx_enter_enclave() be potentially C-compatible with having an ELF-loader. These are not incompabile. __vdso_sgx_enter_enclave() can be C-callable *and* you can have an ELF loader. > > Why is it one or the other? > > Hmm... Not sure I understand what you want to know. > > A raw binary requires a hardcoded way to interpret it. Obviously ELF > is for better. We (Enarx) implement an ELF loader.
