Commit Graph

97 Commits

Author SHA1 Message Date
Rob Bradford
2cc312b65c build: Fix cargo fuzz build
Unfortunately it seems patch entries are ignored when obtaining
dependencies from another workspace.

Remove the problematic kvm-ioctls and kvm-bindings patch entries and use
the forked repository directly.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-12-04 11:38:44 +00:00
Wei Liu
90ae4676c8 hypervisor: x86: emulator: drop unneeded curly brackets
There is no need to have a pair of curly brackets for structures without
any member.

No functional change.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
2020-12-03 14:52:00 +01:00
Wei Liu
c352b59857 hypervisor: emulator: drop InstructionMap
It is not needed anymore.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
2020-12-02 19:45:36 +01:00
Wei Liu
ab89b48143 hypervisor: emulator: use static mapping to dispatch emulation
The mapping between code and its handler is static. We can drop the
HashMap in favour of a static match expression.

This has two benefits:
1. No more memory allocation and deallocation for the HashMap.
2. Shorter look-up time.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
2020-12-02 19:45:36 +01:00
Rob Bradford
ffaab46934 misc: Use a more relaxed memory model when possible
When a total ordering between multiple atomic variables is not required
then use Ordering::Acquire with atomic loads and Ordering::Release with
atomic stores.

This will improve performance as this does not require a memory fence
on x86_64 which Ordering::SeqCst will use.

Add a comment to the code in the vCPU handling code where it operates on
multiple atomics to explain why Ordering::SeqCst is required.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-12-02 19:04:30 +01:00
Samuel Ortiz
15f0451c34 hypervisor: emulator: Format instructions on error paths
Formatting instructions might be costly, so we only want to do that on
the emulation error paths.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-12-01 11:13:54 +01:00
Samuel Ortiz
3fa6e17cc4 hypervisor: x86: Make imm_op available to all emulated instructions
By moving it to the instructions root module.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-12-01 11:13:54 +01:00
Wei Liu
e2c81f9ed8 hypervisor: x86: Extend the imm_op() macro
To support every kind of immediate operands.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-12-01 11:13:54 +01:00
Wei Liu
ba71390d6a hypervisor: drop arc-swap dependency
It is no longer needed after 0fec32658.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
2020-11-30 16:47:32 +00:00
Samuel Ortiz
f6892f2ede hypervisor: emulator: Add a negative instruction fetch unit test
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-11-30 13:09:19 +00:00
Samuel Ortiz
251434862f hypervisor: emulator: Fix logic bug in MockVMM
emulate_first_insn() really means we want only the first instruction to
be emulated.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-11-30 13:09:19 +00:00
Samuel Ortiz
1ac4c42a84 hypervisor: emulator: Add initial fetch support
When the x86 instruction decoder tells us about some missing bytes from
the instruction stream, we call into the platform fetch method and
emulate one last instruction.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-11-30 13:09:19 +00:00
Samuel Ortiz
7dc92d1dac hypervisor: emulator: Decoding loop restructuring
In preparation for the instruction fetching step, we modify the decoding
loop so that we can check what the last decoding error is.

We also switch to explictly using decode_out() which removes a 32 bytes
copy compared to decode().

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-11-30 13:09:19 +00:00
Samuel Ortiz
7a4edecd29 hypervisor: x86: Add an address linearization method to CpuStateManager
From a CPU state and a segment, we can translate a logical (segmented)
address into a linear one.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-11-30 13:09:19 +00:00
Samuel Ortiz
d419e30df1 hypervisor: x86: Add a SegmentRegistorOps trait
In order to validate emulated memory accesses, we need to be able to get
all the segments descriptor attributes.

This is done by abstracting the SegmentRegister attributes through a
trait that each hypervisor will have to implement.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-11-30 13:09:19 +00:00
Samuel Ortiz
3faffcb087 hypervisor: emulator: Add a CPU mode getter to CpuStateManager
We need to be able to build the CPU mode from its state in order to
start implementing mode related checks in the x86 emulator.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-11-30 13:09:19 +00:00
Samuel Ortiz
b3a1f5f1be hypervisor: Start moving register definitions to the hypervisor crate
Most of arch/src/x86_64/regs.rs will eventually move unde
hypervisor/src/x86/regs.rs.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-11-30 13:09:19 +00:00
Samuel Ortiz
a3d957ba4d hypervisor: emulator: Clean the MockVMM initialization path
Separate the standalone _init_and_run() function into 2 MockVMM methods:
new() and emulate_insn().

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-11-27 19:17:35 +01:00
Samuel Ortiz
5bd63efa45 hypervisor: emulator: Move MockVMM into a shared test module
The MockVMM platform will be used by other instructions emulation
implementations, but also by the emulator framework.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-11-27 19:17:35 +01:00
Wei Liu
66b001665f hypervisor: x86: only calculate address when necessary in MOV emulation
Only calculate the address when the operand is memory.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
2020-11-26 16:12:23 +01:00
Wei Liu
c6aea5af5d hypervisor: x86: drop an extraneous box indirection
There is no need to put a box into another box.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
2020-11-26 12:02:47 +00:00
Wei Liu
c8b655490e hypervisor: x86: reference PlatformEmulator in Emulator
The observation here is PlatformEmulator can be seen as the context for
emulation to take place. It should be rather easy to construct a context
that satisfies the lifetime constraints for instruction emulation.

The thread doing the emulation will have full ownership over the
context, so this removes the need to wrap PlatformEmulator in Arc and
Mutex, as well as the need for the context to be either Clone or Copy.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
2020-11-26 12:14:40 +01:00
Wei Liu
a6ad85dc40 hypervisor: x86: handle more registers in emulator
List all GPRs documented in Intel's SDM.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
2020-11-25 20:18:40 +01:00
Samuel Ortiz
0f4f30dbde hypervisor: x86: Add MOV to the emulator instruction map
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-11-25 17:02:11 +01:00
Samuel Ortiz
fe5401223b hypervisor: x86: Emulate MOV
And add a few unit tests based on a Mock platform.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-11-25 17:02:11 +01:00
Samuel Ortiz
12b7a498d9 hypervisor: x86: Add an instruction emulator
The emulator gets a CPU state from a CpuStateManager instance, emulates
the passed instructions stream and returns the modified CPU state.

The emulator is a skeleton for now since it comes with an empty
instruction mnemonic map.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-11-25 17:02:11 +01:00
Samuel Ortiz
1fc97e91a4 hypervisor: x86: Add an InstructionHandler interface
And an InstructionMap helper structure to map x86 mnemonic codes
to instruction handlers.

Any instruction emulation implementation should then boil down with
implementing InstructionHandler for any supported mnemonic.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-11-25 17:02:11 +01:00
Samuel Ortiz
fc5d6c96be hypervisor: x86: Add a minimal CpuStateManager implementation
Minimal will be defined by the amount of emulated instructions.
Carrying all GPRs, all CRs, segment registers and table registers should
cover quite a few instructions.

Co-developed-by: Wei Liu <liuwe@microsoft.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-11-25 17:02:11 +01:00
Samuel Ortiz
546778ebfb hypervisor: x86: Add a CpuStateManager interface
For efficiently emulating x86 instructions, we need to build and pass a
CPU state copy/reference to instruction emulation handlers. Those handlers
will typically modify the CPU state and let the caller commit those
changes back through the PlatformEmulator trait set_cpu_state method.

Hypervisors typically have internal CPU state structures, that maps back
to the correspinding kernel APIs. By implementing the CpuState trait,
instruction emulators will be able to directly work on CPU state
instances that are directly consumable by the underlying hypervisor and
its kernel APIs.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-11-25 17:02:11 +01:00
Samuel Ortiz
f0360aff83 hypervisor: Architecture agnostic instruction emulation interface
In order to emulate instructions, we need a way to get access to some of
the guest resources. The PlatformEmulator interface provides guest
memory and CPU state access to emulator implementations.

Typically, an hypervisor will implement PlatformEmulator for architecture
specific instruction emulators to build their framework on top of.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-11-25 17:02:11 +01:00
Samuel Ortiz
ae96aeda65 arch: Move the gdt module to the hypervisor crate
We will need the GDT API for the hypervisor's x86 instruction
emulator implementation, it's better if the arch crate depends on the
hypervisor one rather than the other way around.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-11-25 17:02:11 +01:00
Samuel Ortiz
1fc0b94fcd cargo: Move to crates.io vm-memory 0.4.0
vm-memory 0.4.0 now contains all our fixes.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-11-23 10:55:13 +01:00
Rob Bradford
0fec326582 hypervisor, vmm: Remove shared ownership of VmmOps
This interface is used by the vCPU thread to delegate responsibility for
handling MMIO/PIO operations and to support different approaches than a
VM exit.

During profiling I found that we were spending 13.75% of the boot CPU
uage acquiring access to the object holding the VmmOps via
ArcSwap::load_full()

    13.75%     6.02%  vcpu0            cloud-hypervisor    [.] arc_swap::ArcSwapAny<T,S>::load_full
            |
            ---arc_swap::ArcSwapAny<T,S>::load_full
               |
                --13.43%--<hypervisor::kvm::KvmVcpu as hypervisor::cpu::Vcpu>::run
                          std::sys_common::backtrace::__rust_begin_short_backtrace
                          core::ops::function::FnOnce::call_once{{vtable-shim}}
                          std::sys::unix:🧵:Thread:🆕:thread_start

However since the object implementing VmmOps does not need to be mutable
and it is only used from the vCPU side we can change the ownership to
being a simple Arc<> that is passed in when calling create_vcpu().

This completely removes the above CPU usage from subsequent profiles.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-11-19 00:16:02 +01:00
Rob Bradford
041724a7cf hypervisor: Add ability to get dirty logged pages
Return a bitmap of pages that have been dirtied (written to) since it
was last called.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-11-17 16:57:11 +00:00
Rob Bradford
8baa244ec1 hypervisor: Add control for dirty page logging
When creating a userspace mapping provide a control for enabling the
logging of dirty pages.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-11-17 16:57:11 +00:00
Rob Bradford
c0827e01b1 build: Update arc-swap dependency to 1.0.0
Also bump the vm-memory version to one that depends on arc-swap 1.0.0
too.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-11-16 14:10:09 +01:00
Rob Bradford
46e736973e hypervisor: kvm: Correctly share VmmOps between Kvm{Vm,Vcpu}
Cloning the ArcSwapOption (like the ArcSwap) does not act like a
.clone() on an Arc, instead an entirely new ArcSwap is created with the
same contents. To correctly share the ArcSwap needs to be placed inside
an Arc.

See: 2433d5719b (diff-6c6d94533c44c19bd1416ef17bad1a878e63dca6e98d59181228fbe8f967c62bR6)

Due to this being wrongly used ::clone() was removed from
ArcSwap/ArcSwapOption in 1.0.0.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-11-16 14:10:09 +01:00
Michael Zhao
093a581ee1 vmm: Implement VM rebooting on AArch64
The logic to handle AArch64 system event was: SHUTDOWN and RESET were
all treated as RESET.

Now we handle them differently:
- RESET event will trigger Vmm::vm_reboot(),
- SHUTDOWN event will trigger Vmm::vm_shutdown().

Signed-off-by: Michael Zhao <michael.zhao@arm.com>
2020-10-30 17:14:44 +00:00
dependabot-preview[bot]
f0d0d8ccaf build(deps): bump libc from 0.2.79 to 0.2.80
Bumps [libc](https://github.com/rust-lang/libc) from 0.2.79 to 0.2.80.
- [Release notes](https://github.com/rust-lang/libc/releases)
- [Commits](https://github.com/rust-lang/libc/compare/0.2.79...0.2.80)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-26 09:31:03 +00:00
Sebastien Boeuf
28e12e9f3a vmm, hypervisor: Fix snapshot/restore for Windows guest
The snasphot/restore feature is not working because some CPU states are
not properly saved, which means they can't be restored later on.

First thing, we ensure the CPUID is stored so that it can be properly
restored later. The code is simplified and pushed down to the hypervisor
crate.

Second thing, we identify for each vCPU if the Hyper-V SynIC device is
emulated or not. In case it is, that means some specific MSRs will be
set by the guest. These MSRs must be saved in order to properly restore
the VM.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-21 19:11:03 +01:00
Muminul Islam
5867a1af4a hypervisor: Update Author and license info in the toml file
Signed-off-by: Muminul Islam <muislam@microsoft.com>
2020-10-20 11:51:32 +01:00
Rob Bradford
c4dc25de09 hypervisor: kvm: aarch64: Trigger reset upon KVM_SYSTEM_EVENT_RESET
This will trigger Vm::vm_reboot to make the VM reboot.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-20 12:46:35 +08:00
Rob Bradford
573a5c63cf hypervisor: kvm: Use unstable_sort() to keep clippy happy
"Using a stable sort consumes more memory and cpu cycles. Because values
which compare equal are identical, preserving their relative order (the
guarantee that a stable sort provides) means nothing, while the extra
costs still apply."

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-09 10:49:54 +02:00
Sebastien Boeuf
2fac41f154 deps: Patch kvm-ioctls and kvm-bindings crates
Instead of having the hypervisor crate embedding Cloud-Hypervisor forks
from the rust-vmm project, it's more appropriate to leave the rust-vmm
references in the hypervisor crate, and have the root Cargo.toml being
patched.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-08 18:45:42 +02:00
Wei Liu
ed1fdd1f7d hypervisor, arch: rename "OneRegister" and relevant code
The OneRegister literally means "one (arbitrary) register". Just call it
"Register" instead. There is no need to inherit KVM's naming scheme in
the hypervisor agnostic code.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
2020-10-08 08:55:10 +02:00
Wei Liu
9ad14e6b3a aarch64: Add OneReg to the list required extensions for KVM
Without that capability save / restore for aarch64 won't work.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
2020-10-08 08:54:38 +02:00
dependabot-preview[bot]
c2cc26fc82 build(deps): bump libc from 0.2.78 to 0.2.79
Bumps [libc](https://github.com/rust-lang/libc) from 0.2.78 to 0.2.79.
- [Release notes](https://github.com/rust-lang/libc/releases)
- [Commits](https://github.com/rust-lang/libc/compare/0.2.78...0.2.79)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-05 07:02:05 +00:00
Praveen Paladugu
71c435ce91 hypervisor, vmm: Introduce VmmOps trait
Run loop in hypervisor needs a callback mechanism to access resources
like guest memory, mmio, pio etc.

VmmOps trait is introduced here, which is implemented by vmm module.
While handling vcpuexits in run loop, this trait allows hypervisor
module access to the above mentioned resources via callbacks.

Signed-off-by: Praveen Paladugu <prapal@microsoft.com>
Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-02 16:42:55 +01:00
Rob Bradford
6a9934d933 build: Fix vm-memory bump build error
A new version of vm-memory was released upstream which resulted in some
components pulling in that new version. Update the version number used
to point to the latest version but continue to use our patched version
due to the fix for #1258

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-02 16:38:02 +01:00
dependabot-preview[bot]
76c3230e08 build(deps): bump libc from 0.2.77 to 0.2.78
Bumps [libc](https://github.com/rust-lang/libc) from 0.2.77 to 0.2.78.
- [Release notes](https://github.com/rust-lang/libc/releases)
- [Commits](https://github.com/rust-lang/libc/compare/0.2.77...0.2.78)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-01 05:40:02 +00:00