Re: [PATCH v17 22/23] x86/sgx: SGX documentation

[Jarkko Sakkinen]

On Sun, Dec 02, 2018 at 07:28:55PM -0800, Randy Dunlap wrote:
> Hi,
> I have more editing comments below.
> 
> 
> On 11/15/18 5:01 PM, Jarkko Sakkinen wrote:
> > Documentation of the features of the Software Guard eXtensions used
> > by the Linux kernel and basic design choices for the core and driver
> > and functionality.
> > 
> > Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@xxxxxxxxxxxxxxx>
> > Co-developed-by: Sean Christopherson <sean.j.christopherson@xxxxxxxxx>
> > Signed-off-by: Sean Christopherson <sean.j.christopherson@xxxxxxxxx>
> > ---
> >  Documentation/index.rst         |   1 +
> >  Documentation/x86/index.rst     |   8 ++
> >  Documentation/x86/intel_sgx.rst | 233 ++++++++++++++++++++++++++++++++
> >  3 files changed, 242 insertions(+)
> >  create mode 100644 Documentation/x86/index.rst
> >  create mode 100644 Documentation/x86/intel_sgx.rst
> 
> > diff --git a/Documentation/x86/intel_sgx.rst b/Documentation/x86/intel_sgx.rst
> > new file mode 100644
> > index 000000000000..f51b43f9e125
> > --- /dev/null
> > +++ b/Documentation/x86/intel_sgx.rst
> > @@ -0,0 +1,233 @@
> > +===================
> > +Intel(R) SGX driver
> > +===================
> > +
> > +Introduction
> > +============
> > +
> > +Intel(R) SGX is a set of CPU instructions that can be used by applications to
> > +set aside private regions of code and data. The code outside the enclave is
> > +disallowed to access the memory inside the enclave by the CPU access control.
> > +In a way you can think that SGX provides inverted sandbox. It protects the
> 
>                                    provides an inverted sandbox.
> 
> > +application from a malicious host.
> > +
> > +You can tell if your CPU supports SGX by looking into ``/proc/cpuinfo``:
> > +
> > +	``cat /proc/cpuinfo  | grep sgx``
> > +
> > +Overview of SGX
> > +===============
> > +
> > +SGX has a set of data structures to maintain information about the enclaves and
> > +their security properties. BIOS reserves a fixed size region of physical memory
> > +for these structures by setting Processor Reserved Memory Range Registers
> > +(PRMRR).
> > +
> > +This memory range is protected from outside access by the CPU and all the data
> > +coming in and out of the CPU package is encrypted by a key that is generated for
> > +each boot cycle.
> > +
> > +Enclaves execute in ring-3 in a special enclave submode using pages from the
> 
>                        ring 3
> 
> > +reserved memory range. A fixed logical address range for the enclave is reserved
> > +by ENCLS(ECREATE), a leaf instruction used to create enclaves. It is referred in
> 
>                                                                         referred to in
> 
> > +the documentation commonly as the ELRANGE.
> > +
> > +Every memory access to the ELRANGE is asserted by the CPU. If the CPU is not
> > +executing in the enclave mode inside the enclave, #GP is raised. On the other
> > +hand, enclave code can make memory accesses both inside and outside of the
> > +ELRANGE.
> > +
> > +Enclave can only execute code inside the ELRANGE. Instructions that may cause
> 
>    An enclave can only
> 
> > +VMEXIT, IO instructions and instructions that require a privilege change are
> > +prohibited inside the enclave. Interrupts and exceptions always cause enclave
> 
>                                                             always cause an enclave
> 
> > +to exit and jump to an address outside the enclave given when the enclave is
> > +entered by using the leaf instruction ENCLS(EENTER).
> > +
> > +Protected memory
> > +----------------
> > +
> > +Enclave Page Cache (EPC)
> > +    Physical pages used with enclaves that are protected by the CPU from
> > +    unauthorized access.
> > +
> > +Enclave Page Cache Map (EPCM)
> > +    A database that describes the properties and state of the pages e.g. their
> > +    permissions or to which enclave they belong to.
> 
> Drop one of those "to" words (either one).
> 
> > +
> > +Memory Encryption Engine (MEE) integrity tree
> > +    Autonomously updated integrity tree. The root of the tree located in on-die
> > +    SRAM.
> > +
> > +EPC data types
> > +--------------
> > +
> > +SGX Enclave Control Structure (SECS)
> > +    Describes the global properties of an enclave. Will not be mapped to the
> > +    ELRANGE.
> > +
> > +Regular (REG)
> > +    These pages contain code and data.
> > +
> > +Thread Control Structure (TCS)
> > +    The pages that define the entry points inside an enclave. An enclave can
> > +    only be entered through these entry points and each can host a single
> > +    hardware thread at a time.
> > +
> > +Version Array (VA)
> > +   The pages contain 64-bit version numbers for pages that have been swapped
> > +   outside the enclave. Each page has the capacity of 512 version numbers.
> > +
> > +Launch control
> > +--------------
> > +
> > +To launch an enclave, two structures must be provided for ENCLS(EINIT):
> > +
> > +1. **SIGSTRUCT:** signed measurement of the enclave binary.
> > +2. **EINITTOKEN:** a cryptographic token CMAC-signed with a AES256-key called
> 
>                                                         with an
> 
> > +   *launch key*, which is re-generated for each boot cycle.
> 
> (prefer)                     regenerated
> 
> > +
> > +The CPU holds a SHA256 hash of a 3072-bit RSA public key inside
> > +IA32_SGXLEPUBKEYHASHn MSRs. Enclaves with a SIGSTRUCT that is signed with this
> > +key do not require a valid EINITTOKEN and can be authorized with special
> > +privileges. One of those privileges is ability to acquire the launch key with
> > +ENCLS(EGETKEY).
> > +
> > +**IA32_FEATURE_CONTROL[17]** is used by the BIOS configure whether
> 
>                                         by the BIOS to configure whether
> 
> > +IA32_SGXLEPUBKEYHASH MSRs are read-only or read-write before locking the
> > +feature control register and handing over control to the operating system.
> > +
> > +Enclave construction
> > +--------------------
> > +
> > +The construction is started by filling out the SECS that contains enclave
> > +address range, privileged attributes and measurement of TCS and REG pages (pages
> > +that will be mapped to the address range) among the other things. This structure
> > +is passed out to the ENCLS(ECREATE) together with a physical address of a page
> 
> This would make more sense to me:
> 
>    is passed to the ENCLS(ECREATE) instruction together with ...
> 
> > +in EPC that will hold the SECS.
> > +
> > +The pages are added with ENCLS(EADD) and measured with ENCLS(EEXTEND) i.e.
> 
>                                                      with ENCLS(EEXTEND), i.e.
> 
> > +SHA256 hash MRENCLAVE residing in the SECS is extended with the page data.
> > +
> > +After all of the pages have been added, the enclave is initialized with
> > +ENCLS(EINIT). It will check that the SIGSTRUCT is signed with the contained
> > +public key. If the given EINITTOKEN has the valid bit set, the CPU checks that
> > +the token is valid (CMAC'd with the launch key). If the token is not valid,
> > +the CPU will check whether the enclave is signed with a key matching to the
> > +IA32_SGXLEPUBKEYHASHn MSRs.
> > +
> > +Swapping pages
> > +--------------
> > +
> > +Enclave pages can be swapped out with ENCLS(EWB) to the unprotected memory. In
> > +addition to the EPC page, ENCLS(EWB) takes in a VA page and address for PCMD
> > +structure (Page Crypto MetaData) as input. The VA page will seal a version
> > +number for the page. PCMD is 128 byte structure that contains tracking
> 
>                                 128-byte

Is having a space instead of dash always grammatically wrong or is this
just to have a coherent style? Just asking for plain curiosity...

> 
> > +information for the page, most importantly its MAC. With these structures the
> > +enclave is sealed and rollback protected while it resides in the unprotected
> > +memory.
> > +
> > +Before the page can be swapped out it must not have any active TLB references.
> > +ENCLS(EBLOCK) instruction moves a page to the *blocked* state, which means
> 
>  The ENCLS(EBLOCK) instruction
> 
> > +that no new TLB entries can be created to it by the hardware threads.
> > +
> > +After this a shootdown sequence is started with ENCLS(ETRACK), which sets an
> > +increased counter value to the entering hardware threads. ENCLS(EWB) will
> > +return SGX_NOT_TRACKED error while there are still threads with the earlier
> > +couner value because that means that there might be hardware thread inside
> 
>    counter                                                      threads
> 
> 
> > +the enclave with TLB entries to pages that are to be swapped.
> > +
> > +Kernel internals
> > +================
> > +
> > +Requirements
> > +------------
> > +
> > +Because SGX has an ever evolving and expanding feature set, it's possible for
> > +a BIOS or VMM to configure a system in such a way that not all CPUs are equal,
> > +e.g. where Launch Control is only enabled on a subset of CPUs.  Linux does
> > +*not* support such a heterogeneous system configuration, nor does it even
> > +attempt to play nice in the face of a misconfigured system.  With the exception
> > +of Launch Control's hash MSRs, which can vary per CPU, Linux assumes that all
> > +CPUs have a configuration that is identical to the boot CPU.
> > +
> > +
> > +Roles and responsibilities
> > +--------------------------
> > +
> > +SGX introduces system resources, e.g. EPC memory, that must be accessible to
> > +multiple entities, e.g. the native kernel driver (to expose SGX to userspace)
> > +and KVM (to expose SGX to VMs), ideally without introducing any dependencies
> > +between each SGX entity.  To that end, the kernel owns and manages the shared
> > +system resources, i.e. the EPC and Launch Control MSRs, and defines functions
> > +that provide appropriate access to the shared resources.  SGX support for
> > +user space and VMs is left to the SGX platform driver and KVM respectively.
> > +
> > +Launching enclaves
> > +------------------
> > +
> > +The current kernel implementation supports only unlocked MSRs i.e.
> 
>                                                             MSRs, i.e.
> 
> > +FEATURE_CONTROL_SGX_LE_WR must be set. The launch is performed by setting the
> > +MSRs to the hash of the public key modulus of the enclave signer, which is one
> > +of the fields in the SIGSTRUCT.
> > +
> > +EPC management
> > +--------------
> > +
> > +Due to the unique requirements for swapping EPC pages, and because EPC pages
> > +(currently) do not have associated page structures, management of the EPC is
> > +not handled by the standard Linux swapper.  SGX directly handles swapping
> > +of EPC pages, including a kthread to initiate reclaim and a rudimentary LRU
> > +mechanism. The consumers of EPC pages, e.g. the SGX driver, are required to
> > +implement function callbacks that can be invoked by the kernel to age,
> > +swap, and/or forcefully reclaim a target EPC page.  In effect, the kernel
> > +controls what happens and when, while the consumers (driver, KVM, etc..) do
> > +the actual work.
> > +
> > +Exception handling
> > +------------------
> > +
> > +The PF_SGX bit is set if and only if the #PF is detected by the SGX Enclave Page
> > +Cache Map (EPCM). The EPCM is a hardware-managed table that enforces accesses to
> > +an enclave's EPC pages in addition to the software-managed kernel page tables,
> > +i.e. the effective permissions for an EPC page are a logical AND of the kernel's
> > +page tables and the corresponding EPCM entry.
> > +
> > +The EPCM is consulted only after an access walks the kernel's page tables, i.e.:
> > +
> > +1. the access was allowed by the kernel
> > +2. the kernel's tables have become less restrictive than the EPCM
> > +3. the kernel cannot fixup the cause of the fault
> > +
> > +Noteably, (2) implies that either the kernel has botched the EPC mappings or the
> 
>    Notably,
> 
> > +EPCM has been invalidated (see below).  Regardless of why the fault occurred,
> > +userspace needs to be alerted so that it can take appropriate action, e.g.
> > +restart the enclave. This is reinforced by (3) as the kernel doesn't really
> > +have any other reasonable option, i.e. signalling SIGSEGV is actually the least
> > +severe action possible.
> > +
> > +Although the primary purpose of the EPCM is to prevent a malicious or
> > +compromised kernel from attacking an enclave, e.g. by modifying the enclave's
> > +page tables, do not WARN on a #PF w/ PF_SGX set.  The SGX architecture
> 
>                                      with
> 
> > +effectively allows the CPU to invalidate all EPCM entries at will and requires
> > +that software be prepared to handle an EPCM fault at any time.  The architecture
> > +defines this behavior because the EPCM is encrypted with an ephemeral key that
> > +isn't exposed to software.  As such, the EPCM entries cannot be preserved across
> > +transitions that result in a new key being used, e.g. CPU power down as part of
> > +an S3 transition or when a VM is live migrated to a new physical system.
> > +
> > +SGX uapi
> 
>        UAPI
> 
> > +========
> > +
> > +.. kernel-doc:: drivers/platform/x86/intel_sgx/sgx_ioctl.c
> > +   :functions: sgx_ioc_enclave_create
> > +               sgx_ioc_enclave_add_page
> > +               sgx_ioc_enclave_init
> > +
> > +.. kernel-doc:: arch/x86/include/uapi/asm/sgx.h
> > +
> > +References
> > +==========
> > +
> > +* A Memory Encryption Engine Suitable for General Purpose Processors
> > +  <https://eprint.iacr.org/2016/204.pdf>
> > +* System Programming Manual: 39.1.4 Intel® SGX Launch Control Configuration
> 
> 
> ciao.
> -- 
> ~Randy

Great, thanks Randy, highly appreciated!

/Jarkko