Rework __vdso_sgx_enter_enclave() to move all input/output params, except
for pass-through GPRs, into a single struct. With the new struct, add
a pass-through param requested by Nathaniel[1], fix a long-standing nit
from Andy[2], and add a flags field to allow for future extensions.
1. Add an opaque param to pass data from the runtime to its handler.
https://lkml.kernel.org/r/CAOASepOFh-vOrNZEVDFrDSuHs+9GEzzpXUTG-fZMuyjWAkpRWw@xxxxxxxxxxxxxx
2. Use a dedicated exit reason instead of using -EFAULT for "exception"
(and effectively -EINTR for interrupts, too).
https://lkml.kernel.org/r/90D05734-1583-4306-A9A4-18E4A1390F3B@xxxxxxxxxxxxxx
RFC->V1:
- Drop the EXIT_ON_INTERRUPT patch. [Andy]
- Fix the macro names in the assembly code. [Jarkko]
- Move the leaf back into the exception sub-struct. The leaf is fully
redundant with SGX_SYNCHRONOUS_EXIT.
- Add selftest support. [Jarkko]
Jarkko, I didn't address you comment about moving the vDSO kernel docs
comments to the .rst file because I have an question on that and didn't
want to hold this up. But Intel's mail servers appear to be on the fritz,
so it might be a moot point...
Sean Christopherson (6):
x86/vdso: x86/sgx: Explicitly force 8-byte CMP for detecting user
handler
x86/vdso: x86/sgx: Rework __vdso_sgx_enter_enclave() API
x86/vdso: x86/sgx: Introduce dedicated SGX exit reasons for vDSO
selftests/sgx: Update the SGX selftest to match the reworked vDSO API
selftests/sgx: Sanity check the return value of the vDSO call
selftests/sgx: Add a smoke test to ensure the user handler is invoked
arch/x86/entry/vdso/vsgx_enter_enclave.S | 76 +++++++++++++-------
arch/x86/include/uapi/asm/sgx.h | 88 ++++++++++++++++--------
tools/testing/selftests/sgx/call.S | 10 +--
tools/testing/selftests/sgx/main.c | 61 ++++++++++++----
tools/testing/selftests/sgx/main.h | 2 +-
5 files changed, 160 insertions(+), 77 deletions(-)
--
2.28.0
