Compare commits

..

208 Commits

Author SHA1 Message Date
Rob Bradford
7dcd3aff05 build, release-notes.md: Document 0.11.0 release
Update release notes and version number for the new release.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-29 17:36:02 +00:00
Bo Chen
8c268e351c tests: Port more tests to use wait_vm_boot
Tests not ported include 1) the ones that start guest VMs without
network (e.g. test_net_hotplug, test_initramfs), 2) test_vfio that
involves l2 guest. Also, some tests that use bionic guest image are
given extended timeout (120s) for 'wait_vm_boot'.

Signed-off-by: Bo Chen <chen.bo@intel.com>
2020-10-29 15:30:17 +00:00
Bo Chen
a79f058a90 tests: Accept custom timeout for wait_vm_boot
This allows known slow tests to have a longer/customized timeout while
keeping the default timeout short.

Signed-off-by: Bo Chen <chen.bo@intel.com>
2020-10-29 15:30:17 +00:00
Bo Chen
26783fea89 tests: Wait explicitly for vm shutdown w/ the 'wait-timeout' crate
Instead of waiting blindly with fixed amount of sleeping time, we can
use the `wait-timeout` crate to explicitly wait VM shutdown (with a
timeout). It can reduces the execution time of some tests
substantially. Also, this patch increases the `shutdown` timeout for
'test_reboot', which should fix the recent sporadic failures on this
test.

Signed-off-by: Bo Chen <chen.bo@intel.com>
2020-10-29 15:30:17 +00:00
Bo Chen
cd1c2ed31e build: Add the 'v' prefix when using the crate version
Signed-off-by: Bo Chen <chen.bo@intel.com>
2020-10-29 08:19:25 -07:00
dependabot-preview[bot]
b947320feb build(deps): bump memchr from 2.3.3 to 2.3.4
Bumps [memchr](https://github.com/BurntSushi/rust-memchr) from 2.3.3 to 2.3.4.
- [Release notes](https://github.com/BurntSushi/rust-memchr/releases)
- [Commits](https://github.com/BurntSushi/rust-memchr/compare/2.3.3...2.3.4)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-29 09:24:32 +00:00
Rob Bradford
d54544046c build: Use stable virtio-fs QEMU branch
Address build failure from activity in the development virtio-fs branch
by using the stable fork.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-29 08:52:25 +00:00
Bo Chen
8c92d1dbdc tests: Wait explicitly for the guest vm to boot
Instead of blindly waiting for 20-40s for the guest VM to boot, this
patch waits the notification from the guest VM explicitly by using a
simple TcpListener on the host and a custom systemd service in the
guest.

This patch also ported few tests to use this new machanism, while more
tests are to be ported.

Signed-off-by: Bo Chen <chen.bo@intel.com>
2020-10-28 11:27:25 -07:00
dependabot-preview[bot]
cc67a05474 build(deps): bump proc-macro-hack from 0.5.18 to 0.5.19
Bumps [proc-macro-hack](https://github.com/dtolnay/proc-macro-hack) from 0.5.18 to 0.5.19.
- [Release notes](https://github.com/dtolnay/proc-macro-hack/releases)
- [Commits](https://github.com/dtolnay/proc-macro-hack/compare/0.5.18...0.5.19)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-28 08:04:11 +00:00
Rob Bradford
21db6f53c8 vmm: memory_manager: Write all guest region to disk
As a mirror of bdbea19e23 which ensured
that GuestMemoryMmap::read_exact_from() was used to read all the file to
the region ensure that all the guest memory region is written to disk.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-27 12:11:31 -07:00
Rob Bradford
be1b6bc1e1 main: Remove API socket when exiting
When exiting remove the API socket from the filesystem.

Fixes: #1241

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-27 13:27:23 +00:00
Rob Bradford
c22b788b47 main: Simplify error and return handling in start_vmm
Use a Result<> type with an error to simplify the code in start_vmm().
This will also make it easier to add cleanup funtionality.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-27 13:27:23 +00:00
Rob Bradford
dfd21cbfc5 vmm: Use thiserror/anyhow for vmm::Error
This gives a nicer user experience and this error can now be used as the
source for other errors based off this.

See: #1910

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-27 13:27:23 +00:00
Rob Bradford
6d4656c68f virtio-devices: seccomp_filters: Add fsync to block io_uring filter
This is required when booting with hypervisor-fw.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-27 10:53:54 +00:00
Rob Bradford
338521c700 build: Enable io_uring feature by default
Now that we that our CI is running with a kernel that is new enough to
support io_uring we can turn this feature on by default.

See: #1561

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-27 10:53:54 +00:00
Muminul Islam
2e1866a5db scripts: dev_cli: Add an option to dev_cli to export volumes
Signed-off-by: Muminul Islam <muislam@microsoft.com>
2020-10-27 09:52:35 +00:00
Rob Bradford
8bd3fdea8f build: Move on-demand workers from bionic to groovy
This is an interim step to be able to test io_uring on our CI system.

See: #1561

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-26 14:27:20 -07:00
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
dependabot-preview[bot]
0415146018 build(deps): bump syn from 1.0.46 to 1.0.48
Bumps [syn](https://github.com/dtolnay/syn) from 1.0.46 to 1.0.48.
- [Release notes](https://github.com/dtolnay/syn/releases)
- [Commits](https://github.com/dtolnay/syn/compare/1.0.46...1.0.48)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-24 21:58:39 +00:00
dependabot-preview[bot]
499fbd0367 build(deps): bump futures from 0.3.6 to 0.3.7
Bumps [futures](https://github.com/rust-lang/futures-rs) from 0.3.6 to 0.3.7.
- [Release notes](https://github.com/rust-lang/futures-rs/releases)
- [Changelog](https://github.com/rust-lang/futures-rs/blob/master/CHANGELOG.md)
- [Commits](https://github.com/rust-lang/futures-rs/compare/0.3.6...0.3.7)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-24 00:08:41 +00:00
dependabot-preview[bot]
e70674943f build(deps): bump futures-sink from 0.3.6 to 0.3.7
Bumps [futures-sink](https://github.com/rust-lang/futures-rs) from 0.3.6 to 0.3.7.
- [Release notes](https://github.com/rust-lang/futures-rs/releases)
- [Changelog](https://github.com/rust-lang/futures-rs/blob/master/CHANGELOG.md)
- [Commits](https://github.com/rust-lang/futures-rs/compare/0.3.6...0.3.7)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-23 20:00:07 +00:00
Sebastien Boeuf
7e127df415 vmm: memory_manager: Replace 'ext_region' by 'saved_region'
Any occurrence of of a variable containing `ext_region` is replaced with
the less confusing name `saved_region`. The point is to clearly identify
the memory regions that might have been saved during a snapshot, while
the `ext` standing for `external` was pretty unclear.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-23 21:59:52 +02:00
Sebastien Boeuf
c0e8e5b53f vmm: memory_manager: Replace 'backing_file' variable names
In the context of saving the memory regions content through snapshot,
using the term "backing file" brings confusion with the actual backing
file that might back the memory mapping.

To avoid such conflicting naming, the 'backing_file' field from the
MemoryRegion structure gets replaced with 'content', as this is
designating the potential file containing the memory region data.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-23 21:59:52 +02:00
dependabot-preview[bot]
1409b36cad build(deps): bump futures-task from 0.3.6 to 0.3.7
Bumps [futures-task](https://github.com/rust-lang/futures-rs) from 0.3.6 to 0.3.7.
- [Release notes](https://github.com/rust-lang/futures-rs/releases)
- [Changelog](https://github.com/rust-lang/futures-rs/blob/master/CHANGELOG.md)
- [Commits](https://github.com/rust-lang/futures-rs/compare/0.3.6...0.3.7)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-23 18:58:11 +00:00
dependabot-preview[bot]
cf067d94c3 build(deps): bump futures-io from 0.3.6 to 0.3.7
Bumps [futures-io](https://github.com/rust-lang/futures-rs) from 0.3.6 to 0.3.7.
- [Release notes](https://github.com/rust-lang/futures-rs/releases)
- [Changelog](https://github.com/rust-lang/futures-rs/blob/master/CHANGELOG.md)
- [Commits](https://github.com/rust-lang/futures-rs/compare/0.3.6...0.3.7)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-23 18:57:45 +00:00
dependabot-preview[bot]
c377c1398f build(deps): bump futures-macro from 0.3.6 to 0.3.7
Bumps [futures-macro](https://github.com/rust-lang/futures-rs) from 0.3.6 to 0.3.7.
- [Release notes](https://github.com/rust-lang/futures-rs/releases)
- [Changelog](https://github.com/rust-lang/futures-rs/blob/master/CHANGELOG.md)
- [Commits](https://github.com/rust-lang/futures-rs/compare/0.3.6...0.3.7)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-23 18:57:26 +00:00
dependabot-preview[bot]
b64a48adb8 build(deps): bump arrayvec from 0.5.1 to 0.5.2
Bumps [arrayvec](https://github.com/bluss/arrayvec) from 0.5.1 to 0.5.2.
- [Release notes](https://github.com/bluss/arrayvec/releases)
- [Changelog](https://github.com/bluss/arrayvec/blob/master/CHANGELOG.md)
- [Commits](https://github.com/bluss/arrayvec/commits)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-23 17:27:29 +00:00
Rob Bradford
bdbea19e23 vmm: memory_manager: Completely fill guest ram from snapshot
Use GuestRegionMmap::read_exact_from() to ensure that all of the file is
read into the guest. This addresses an issue where
GuestRegionMmap::read_from() was only copying the first 2GiB of the
memory and so lead to snapshot-restore was failing when the guest RAM
was 2GiB or greater.

This change also propagates any error from the copying upwards.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-23 17:56:19 +01:00
Rob Bradford
8de3bd728c ch-remote, api_client: Split HTTP/API client code into new crate
Split out the HTTP request handling code from ch-remote into a new
crate which can be used in other places where talking to the API server
by HTTP is necessary.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-23 14:50:36 +02:00
Rob Bradford
654846f433 bin: ch-remote: Generalise functions that work on a HTTP socket
Use the trait for Read/Write rather than specifying the concrete type.
This allows for the functionality to be used for different socket types.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-23 14:50:36 +02:00
Rob Bradford
a60b437f89 vmm: memory_manager: Always copy anonymous RAM regions from disk
When restoring if a region of RAM is backed by anonymous memory i.e from
memfd_create() then copy the contents of the ram from the file that has
been saved to disk.

Previously the code would map the memory from that file into the guest
using a MAP_PRIVATE mapping. This has the effect of
minimising the restore time but provides an issue where the restored VM
does not have the same structure as the snapshotted VM, in particular
memory is backed by files in the restored VM that were anonymously
backed in the original.

This creates two problems:

* The snapshot data is mapped from files for the pages of the guest
  which prevents the storage from being reclaimed.
* When snapshotting again the guest memory will not be correctly saved
  as it will have looked like it was backed by a file so it will not be
  written to disk but as it is a MAP_PRIVATE mapping the changes will
  never be written to the disk again. This results in incorrect
  behaviour.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-23 12:34:32 +02:00
Wei Liu
255dbd29ef tests: factor out windows_auth
This avoid repetition once we have more Windows tests.

No functional change.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
2020-10-22 18:23:08 +02:00
Wei Liu
2902a96ae6 tests: add Windows guest snapshot test
We also need to restrict the number of test threads to 1 to avoid tests
interfere with each other.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
2020-10-22 18:23:08 +02:00
Sebastien Boeuf
89b3d7b49b ci: Update integration tests with latest balloon changes
Now that virtio-balloon is not declared as part of the --memory
parameter, the integration tests are updated to keep the correct
behavior.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-22 16:33:16 +02:00
Sebastien Boeuf
f4e391922f vmm: Remove balloon options from --memory parameter
The standalone `--balloon` parameter being fully functional at this
point, we can get rid of the balloon options from the --memory
parameter.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-22 16:33:16 +02:00
Sebastien Boeuf
3594685279 vmm: Move balloon code from MemoryManager to DeviceManager
Now that we have a new dedicated way of asking for a balloon through the
CLI and the REST API, we can move all the balloon code to the device
manager. This allows us to simplify the memory manager, which is already
quite complex.

It also simplifies the behavior of the balloon resizing command. Instead
of providing the expected size for the RAM, which is complex when memory
zones are involved, it now expects the balloon size. This is a much more
straightforward behavior as it really resizes the balloon to the desired
size. Additionally to the simplication, the benefit of this approach is
that it does not need to be tied to the memory manager at all.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-22 16:33:16 +02:00
Sebastien Boeuf
1d479e5e08 vmm: Introduce new --balloon parameter
This introduces a new way of defining the virtio-balloon device. Instead
of going through the --memory parameter, the idea is to consider balloon
as a standalone virtio device.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-22 16:33:16 +02:00
dependabot-preview[bot]
ae296d2229 build(deps): bump syn from 1.0.45 to 1.0.46
Bumps [syn](https://github.com/dtolnay/syn) from 1.0.45 to 1.0.46.
- [Release notes](https://github.com/dtolnay/syn/releases)
- [Commits](https://github.com/dtolnay/syn/compare/1.0.45...1.0.46)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-22 06:53:06 +00:00
dependabot-preview[bot]
555610181b build(deps): bump rustc-demangle from 0.1.17 to 0.1.18
Bumps [rustc-demangle](https://github.com/alexcrichton/rustc-demangle) from 0.1.17 to 0.1.18.
- [Release notes](https://github.com/alexcrichton/rustc-demangle/releases)
- [Commits](https://github.com/alexcrichton/rustc-demangle/compare/0.1.17...0.1.18)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-21 21:26:57 +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
Rob Bradford
700f63fad8 tests: Add integration test for virtio-watchdog
Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-21 16:02:39 +01:00
Rob Bradford
7ee63134d2 resources: Update Linux config to enable virtio-watchdog
Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-21 16:02:39 +01:00
Rob Bradford
885ee9567b vmm: Add support for creating virtio-watchdog
The watchdog device is created through the "--watchdog" parameter. At
most a single watchdog can be created per VM.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-21 16:02:39 +01:00
Rob Bradford
d2c7645731 virtio-devices: Add simple virtio-watchdog device
This device operates a single virtq. When the driver offers a descriptor
to the device it is interpreted as a "ping" to indicate that the guest
is alive. A periodic timer fires and if when the timer is fired there
has not been a "ping" from the guest then the device will reset the VM.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-21 16:02:39 +01:00
Rob Bradford
ae375434cd vm-virtio: Add (temporary) number allocation for virtio-watchdog
The next available number for allocation is 35.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-21 16:02:39 +01:00
Michael Zhao
0b0596ef30 arch: Simplify PCI space address handling in AArch64 FDT
Before Virtio-mmio was removed, we passed an optional PCI space address
parameter to AArch64 code for generating FDT. The address is none if the
transport is MMIO.
Now Virtio-PCI is the only option, the parameter is mandatory.

Signed-off-by: Michael Zhao <michael.zhao@arm.com>
2020-10-21 12:20:30 +01:00
Bo Chen
e83d99ab10 tests: Run binary with INFO (-v) level logging
Given the increased amount of output from cloud-hypervisor, this patch
also increased the PIPE_SIZE to 32MB (from 256KB).

Signed-off-by: Bo Chen <chen.bo@intel.com>
2020-10-21 08:48:28 +01:00
Bo Chen
e869d2f1bc tests: Print full commandline for better debugging
This patch prints the complete commandline when launching
cloud-hypervisor. It also prints the details of the `ssh` command if
the command is failing.

Signed-off-by: Bo Chen <chen.bo@intel.com>
2020-10-21 08:48:28 +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
e9880ab0d1 vhost_user_fs: seccomp: Propagate error correctly
Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-19 21:53:38 -07:00
Rob Bradford
02621c6150 vhost_user_fs: sandbox: Fix clippy errors
Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-19 21:53:38 -07:00
Rob Bradford
cf86ca15c2 virtio-devices: vsock: Fix clippy issue in tests
Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-19 21:53:38 -07:00
Rob Bradford
05c79119b9 virtio-devices: pci_common_config: Fix clippy issues in tests
Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-19 21:53:38 -07:00
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
Anatol Belski
b076602b87 doc: Add Windows debug instructions
Signed-off-by: Anatol Belski <anbelski@linux.microsoft.com>
2020-10-19 16:31:26 +01:00
Michael Zhao
2f2e10ea35 arch: Remove GICv2
Virtio-mmio is removed, now virtio-pci is the only option for virtio
transport layer. We use MSI for PCI device interrupt. While GICv2, the
legacy interrupt controller, doesn't support MSI. So GICv2 is not very
practical for Cloud-hypervisor, we can remove it.

Signed-off-by: Michael Zhao <michael.zhao@arm.com>
2020-10-19 14:58:48 +01:00
Sebastien Boeuf
58474854f2 docs: Remove all virtio-mmio references
Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-19 14:58:48 +01:00
Sebastien Boeuf
8b9c9bc97f ci: Remove all references to 'mmio' feature
Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-19 14:58:48 +01:00
Sebastien Boeuf
f73a345107 scripts: Remove all references to 'mmio' feature
Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-19 14:58:48 +01:00
Sebastien Boeuf
74c08e5669 github-actions: Remove all references to 'mmio' feature
Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-19 14:58:48 +01:00
Sebastien Boeuf
cc8b553e86 virtio-devices: Remove mmio and pci differentiation
Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-19 14:58:48 +01:00
Sebastien Boeuf
af3c6c34c3 vmm: Remove mmio and pci differentiation
Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-19 14:58:48 +01:00
pierwill
d2741fdc84 docs: Edit README.md
Add link to issue and fix two typos.

Signed-off-by: pierwill <19642016+pierwill@users.noreply.github.com>
2020-10-19 07:30:13 +02:00
dependabot-preview[bot]
9c6b03927a build(deps): bump syn from 1.0.44 to 1.0.45
Bumps [syn](https://github.com/dtolnay/syn) from 1.0.44 to 1.0.45.
- [Release notes](https://github.com/dtolnay/syn/releases)
- [Commits](https://github.com/dtolnay/syn/compare/1.0.44...1.0.45)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-19 05:28:55 +00:00
dependabot-preview[bot]
d986f74e27 build(deps): bump serde_derive from 1.0.116 to 1.0.117
Bumps [serde_derive](https://github.com/serde-rs/serde) from 1.0.116 to 1.0.117.
- [Release notes](https://github.com/serde-rs/serde/releases)
- [Commits](https://github.com/serde-rs/serde/compare/v1.0.116...v1.0.117)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-16 09:02:37 +00:00
dependabot-preview[bot]
fee873e325 build(deps): bump serde from 1.0.116 to 1.0.117
Bumps [serde](https://github.com/serde-rs/serde) from 1.0.116 to 1.0.117.
- [Release notes](https://github.com/serde-rs/serde/releases)
- [Commits](https://github.com/serde-rs/serde/compare/v1.0.116...v1.0.117)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-16 08:21:05 +00:00
Sebastien Boeuf
7cbd47a71a vmm: Prevent KVM device fd from being unusable from VfioContainer
When shutting down a VM using VFIO, the following error has been
detected:

vfio-ioctls/src/vfio_device.rs:312 -- Could not delete VFIO group:
KvmSetDeviceAttr(Error(9))

After some investigation, it appears the KVM device file descriptor used
for removing a VFIO group was already closed. This is coming from the
Rust sequence of Drop, from the DeviceManager all the way down to
VfioDevice.

Because the DeviceManager owns passthrough_device, which is effectively
a KVM device file descriptor, when the DeviceManager is dropped, the
passthrough_device follows, with the effect of closing the KVM device
file descriptor. Problem is, VfioDevice has not been dropped yet and it
still needs a valid KVM device file descriptor.

That's why the simple way to fix this issue coming from Rust dropping
all resources is to make Linux accountable for it by duplicating the
file descriptor. This way, even when the passthrough_device is dropped,
the KVM file descriptor is closed, but a duplicated instance is still
valid and owned by the VfioContainer.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-15 19:15:28 +02:00
Wei Liu
d667ed0c70 vmm: don't call notify_guest_clock_paused when Hyper-V emulation is on
We turn on that emulation for Windows. Windows does not have KVM's PV
clock, so calling notify_guest_clock_paused results in an error.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
2020-10-15 19:14:25 +02:00
Sebastien Boeuf
0c967e1aa0 virtio-devices: iommu: Update the list of seccomp filters
While using the virtio-iommu device involving L2 scenario, and tearing
things down all the way from L2 back to L0 exposed some bad syscalls
that were not part of the authorized list.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-14 19:15:09 +02:00
Wei Liu
57f81d0375 scripts: dev_cli: clarify which integration test is running
This makes it easier to grep. No functional change obviously.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
2020-10-14 14:23:11 +02:00
Rob Bradford
1bae38ceb4 build: Add Jenkinsfile entry for Windows guest
Add one more parallel stage to run Windows guest tests in a dedicated
VM.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-14 11:19:29 +02:00
Rob Bradford
934f992536 scripts: dev_cli: Add support for Windows guest integration tests
Now that we've written Windows integration tests and the associated
script to launch them, this patch enables the support for Windows tests
in dev_cli.sh, so that we can run it in our Cloud Hypervisor container.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-14 11:19:29 +02:00
Rob Bradford
76c04878fa scripts: Add wrapper script to run Windows guest integration tests
Adding a dedicated script to avoid confusion with other Linux
integration tests scripts.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-14 11:19:29 +02:00
Rob Bradford
d598341534 tests: Add Windows guest integration test
This is a new integration test running Windows as a guest with Cloud
Hypervisor. Once the VM is booted, the test connects to the guest
through SSH and shutdown the VM. If this succeeds, this means the VM
was properly booted to userspace and that the network was functional.

Important to note that because this test generates lots of logs, it
requires a large pipe size for both stdout and stderr.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-14 11:19:29 +02:00
Rob Bradford
0eabc00b27 tests: Allow controlling the password for SSH auth
The login details for the Windows VM are different.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-14 11:19:29 +02:00
dependabot-preview[bot]
08ded4b449 build(deps): bump regex from 1.4.0 to 1.4.1
Bumps [regex](https://github.com/rust-lang/regex) from 1.4.0 to 1.4.1.
- [Release notes](https://github.com/rust-lang/regex/releases)
- [Changelog](https://github.com/rust-lang/regex/blob/master/CHANGELOG.md)
- [Commits](https://github.com/rust-lang/regex/compare/1.4.0...1.4.1)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-13 17:41:10 +00:00
Sebastien Boeuf
9d882bc8fe ci: Add integration test for testing maximum physical bits
Introduce a new test that will validate the new option `max_phys_bits`
from the `--cpus` parameter behaves as expected.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-13 18:58:36 +02:00
Sebastien Boeuf
1b9890b807 vmm: cpu: Set CPU physical bits based on user input
If the user specified a maximum physical bits value through the
`max_phys_bits` option from `--cpus` parameter, the guest CPUID
will be patched accordingly to ensure the guest will find the
right amount of physical bits.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-13 18:58:36 +02:00
Sebastien Boeuf
aec88e20d7 vmm: memory_manager: Rely on physical bits for address space size
If the user provided a maximum physical bits value for the vCPUs, the
memory manager will adapt the guest physical address space accordingly
so that devices are not placed further than the specified value.

It's important to note that if the number exceed what is available on
the host, the smaller number will be picked.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-13 18:58:36 +02:00
Sebastien Boeuf
52ad78886c vmm: Introduce new CPU option to set maximum physical bits
In order to let the user choose maximum address space size, this patch
introduces a new option `max_phys_bits` to the `--cpus` parameter.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-13 18:58:36 +02:00
dependabot-preview[bot]
c97824ab76 build(deps): bump pkg-config from 0.3.18 to 0.3.19
Bumps [pkg-config](https://github.com/rust-lang/pkg-config-rs) from 0.3.18 to 0.3.19.
- [Release notes](https://github.com/rust-lang/pkg-config-rs/releases)
- [Changelog](https://github.com/rust-lang/pkg-config-rs/blob/master/CHANGELOG.md)
- [Commits](https://github.com/rust-lang/pkg-config-rs/compare/0.3.18...0.3.19)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-13 13:14:47 +00:00
Bo Chen
e9738a4a49 vmm: Replace the use of 'unchecked_add' with 'checked_add'
The 'GuestAddress::unchecked_add' function has undefined behavior when
an overflow occurs. Its alternative 'checked_add' requires use to handle
the overflow explicitly.

Signed-off-by: Bo Chen <chen.bo@intel.com>
2020-10-13 12:09:22 +02:00
Bo Chen
9ab2a34b40 vmm: Remove reserved 256M gaps for hotplugging memory with ACPI
We are now reserving a 256M gap in the guest address space each time
when hotplugging memory with ACPI, which prevents users from hotplugging
memory to the maximum size they requested. We confirm that there is no
need to reserve this gap.

This patch removes the 'reserved gaps'. It also refactors the
'MemoryManager::start_addr' so that it is rounding-up to 128M alignment
when hotplugged memory is allowed with ACPI.

Signed-off-by: Bo Chen <chen.bo@intel.com>
2020-10-13 12:09:22 +02:00
Henry Wang
8bf0bac5a7 dev_cli: Fix TARGETARCH inconsistancy
In previous dev_cli.sh, the `uname -m` command will generate
either `x86_64` or `aarch64`, which is inconsistent with the
architectures in the Dockerfile, namely `amd64` and `arm64`.

This will cause some dependancy missing in the docker container
when the docker image is built locally.

This commit fixes this inconsistancy.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-10-13 11:57:01 +02:00
dependabot-preview[bot]
ef1e5fd953 build(deps): bump serde_json from 1.0.58 to 1.0.59
Bumps [serde_json](https://github.com/serde-rs/json) from 1.0.58 to 1.0.59.
- [Release notes](https://github.com/serde-rs/json/releases)
- [Commits](https://github.com/serde-rs/json/compare/v1.0.58...v1.0.59)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-13 06:57:48 +00:00
dependabot-preview[bot]
a06a6e4696 build(deps): bump aho-corasick from 0.7.13 to 0.7.14
Bumps [aho-corasick](https://github.com/BurntSushi/aho-corasick) from 0.7.13 to 0.7.14.
- [Release notes](https://github.com/BurntSushi/aho-corasick/releases)
- [Commits](https://github.com/BurntSushi/aho-corasick/compare/0.7.13...0.7.14)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-13 06:57:37 +00:00
Bo Chen
10f380f95b vmm: Report no error when resizing to current memory size with ACPI
We now try to create a ram region of size 0 when the requested memory
size is the same as current memory size. It results in an error of
`GuestMemoryRegion(Mmap(Os { code: 22, kind: InvalidInput, message:
"Invalid argument" }))`. This error is not meaningful to users and we
should not report it.

Signed-off-by: Bo Chen <chen.bo@intel.com>
2020-10-12 08:46:38 +02:00
dependabot-preview[bot]
299e83bf43 build(deps): bump regex from 1.3.9 to 1.4.0
Bumps [regex](https://github.com/rust-lang/regex) from 1.3.9 to 1.4.0.
- [Release notes](https://github.com/rust-lang/regex/releases)
- [Changelog](https://github.com/rust-lang/regex/blob/master/CHANGELOG.md)
- [Commits](https://github.com/rust-lang/regex/compare/1.3.9...1.4.0)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-12 05:17:07 +00:00
dependabot-preview[bot]
54d2298a93 build(deps): bump regex-syntax from 0.6.18 to 0.6.19
Bumps [regex-syntax](https://github.com/rust-lang/regex) from 0.6.18 to 0.6.19.
- [Release notes](https://github.com/rust-lang/regex/releases)
- [Changelog](https://github.com/rust-lang/regex/blob/master/CHANGELOG.md)
- [Commits](https://github.com/rust-lang/regex/commits)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-12 05:16:46 +00:00
dependabot-preview[bot]
74627d2de5 build(deps): bump backtrace from 0.3.52 to 0.3.53
Bumps [backtrace](https://github.com/rust-lang/backtrace-rs) from 0.3.52 to 0.3.53.
- [Release notes](https://github.com/rust-lang/backtrace-rs/releases)
- [Commits](https://github.com/rust-lang/backtrace-rs/compare/0.3.52...0.3.53)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-12 05:16:17 +00:00
dependabot-preview[bot]
818dc72e5f build(deps): bump pin-project from 0.4.26 to 0.4.27
Bumps [pin-project](https://github.com/taiki-e/pin-project) from 0.4.26 to 0.4.27.
- [Release notes](https://github.com/taiki-e/pin-project/releases)
- [Changelog](https://github.com/taiki-e/pin-project/blob/master/CHANGELOG.md)
- [Commits](https://github.com/taiki-e/pin-project/compare/v0.4.26...v0.4.27)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-12 05:12:45 +00:00
dependabot-preview[bot]
a4be05ac40 build(deps): bump syn from 1.0.43 to 1.0.44
Bumps [syn](https://github.com/dtolnay/syn) from 1.0.43 to 1.0.44.
- [Release notes](https://github.com/dtolnay/syn/releases)
- [Commits](https://github.com/dtolnay/syn/compare/1.0.43...1.0.44)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-12 05:12:12 +00:00
Muminul Islam
029a6d8ab9 README: Add steps to install prerequisites
Signed-off-by: Muminul Islam <muislam@microsoft.com>
2020-10-10 09:52:06 +02:00
Muminul Islam
0f48b61265 README: Syntax cleanups
Signed-off-by: Muminul Islam <muislam@microsoft.com>
2020-10-10 09:52:06 +02:00
Bo Chen
789ee7b3e4 vmm: Support resizing memory up to and including hotplug size
The start address after the hottplugged memory can be the start address of
device area.

Fixes: #1803

Signed-off-by: Bo Chen <chen.bo@intel.com>
2020-10-10 09:51:32 +02:00
dependabot-preview[bot]
f639a8a0fd build(deps): bump syn from 1.0.42 to 1.0.43
Bumps [syn](https://github.com/dtolnay/syn) from 1.0.42 to 1.0.43.
- [Release notes](https://github.com/dtolnay/syn/releases)
- [Commits](https://github.com/dtolnay/syn/compare/1.0.42...1.0.43)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-09 17:56:30 +00:00
dependabot-preview[bot]
eabfd642e0 build(deps): bump cc from 1.0.60 to 1.0.61
Bumps [cc](https://github.com/alexcrichton/cc-rs) from 1.0.60 to 1.0.61.
- [Release notes](https://github.com/alexcrichton/cc-rs/releases)
- [Commits](https://github.com/alexcrichton/cc-rs/compare/1.0.60...1.0.61)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-09 10:23:32 +00:00
dependabot-preview[bot]
ab32375666 build(deps): bump backtrace from 0.3.51 to 0.3.52
Bumps [backtrace](https://github.com/rust-lang/backtrace-rs) from 0.3.51 to 0.3.52.
- [Release notes](https://github.com/rust-lang/backtrace-rs/releases)
- [Commits](https://github.com/rust-lang/backtrace-rs/compare/0.3.51...0.3.52)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-09 09:33:32 +00:00
Michael Zhao
2a7b1d78ff build: Fix AArch64 musl cross build failure
A failure appeared in AArch64 musl cross build, after upgrading rust
to v1.47.0. A symbol "strrchr" was missing while linking against
static libfdt.a.

The issue could be caused by missing symbol(s) in new rust toolchain.

This fix pins the rust version in this cross build action to a stable-
enough version. Further upgrade will be done manually after testing.

Signed-off-by: Michael Zhao <michael.zhao@arm.com>
2020-10-09 10:49:54 +02:00
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
Rob Bradford
bb5b9584d2 pci, ch-remote, vmm: Replace simple match blocks with matches!
This is a new clippy check introduced in 1.47 which requires the use of
the matches!() macro for simple match blocks that return a boolean.

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
Anatol Belski
095ea049e2 doc: Add UEFI and Windows consumer pages
This changeset extends the documentation with the UEFI and Windows
related info. The focus is on providing consumer with a minimum
necessary and proper piece of the information to enter the features
quickly. While UEFI is a cross platform topic, it is a required
prerequisite for the Windows usage.

Signed-off-by: Anatol Belski <anbelski@linux.microsoft.com>
2020-10-08 15:45:30 +01:00
dependabot-preview[bot]
ff0d2431d0 build(deps): bump rustc-demangle from 0.1.16 to 0.1.17
Bumps [rustc-demangle](https://github.com/alexcrichton/rustc-demangle) from 0.1.16 to 0.1.17.
- [Release notes](https://github.com/alexcrichton/rustc-demangle/releases)
- [Commits](https://github.com/alexcrichton/rustc-demangle/compare/0.1.16...0.1.17)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-08 08:59:34 +00:00
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
Sebastien Boeuf
30e1162d7c virtio: Replace some repr(C, packed) with repr(C)
The goal here is to replace anywhere possible a virtio structure
with a "C, packed" representation by a "C" representation. Some
virtio structures are not expected to be packed, therefore there's
no reason for using the more restrictive "C, packed" representation.

This is important since "packed" representation can still cause
undefined behaviors with Rust 2018.

By removing the need for "packed" representation, we can simplify a
bit of code by deriving the Serialize trait without writing the
implementation ourselves.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-07 22:11:36 +02:00
dependabot-preview[bot]
2333c5e233 build(deps): bump miniz_oxide from 0.4.2 to 0.4.3
Bumps [miniz_oxide](https://github.com/Frommi/miniz_oxide) from 0.4.2 to 0.4.3.
- [Release notes](https://github.com/Frommi/miniz_oxide/releases)
- [Changelog](https://github.com/Frommi/miniz_oxide/blob/master/CHANGELOG.md)
- [Commits](https://github.com/Frommi/miniz_oxide/compare/0.4.2...0.4.3)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-07 15:55:10 +00:00
Sebastien Boeuf
1e3a6cb450 vmm: Simplify some of the io_uring code
Small patch creating a dedicated `block_io_uring_is_supported()`
function for the non-io_uring case, so that we can simplify the
code in the DeviceManager.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-07 14:26:49 +02:00
Wei Liu
ac32b1e35f ci: use cargo check to validate building commits
The purpose of that step is to make sure each commit builds. The `check`
command is much faster for that purpose.

On my 32-core machine `cargo check --all` takes around 25 seconds while
`cargo build --all` takes around 35 seconds, so that's quite a bit of
time saving there.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
2020-10-07 14:26:32 +02:00
Sebastien Boeuf
67025948dc ci: Simplify test scripts
Since all unit and integration tests are run inside containers because
they are called from dev_cli.sh, they always run as root. That's why
both unit and integration scripts can be simplified as they don't need
to apply specific capabilities and run cargo tests in a dedicated 'kvm'
group.

Fixes #1683

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-07 14:26:13 +02:00
dependabot-preview[bot]
a5d3604272 build(deps): bump anyhow from 1.0.32 to 1.0.33
Bumps [anyhow](https://github.com/dtolnay/anyhow) from 1.0.32 to 1.0.33.
- [Release notes](https://github.com/dtolnay/anyhow/releases)
- [Commits](https://github.com/dtolnay/anyhow/compare/1.0.32...1.0.33)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-07 06:58:14 +00:00
dependabot-preview[bot]
f48398083c build(deps): bump thiserror from 1.0.20 to 1.0.21
Bumps [thiserror](https://github.com/dtolnay/thiserror) from 1.0.20 to 1.0.21.
- [Release notes](https://github.com/dtolnay/thiserror/releases)
- [Commits](https://github.com/dtolnay/thiserror/compare/1.0.20...1.0.21)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-07 05:38:10 +00:00
Rob Bradford
00c0b9e170 devices: ioapic: Tolerate Windows writing zero to version register
Windows will write zero to the IOAPIC version register as part of
probing if the device is there.

Fixes: #1791

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-06 18:13:33 +01:00
Sebastien Boeuf
162789a2d1 ci: Extend virtio-blk hotplug integration test
The existing virtio-blk hotplug test is extended by removing and
re-adding the virtio-blk device. This ensures the unplug/re-plug
feature is properly tested.

Fixes #1809

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-06 17:22:29 +01:00
Sebastien Boeuf
14163c5200 ci: Extend virtio-net hotplug integration test
The existing virtio-net hotplug test is extended by removing and
re-adding the virtio-net device. This ensures the unplug/re-plug
feature is properly tested.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-06 17:22:29 +01:00
Rob Bradford
8824da27cb docs: Add documentation about the log levels and when to use them
Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-06 16:52:29 +01:00
Rob Bradford
03f7d39ce5 main: Set default log level to warn!() equivalent.
Using our standard configuration and default kernel we trigger no
messages at this level.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-06 16:52:29 +01:00
Sebastien Boeuf
c02a02edfc vmm: Allow unlink syscall for vCPU threads
Without the unlink(2) syscall being allowed, Cloud-Hypervisor crashes
when we remove a virtio-vsock device that has been previously added.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-06 16:05:59 +01:00
Sebastien Boeuf
cc7a5aca6c ci: Validate each commit can be built
It's important to ensure Cloud-Hypervisor's Git tree is bisectable, and
that's why this commit extends the existing build.yaml Github action.
This will validate that each commit from an incoming pull request can be
built.

Fixes #1808

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-06 14:28:54 +01:00
Sebastien Boeuf
b33969896b virtio-devices: Rely on vhost crate from Cloud-Hypervisor fork
The Cloud-Hypervisor fork of the vhost crate contains one small
additional patch compared to the rust-vmm upstream version, meant for
increasing the connection timeout.

This patch is intended to be merged in order to check if it helps our CI
fixing the vhost-user-blk flakes that we've been observing recently.

If it fixes it, we'll submit a similar patch upstream and switch back to
the upstream vhost crate, otherwise we'll simply switch back to the
upstream crate, discarding this patch.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-06 12:56:33 +02:00
Sebastien Boeuf
6aa5e21212 vmm: device_manager: Fix PCI device unplug issues
Because of the PCI refactoring that happened in the previous commit
d793cc4da3, the ability to fully remove a
PCI device was altered.

The refactoring was correct, but the usage of a generic function to pass
the same reference for both BusDevice, PciDevice and Any + Send + Sync
causes the Arc::ptr_eq() function to behave differently than expected,
as it does not match the references later in the code. That means we
were not able to remove the device reference from the MMIO and/or PIO
buses, which was leading to some bus range overlapping error once we
were trying to add a device again to the previous range that should have
been removed.

Fixes #1802

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-06 12:56:17 +02:00
dependabot-preview[bot]
a32bd2fae1 build(deps): bump hermit-abi from 0.1.16 to 0.1.17
Bumps [hermit-abi](https://github.com/hermitcore/libhermit-rs) from 0.1.16 to 0.1.17.
- [Release notes](https://github.com/hermitcore/libhermit-rs/releases)
- [Commits](https://github.com/hermitcore/libhermit-rs/commits)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-06 09:25:31 +00:00
dependabot-preview[bot]
e553c053af build(deps): bump futures from 0.3.5 to 0.3.6
Bumps [futures](https://github.com/rust-lang/futures-rs) from 0.3.5 to 0.3.6.
- [Release notes](https://github.com/rust-lang/futures-rs/releases)
- [Changelog](https://github.com/rust-lang/futures-rs/blob/master/CHANGELOG.md)
- [Commits](https://github.com/rust-lang/futures-rs/commits)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-06 06:34:18 +00:00
dependabot-preview[bot]
83f6aa47c9 build(deps): bump futures-core from 0.3.5 to 0.3.6
Bumps [futures-core](https://github.com/rust-lang/futures-rs) from 0.3.5 to 0.3.6.
- [Release notes](https://github.com/rust-lang/futures-rs/releases)
- [Changelog](https://github.com/rust-lang/futures-rs/blob/master/CHANGELOG.md)
- [Commits](https://github.com/rust-lang/futures-rs/commits)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-05 20:58:50 +00:00
dependabot-preview[bot]
a4ebb6399a build(deps): bump futures-task from 0.3.5 to 0.3.6
Bumps [futures-task](https://github.com/rust-lang/futures-rs) from 0.3.5 to 0.3.6.
- [Release notes](https://github.com/rust-lang/futures-rs/releases)
- [Changelog](https://github.com/rust-lang/futures-rs/blob/master/CHANGELOG.md)
- [Commits](https://github.com/rust-lang/futures-rs/commits)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-05 20:58:24 +00:00
dependabot-preview[bot]
0c25b4d9fb build(deps): bump futures-io from 0.3.5 to 0.3.6
Bumps [futures-io](https://github.com/rust-lang/futures-rs) from 0.3.5 to 0.3.6.
- [Release notes](https://github.com/rust-lang/futures-rs/releases)
- [Changelog](https://github.com/rust-lang/futures-rs/blob/master/CHANGELOG.md)
- [Commits](https://github.com/rust-lang/futures-rs/commits)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-05 20:57:51 +00:00
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
dependabot-preview[bot]
03eebcede2 build(deps): bump pin-project from 0.4.25 to 0.4.26
Bumps [pin-project](https://github.com/taiki-e/pin-project) from 0.4.25 to 0.4.26.
- [Release notes](https://github.com/taiki-e/pin-project/releases)
- [Changelog](https://github.com/taiki-e/pin-project/blob/v0.4.26/CHANGELOG.md)
- [Commits](https://github.com/taiki-e/pin-project/compare/v0.4.25...v0.4.26)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-04 21:22:01 +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
Rob Bradford
2d457ab974 vmm: device_manager: Make PMEM "discard_writes" mode true CoW
The PMEM support has an option called "discard_writes" which when true
will prevent changes to the device from hitting the backing file. This
is trying to be the equivalent of "readonly" support of the block
device.

Previously the memory of the device was marked as KVM_READONLY. This
resulted in a trap when the guest attempted to write to it resulting a
VM exit (and recently a warning). This has a very detrimental effect on
the performance so instead make "discard_writes" truly CoW by mapping
the memory as `PROT_READ | PROT_WRITE` and using `MAP_PRIVATE` to
establish the CoW mapping.

Fixes: #1795

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-02 14:26:15 +02:00
Daniel Verkamp
6eeab85db0 qcow: avoid out-of-bounds access in alloc_refblocks
When all refblocks are consumed, the loop looking for the first free
cluster would access the element at refcounts[refcounts.len()], which is
out of bounds.  Modify the free cluster search loop to check that the
index is in bounds before accessing it.

Tested-by: Daniel Verkamp <dverkamp@chromium.org>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Dylan Reid <dgreid@chromium.org>
Commit-Queue: Daniel Verkamp <dverkamp@chromium.org>

(cherry picked from crosvm commit f21572c7187c8beb9c6bfea6446351ae93200d01)

Fixes: #1792

Signed-off-by: Bo Chen <chen.bo@intel.com>
2020-10-02 09:43:09 +02:00
Hui Zhu
c75f8b2f89 virtio-balloon: Add memory_actual_size to vm.info to show memory actual size
The virtio-balloon change the memory size is asynchronous.
VirtioBalloonConfig.actual of balloon device show current balloon size.

This commit add memory_actual_size to vm.info to show memory actual size.

Signed-off-by: Hui Zhu <teawater@antfin.com>
2020-10-01 17:46:30 +02:00
Sebastien Boeuf
9dffc5da5c devices: ioapic: Always retrieve destination field on 8 bits
When the destination mode is physical, the destination field should
only be defined through bits 56-59, as defined in the IOAPIC spec. But
from the APIC specification, the APIC ID is always defined on 8 bits no
matter which destination mode is selected. That's why we always retrieve
the destination field based on bits 56-63.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-01 14:35:38 +01:00
dependabot-preview[bot]
a8c2d70869 build(deps): bump pin-project from 0.4.24 to 0.4.25
Bumps [pin-project](https://github.com/taiki-e/pin-project) from 0.4.24 to 0.4.25.
- [Release notes](https://github.com/taiki-e/pin-project/releases)
- [Changelog](https://github.com/taiki-e/pin-project/blob/master/CHANGELOG.md)
- [Commits](https://github.com/taiki-e/pin-project/compare/v0.4.24...v0.4.25)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-01 11:15:43 +00:00
Bo Chen
a070f44a51 build: github: Add a new action to build cargo fuzz
Signed-off-by: Bo Chen <chen.bo@intel.com>
2020-10-01 11:16:35 +01:00
Bo Chen
dda1df0a58 fuzz: Fix the building errors with recent changes
This patch adds two required dependencies to fuzz/Cargo.toml, and fixes
the building error on the 'block' fuzzer.

Signed-off-by: Bo Chen <chen.bo@intel.com>
2020-10-01 11:16:35 +01:00
Rob Bradford
2c2e7016c7 main: Improve documentation for --kernel
Fixes: #1712

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-10-01 10:08:25 +01:00
Sebastien Boeuf
532b3063be ci: No need to build Docker image locally for arm64
Now that Docker images are automatically generated for both amd64 and
arm64 architectures, there's no need to generate the arm64 image locally
on the ARM CI during a CI run. The image should be available from
DockerHub instead.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-10-01 10:04:32 +01:00
dependabot-preview[bot]
91314e7b6c build(deps): bump serde_json from 1.0.57 to 1.0.58
Bumps [serde_json](https://github.com/serde-rs/json) from 1.0.57 to 1.0.58.
- [Release notes](https://github.com/serde-rs/json/releases)
- [Commits](https://github.com/serde-rs/json/compare/v1.0.57...v1.0.58)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-01 08:07:53 +00:00
dependabot-preview[bot]
8851b48ce6 build(deps): bump proc-macro2 from 1.0.23 to 1.0.24
Bumps [proc-macro2](https://github.com/alexcrichton/proc-macro2) from 1.0.23 to 1.0.24.
- [Release notes](https://github.com/alexcrichton/proc-macro2/releases)
- [Commits](https://github.com/alexcrichton/proc-macro2/compare/1.0.23...1.0.24)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-10-01 05:40:28 +00:00
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
Rob Bradford
664c3ceda6 vmm: device_manager: Warn that vhost-user self spawning is deprecated
See #1724 for details.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-09-30 18:32:50 +02:00
Rob Bradford
0a4be7ddf5 vmm: "Cleanly" shutdown on SIGTERM
Write to the exit_evt EventFD which will trigger all the devices and
vCPUs to exit. This is slightly cleaner than just exiting the process as
any temporary files will be removed.

Fixes: #1242

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-09-30 18:32:16 +02:00
Sebastien Boeuf
8a11ef5ba6 ci: Move back to latest virtiofsd version
This commit moves back to the branch "virtio-fs-dev" from virtiofsd, as
we figured the changes needed to use this branch and the requirements
from the new meson build from QEMU.

It updates the container version to ensure the dev_cli.sh script will
rely on the latest container which contains the needed packages.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-09-30 13:10:23 +01:00
Sebastien Boeuf
1e5d35d13a resources: Finalize automated generation of the Docker image
By fixing the Dockerfile, we have now finalized the automated generation
of the Docker images for both architectures (amd64 and arm64).

Fixes #953

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-09-30 10:03:06 +01:00
Rob Bradford
df26b428b1 arch: x86_64: Copy CPU identification string into guest cpuid
This is stored in leaves 0x80000002 to 0x80000004 inclusive.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-09-30 08:18:08 +01:00
Bo Chen
6d30fe05e4 vmm: openapi: Add the 'iommu' and 'id' option to 'VmAddDevice'
This patch adds the missing the `iommu` and `id` option for
`VmAddDevice` in the openApi yaml to respect the internal data structure
in the code base. Also, setting the `id` explicitly for VFIO device
hotplug is required for VFIO device unplug through openAPI calls.

Signed-off-by: Bo Chen <chen.bo@intel.com>
2020-09-30 08:17:44 +01:00
dependabot-preview[bot]
af0c4c1659 build(deps): bump backtrace from 0.3.50 to 0.3.51
Bumps [backtrace](https://github.com/rust-lang/backtrace-rs) from 0.3.50 to 0.3.51.
- [Release notes](https://github.com/rust-lang/backtrace-rs/releases)
- [Commits](https://github.com/rust-lang/backtrace-rs/compare/0.3.50...0.3.51)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-09-29 20:25:50 +00:00
Julio Montes
668c563dac vmm: openapi: fix integers format
According to openAPI specification [1], the format for `integer` types
can be only `int32` or `int64`, unsigned and 8-bits integers are not
supported.
This patch replaces `uint64` with `int64`, `uint32` with `int32` and
`uint8` with `int32`.

[1]: https://swagger.io/specification/#data-types

Signed-off-by: Julio Montes <julio.montes@intel.com>
2020-09-29 12:55:40 -07:00
Rob Bradford
5a0d3277c8 vmm: vm: Replace \n newline character with \r
This allows the CMD prompt under SAC to be used without affecting getty
on Linux.

Fixes: #1770

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-09-29 16:10:12 +02:00
Wei Liu
4ef97d8ddb vmm: interrupts: clearly separate MsiInterruptGroup and InterruptRoute
MsiInterruptGroup doesn't need to know the internal field names of
InterruptRoute. Introduce two helper functions to eliminate references
to irq_fd. This is done similarly to the enable and disable helper
functions.

Also drop the pub keyword from InterruptRoute fields. It is not needed
anymore.

No functional change.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
2020-09-29 13:51:35 +02:00
Sebastien Boeuf
db5d42ad41 resources: Fix Dockerfile to support multi-architecture
In order to support both amd64 and arm64, we rely on the TARGETARCH
variable that is passed from the docker buildx command, based on the
platform used to build the container image.

There is no way to rely directly on $(uname -m) to assign a variable
with the correct x86_64 or aarch64 values we're looking for. Both ENV
and ARG don't evaluate the command, which means they see it as a simple
string. Using RUN is the only way to evaluate a command.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-09-29 09:29:36 +02:00
dependabot-preview[bot]
315cd5a054 build(deps): bump pin-project from 0.4.23 to 0.4.24
Bumps [pin-project](https://github.com/taiki-e/pin-project) from 0.4.23 to 0.4.24.
- [Release notes](https://github.com/taiki-e/pin-project/releases)
- [Changelog](https://github.com/taiki-e/pin-project/blob/master/CHANGELOG.md)
- [Commits](https://github.com/taiki-e/pin-project/compare/v0.4.23...v0.4.24)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-09-28 14:34:47 +00:00
Sebastien Boeuf
77e9fcdebd resources: Update Dockerfile to install missing Python packages
In order to build virtiofsd from the latest build system, the Python
package python3-setuptools is required.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-09-28 09:28:37 +02:00
Sebastien Boeuf
3534b6932a github: Add a new action to update the Docker image
The idea is to rely on this new Github Action to detect when the
Dockerfile is updated after a push to the master branch on the
repository.

Once triggered, this action builds the Docker image for both
linux/amd64 and linux/arm64 platforms, and updates it directly
on Docker Hub.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-09-28 08:25:11 +02:00
dependabot-preview[bot]
34dd3b83af build(deps): bump syn from 1.0.41 to 1.0.42
Bumps [syn](https://github.com/dtolnay/syn) from 1.0.41 to 1.0.42.
- [Release notes](https://github.com/dtolnay/syn/releases)
- [Commits](https://github.com/dtolnay/syn/compare/1.0.41...1.0.42)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-09-28 05:20:09 +00:00
dependabot-preview[bot]
e105611eb6 build(deps): bump proc-macro2 from 1.0.21 to 1.0.23
Bumps [proc-macro2](https://github.com/alexcrichton/proc-macro2) from 1.0.21 to 1.0.23.
- [Release notes](https://github.com/alexcrichton/proc-macro2/releases)
- [Commits](https://github.com/alexcrichton/proc-macro2/compare/1.0.21...1.0.23)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-09-28 05:19:46 +00:00
Praveen Paladugu
f10872e706 vmm: fix clippy warnings
Signed-off-by: Praveen Paladugu <prapal@microsoft.com>
2020-09-26 14:07:12 +01:00
Praveen Paladugu
4b32252028 hypervisor, vmm: fix clippy warnings
Signed-off-by: Praveen Paladugu <prapal@microsoft.com>
2020-09-26 14:07:12 +01:00
Praveen Paladugu
6ebeaa9246 option_parser: fix clippy warnings
Signed-off-by: Praveen Paladugu <prapal@microsoft.com>
2020-09-26 14:07:12 +01:00
Julio Montes
c54452c08a vmm: openapi: fix integers format
According to openAPI specification[1], the format for `integer` types
can be only `int32` or `int64`, unsigned integers are not supported.
This patch replaces `uint64` with `int64`.

[1]: https://swagger.io/specification/#data-types

Signed-off-by: Julio Montes <julio.montes@intel.com>
2020-09-26 14:05:51 +01:00
Rob Bradford
f762bc7573 arch: x86_64: Create MP table after SMBIOS table if space
In order to speed up the Linux boot (so as to avoid it having to scan a
large number of pages) place the MP table directly after the SMBIOS
table if there is sufficient room. The start address of the SMBIOS table
is one of the three (and the largest) location that the MP table can
also be located at.

Before:
[    0.000399] x86/PAT: Configuration [0-7]: WB  WC  UC- UC  WB  WP  UC- WT
[    0.014945] check: Scanning 1 areas for low memory corruption

After:
[    0.000284] x86/PAT: Configuration [0-7]: WB  WC  UC- UC  WB  WP  UC- WT
[    0.000421] found SMP MP-table at [mem 0x000f0090-0x000f009f]

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-09-25 19:41:24 +02:00
Wei Liu
7e130a65ba vmm: interrupts: adjust set_gsi_routes
There is no point in manually dropping the lock for gsi_msi_routes then
instantly grabbing it again in set_gsi_routes.

Make set_gsi_routes take a reference to the routing hashmap instead.

No functional change intended.

Signed-off-by: Wei Liu <liuwe@microsoft.com>
2020-09-25 17:17:35 +02:00
Hui Zhu
4913acc05e vmm: Add 'balloon' to memory parameters
Add the option 'balloon' to --memory.

Signed-off-by: Hui Zhu <teawater@antfin.com>
2020-09-25 17:13:39 +02:00
Sebastien Boeuf
c85e396ce5 vmm: cpu: x86: Enable MTRR feature in CPUID
The MTRR feature was missing from the CPUID, which is causing the guest
to ignore the MTRR settings exposed through dedicated MSRs.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-09-25 15:03:52 +02:00
Rob Bradford
f4ec915c5d resources: Remove unused PPS features from kernel config
In particular this removes the annoying PPS messages that fill up the
dmesg log.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-09-25 14:07:41 +02:00
Sebastien Boeuf
ae44e9c076 resources: Reduce x86_64 kernel configuration to fix warnings
Removing the ISA DMA configurations prevents the kernel from accessing
the port I/O 0x87, which was generating the following warning:

WARN:vmm/src/cpu.rs:378 -- Guest PIO read to unregistered address 0x87

Removing the TELCLOCK configuration prevents the kernel from accessing
the port I/O reserved for the memory manager, which was causing the
following warning:

WARN:vmm/src/memory_manager.rs:289 -- Unexpected offset for accessing
memory manager device: 15

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-09-25 10:22:28 +01:00
Rob Bradford
29b74804e1 main: Improve the error reporting when creating the hypervisor object
The ::new() does very little beyond trying to open the /dev/kvm device
so provide a hint to the user about what has gone wrong.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-09-25 11:08:01 +02:00
Bo Chen
1d3c3bc6ec tests: Capture child process stdout/err in 'test_memory_mergeable'
Signed-off-by: Bo Chen <chen.bo@intel.com>
2020-09-25 08:40:23 +02:00
Bo Chen
2441798fe4 tests: Resize the pipe size to 256K for capturing child stdout/err
As discussed in #1707, the `vcpu` thread can be stalled when using
`--serial tty`. To workaround that issue, this patch enforces to resize
the pipe size to 256K when we capture the stdout/stderr of the
cloud-hypervisor child process in the integration tests. Note that the
pipe size (256K) is chosen based on the output size of our integration
tests at this point, which may need to be increased in the future.

Signed-off-by: Bo Chen <chen.bo@intel.com>
2020-09-25 08:40:23 +02:00
Bo Chen
365b947023 tests: Port test_simple_launch to the new methodology
This is the last test to be ported to the new methodology.

Signed-off-by: Bo Chen <chen.bo@intel.com>
2020-09-25 08:40:23 +02:00
Bo Chen
cb2f11724a tests: Port test_reboot to the new methodology
Signed-off-by: Bo Chen <chen.bo@intel.com>
2020-09-25 08:40:23 +02:00
Hui Zhu
d03a48162f balloon.rs: BalloonEpollHandler: Fix wrong error in handle_event
error!("Unknown event for virtio-mem");
This error should be
error!("Unknown event for virtio-balloon");

Signed-off-by: Hui Zhu <teawater@antfin.com>
2020-09-25 08:36:23 +02:00
Sebastien Boeuf
de88bef429 pci: msix: Fix masking/enabling semantics
By looking at Linux kernel boot time, we identified that a lot of time
was spent registering and unregistering IRQ fds to KVM. This is not
efficient and certainly not a wrong behavior from the Linux kernel,
but rather a problem with the Cloud-Hypervisor's implementation of
MSI-X.

The way to fix this issue is by ensuring the initial conditions are
correct, which means the entire MSI-X vector table must be disabled
and masked. Additionally, each vector must be individually masked.

With these correct conditions, Linux won't start masking interrupt
vectors, and later unmask them since they will be seen as masked from
the beginning. This means the OS will simply have to unmask them when
needed, avoiding the extra operation.

Another aspect of this patch is to prevent Cloud-Hypervisor from
enabling (by registering IRQ fd) all vectors when either the global
'mask' or 'enable' bits are set. Instead, we can simply let the mask()
and unmask() operations take care of it if needed.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-09-24 22:29:16 +02:00
Sebastien Boeuf
64351c1f3f build: Update Cargo.lock
Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-09-23 19:03:19 +02:00
Sebastien Boeuf
2eaf1c70c0 vmm: acpi: Advertise the correct PCI bus range
Since Cloud-Hypervisor currently support one single PCI bus, we must
reflect this through the MCFG table, as it advertises the first bus and
the last bus available. In this case both are bus 0.

This patch saves quite some time during guest kernel boot, as it
prevents from checking each bus for available devices.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-09-23 19:03:19 +02:00
Sebastien Boeuf
ec56710c9b devices: ioapic: Mask entries by default
When created, the IOAPIC entries should be masked, as it is the guest's
responsibility (FW and/or OS) to unmask them if/when necessary.

This patch saves a full round of port I/O writes from the guest to the
IOAPIC, meant for masking the unmasked entries.

Because they're now masked, the entries are not enabled, which means
they are not connected from a KVM perspective, saving from unneeded
registration/unregistration of the irq fds.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-09-23 19:03:19 +02:00
Sebastien Boeuf
827810dbd5 ci: Fix virtiofsd build by staying on older branch
While we figure out the details on how to correctly build virtiofsd from
the latest rebase from the branch "virtio-fs-dev" (which now relies on
QEMU's new build system), let's fix the CI by relying on an older branch
which still relies on the previous build system.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-09-23 17:06:34 +01:00
Henry Wang
c85c1f0d76 ci: AArch64: enable snapshot/restore integration test case
Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
961c5f2cb2 vmm: AArch64: enable VM states save/restore for AArch64
The states of GIC should be part of the VM states. This commit
enables the AArch64 VM states save/restore by adding save/restore
of GIC states.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
7c40a78b66 arch: Fix wrong trial of creating GICv3-ITS for non-PCI use cases
Currently for AArch64, the GICv3-ITS is tried to be created first
when PCI is not needed, which is unnecessary. This commit fixes
the problem.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
3ea4a0797d vmm: seccomp: unify AArch64 and x86_64 FTRUNCATE syscall
The definition of libc::SYS_ftruncate on AArch64 is different
from that on x86_64. This commit unifies the previously hard-coded
syscall number for AArch64.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
48544e4e82 vmm: seccomp: whitelist KVM_GET_REG_LIST in seccomp
`KVM_GET_REG_LIST` ioctl is needed in save/restore AArch64 vCPU.
Therefore we whitelist this ioctl in seccomp.

Also this commit unifies the `SYS_FTRUNCATE` syscall for x86_64
and AArch64.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
c6b47d39e0 vmm: refactor vCPU save/restore code in restoring VM
Similarly as the VM booting process, on AArch64 systems,
the vCPUs should be created before the creation of GIC. This
commit refactors the vCPU save/restore code to achieve the
above-mentioned restoring order.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
970a5a410d vmm: decouple vCPU init from configure_vcpus
Since calling `KVM_GET_ONE_REG` before `KVM_VCPU_INIT` will
result in an error: Exec format error (os error 8). This commit
decouples the vCPU init process from `configure_vcpus`. Therefore
in the process of restoring the vCPUs, these vCPUs can be
initialized separately before started.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
47e65cd341 vmm: AArch64: add methods to get saved vCPU states
The construction of `GICR_TYPER` register will need vCPU states.
Therefore this commit adds methods to extract saved vCPU states
from the cpu manager.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
381d0b4372 devices: remove the migration traits for the Gic struct
Unlike x86_64, the "interrupt_controller" in the device manager
for AArch64 is only a `Gic` object that implements the
`InterruptController` to provide the interrupt delivery service.
This is not the real GIC device so that we do not need to save
its states. Also, we do not need to insert it to the device_tree.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
39c9583b48 arch: AArch64: implement save/restore for GICv3
This commit implements the save/restore for GICv3.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
7ddcad1d8b arch: AArch64: add a field gicr_typers for GIC implementations
The value of GIC register `GICR_TYPER` is needed in restoring
the GIC states. This commit adds a field in the GIC device struct
and a method to construct its value.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
dcf6d9d731 device_manager: AArch64: add a field to set/get GIC device entity
In AArch64 systems, the state of GIC device can only be
retrieved from `KVM_GET_DEVICE_ATTR` ioctl. Therefore to implement
saving/restoring the GIC states, we need to make sure that the
GIC object (either the file descriptor or the device itself) can
be extracted after the VM is started.

This commit refactors the code of GIC creation by adding a new
field `gic_device_entity` in device manager and methods to set/get
this field. The GIC object can be therefore saved in the device
manager after calling `arch::configure_system`.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
e7acbcc184 arch: AArch64: support saving RDIST pending tables into guest RAM
This commit adds a function which allows to save RDIST pending
tables to the guest RAM, as well as unit test case for it.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
29ce3076c2 tests: AArch64: Add unit test cases for accessing GIC registers
This commit adds the unit test cases for getting/setting the GIC
distributor, redistributor and ICC registers.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
09d53aad11 arch: AArch64: Porting GIC icc_regs implementation
This commit ports the implementation of GIC ICC registers
from Firecracker.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
bfde6977c8 arch: AArch64: Porting GIC redist_regs implementation
This commit ports the implementation of GIC redistributor registers
from Firecracker.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
f53990c7e7 arch: AArch64: Porting GIC dist_regs implementation
This commit ports the implementation of GIC distributor registers
from Firecracker.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
89a6b63e6e hypervisor: Implement get_device_attr method for AArch64
This commit implements the `get_device_attr` method for the
`KVM_GET_DEVICE_ATTR` ioctl. This ioctl will be used in retrieving
the GIC status.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
b1285cf528 arch: AArch64: move GIC implementations to a separate module
This commit moves the GIC-related code to a separate module.
Therefore the implementation of GIC registers can be introduced
to the new module.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
9dd188a8e8 tests: AArch64: Add unit test cases for vCPU save/restore
Adds 3 more unit test cases for AArch64:

*save_restore_core_regs
*save_restore_system_regs
*get_set_mpstate

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
ffafeda4b6 AArch64: Implement AArch64 vCPU states save/restore
This commit adds methods to save/restore AArch64 vCPU registers,
including:

1. The AArch64 `VcpuKvmState` structure.

2. Some `Vcpu` trait methods of the `KvmVcpu` structure to
enable the save/restore of the AArch64 vCPU states.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Henry Wang
e3d45be6f7 AArch64: Preparation for vCPU save/restore
This commit ports code from firecracker and refactors the existing
AArch64 code as the preparation for implementing save/restore
AArch64 vCPU, including:

1. Modification of `arm64_core_reg` macro to retrive the index of
arm64 core register and implemention of a helper to determine if
a register is a system register.

2. Move some macros and helpers in `arch` crate to the `hypervisor`
crate.

3. Added related unit tests for above functions and macros.

Signed-off-by: Henry Wang <Henry.Wang@arm.com>
2020-09-23 12:37:25 +01:00
Josh Soref
5c3f4dbe6f ch: Fix various misspelled words
Misspellings were identified by https://github.com/marketplace/actions/check-spelling
* Initial corrections suggested by Google Sheets
* Additional corrections by Google Chrome auto-suggest
* Some manual corrections

Signed-off-by: Josh Soref <jsoref@users.noreply.github.com>
2020-09-23 08:59:31 +01:00
Jiangbo Wu
22a2a99e5f acpi: Add hotplug numa node
virtio-mem device would use 'VIRTIO_MEM_F_ACPI_PXM' to add memory to NUMA
node, which MUST be existed, otherwise it will be assigned to node id 0,
even if user specify different node id.

According ACPI spec about Memory Affinity Structure, system hardware
supports hot-add memory region using 'Hot Pluggable | Enabled' flags.

Signed-off-by: Jiangbo Wu <jiangbo.wu@intel.com>
2020-09-22 13:11:39 +02:00
Jiangbo Wu
223189c063 mm: Apply zone's property instread of global config
Apply memory zone's property for associated virtio-mem regions.

Signed-off-by: Jiangbo Wu <jiangbo.wu@intel.com>
2020-09-22 09:56:37 +02:00
Jiangbo Wu
80be8ac0dc mm: Apply memory policy for virtio-mem region
Use zone.host_numa_node to create memory zone, so that memory zone
can apply memory policy in according with host numa node ID

Signed-off-by: Jiangbo Wu <jiangbo.wu@intel.com>
2020-09-22 09:56:37 +02:00
dependabot-preview[bot]
097ba3b191 build(deps): bump hermit-abi from 0.1.15 to 0.1.16
Bumps [hermit-abi](https://github.com/hermitcore/libhermit-rs) from 0.1.15 to 0.1.16.
- [Release notes](https://github.com/hermitcore/libhermit-rs/releases)
- [Commits](https://github.com/hermitcore/libhermit-rs/commits)

Signed-off-by: dependabot-preview[bot] <support@dependabot.com>
2020-09-22 06:00:35 +00:00
Sebastien Boeuf
7c346c3844 vmm: Kill vhost-user self-spawned process on failure
If after the creation of the self-spawned backend, the VMM cannot create
the corresponding vhost-user frontend, the VMM must kill the freshly
spawned process in order to ensure the error propagation can happen.

In case the child process would still be around, the VMM cannot return
the error as it waits onto the child to terminate.

This should help us identify when self-spawned failures are caused by a
connection being refused between the VMM and the backend.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-09-18 17:26:25 +01:00
128 changed files with 5886 additions and 3157 deletions

View File

@@ -18,6 +18,9 @@ jobs:
steps:
- name: Code checkout
uses: actions/checkout@v2
with:
fetch-depth: 0
- name: Install Rust toolchain (${{ matrix.rust }})
uses: actions-rs/toolchain@v1
with:
@@ -25,17 +28,16 @@ jobs:
target: ${{ matrix.target }}
override: true
- name: Debug Build (default features)
run: cargo build --all --target=${{ matrix.target }}
- name: Debug Check (default features)
run: |
git rev-list origin/master..$GITHUB_SHA | xargs -t -I % sh -c 'git checkout %; cargo check --all --target=${{ matrix.target }}'
git checkout $GITHUB_SHA
- name: Build (pci,acpi,kvm)
run: cargo rustc --bin cloud-hypervisor --no-default-features --features "pci,acpi,kvm" -- -D warnings
- name: Build (acpi,kvm)
run: cargo rustc --bin cloud-hypervisor --no-default-features --features "acpi,kvm" -- -D warnings
- name: Build (pci,kvm)
run: cargo rustc --bin cloud-hypervisor --no-default-features --features "pci,kvm" -- -D warnings
- name: Build (kvm)
run: cargo rustc --bin cloud-hypervisor --no-default-features --features "kvm" -- -D warnings
- name: Build (mmio,kvm)
run: cargo rustc --bin cloud-hypervisor --no-default-features --features "mmio,kvm" -- -D warnings
- name: Release Build (default features)
run: cargo build --all --release --target=${{ matrix.target }}

View File

@@ -9,7 +9,7 @@ jobs:
strategy:
matrix:
rust:
- stable
- 1.46.0
target:
- aarch64-unknown-linux-gnu
- aarch64-unknown-linux-musl
@@ -29,4 +29,4 @@ jobs:
with:
use-cross: true
command: build
args: --target=${{ matrix.target }} --no-default-features --features "mmio,kvm"
args: --target=${{ matrix.target }} --no-default-features --features "kvm"

36
.github/workflows/docker-image.yaml vendored Normal file
View File

@@ -0,0 +1,36 @@
name: Cloud-Hypervisor's Docker image update
on:
push:
branches: master
paths: resources/Dockerfile
jobs:
main:
runs-on: ubuntu-latest
steps:
- name: Code checkout
uses: actions/checkout@v2
- name: Set up QEMU
uses: docker/setup-qemu-action@v1
- name: Set up Docker Buildx
uses: docker/setup-buildx-action@v1
- name: Login to DockerHub
uses: docker/login-action@v1
with:
username: ${{ secrets.DOCKERHUB_USERNAME }}
password: ${{ secrets.DOCKERHUB_TOKEN }}
- name: Build and push
uses: docker/build-push-action@v2
with:
file: ./resources/Dockerfile
platforms: linux/amd64,linux/arm64
push: true
tags: cloudhypervisor/dev:latest
- name: Image digest
run: echo ${{ steps.docker_build.outputs.digest }}

27
.github/workflows/fuzz-build.yaml vendored Normal file
View File

@@ -0,0 +1,27 @@
name: Cloud Hypervisor Cargo Fuzz Build
on: [pull_request, create]
jobs:
build:
if: github.event_name == 'pull_request'
name: Cargo Fuzz Build
runs-on: ubuntu-latest
strategy:
matrix:
rust:
- nightly
target:
- x86_64-unknown-linux-gnu
steps:
- name: Code checkout
uses: actions/checkout@v2
- name: Install Rust toolchain (${{ matrix.rust }})
uses: actions-rs/toolchain@v1
with:
toolchain: ${{ matrix.rust }}
target: ${{ matrix.target }}
override: true
- name: Install Cargo fuzz
run: cargo install -f cargo-fuzz
- name: Cargo Fuzz Build
run: cargo fuzz build

View File

@@ -26,15 +26,9 @@ jobs:
run: wget http://ftp.us.debian.org/debian/pool/main/d/device-tree-compiler/libfdt-dev_1.6.0-1_arm64.deb && dpkg-deb -xv libfdt-dev_1.6.0-1_arm64.deb ./tlibfdtdev && sudo mkdir /tmmmp && mkdir target && mkdir target/debug && mkdir target/debug/deps && sudo cp ./tlibfdtdev/usr/lib/aarch64-linux-gnu/libfdt.a target/debug/deps/libfdt.a && echo "libfdt installed"
- name: Formatting (rustfmt)
run: cargo fmt -- --check
- name: Clippy (mmio,kvm)
- name: Clippy (kvm)
uses: actions-rs/cargo@v1
with:
use-cross: true
command: clippy
args: --target=${{ matrix.target }} --no-default-features --features "mmio,kvm" -- -D warnings
- name: Clippy (pci,kvm)
uses: actions-rs/cargo@v1
with:
use-cross: true
command: clippy
args: --target=${{ matrix.target }} --no-default-features --features "pci,kvm" -- -D warnings
args: --target=${{ matrix.target }} --no-default-features --features "kvm" -- -D warnings

View File

@@ -29,11 +29,8 @@ jobs:
- name: Clippy (all features)
run: cargo clippy --all-targets --all-features -- -D warnings
- name: Clippy (pci,acpi,kvm)
run: cargo clippy --all-targets --no-default-features --features "pci,acpi,kvm" -- -D warnings
- name: Clippy (acpi,kvm)
run: cargo clippy --all-targets --no-default-features --features "acpi,kvm" -- -D warnings
- name: Clippy (pci,kvm)
run: cargo clippy --all-targets --no-default-features --features "pci,kvm" -- -D warnings
- name: Clippy (mmio,kvm)
run: cargo clippy --all-targets --no-default-features --features "mmio,kvm" -- -D warnings
- name: Clippy (kvm)
run: cargo clippy --all-targets --no-default-features --features "kvm" -- -D warnings

View File

@@ -88,5 +88,5 @@ Fixes #88
Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
```
Then, after the correspoding PR is merged, Github will automatically close that issue when parsing the
Then, after the corresponding PR is merged, Github will automatically close that issue when parsing the
[commit message](https://help.github.com/articles/closing-issues-via-commit-messages/).

183
Cargo.lock generated
View File

@@ -24,9 +24,9 @@ checksum = "ee2a4ec343196209d6594e19543ae87a39f96d5534d7174822a3ad825dd6ed7e"
[[package]]
name = "aho-corasick"
version = "0.7.13"
version = "0.7.14"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "043164d8ba5c4c3035fec9bbee8647c0261d788f3474306f93bb65901cae0e86"
checksum = "b476ce7103678b0c6d3d395dbbae31d48ff910bd28be979ba5d48c6351131d0d"
dependencies = [
"memchr",
]
@@ -42,9 +42,13 @@ dependencies = [
[[package]]
name = "anyhow"
version = "1.0.32"
version = "1.0.33"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "6b602bfe940d21c130f3895acd65221e8a61270debe89d628b9cb4e3ccb8569b"
checksum = "a1fd36ffbb1fb7c834eac128ea8d0e310c5aeb635548f9d58861e1308d46e71c"
[[package]]
name = "api_client"
version = "0.1.0"
[[package]]
name = "arc-swap"
@@ -65,7 +69,11 @@ dependencies = [
"linux-loader",
"log 0.4.11",
"rand 0.7.3",
"serde",
"serde_derive",
"serde_json",
"vm-memory",
"vm-migration",
]
[[package]]
@@ -80,9 +88,9 @@ checksum = "a4c527152e37cf757a3f78aae5a06fbeefdb07ccc535c980a3208ee3060dd544"
[[package]]
name = "arrayvec"
version = "0.5.1"
version = "0.5.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "cff77d8686867eceff3105329d4698d96c2391c176d5d03adc90c7389162b5b8"
checksum = "23b62fc65de8e4e7f52534fb52b0f3ed04746ae267519eef2a83941e8085068b"
[[package]]
name = "atty"
@@ -103,12 +111,12 @@ checksum = "cdb031dd78e28731d87d56cc8ffef4a8f36ca26c38fe2de700543e627f8a464a"
[[package]]
name = "backtrace"
version = "0.3.50"
version = "0.3.53"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "46254cf2fdcdf1badb5934448c1bcbe046a56537b3987d96c51a7afc5d03f293"
checksum = "707b586e0e2f247cbde68cdd2c3ce69ea7b7be43e1c5b426e37c9319c4b9838e"
dependencies = [
"addr2line",
"cfg-if",
"cfg-if 1.0.0",
"libc",
"miniz_oxide",
"object",
@@ -168,9 +176,9 @@ checksum = "08c48aae112d48ed9f069b33538ea9e3e90aa263cfa3d1c24309612b1f7472de"
[[package]]
name = "cc"
version = "1.0.60"
version = "1.0.61"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "ef611cc68ff783f18535d77ddd080185275713d852c4f5cbb6122c462a7a825c"
checksum = "ed67cbde08356238e75fc4656be4749481eeffb09e19f320a25237d5221c985d"
[[package]]
name = "cfg-if"
@@ -178,6 +186,12 @@ version = "0.1.10"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "4785bdd1c96b2a846b2bd7cc02e86b6b3dbf14e7e53446c4f54c92a361040822"
[[package]]
name = "cfg-if"
version = "1.0.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "baf1de4339761588bc0619e3cbc0120ee582ebb74b53b4efbf79117bd2da40fd"
[[package]]
name = "clap"
version = "2.33.3"
@@ -196,11 +210,14 @@ dependencies = [
[[package]]
name = "cloud-hypervisor"
version = "0.10.0"
version = "0.11.0"
dependencies = [
"anyhow",
"api_client",
"clap",
"credibility",
"dirs",
"epoll",
"hypervisor",
"lazy_static",
"libc",
@@ -212,10 +229,12 @@ dependencies = [
"ssh2",
"tempdir",
"tempfile",
"thiserror",
"vhost_user_block",
"vhost_user_net",
"vmm",
"vmm-sys-util",
"wait-timeout",
]
[[package]]
@@ -250,7 +269,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "c3c7c73a2d1e9fc0886a08b93e98eb643461230d5f1925e4036204d5f2e261a8"
dependencies = [
"autocfg",
"cfg-if",
"cfg-if 0.1.10",
"lazy_static",
]
@@ -335,9 +354,9 @@ checksum = "a06f77d526c1a601b7c4cdd98f54b5eaabffc14d5f2f0296febdc7f357c6d3ba"
[[package]]
name = "futures"
version = "0.3.5"
version = "0.3.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "1e05b85ec287aac0dc34db7d4a569323df697f9c55b99b15d6b4ef8cde49f613"
checksum = "95314d38584ffbfda215621d723e0a3906f032e03ae5551e650058dac83d4797"
dependencies = [
"futures-channel",
"futures-core",
@@ -350,9 +369,9 @@ dependencies = [
[[package]]
name = "futures-channel"
version = "0.3.5"
version = "0.3.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f366ad74c28cca6ba456d95e6422883cfb4b252a83bed929c83abfdbbf2967d5"
checksum = "0448174b01148032eed37ac4aed28963aaaa8cfa93569a08e5b479bbc6c2c151"
dependencies = [
"futures-core",
"futures-sink",
@@ -360,15 +379,15 @@ dependencies = [
[[package]]
name = "futures-core"
version = "0.3.5"
version = "0.3.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "59f5fff90fd5d971f936ad674802482ba441b6f09ba5e15fd8b39145582ca399"
checksum = "18eaa56102984bed2c88ea39026cff3ce3b4c7f508ca970cedf2450ea10d4e46"
[[package]]
name = "futures-executor"
version = "0.3.5"
version = "0.3.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "10d6bb888be1153d3abeb9006b11b02cf5e9b209fda28693c31ae1e4e012e314"
checksum = "f5f8e0c9258abaea85e78ebdda17ef9666d390e987f006be6080dfe354b708cb"
dependencies = [
"futures-core",
"futures-task",
@@ -378,15 +397,15 @@ dependencies = [
[[package]]
name = "futures-io"
version = "0.3.5"
version = "0.3.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "de27142b013a8e869c14957e6d2edeef89e97c289e69d042ee3a49acd8b51789"
checksum = "6e1798854a4727ff944a7b12aa999f58ce7aa81db80d2dfaaf2ba06f065ddd2b"
[[package]]
name = "futures-macro"
version = "0.3.5"
version = "0.3.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d0b5a30a4328ab5473878237c447333c093297bded83a4983d10f4deea240d39"
checksum = "e36fccf3fc58563b4a14d265027c627c3b665d7fed489427e88e7cc929559efe"
dependencies = [
"proc-macro-hack",
"proc-macro2",
@@ -396,24 +415,24 @@ dependencies = [
[[package]]
name = "futures-sink"
version = "0.3.5"
version = "0.3.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "3f2032893cb734c7a05d85ce0cc8b8c4075278e93b24b66f9de99d6eb0fa8acc"
checksum = "0e3ca3f17d6e8804ae5d3df7a7d35b2b3a6fe89dac84b31872720fc3060a0b11"
[[package]]
name = "futures-task"
version = "0.3.5"
version = "0.3.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "bdb66b5f09e22019b1ab0830f7785bcea8e7a42148683f99214f73f8ec21a626"
checksum = "96d502af37186c4fef99453df03e374683f8a1eec9dcc1e66b3b82dc8278ce3c"
dependencies = [
"once_cell",
]
[[package]]
name = "futures-util"
version = "0.3.5"
version = "0.3.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "8764574ff08b701a084482c3c7031349104b07ac897393010494beaa18ce32c6"
checksum = "abcb44342f62e6f3e8ac427b8aa815f724fd705dfad060b18ac7866c15bb8e34"
dependencies = [
"futures-channel",
"futures-core",
@@ -435,7 +454,7 @@ version = "0.1.15"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "fc587bc0ec293155d5bfa6b9891ec18a1e330c234f896ea47fbada4cadbe47e6"
dependencies = [
"cfg-if",
"cfg-if 0.1.10",
"libc",
"wasi",
]
@@ -454,9 +473,9 @@ checksum = "8be18de09a56b60ed0edf84bc9df007e30040691af7acd1c41874faac5895bfb"
[[package]]
name = "hermit-abi"
version = "0.1.15"
version = "0.1.17"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "3deed196b6e7f9e44a2ae8d94225d80302d81208b1bb673fd21fe634645c85a9"
checksum = "5aca5565f760fb5b220e499d72710ed156fdb74e631659e99377d9ebfbd13ae8"
dependencies = [
"libc",
]
@@ -466,6 +485,7 @@ name = "hypervisor"
version = "0.1.0"
dependencies = [
"anyhow",
"arc-swap",
"kvm-bindings",
"kvm-ioctls",
"libc",
@@ -553,9 +573,9 @@ checksum = "e2abad23fbc42b3700f2f279844dc832adb2b2eb069b2df918f455c4e18cc646"
[[package]]
name = "libc"
version = "0.2.77"
version = "0.2.80"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f2f96b10ec2560088a8e76961b00d47107b3a625fecb76dedb29ee7ccbf98235"
checksum = "4d58d1b70b004888f764dfbf6a26a3b0342a1632d33968e4a179d8011c760614"
[[package]]
name = "libssh2-sys"
@@ -615,7 +635,7 @@ version = "0.4.11"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "4fabed175da42fed1fa0746b0ea71f412aa9d35e76e95e59b192c64b9dc2bf8b"
dependencies = [
"cfg-if",
"cfg-if 0.1.10",
]
[[package]]
@@ -626,9 +646,9 @@ checksum = "7ffc5c5338469d4d3ea17d269fa8ea3512ad247247c30bd2df69e68309ed0a08"
[[package]]
name = "memchr"
version = "2.3.3"
version = "2.3.4"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "3728d817d99e5ac407411fa471ff9800a778d88a24685968b36824eaf4bee400"
checksum = "0ee1c47aaa256ecabcaea351eae4a9b01ef39ed810004e298d2511ed284b1525"
[[package]]
name = "micro_http"
@@ -640,9 +660,9 @@ dependencies = [
[[package]]
name = "miniz_oxide"
version = "0.4.2"
version = "0.4.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "c60c0dfe32c10b43a144bad8fc83538c52f58302c92300ea7ec7bf7b38d5a7b9"
checksum = "0f2d26ec3309788e423cfbf68ad1800f061638098d76a83681af979dc4eda19d"
dependencies = [
"adler",
"autocfg",
@@ -686,9 +706,9 @@ dependencies = [
[[package]]
name = "object"
version = "0.20.0"
version = "0.21.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "1ab52be62400ca80aa00285d25253d7f7c437b7375c4de678f5405d3afe82ca5"
checksum = "37fd5004feb2ce328a52b0b3d01dbf4ffff72583493900ed15f22d4111c51693"
[[package]]
name = "once_cell"
@@ -729,7 +749,7 @@ version = "0.7.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d58c7c768d4ba344e3e8d72518ac13e259d7c7ade24167003b8488e10b6740a3"
dependencies = [
"cfg-if",
"cfg-if 0.1.10",
"cloudabi",
"libc",
"redox_syscall",
@@ -766,18 +786,18 @@ checksum = "d4fd5641d01c8f18a23da7b6fe29298ff4b55afcccdf78973b24cf3175fee32e"
[[package]]
name = "pin-project"
version = "0.4.23"
version = "1.0.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "ca4433fff2ae79342e497d9f8ee990d174071408f28f726d6d83af93e58e48aa"
checksum = "ee41d838744f60d959d7074e3afb6b35c7456d0f61cad38a24e35e6553f73841"
dependencies = [
"pin-project-internal",
]
[[package]]
name = "pin-project-internal"
version = "0.4.23"
version = "1.0.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "2c0e815c3ee9a031fdf5af21c10aa17c573c9c6a566328d99e3936c34e36461f"
checksum = "81a4ffa594b66bff340084d4081df649a7dc049ac8d7fc458d8e628bfbbb2f86"
dependencies = [
"proc-macro2",
"quote",
@@ -792,9 +812,9 @@ checksum = "8b870d8c151b6f2fb93e84a13146138f05d02ed11c7e7c54f8826aaaf7c9f184"
[[package]]
name = "pkg-config"
version = "0.3.18"
version = "0.3.19"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d36492546b6af1463394d46f0c834346f31548646f6ba10849802c9c9a27ac33"
checksum = "3831453b3449ceb48b6d9c7ad7c96d5ea673e9b470a1dc578c2ce6521230884c"
[[package]]
name = "pnet"
@@ -893,9 +913,9 @@ checksum = "c36fa947111f5c62a733b652544dd0016a43ce89619538a8ef92724a6f501a20"
[[package]]
name = "proc-macro-hack"
version = "0.5.18"
version = "0.5.19"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "99c605b9a0adc77b7211c6b1f722dcb613d68d66859a44f3d485a6da332b0598"
checksum = "dbf0c48bc1d91375ae5c3cd81e3722dff1abcf81a30960240640d223f59fe0e5"
[[package]]
name = "proc-macro-nested"
@@ -905,9 +925,9 @@ checksum = "eba180dafb9038b050a4c280019bbedf9f2467b61e5d892dcad585bb57aadc5a"
[[package]]
name = "proc-macro2"
version = "1.0.21"
version = "1.0.24"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "36e28516df94f3dd551a587da5357459d9b36d945a7c37c3557928c1c2ff2a2c"
checksum = "1e0704ee1a7e00d7bb417d0770ea303c1bccbabf0ef1667dae92b5967f5f8a71"
dependencies = [
"unicode-xid 0.2.1",
]
@@ -1030,9 +1050,9 @@ dependencies = [
[[package]]
name = "regex"
version = "1.3.9"
version = "1.4.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9c3780fcf44b193bc4d09f36d2a3c87b251da4a046c87795a0d35f4f927ad8e6"
checksum = "8963b85b8ce3074fecffde43b4b0dded83ce2f367dc8d363afc56679f3ee820b"
dependencies = [
"aho-corasick",
"memchr",
@@ -1042,9 +1062,9 @@ dependencies = [
[[package]]
name = "regex-syntax"
version = "0.6.18"
version = "0.6.20"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "26412eb97c6b088a6997e05f69403a802a92d520de2f8e63c2b65f9e0f47c4e8"
checksum = "8cab7a364d15cde1e505267766a2d3c4e22a843e1a601f0fa7564c0f82ced11c"
[[package]]
name = "remain"
@@ -1080,9 +1100,9 @@ dependencies = [
[[package]]
name = "rustc-demangle"
version = "0.1.16"
version = "0.1.18"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "4c691c0e608126e00913e33f0ccf3727d5fc84573623b8d65b2df340b5201783"
checksum = "6e3bad0ee36814ca07d7968269dd4b7ec89ec2da10c4bb613928d3077083c232"
[[package]]
name = "rustc-serialize"
@@ -1112,15 +1132,15 @@ dependencies = [
[[package]]
name = "serde"
version = "1.0.116"
version = "1.0.117"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "96fe57af81d28386a513cbc6858332abc6117cfdb5999647c6444b8f43a370a5"
checksum = "b88fa983de7720629c9387e9f517353ed404164b1e482c970a90c1a4aaf7dc1a"
[[package]]
name = "serde_derive"
version = "1.0.116"
version = "1.0.117"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f630a6370fd8e457873b4bd2ffdae75408bc291ba72be773772a4c2a065d9ae8"
checksum = "cbd1ae72adb44aab48f325a02444a5fc079349a8d804c1fc922aed3f7454c74e"
dependencies = [
"proc-macro2",
"quote",
@@ -1129,9 +1149,9 @@ dependencies = [
[[package]]
name = "serde_json"
version = "1.0.57"
version = "1.0.59"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "164eacbdb13512ec2745fb09d51fd5b22b0d65ed294a1dcf7285a360c80a675c"
checksum = "dcac07dbffa1c65e7f816ab9eba78eb142c6d44410f4eeba1e26e4f5dfa56b95"
dependencies = [
"itoa",
"ryu",
@@ -1190,9 +1210,9 @@ checksum = "8ea5119cdb4c55b55d432abb513a0429384878c15dde60cc77b1c99de1a95a6a"
[[package]]
name = "syn"
version = "1.0.41"
version = "1.0.48"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "6690e3e9f692504b941dc6c3b188fd28df054f7fb8469ab40680df52fdcc842b"
checksum = "cc371affeffc477f42a221a1e4297aedcea33d47d19b61455588bd9d8f6b19ac"
dependencies = [
"proc-macro2",
"quote",
@@ -1276,7 +1296,7 @@ version = "3.1.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "7a6e24d9338a0a5be79593e2fa15a648add6138caa803e2d5bc782c371732ca9"
dependencies = [
"cfg-if",
"cfg-if 0.1.10",
"libc",
"rand 0.7.3",
"redox_syscall",
@@ -1316,18 +1336,18 @@ dependencies = [
[[package]]
name = "thiserror"
version = "1.0.20"
version = "1.0.21"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "7dfdd070ccd8ccb78f4ad66bf1982dc37f620ef696c6b5028fe2ed83dd3d0d08"
checksum = "318234ffa22e0920fe9a40d7b8369b5f649d490980cf7aadcf1eb91594869b42"
dependencies = [
"thiserror-impl",
]
[[package]]
name = "thiserror-impl"
version = "1.0.20"
version = "1.0.21"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "bd80fc12f73063ac132ac92aceea36734f04a1d93c1240c6944e23a3b8841793"
checksum = "cae2447b6282786c3493999f40a9be2a6ad20cb8bd268b0a0dbf5a065535c0ab"
dependencies = [
"proc-macro2",
"quote",
@@ -1434,7 +1454,7 @@ dependencies = [
[[package]]
name = "vhost"
version = "0.1.0"
source = "git+https://github.com/rust-vmm/vhost#d257479a40b42895b9adab11befa5385bbee0698"
source = "git+https://github.com/cloud-hypervisor/vhost?branch=ch#6fc980bf5083d67586ab91d6a965c3a593c33a78"
dependencies = [
"bitflags 1.2.1",
"libc",
@@ -1572,8 +1592,8 @@ dependencies = [
[[package]]
name = "vm-memory"
version = "0.2.2"
source = "git+https://github.com/cloud-hypervisor/vm-memory?branch=ch#b28148d201b464acdef3124116ea43bb645cfd2f"
version = "0.3.0"
source = "git+https://github.com/cloud-hypervisor/vm-memory?branch=ch#6f2f7107f29a8f2b66b64a849aeace20783ec8f3"
dependencies = [
"arc-swap",
"libc",
@@ -1611,6 +1631,7 @@ dependencies = [
"anyhow",
"arc-swap",
"arch",
"bitflags 1.2.1",
"block_util",
"clap",
"credibility",
@@ -1632,6 +1653,7 @@ dependencies = [
"serde_json",
"signal-hook",
"tempfile",
"thiserror",
"url",
"vfio-ioctls",
"virtio-devices",
@@ -1655,6 +1677,15 @@ dependencies = [
"serde_derive",
]
[[package]]
name = "wait-timeout"
version = "0.2.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9f200f5b12eb75f8c1ed65abd4b2db8a6e1b138a20de009dacee265a2498f3f6"
dependencies = [
"libc",
]
[[package]]
name = "wasi"
version = "0.9.0+wasi-snapshot-preview1"

View File

@@ -1,6 +1,6 @@
[package]
name = "cloud-hypervisor"
version = "0.10.0"
version = "0.11.0"
authors = ["The Cloud Hypervisor Authors"]
edition = "2018"
default-run = "cloud-hypervisor"
@@ -13,23 +13,35 @@ homepage = "https://github.com/cloud-hypervisor/cloud-hypervisor"
lto = true
[dependencies]
anyhow = "1.0"
api_client = { path = "api_client" }
clap = { version = "2.33.3", features = ["wrap_help"] }
epoll = ">=4.0.1"
hypervisor = { path = "hypervisor" }
libc = "0.2.77"
libc = "0.2.80"
log = { version = "0.4.11", features = ["std"] }
option_parser = { path = "option_parser" }
seccomp = { git = "https://github.com/firecracker-microvm/firecracker", tag = "v0.22.0" }
serde_json = "1.0.57"
serde_json = "1.0.59"
thiserror = "1.0"
vhost_user_block = { path = "vhost_user_block"}
vhost_user_net = { path = "vhost_user_net"}
vmm = { path = "vmm" }
vmm-sys-util = "0.6.1"
wait-timeout = "0.2.0"
[build-dependencies]
clap = { version = "2.33.3", features = ["wrap_help"] }
# List of patched crates
[patch.crates-io]
vm-memory = { git = "https://github.com/cloud-hypervisor/vm-memory", branch = "ch" }
[patch.'https://github.com/rust-vmm/vhost']
vhost_rs = { git = "https://github.com/cloud-hypervisor/vhost", branch = "ch", package = "vhost", features = ["vhost-user-master", "vhost-user-slave"] }
[patch.'https://github.com/rust-vmm/kvm-ioctls']
kvm-ioctls = { git = "https://github.com/cloud-hypervisor/kvm-ioctls", branch = "ch" }
[patch.'https://github.com/rust-vmm/kvm-bindings']
kvm-bindings = { git = "https://github.com/cloud-hypervisor/kvm-bindings", branch = "ch", features = ["with-serde", "fam-wrappers"] }
[dev-dependencies]
ssh2 = "0.8.2"
@@ -38,14 +50,12 @@ credibility = "0.1.3"
tempdir = "0.3.7"
lazy_static= "1.4.0"
tempfile = "3.1.0"
serde_json = "1.0.57"
serde_json = "1.0.59"
net_util = { path = "net_util" }
[features]
default = ["acpi", "pci", "cmos", "kvm"]
default = ["acpi", "cmos", "io_uring", "kvm"]
acpi = ["vmm/acpi"]
pci = ["vmm/pci_support"]
mmio = ["vmm/mmio_support"]
cmos = ["vmm/cmos"]
fwdebug = ["vmm/fwdebug"]
kvm = ["vmm/kvm"]
@@ -57,6 +67,7 @@ integration_tests = []
[workspace]
members = [
"acpi_tables",
"api_client",
"arch",
"arch_gen",
"block_util",

42
Jenkinsfile vendored
View File

@@ -25,7 +25,7 @@ pipeline{
failFast true
parallel {
stage ('Worker build') {
agent { node { label 'bionic' } }
agent { node { label 'groovy' } }
options {
timeout(time: 1, unit: 'HOURS')
}
@@ -82,7 +82,7 @@ pipeline{
}
}
stage ('Worker build (musl)') {
agent { node { label 'bionic' } }
agent { node { label 'groovy' } }
options {
timeout(time: 1, unit: 'HOURS')
}
@@ -134,6 +134,44 @@ pipeline{
}
}
}
stage ('Worker build - Windows guest') {
agent { node { label 'groovy-win' } }
options {
timeout(time: 1, unit: 'HOURS')
}
stages {
stage ('Checkout') {
steps {
checkout scm
}
}
stage ('Download assets') {
steps {
sh "mkdir ${env.HOME}/workloads"
azureDownload(storageCredentialId: 'ch-image-store',
containerName: 'private-images',
includeFilesPattern: 'OVMF.fd',
downloadType: 'container',
downloadDirLoc: "${env.HOME}/workloads")
azureDownload(storageCredentialId: 'ch-image-store',
containerName: 'private-images',
includeFilesPattern: 'windows-server-2019.raw',
downloadType: 'container',
downloadDirLoc: "${env.HOME}/workloads")
}
}
stage ('Run Windows guest integration tests') {
steps {
sh "scripts/dev_cli.sh tests --integration-windows"
}
}
stage ('Run Windows guest integration tests for musl') {
steps {
sh "scripts/dev_cli.sh tests --integration-windows --libc musl"
}
}
}
}
}
}
}

View File

@@ -1,28 +1,28 @@
[![Build Status](https://travis-ci.com/cloud-hypervisor/cloud-hypervisor.svg?branch=master)](https://travis-ci.com/cloud-hypervisor/cloud-hypervisor)
- [1. What is Cloud Hypervisor?](#1-what-is-cloud-hypervisor)
- [Objectives](#objectives)
- [High Level](#high-level)
- [Architectures](#architectures)
- [Guest OS](#guest-os)
- [Objectives](#objectives)
- [High Level](#high-level)
- [Architectures](#architectures)
- [Guest OS](#guest-os)
- [2. Getting Started](#2-getting-started)
- [Clone and build](#clone-and-build)
- [Containerized builds and tests](#containerized-builds-and-tests)
- [Run](#run)
- [Cloud image](#cloud-image)
- [Custom kernel and disk image](#custom-kernel-and-disk-image)
- [Building your kernel](#building-your-kernel)
- [Disk image](#disk-image)
- [Booting the guest VM](#booting-the-guest-vm)
- [Clone and build](#clone-and-build)
- [Containerized builds and tests](#containerized-builds-and-tests)
- [Run](#run)
- [Cloud image](#cloud-image)
- [Custom kernel and disk image](#custom-kernel-and-disk-image)
- [Building your kernel](#building-your-kernel)
- [Disk image](#disk-image)
- [Booting the guest VM](#booting-the-guest-vm)
- [3. Status](#3-status)
- [Hot Plug](#hot-plug)
- [Device Model](#device-model)
- [TODO](#todo)
- [Hot Plug](#hot-plug)
- [Device Model](#device-model)
- [TODO](#todo)
- [4. `rust-vmm` project dependency](#4-rust-vmm-project-dependency)
- [Firecracker and crosvm](#firecracker-and-crosvm)
- [Firecracker and crosvm](#firecracker-and-crosvm)
- [5. Community](#5-community)
- [Contribute](#contribute)
- [Join us](#join-us)
- [Contribute](#contribute)
- [Join us](#join-us)
# 1. What is Cloud Hypervisor?
@@ -36,24 +36,24 @@ Cloud Hypervisor is implemented in [Rust](https://www.rust-lang.org/) and is bas
### High Level
* KVM based
* Minimal emulation
* Low latency
* Low memory footprint
* Low complexity
* High performance
* Small attack surface
* 64-bit support only
* CPU, memory, PCI hotplug
* Machine to machine migration
- KVM based
- Minimal emulation
- Low latency
- Low memory footprint
- Low complexity
- High performance
- Small attack surface
- 64-bit support only
- CPU, memory, PCI hotplug
- Machine to machine migration
### Architectures
Cloud Hypervisor supports the `x86-64` and `AArch64` architecture. There are some small differences in functionality between the two architecture (see #1125).
Cloud Hypervisor supports the `x86-64` and `AArch64` architectures. There are some small differences in functionality between the two architectures (see [#1125](https://github.com/cloud-hypervisor/cloud-hypervisor/issues/1125)).
### Guest OS
Cloud Hypervisor supports `64-bit Linux` with support for *modern* 64-bit Windows guests currently under development.
Cloud Hypervisor supports `64-bit Linux` with support for _modern_ 64-bit Windows guests currently under development.
# 2. Getting Started
@@ -64,6 +64,23 @@ $ export CLOUDH=$HOME/cloud-hypervisor
$ mkdir $CLOUDH
```
## Install prerequisites
You need to install some prerequisite packages in order to build and test Cloud Hypervisor.
Here, all the steps are based on Ubuntu, for other Linux distributions please replace the
package manager and package name.
```shell
# Install git
$ sudo apt install git
# Install rust tool chain
$ curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
# Install build-essential
$ sudo apt install build-essential
# If you want to build statically linked binary please add musl target
$ rustup target add x86_64-unknown-linux-musl
```
## Clone and build
First you need to clone and build the cloud-hypervisor repo:
@@ -77,6 +94,8 @@ $ cargo build --release
# We need to give the cloud-hypervisor binary the NET_ADMIN capabilities for it to set TAP interfaces up on the host.
$ sudo setcap cap_net_admin+ep ./target/release/cloud-hypervisor
# If you want to build statically linked binary
$ cargo build --release --target=x86_64-unknown-linux-musl --all
$ popd
```
@@ -117,7 +136,7 @@ You can run a guest VM by either using an existing cloud image or booting into y
### Cloud image
Cloud Hypervisor supports booting disk images containing all needed
components to run cloud workloads, a.k.a. cloud images. To do that we rely on
components to run cloud workloads, a.k.a. cloud images. To do that we rely on
the [Rust Hypervisor
Firmware](https://github.com/cloud-hypervisor/rust-hypervisor-firmware) project to provide
an ELF
@@ -225,8 +244,9 @@ $ ./cloud-hypervisor/target/release/cloud-hypervisor \
Cloud Hypervisor is under active development. No API or feature stability is guaranteed.
As of 2020-07-02, the following cloud images are supported:
* [Ubuntu Bionic](https://cloud-images.ubuntu.com/bionic/current/) (cloudimg)
* [Ubuntu Focal](https://cloud-images.ubuntu.com/focal/current/) (cloudimg)
- [Ubuntu Bionic](https://cloud-images.ubuntu.com/bionic/current/) (cloudimg)
- [Ubuntu Focal](https://cloud-images.ubuntu.com/focal/current/) (cloudimg)
Direct kernel boot to userspace should work with a rootfs from most distributions.
@@ -297,4 +317,3 @@ etc, are all equal and welcome means of contribution. See the [CONTRIBUTING](CON
Get an [invite to our Slack channel](https://join.slack.com/t/cloud-hypervisor/shared_invite/enQtNjY3MTE3MDkwNDQ4LWQ1MTA1ZDVmODkwMWQ1MTRhYzk4ZGNlN2UwNTI3ZmFlODU0OTcwOWZjMTkwZDExYWE3YjFmNzgzY2FmNDAyMjI)
and [join us on Slack](https://cloud-hypervisor.slack.com/).

View File

@@ -5,5 +5,5 @@ authors = ["The Cloud Hypervisor Authors"]
edition = "2018"
[dependencies]
vm-memory = "0.2.1"
vm-memory = "0.3.0"

5
api_client/Cargo.toml Normal file
View File

@@ -0,0 +1,5 @@
[package]
name = "api_client"
version = "0.1.0"
authors = ["The Cloud Hypervisor Authors"]
edition = "2018"

175
api_client/src/lib.rs Normal file
View File

@@ -0,0 +1,175 @@
// Copyright © 2020 Intel Corporation
//
// SPDX-License-Identifier: Apache-2.0
//
use std::fmt;
use std::io::{Read, Write};
#[derive(Debug)]
pub enum Error {
Socket(std::io::Error),
StatusCodeParsing(std::num::ParseIntError),
MissingProtocol,
ContentLengthParsing(std::num::ParseIntError),
ServerResponse(StatusCode, Option<String>),
}
impl fmt::Display for Error {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
use Error::*;
match self {
Socket(e) => write!(f, "Error writing to or reading from HTTP socket: {}", e),
StatusCodeParsing(e) => write!(f, "Error parsing HTTP status code: {}", e),
MissingProtocol => write!(f, "HTTP output is missing protocol statement"),
ContentLengthParsing(e) => write!(f, "Error parsing HTTP Content-Length field: {}", e),
ServerResponse(s, o) => {
if let Some(o) = o {
write!(f, "Server responded with an error: {:?}: {}", s, o)
} else {
write!(f, "Server responded with an error: {:?}", s)
}
}
}
}
}
#[derive(Clone, Copy, Debug)]
pub enum StatusCode {
Continue,
OK,
NoContent,
BadRequest,
NotFound,
InternalServerError,
NotImplemented,
Unknown,
}
impl StatusCode {
fn from_raw(code: usize) -> StatusCode {
match code {
100 => StatusCode::Continue,
200 => StatusCode::OK,
204 => StatusCode::NoContent,
400 => StatusCode::BadRequest,
404 => StatusCode::NotFound,
500 => StatusCode::InternalServerError,
501 => StatusCode::NotImplemented,
_ => StatusCode::Unknown,
}
}
fn parse(code: &str) -> Result<StatusCode, Error> {
Ok(StatusCode::from_raw(
code.trim().parse().map_err(Error::StatusCodeParsing)?,
))
}
fn is_server_error(self) -> bool {
!matches!(
self,
StatusCode::OK | StatusCode::Continue | StatusCode::NoContent
)
}
}
fn get_header<'a>(res: &'a str, header: &'a str) -> Option<&'a str> {
let header_str = format!("{}: ", header);
if let Some(o) = res.find(&header_str) {
Some(&res[o + header_str.len()..o + res[o..].find('\r').unwrap()])
} else {
None
}
}
fn get_status_code(res: &str) -> Result<StatusCode, Error> {
if let Some(o) = res.find("HTTP/1.1") {
Ok(StatusCode::parse(
&res[o + "HTTP/1.1 ".len()..res[o..].find('\r').unwrap()],
)?)
} else {
Err(Error::MissingProtocol)
}
}
fn parse_http_response(socket: &mut dyn Read) -> Result<Option<String>, Error> {
let mut res = String::new();
let mut body_offset = None;
let mut content_length: Option<usize> = None;
loop {
let mut bytes = vec![0; 256];
let count = socket.read(&mut bytes).map_err(Error::Socket)?;
res.push_str(std::str::from_utf8(&bytes[0..count]).unwrap());
// End of headers
if let Some(o) = res.find("\r\n\r\n") {
body_offset = Some(o + "\r\n\r\n".len());
// With all headers available we can see if there is any body
content_length = if let Some(length) = get_header(&res, "Content-Length") {
Some(length.trim().parse().map_err(Error::ContentLengthParsing)?)
} else {
None
};
if content_length.is_none() {
break;
}
}
if let Some(body_offset) = body_offset {
if let Some(content_length) = content_length {
if res.len() >= content_length + body_offset {
break;
}
}
}
}
let body_string = content_length.and(Some(String::from(&res[body_offset.unwrap()..])));
let status_code = get_status_code(&res)?;
if status_code.is_server_error() {
Err(Error::ServerResponse(status_code, body_string))
} else {
Ok(body_string)
}
}
pub fn simple_api_command<T: Read + Write>(
socket: &mut T,
method: &str,
c: &str,
request_body: Option<&str>,
) -> Result<(), Error> {
socket
.write_all(
format!(
"{} /api/v1/vm.{} HTTP/1.1\r\nHost: localhost\r\nAccept: */*\r\n",
method, c
)
.as_bytes(),
)
.map_err(Error::Socket)?;
if let Some(request_body) = request_body {
socket
.write_all(format!("Content-Length: {}\r\n", request_body.len()).as_bytes())
.map_err(Error::Socket)?;
}
socket.write_all(b"\r\n").map_err(Error::Socket)?;
if let Some(request_body) = request_body {
socket
.write_all(request_body.as_bytes())
.map_err(Error::Socket)?;
}
socket.flush().map_err(Error::Socket)?;
if let Some(body) = parse_http_response(socket)? {
println!("{}", body);
}
Ok(())
}

View File

@@ -11,9 +11,13 @@ acpi = ["acpi_tables"]
anyhow = "1.0"
byteorder = "1.3.4"
hypervisor = { path = "../hypervisor" }
libc = "0.2.77"
libc = "0.2.80"
log = "0.4.11"
vm-memory = { version = "0.2.1", features = ["backend-mmap"] }
serde = {version = ">=1.0.27", features = ["rc"] }
serde_derive = ">=1.0.27"
serde_json = ">=1.0.9"
vm-memory = { version = "0.3.0", features = ["backend-mmap"] }
vm-migration = { path = "../vm-migration" }
acpi_tables = { path = "../acpi_tables", optional = true }
arch_gen = { path = "../arch_gen" }

View File

@@ -97,9 +97,9 @@ pub fn create_fdt<T: DeviceInfoForFDT + Clone + Debug, S: ::std::hash::BuildHash
device_info: &HashMap<(DeviceType, String), T, S>,
gic_device: &dyn GICDevice,
initrd: &Option<InitramfsConfig>,
pci_space_address: &Option<(u64, u64)>,
pci_space_address: &(u64, u64),
) -> Result<Vec<u8>> {
// Alocate stuff necessary for the holding the blob.
// Allocate stuff necessary for the holding the blob.
let mut fdt = vec![0; FDT_MAX_SIZE];
allocate_fdt(&mut fdt)?;
@@ -126,9 +126,7 @@ pub fn create_fdt<T: DeviceInfoForFDT + Clone + Debug, S: ::std::hash::BuildHash
create_clock_node(&mut fdt)?;
create_psci_node(&mut fdt)?;
create_devices_node(&mut fdt, device_info)?;
if let Some((pci_device_base, pci_device_size)) = pci_space_address {
create_pci_nodes(&mut fdt, *pci_device_base, *pci_device_size)?;
}
create_pci_nodes(&mut fdt, pci_space_address.0, pci_space_address.1)?;
// End Header node.
append_end_node(&mut fdt)?;
@@ -402,7 +400,7 @@ fn create_gic_node(fdt: &mut Vec<u8>, gic_device: &dyn GICDevice) -> Result<()>
if gic_device.msi_compatible() {
append_begin_node(fdt, "msic")?;
append_property_string(fdt, "compatible", gic_device.msi_compatiblility())?;
append_property_string(fdt, "compatible", gic_device.msi_compatibility())?;
append_property_null(fdt, "msi-controller")?;
append_property_u32(fdt, "phandle", MSI_PHANDLE)?;
let msi_reg_prop = generate_prop64(gic_device.msi_properties());

View File

@@ -0,0 +1,182 @@
// Copyright 2020 Amazon.com, Inc. or its affiliates. All Rights Reserved.
// SPDX-License-Identifier: Apache-2.0
use super::{Error, Result};
use crate::layout::IRQ_BASE;
use hypervisor::kvm::kvm_bindings::{
kvm_device_attr, KVM_DEV_ARM_VGIC_GRP_DIST_REGS, KVM_DEV_ARM_VGIC_GRP_NR_IRQS,
};
use std::sync::Arc;
/*
Distributor registers as detailed at page 456 from
https://static.docs.arm.com/ihi0069/c/IHI0069C_gic_architecture_specification.pdf.
Address offsets are relative to the Distributor base address defined
by the system memory map. Unless otherwise stated in the register description,
all GIC registers are 32-bits wide.
*/
const GICD_CTLR: u32 = 0x0;
const GICD_STATUSR: u32 = 0x0010;
const GICD_IGROUPR: u32 = 0x0080;
const GICD_ISENABLER: u32 = 0x0100;
const GICD_ICENABLER: u32 = 0x0180;
const GICD_ISPENDR: u32 = 0x0200;
const GICD_ICPENDR: u32 = 0x0280;
const GICD_ISACTIVER: u32 = 0x0300;
const GICD_ICACTIVER: u32 = 0x0380;
const GICD_IPRIORITYR: u32 = 0x0400;
const GICD_ICFGR: u32 = 0x0C00;
const GICD_IROUTER: u32 = 0x6000;
/// This is how we represent the registers of the vgic's distributor.
/// Some of the distributor register )(i.e GICD_STATUSR) are simple
/// registers (i.e they are associated to a 32 bit value).
/// However, there are other registers that have variable lengths since
/// they dedicate some of the 32 bits to some specific interrupt. So, their length
/// depends on the number of interrupts (i.e the ones that are represented as GICD_REG<n>)
/// in the documentation mentioned above.
struct DistReg {
/// Offset from distributor address.
base: u32,
/// Bits per interrupt.
/// Relevant for registers that DO share IRQs.
bpi: u8,
/// Length of the register.
/// Relevant for registers that DO NOT share IRQs.
length: u16,
}
// All or at least the registers we are interested in are 32 bit, so
// we use a constant for size(u32).
const REG_SIZE: u8 = 4;
// Creates a vgic distributor register.
macro_rules! VGIC_DIST_REG {
($base:expr, $bpi:expr, $length:expr) => {
DistReg {
base: $base,
bpi: $bpi,
length: $length,
}
};
}
// List with relevant distributor registers that we will be restoring.
// Order is taken from qemu.
static VGIC_DIST_REGS: &'static [DistReg] = &[
VGIC_DIST_REG!(GICD_STATUSR, 0, 4),
VGIC_DIST_REG!(GICD_ICENABLER, 1, 0),
VGIC_DIST_REG!(GICD_ISENABLER, 1, 0),
VGIC_DIST_REG!(GICD_IGROUPR, 1, 0),
VGIC_DIST_REG!(GICD_IROUTER, 64, 0),
VGIC_DIST_REG!(GICD_ICFGR, 2, 0),
VGIC_DIST_REG!(GICD_ICPENDR, 1, 0),
VGIC_DIST_REG!(GICD_ISPENDR, 1, 0),
VGIC_DIST_REG!(GICD_ICACTIVER, 1, 0),
VGIC_DIST_REG!(GICD_ISACTIVER, 1, 0),
VGIC_DIST_REG!(GICD_IPRIORITYR, 8, 0),
];
fn dist_attr_access(
gic: &Arc<dyn hypervisor::Device>,
offset: u32,
val: &u32,
set: bool,
) -> Result<()> {
let mut gic_dist_attr = kvm_device_attr {
group: KVM_DEV_ARM_VGIC_GRP_DIST_REGS,
attr: offset as u64,
addr: val as *const u32 as u64,
flags: 0,
};
if set {
gic.set_device_attr(&gic_dist_attr)
.map_err(Error::SetDeviceAttribute)?;
} else {
gic.get_device_attr(&mut gic_dist_attr)
.map_err(Error::GetDeviceAttribute)?;
}
Ok(())
}
/// Get the distributor control register.
pub fn read_ctlr(gic: &Arc<dyn hypervisor::Device>) -> Result<u32> {
let val: u32 = 0;
dist_attr_access(gic, GICD_CTLR, &val, false)?;
Ok(val)
}
/// Set the distributor control register.
pub fn write_ctlr(gic: &Arc<dyn hypervisor::Device>, val: u32) -> Result<()> {
dist_attr_access(gic, GICD_CTLR, &val, true)
}
fn get_interrupts_num(gic: &Arc<dyn hypervisor::Device>) -> Result<u32> {
let num_irq = 0;
let mut nr_irqs_attr = kvm_device_attr {
group: KVM_DEV_ARM_VGIC_GRP_NR_IRQS,
attr: 0,
addr: &num_irq as *const u32 as u64,
flags: 0,
};
gic.get_device_attr(&mut nr_irqs_attr)
.map_err(Error::GetDeviceAttribute)?;
Ok(num_irq)
}
fn compute_reg_len(gic: &Arc<dyn hypervisor::Device>, reg: &DistReg, base: u32) -> Result<u32> {
let mut end = base;
let num_irq = get_interrupts_num(gic)?;
if reg.length > 0 {
// This is the single type register (i.e one that is not DIST_X<n>) and for which
// the bpi is 0.
// Look in the kernel for REGISTER_DESC_WITH_LENGTH.
end = base + reg.length as u32;
}
if reg.bpi > 0 {
// This is the type of register that takes into account the number of interrupts
// that the model has. It is also the type of register where
// a register relates to multiple interrupts.
end = base + (reg.bpi as u32 * (num_irq - IRQ_BASE) / 8);
if reg.bpi as u32 * (num_irq - IRQ_BASE) % 8 > 0 {
end += REG_SIZE as u32;
}
}
Ok(end)
}
/// Set distributor registers of the GIC.
pub fn set_dist_regs(gic: &Arc<dyn hypervisor::Device>, state: &[u32]) -> Result<()> {
let mut idx = 0;
for dreg in VGIC_DIST_REGS {
let mut base = dreg.base + REG_SIZE as u32 * dreg.bpi as u32;
let end = compute_reg_len(gic, &dreg, base)?;
while base < end {
let val = state[idx];
dist_attr_access(gic, base, &val, true)?;
idx += 1;
base += REG_SIZE as u32;
}
}
Ok(())
}
/// Get distributor registers of the GIC.
pub fn get_dist_regs(gic: &Arc<dyn hypervisor::Device>) -> Result<Vec<u32>> {
let mut state = Vec::new();
for dreg in VGIC_DIST_REGS {
let mut base = dreg.base + REG_SIZE as u32 * dreg.bpi as u32;
let end = compute_reg_len(gic, &dreg, base)?;
while base < end {
let val: u32 = 0;
dist_attr_access(gic, base, &val, false)?;
state.push(val);
base += REG_SIZE as u32;
}
}
Ok(state)
}

View File

@@ -0,0 +1,275 @@
// Copyright 2019 Amazon.com, Inc. or its affiliates. All Rights Reserved.
// SPDX-License-Identifier: Apache-2.0
pub mod kvm {
use crate::aarch64::gic::dist_regs::{get_dist_regs, read_ctlr, set_dist_regs, write_ctlr};
use crate::aarch64::gic::icc_regs::{get_icc_regs, set_icc_regs};
use crate::aarch64::gic::kvm::{save_pending_tables, KvmGICDevice};
use crate::aarch64::gic::redist_regs::{get_redist_regs, set_redist_regs};
use crate::aarch64::gic::GICDevice;
use crate::layout;
use anyhow::anyhow;
use hypervisor::kvm::kvm_bindings;
use std::any::Any;
use std::convert::TryInto;
use std::sync::Arc;
use std::{boxed::Box, result};
use vm_migration::{
Migratable, MigratableError, Pausable, Snapshot, SnapshotDataSection, Snapshottable,
Transportable,
};
/// Errors thrown while saving/restoring the GICv3.
#[derive(Debug)]
pub enum Error {
/// Error in saving RDIST pending tables into guest RAM.
SavePendingTables(crate::aarch64::gic::Error),
/// Error in saving GIC distributor registers.
SaveDistributorRegisters(crate::aarch64::gic::Error),
/// Error in restoring GIC distributor registers.
RestoreDistributorRegisters(crate::aarch64::gic::Error),
/// Error in saving GIC distributor control registers.
SaveDistributorCtrlRegisters(crate::aarch64::gic::Error),
/// Error in restoring GIC distributor control registers.
RestoreDistributorCtrlRegisters(crate::aarch64::gic::Error),
/// Error in saving GIC redistributor registers.
SaveRedistributorRegisters(crate::aarch64::gic::Error),
/// Error in restoring GIC redistributor registers.
RestoreRedistributorRegisters(crate::aarch64::gic::Error),
/// Error in saving GIC CPU interface registers.
SaveICCRegisters(crate::aarch64::gic::Error),
/// Error in restoring GIC CPU interface registers.
RestoreICCRegisters(crate::aarch64::gic::Error),
}
type Result<T> = result::Result<T, Error>;
pub struct KvmGICv3 {
/// The hypervisor agnostic device
device: Arc<dyn hypervisor::Device>,
/// Vector holding values of GICR_TYPER for each vCPU
gicr_typers: Vec<u64>,
/// GIC device properties, to be used for setting up the fdt entry
properties: [u64; 4],
/// Number of CPUs handled by the device
vcpu_count: u64,
}
#[derive(Serialize, Deserialize)]
pub struct Gicv3State {
dist: Vec<u32>,
rdist: Vec<u32>,
icc: Vec<u32>,
// special register that enables interrupts and affinity routing
gicd_ctlr: u32,
}
impl KvmGICv3 {
// Unfortunately bindgen omits defines that are based on other defines.
// See arch/arm64/include/uapi/asm/kvm.h file from the linux kernel.
pub const SZ_64K: u64 = 0x0001_0000;
const KVM_VGIC_V3_DIST_SIZE: u64 = KvmGICv3::SZ_64K;
const KVM_VGIC_V3_REDIST_SIZE: u64 = (2 * KvmGICv3::SZ_64K);
// Device trees specific constants
pub const ARCH_GIC_V3_MAINT_IRQ: u32 = 9;
/// Get the address of the GIC distributor.
pub fn get_dist_addr() -> u64 {
layout::MAPPED_IO_START - KvmGICv3::KVM_VGIC_V3_DIST_SIZE
}
/// Get the size of the GIC distributor.
pub fn get_dist_size() -> u64 {
KvmGICv3::KVM_VGIC_V3_DIST_SIZE
}
/// Get the address of the GIC redistributors.
pub fn get_redists_addr(vcpu_count: u64) -> u64 {
KvmGICv3::get_dist_addr() - KvmGICv3::get_redists_size(vcpu_count)
}
/// Get the size of the GIC redistributors.
pub fn get_redists_size(vcpu_count: u64) -> u64 {
vcpu_count * KvmGICv3::KVM_VGIC_V3_REDIST_SIZE
}
/// Save the state of GIC.
fn state(&self, gicr_typers: &[u64]) -> Result<Gicv3State> {
// Flush redistributors pending tables to guest RAM.
save_pending_tables(&self.device()).map_err(Error::SavePendingTables)?;
let gicd_ctlr =
read_ctlr(&self.device()).map_err(Error::SaveDistributorCtrlRegisters)?;
let dist_state =
get_dist_regs(&self.device()).map_err(Error::SaveDistributorRegisters)?;
let rdist_state = get_redist_regs(&self.device(), &gicr_typers)
.map_err(Error::SaveRedistributorRegisters)?;
let icc_state =
get_icc_regs(&self.device(), &gicr_typers).map_err(Error::SaveICCRegisters)?;
Ok(Gicv3State {
dist: dist_state,
rdist: rdist_state,
icc: icc_state,
gicd_ctlr,
})
}
/// Restore the state of GIC.
fn set_state(&mut self, gicr_typers: &[u64], state: &Gicv3State) -> Result<()> {
write_ctlr(&self.device(), state.gicd_ctlr)
.map_err(Error::RestoreDistributorCtrlRegisters)?;
set_dist_regs(&self.device(), &state.dist)
.map_err(Error::RestoreDistributorRegisters)?;
set_redist_regs(&self.device(), gicr_typers, &state.rdist)
.map_err(Error::RestoreRedistributorRegisters)?;
set_icc_regs(&self.device(), &gicr_typers, &state.icc)
.map_err(Error::RestoreICCRegisters)?;
Ok(())
}
}
impl GICDevice for KvmGICv3 {
fn device(&self) -> &Arc<dyn hypervisor::Device> {
&self.device
}
fn fdt_compatibility(&self) -> &str {
"arm,gic-v3"
}
fn fdt_maint_irq(&self) -> u32 {
KvmGICv3::ARCH_GIC_V3_MAINT_IRQ
}
fn device_properties(&self) -> &[u64] {
&self.properties
}
fn vcpu_count(&self) -> u64 {
self.vcpu_count
}
fn set_gicr_typers(&mut self, gicr_typers: Vec<u64>) {
self.gicr_typers = gicr_typers;
}
fn as_any_concrete_mut(&mut self) -> &mut dyn Any {
self
}
}
impl KvmGICDevice for KvmGICv3 {
fn version() -> u32 {
kvm_bindings::kvm_device_type_KVM_DEV_TYPE_ARM_VGIC_V3
}
fn create_device(
device: Arc<dyn hypervisor::Device>,
vcpu_count: u64,
) -> Box<dyn GICDevice> {
Box::new(KvmGICv3 {
device,
gicr_typers: vec![0; vcpu_count.try_into().unwrap()],
properties: [
KvmGICv3::get_dist_addr(),
KvmGICv3::get_dist_size(),
KvmGICv3::get_redists_addr(vcpu_count),
KvmGICv3::get_redists_size(vcpu_count),
],
vcpu_count,
})
}
fn init_device_attributes(
_vm: &Arc<dyn hypervisor::Vm>,
gic_device: &dyn GICDevice,
) -> crate::aarch64::gic::Result<()> {
/* Setting up the distributor attribute.
We are placing the GIC below 1GB so we need to substract the size of the distributor.
*/
Self::set_device_attribute(
gic_device.device(),
kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ADDR,
u64::from(kvm_bindings::KVM_VGIC_V3_ADDR_TYPE_DIST),
&KvmGICv3::get_dist_addr() as *const u64 as u64,
0,
)?;
/* Setting up the redistributors' attribute.
We are calculating here the start of the redistributors address. We have one per CPU.
*/
Self::set_device_attribute(
gic_device.device(),
kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ADDR,
u64::from(kvm_bindings::KVM_VGIC_V3_ADDR_TYPE_REDIST),
&KvmGICv3::get_redists_addr(gic_device.vcpu_count()) as *const u64 as u64,
0,
)?;
Ok(())
}
}
pub const GIC_V3_SNAPSHOT_ID: &str = "gic-v3";
impl Snapshottable for KvmGICv3 {
fn id(&self) -> String {
GIC_V3_SNAPSHOT_ID.to_string()
}
fn snapshot(&mut self) -> std::result::Result<Snapshot, MigratableError> {
let gicr_typers = self.gicr_typers.clone();
let snapshot = serde_json::to_vec(&self.state(&gicr_typers).unwrap())
.map_err(|e| MigratableError::Snapshot(e.into()))?;
let mut gic_v3_snapshot = Snapshot::new(self.id().as_str());
gic_v3_snapshot.add_data_section(SnapshotDataSection {
id: format!("{}-section", self.id()),
snapshot,
});
Ok(gic_v3_snapshot)
}
fn restore(&mut self, snapshot: Snapshot) -> std::result::Result<(), MigratableError> {
if let Some(gic_v3_section) = snapshot
.snapshot_data
.get(&format!("{}-section", self.id()))
{
let gic_v3_state = match serde_json::from_slice(&gic_v3_section.snapshot) {
Ok(state) => state,
Err(error) => {
return Err(MigratableError::Restore(anyhow!(
"Could not deserialize GICv3 {}",
error
)))
}
};
let gicr_typers = self.gicr_typers.clone();
return self.set_state(&gicr_typers, &gic_v3_state).map_err(|e| {
MigratableError::Restore(anyhow!("Could not restore GICv3 state {:?}", e))
});
}
Err(MigratableError::Restore(anyhow!(
"Could not find GICv3 snapshot section"
)))
}
}
impl Pausable for KvmGICv3 {}
impl Transportable for KvmGICv3 {}
impl Migratable for KvmGICv3 {}
}

View File

@@ -2,18 +2,23 @@
// SPDX-License-Identifier: Apache-2.0
pub mod kvm {
use std::any::Any;
use std::convert::TryInto;
use std::sync::Arc;
use std::{boxed::Box, result};
type Result<T> = result::Result<T, Error>;
use crate::aarch64::gic::gicv3::kvm::KvmGICv3;
use crate::aarch64::gic::kvm::KvmGICDevice;
use crate::aarch64::gic::{Error, GICDevice};
use crate::aarch64::gicv3::kvm::KvmGICv3;
use hypervisor::kvm::kvm_bindings;
pub struct KvmGICv3ITS {
/// The hypervisor agnostic device
device: Arc<dyn hypervisor::Device>,
/// Vector holding values of GICR_TYPER for each vCPU
gicr_typers: Vec<u64>,
/// GIC device properties, to be used for setting up the fdt entry
gic_properties: [u64; 4],
@@ -49,7 +54,7 @@ pub mod kvm {
true
}
fn msi_compatiblility(&self) -> &str {
fn msi_compatibility(&self) -> &str {
"arm,gic-v3-its"
}
@@ -68,6 +73,14 @@ pub mod kvm {
fn vcpu_count(&self) -> u64 {
self.vcpu_count
}
fn set_gicr_typers(&mut self, gicr_typers: Vec<u64>) {
self.gicr_typers = gicr_typers;
}
fn as_any_concrete_mut(&mut self) -> &mut dyn Any {
self
}
}
impl KvmGICDevice for KvmGICv3ITS {
@@ -81,6 +94,7 @@ pub mod kvm {
) -> Box<dyn GICDevice> {
Box::new(KvmGICv3ITS {
device,
gicr_typers: vec![0; vcpu_count.try_into().unwrap()],
gic_properties: [
KvmGICv3::get_dist_addr(),
KvmGICv3::get_dist_size(),

View File

@@ -0,0 +1,196 @@
// Copyright 2020 Amazon.com, Inc. or its affiliates. All Rights Reserved.
// SPDX-License-Identifier: Apache-2.0
use super::{Error, Result};
use hypervisor::kvm::kvm_bindings::{
kvm_device_attr, KVM_DEV_ARM_VGIC_GRP_CPU_SYSREGS, KVM_REG_ARM64_SYSREG_CRM_MASK,
KVM_REG_ARM64_SYSREG_CRM_SHIFT, KVM_REG_ARM64_SYSREG_CRN_MASK, KVM_REG_ARM64_SYSREG_CRN_SHIFT,
KVM_REG_ARM64_SYSREG_OP0_MASK, KVM_REG_ARM64_SYSREG_OP0_SHIFT, KVM_REG_ARM64_SYSREG_OP1_MASK,
KVM_REG_ARM64_SYSREG_OP1_SHIFT, KVM_REG_ARM64_SYSREG_OP2_MASK, KVM_REG_ARM64_SYSREG_OP2_SHIFT,
};
use std::sync::Arc;
const KVM_DEV_ARM_VGIC_V3_MPIDR_SHIFT: u32 = 32;
const KVM_DEV_ARM_VGIC_V3_MPIDR_MASK: u64 = 0xffffffff << KVM_DEV_ARM_VGIC_V3_MPIDR_SHIFT as u64;
const ICC_CTLR_EL1_PRIBITS_SHIFT: u32 = 8;
const ICC_CTLR_EL1_PRIBITS_MASK: u32 = 7 << ICC_CTLR_EL1_PRIBITS_SHIFT;
macro_rules! arm64_vgic_sys_reg {
($name: tt, $op0: tt, $op1: tt, $crn: tt, $crm: tt, $op2: expr) => {
const $name: u64 = ((($op0 as u64) << KVM_REG_ARM64_SYSREG_OP0_SHIFT)
& KVM_REG_ARM64_SYSREG_OP0_MASK as u64)
| ((($op1 as u64) << KVM_REG_ARM64_SYSREG_OP1_SHIFT)
& KVM_REG_ARM64_SYSREG_OP1_MASK as u64)
| ((($crn as u64) << KVM_REG_ARM64_SYSREG_CRN_SHIFT)
& KVM_REG_ARM64_SYSREG_CRN_MASK as u64)
| ((($crm as u64) << KVM_REG_ARM64_SYSREG_CRM_SHIFT)
& KVM_REG_ARM64_SYSREG_CRM_MASK as u64)
| ((($op2 as u64) << KVM_REG_ARM64_SYSREG_OP2_SHIFT)
& KVM_REG_ARM64_SYSREG_OP2_MASK as u64);
};
}
macro_rules! SYS_ICC_AP0Rn_EL1 {
($name: tt, $n: tt) => {
arm64_vgic_sys_reg!($name, 3, 0, 12, 8, (4 | $n));
};
}
macro_rules! SYS_ICC_AP1Rn_EL1 {
($name: tt, $n: tt) => {
arm64_vgic_sys_reg!($name, 3, 0, 12, 9, $n);
};
}
arm64_vgic_sys_reg!(SYS_ICC_SRE_EL1, 3, 0, 12, 12, 5);
arm64_vgic_sys_reg!(SYS_ICC_CTLR_EL1, 3, 0, 12, 12, 4);
arm64_vgic_sys_reg!(SYS_ICC_IGRPEN0_EL1, 3, 0, 12, 12, 6);
arm64_vgic_sys_reg!(SYS_ICC_IGRPEN1_EL1, 3, 0, 12, 12, 7);
arm64_vgic_sys_reg!(SYS_ICC_PMR_EL1, 3, 0, 4, 6, 0);
arm64_vgic_sys_reg!(SYS_ICC_BPR0_EL1, 3, 0, 12, 8, 3);
arm64_vgic_sys_reg!(SYS_ICC_BPR1_EL1, 3, 0, 12, 12, 3);
SYS_ICC_AP0Rn_EL1!(SYS_ICC_AP0R0_EL1, 0);
SYS_ICC_AP0Rn_EL1!(SYS_ICC_AP0R1_EL1, 1);
SYS_ICC_AP0Rn_EL1!(SYS_ICC_AP0R2_EL1, 2);
SYS_ICC_AP0Rn_EL1!(SYS_ICC_AP0R3_EL1, 3);
SYS_ICC_AP1Rn_EL1!(SYS_ICC_AP1R0_EL1, 0);
SYS_ICC_AP1Rn_EL1!(SYS_ICC_AP1R1_EL1, 1);
SYS_ICC_AP1Rn_EL1!(SYS_ICC_AP1R2_EL1, 2);
SYS_ICC_AP1Rn_EL1!(SYS_ICC_AP1R3_EL1, 3);
static VGIC_ICC_REGS: &'static [u64] = &[
SYS_ICC_SRE_EL1,
SYS_ICC_CTLR_EL1,
SYS_ICC_IGRPEN0_EL1,
SYS_ICC_IGRPEN1_EL1,
SYS_ICC_PMR_EL1,
SYS_ICC_BPR0_EL1,
SYS_ICC_BPR1_EL1,
SYS_ICC_AP0R0_EL1,
SYS_ICC_AP0R1_EL1,
SYS_ICC_AP0R2_EL1,
SYS_ICC_AP0R3_EL1,
SYS_ICC_AP1R0_EL1,
SYS_ICC_AP1R1_EL1,
SYS_ICC_AP1R2_EL1,
SYS_ICC_AP1R3_EL1,
];
fn icc_attr_access(
gic: &Arc<dyn hypervisor::Device>,
offset: u64,
typer: u64,
val: &u32,
set: bool,
) -> Result<()> {
let mut gic_icc_attr = kvm_device_attr {
group: KVM_DEV_ARM_VGIC_GRP_CPU_SYSREGS,
attr: ((typer & KVM_DEV_ARM_VGIC_V3_MPIDR_MASK) | offset), // this needs the mpidr
addr: val as *const u32 as u64,
flags: 0,
};
if set {
#[allow(clippy::unnecessary_mut_passed)]
gic.set_device_attr(&mut gic_icc_attr)
.map_err(Error::SetDeviceAttribute)?;
} else {
gic.get_device_attr(&mut gic_icc_attr)
.map_err(Error::GetDeviceAttribute)?;
}
Ok(())
}
/// Get ICC registers.
pub fn get_icc_regs(gic: &Arc<dyn hypervisor::Device>, gicr_typer: &[u64]) -> Result<Vec<u32>> {
let mut state: Vec<u32> = Vec::new();
// We need this for the ICC_AP<m>R<n>_EL1 registers.
let mut num_priority_bits = 0;
for ix in gicr_typer {
let i = *ix;
for icc_offset in VGIC_ICC_REGS {
let val = 0;
if *icc_offset == SYS_ICC_CTLR_EL1 {
// calculate priority bits by reading the ctrl_el1 register.
icc_attr_access(gic, *icc_offset, i, &val, false)?;
// The priority bits are found in the ICC_CTLR_EL1 register (bits from 10:8).
// See page 194 from https://static.docs.arm.com/ihi0069/c/IHI0069C_gic_
// architecture_specification.pdf.
// Citation:
// "Priority bits. Read-only and writes are ignored. The number of priority bits
// implemented, minus one."
num_priority_bits =
((val & ICC_CTLR_EL1_PRIBITS_MASK) >> ICC_CTLR_EL1_PRIBITS_SHIFT) + 1;
state.push(val);
}
// As per ARMv8 documentation: https://static.docs.arm.com/ihi0069/c/IHI0069C_
// gic_architecture_specification.pdf
// page 178,
// ICC_AP0R1_EL1 is only implemented in implementations that support 6 or more bits of
// priority.
// ICC_AP0R2_EL1 and ICC_AP0R3_EL1 are only implemented in implementations that support
// 7 bits of priority.
else if *icc_offset == SYS_ICC_AP0R1_EL1 || *icc_offset == SYS_ICC_AP1R1_EL1 {
if num_priority_bits >= 6 {
icc_attr_access(gic, *icc_offset, i, &val, false)?;
state.push(val);
}
} else if *icc_offset == SYS_ICC_AP0R2_EL1
|| *icc_offset == SYS_ICC_AP0R3_EL1
|| *icc_offset == SYS_ICC_AP1R2_EL1
|| *icc_offset == SYS_ICC_AP1R3_EL1
{
if num_priority_bits == 7 {
icc_attr_access(gic, *icc_offset, i, &val, false)?;
state.push(val);
}
} else {
icc_attr_access(gic, *icc_offset, i, &val, false)?;
state.push(val);
}
}
}
Ok(state)
}
/// Set ICC registers.
pub fn set_icc_regs(
gic: &Arc<dyn hypervisor::Device>,
gicr_typer: &[u64],
state: &[u32],
) -> Result<()> {
let mut num_priority_bits = 0;
let mut idx = 0;
for ix in gicr_typer {
let i = *ix;
for icc_offset in VGIC_ICC_REGS {
if *icc_offset == SYS_ICC_CTLR_EL1 {
let ctrl_el1 = state[idx];
num_priority_bits =
((ctrl_el1 & ICC_CTLR_EL1_PRIBITS_MASK) >> ICC_CTLR_EL1_PRIBITS_SHIFT) + 1;
}
if *icc_offset == SYS_ICC_AP0R1_EL1 || *icc_offset == SYS_ICC_AP1R1_EL1 {
if num_priority_bits >= 6 {
icc_attr_access(gic, *icc_offset, i, &state[idx], true)?;
idx += 1;
}
continue;
}
if *icc_offset == SYS_ICC_AP0R2_EL1
|| *icc_offset == SYS_ICC_AP0R3_EL1
|| *icc_offset == SYS_ICC_AP1R2_EL1
|| *icc_offset == SYS_ICC_AP1R3_EL1
{
if num_priority_bits == 7 {
icc_attr_access(gic, *icc_offset, i, &state[idx], true)?;
idx += 1;
}
continue;
}
icc_attr_access(gic, *icc_offset, i, &state[idx], true)?;
idx += 1;
}
}
Ok(())
}

View File

@@ -1,6 +1,16 @@
// Copyright 2019 Amazon.com, Inc. or its affiliates. All Rights Reserved.
// SPDX-License-Identifier: Apache-2.0
pub mod dist_regs;
pub mod gicv3;
pub mod gicv3_its;
pub mod icc_regs;
pub mod redist_regs;
pub use self::dist_regs::{get_dist_regs, read_ctlr, set_dist_regs, write_ctlr};
pub use self::icc_regs::{get_icc_regs, set_icc_regs};
pub use self::redist_regs::{get_redist_regs, set_redist_regs};
use std::any::Any;
use std::result;
use std::sync::Arc;
@@ -11,10 +21,12 @@ pub enum Error {
CreateGIC(hypervisor::HypervisorVmError),
/// Error while setting device attributes for the GIC.
SetDeviceAttribute(hypervisor::HypervisorDeviceError),
/// Error while getting device attributes for the GIC.
GetDeviceAttribute(hypervisor::HypervisorDeviceError),
}
type Result<T> = result::Result<T, Error>;
pub trait GICDevice {
pub trait GICDevice: Send {
/// Returns the hypervisor agnostic Device of the GIC device
fn device(&self) -> &Arc<dyn hypervisor::Device>;
@@ -36,7 +48,7 @@ pub trait GICDevice {
}
/// Returns the MSI compatibility property of the device
fn msi_compatiblility(&self) -> &str {
fn msi_compatibility(&self) -> &str {
""
}
@@ -44,14 +56,18 @@ pub trait GICDevice {
fn msi_properties(&self) -> &[u64] {
&[]
}
/// Get the values of GICR_TYPER for each vCPU.
fn set_gicr_typers(&mut self, gicr_typers: Vec<u64>);
/// Downcast the trait object to its concrete type.
fn as_any_concrete_mut(&mut self) -> &mut dyn Any;
}
pub mod kvm {
use super::GICDevice;
use super::Result;
use crate::aarch64::gicv2::kvm::KvmGICv2;
use crate::aarch64::gicv3::kvm::KvmGICv3;
use crate::aarch64::gicv3_its::kvm::KvmGICv3ITS;
use crate::aarch64::gic::gicv3_its::kvm::KvmGICv3ITS;
use crate::layout;
use hypervisor::kvm::kvm_bindings;
use std::boxed::Box;
@@ -107,6 +123,27 @@ pub mod kvm {
Ok(())
}
/// Get a GIC device attribute
fn get_device_attribute(
device: &Arc<dyn hypervisor::Device>,
group: u32,
attr: u64,
addr: u64,
flags: u32,
) -> Result<()> {
let mut attr = kvm_bindings::kvm_device_attr {
group,
attr,
addr,
flags,
};
device
.get_device_attr(&mut attr)
.map_err(super::Error::GetDeviceAttribute)?;
Ok(())
}
/// Finalize the setup of a GIC device
fn finalize_device(gic_device: &dyn GICDevice) -> Result<()> {
/* We need to tell the kernel how many irqs to support with this vgic.
@@ -150,25 +187,24 @@ pub mod kvm {
}
}
/// Create a GIC device.
/// Create a GICv3-ITS device.
///
/// It will try to create by default a GICv3 device. If that fails it will try
/// to fall-back to a GICv2 device.
pub fn create_gic(
vm: &Arc<dyn hypervisor::Vm>,
vcpu_count: u64,
its_required: bool,
) -> Result<Box<dyn GICDevice>> {
if its_required {
KvmGICv3ITS::new(vm, vcpu_count)
} else {
KvmGICv3ITS::new(vm, vcpu_count).or_else(|_| {
debug!("Failed to create GICv3-ITS, will try GICv3 instead.");
KvmGICv3::new(vm, vcpu_count).or_else(|_| {
debug!("Failed to create GICv3, will try GICv2 instead.");
KvmGICv2::new(vm, vcpu_count)
})
})
}
pub fn create_gic(vm: &Arc<dyn hypervisor::Vm>, vcpu_count: u64) -> Result<Box<dyn GICDevice>> {
debug!("creating a GICv3-ITS");
KvmGICv3ITS::new(vm, vcpu_count)
}
/// Function that saves RDIST pending tables into guest RAM.
///
/// The tables get flushed to guest RAM whenever the VM gets stopped.
pub fn save_pending_tables(gic: &Arc<dyn hypervisor::Device>) -> Result<()> {
let init_gic_attr = kvm_bindings::kvm_device_attr {
group: kvm_bindings::KVM_DEV_ARM_VGIC_GRP_CTRL,
attr: u64::from(kvm_bindings::KVM_DEV_ARM_VGIC_SAVE_PENDING_TABLES),
addr: 0,
flags: 0,
};
gic.set_device_attr(&init_gic_attr)
.map_err(super::Error::SetDeviceAttribute)
}
}

View File

@@ -0,0 +1,179 @@
// Copyright 2020 Amazon.com, Inc. or its affiliates. All Rights Reserved.
// SPDX-License-Identifier: Apache-2.0
use super::{Error, Result};
use hypervisor::kvm::kvm_bindings::{kvm_device_attr, KVM_DEV_ARM_VGIC_GRP_REDIST_REGS};
use std::sync::Arc;
// Relevant redistributor registers that we want to save/restore.
const GICR_CTLR: u32 = 0x0000;
const GICR_STATUSR: u32 = 0x0010;
const GICR_WAKER: u32 = 0x0014;
const GICR_PROPBASER: u32 = 0x0070;
const GICR_PENDBASER: u32 = 0x0078;
/* SGI and PPI Redistributor registers, offsets from RD_base */
/*
* Redistributor frame offsets from RD_base which is actually SZ_
*/
const GICR_SGI_OFFSET: u32 = 0x0001_0000;
const GICR_IGROUPR0: u32 = GICR_SGI_OFFSET + 0x0080;
const GICR_ICENABLER0: u32 = GICR_SGI_OFFSET + 0x0180;
const GICR_ISENABLER0: u32 = GICR_SGI_OFFSET + 0x0100;
const GICR_ISPENDR0: u32 = GICR_SGI_OFFSET + 0x0200;
const GICR_ICPENDR0: u32 = GICR_SGI_OFFSET + 0x0280;
const GICR_ISACTIVER0: u32 = GICR_SGI_OFFSET + 0x0300;
const GICR_ICACTIVER0: u32 = GICR_SGI_OFFSET + 0x0380;
const GICR_IPRIORITYR0: u32 = GICR_SGI_OFFSET + 0x0400;
const GICR_ICFGR0: u32 = GICR_SGI_OFFSET + 0x0C00;
const KVM_DEV_ARM_VGIC_V3_MPIDR_SHIFT: u32 = 32;
const KVM_DEV_ARM_VGIC_V3_MPIDR_MASK: u64 = 0xffffffff << KVM_DEV_ARM_VGIC_V3_MPIDR_SHIFT as u64;
/// This is how we represent the registers of a distributor.
/// It is relrvant their offset from the base address of the
/// distributor.
/// Each register has a different number
/// of bits_per_irq and is therefore variable length.
/// First 32 interrupts (0-32) are private to each CPU (SGIs and PPIs).
/// and so we save the first irq to identify between the type of the interrupt
/// that the respective register deals with.
struct RdistReg {
/// Offset from distributor address.
base: u32,
/// Length of the register.
length: u8,
}
// All or at least the registers we are interested in are 32 bit, so
// we use a constant for size(u32).
const REG_SIZE: u8 = 4;
// Creates a vgic redistributor register.
macro_rules! VGIC_RDIST_REG {
($base:expr, $len:expr) => {
RdistReg {
base: $base,
length: $len,
}
};
}
// List with relevant distributor registers that we will be restoring.
static VGIC_RDIST_REGS: &'static [RdistReg] = &[
VGIC_RDIST_REG!(GICR_CTLR, 4),
VGIC_RDIST_REG!(GICR_STATUSR, 4),
VGIC_RDIST_REG!(GICR_WAKER, 4),
VGIC_RDIST_REG!(GICR_PROPBASER, 8),
VGIC_RDIST_REG!(GICR_PENDBASER, 8),
];
// List with relevant distributor registers that we will be restoring.
static VGIC_SGI_REGS: &'static [RdistReg] = &[
VGIC_RDIST_REG!(GICR_IGROUPR0, 4),
VGIC_RDIST_REG!(GICR_ICENABLER0, 4),
VGIC_RDIST_REG!(GICR_ISENABLER0, 4),
VGIC_RDIST_REG!(GICR_ICFGR0, 8),
VGIC_RDIST_REG!(GICR_ICPENDR0, 4),
VGIC_RDIST_REG!(GICR_ISPENDR0, 4),
VGIC_RDIST_REG!(GICR_ICACTIVER0, 4),
VGIC_RDIST_REG!(GICR_ISACTIVER0, 4),
VGIC_RDIST_REG!(GICR_IPRIORITYR0, 32),
];
fn redist_attr_access(
gic: &Arc<dyn hypervisor::Device>,
offset: u32,
typer: u64,
val: &u32,
set: bool,
) -> Result<()> {
let mut gic_dist_attr = kvm_device_attr {
group: KVM_DEV_ARM_VGIC_GRP_REDIST_REGS,
attr: (typer & KVM_DEV_ARM_VGIC_V3_MPIDR_MASK) | (offset as u64), // this needs the mpidr
addr: val as *const u32 as u64,
flags: 0,
};
if set {
#[allow(clippy::unnecessary_mut_passed)]
gic.set_device_attr(&mut gic_dist_attr)
.map_err(Error::SetDeviceAttribute)?;
} else {
gic.get_device_attr(&mut gic_dist_attr)
.map_err(Error::GetDeviceAttribute)?;
}
Ok(())
}
fn access_redists_aux(
gic: &Arc<dyn hypervisor::Device>,
gicr_typer: &[u64],
state: &mut Vec<u32>,
reg_list: &'static [RdistReg],
idx: &mut usize,
set: bool,
) -> Result<()> {
for i in gicr_typer {
for rdreg in reg_list {
let mut base = rdreg.base;
let end = base + rdreg.length as u32;
while base < end {
let mut val = 0;
if set {
val = state[*idx];
redist_attr_access(gic, base, *i, &val, true)?;
*idx += 1;
} else {
redist_attr_access(gic, base, *i, &val, false)?;
state.push(val);
}
base += REG_SIZE as u32;
}
}
}
Ok(())
}
/// Get redistributor registers.
pub fn get_redist_regs(gic: &Arc<dyn hypervisor::Device>, gicr_typer: &[u64]) -> Result<Vec<u32>> {
let mut state = Vec::new();
let mut idx: usize = 0;
access_redists_aux(
gic,
&gicr_typer,
&mut state,
VGIC_RDIST_REGS,
&mut idx,
false,
)?;
access_redists_aux(gic, &gicr_typer, &mut state, VGIC_SGI_REGS, &mut idx, false)?;
Ok(state)
}
/// Set redistributor registers.
pub fn set_redist_regs(
gic: &Arc<dyn hypervisor::Device>,
gicr_typer: &[u64],
state: &[u32],
) -> Result<()> {
let mut idx: usize = 0;
let mut mut_state = state.to_owned();
access_redists_aux(
gic,
gicr_typer,
&mut mut_state,
VGIC_RDIST_REGS,
&mut idx,
true,
)?;
access_redists_aux(
gic,
gicr_typer,
&mut mut_state,
VGIC_SGI_REGS,
&mut idx,
true,
)
}

View File

@@ -1,124 +0,0 @@
// Copyright 2019 Amazon.com, Inc. or its affiliates. All Rights Reserved.
// SPDX-License-Identifier: Apache-2.0
pub mod kvm {
use crate::aarch64::gic::kvm::KvmGICDevice;
use crate::aarch64::gic::{Error, GICDevice};
use std::{boxed::Box, result};
type Result<T> = result::Result<T, Error>;
use crate::layout;
use hypervisor::kvm::kvm_bindings;
use std::sync::Arc;
/// Represent a GIC v2 device
pub struct KvmGICv2 {
/// The hypervisor agnostic device
device: Arc<dyn hypervisor::Device>,
/// GIC device properties, to be used for setting up the fdt entry
properties: [u64; 4],
/// Number of CPUs handled by the device
vcpu_count: u64,
}
impl KvmGICv2 {
// Unfortunately bindgen omits defines that are based on other defines.
// See arch/arm64/include/uapi/asm/kvm.h file from the linux kernel.
const KVM_VGIC_V2_DIST_SIZE: u64 = 0x1000;
const KVM_VGIC_V2_CPU_SIZE: u64 = 0x2000;
// Device trees specific constants
const ARCH_GIC_V2_MAINT_IRQ: u32 = 8;
/// Get the address of the GICv2 distributor.
const fn get_dist_addr() -> u64 {
layout::MAPPED_IO_START - KvmGICv2::KVM_VGIC_V2_DIST_SIZE
}
/// Get the size of the GIC_v2 distributor.
const fn get_dist_size() -> u64 {
KvmGICv2::KVM_VGIC_V2_DIST_SIZE
}
/// Get the address of the GIC_v2 CPU.
const fn get_cpu_addr() -> u64 {
KvmGICv2::get_dist_addr() - KvmGICv2::KVM_VGIC_V2_CPU_SIZE
}
/// Get the size of the GIC_v2 CPU.
const fn get_cpu_size() -> u64 {
KvmGICv2::KVM_VGIC_V2_CPU_SIZE
}
}
impl GICDevice for KvmGICv2 {
fn device(&self) -> &Arc<dyn hypervisor::Device> {
&self.device
}
fn device_properties(&self) -> &[u64] {
&self.properties
}
fn fdt_compatibility(&self) -> &str {
"arm,gic-400"
}
fn fdt_maint_irq(&self) -> u32 {
KvmGICv2::ARCH_GIC_V2_MAINT_IRQ
}
fn vcpu_count(&self) -> u64 {
self.vcpu_count
}
}
impl KvmGICDevice for KvmGICv2 {
fn version() -> u32 {
kvm_bindings::kvm_device_type_KVM_DEV_TYPE_ARM_VGIC_V2
}
fn create_device(
device: Arc<dyn hypervisor::Device>,
vcpu_count: u64,
) -> Box<dyn GICDevice> {
Box::new(KvmGICv2 {
device,
properties: [
KvmGICv2::get_dist_addr(),
KvmGICv2::get_dist_size(),
KvmGICv2::get_cpu_addr(),
KvmGICv2::get_cpu_size(),
],
vcpu_count,
})
}
fn init_device_attributes(
_vm: &Arc<dyn hypervisor::Vm>,
gic_device: &dyn GICDevice,
) -> Result<()> {
/* Setting up the distributor attribute.
We are placing the GIC below 1GB so we need to substract the size of the distributor. */
Self::set_device_attribute(
gic_device.device(),
kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ADDR,
u64::from(kvm_bindings::KVM_VGIC_V2_ADDR_TYPE_DIST),
&KvmGICv2::get_dist_addr() as *const u64 as u64,
0,
)?;
/* Setting up the CPU attribute. */
Self::set_device_attribute(
gic_device.device(),
kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ADDR,
u64::from(kvm_bindings::KVM_VGIC_V2_ADDR_TYPE_CPU),
&KvmGICv2::get_cpu_addr() as *const u64 as u64,
0,
)?;
Ok(())
}
}
}

View File

@@ -1,127 +0,0 @@
// Copyright 2019 Amazon.com, Inc. or its affiliates. All Rights Reserved.
// SPDX-License-Identifier: Apache-2.0
pub mod kvm {
use crate::aarch64::gic::kvm::KvmGICDevice;
use crate::aarch64::gic::{Error, GICDevice};
use crate::layout;
use hypervisor::kvm::kvm_bindings;
use std::sync::Arc;
use std::{boxed::Box, result};
type Result<T> = result::Result<T, Error>;
pub struct KvmGICv3 {
/// The hypervisor agnostic device
device: Arc<dyn hypervisor::Device>,
/// GIC device properties, to be used for setting up the fdt entry
properties: [u64; 4],
/// Number of CPUs handled by the device
vcpu_count: u64,
}
impl KvmGICv3 {
// Unfortunately bindgen omits defines that are based on other defines.
// See arch/arm64/include/uapi/asm/kvm.h file from the linux kernel.
pub const SZ_64K: u64 = 0x0001_0000;
const KVM_VGIC_V3_DIST_SIZE: u64 = KvmGICv3::SZ_64K;
const KVM_VGIC_V3_REDIST_SIZE: u64 = (2 * KvmGICv3::SZ_64K);
// Device trees specific constants
pub const ARCH_GIC_V3_MAINT_IRQ: u32 = 9;
/// Get the address of the GIC distributor.
pub fn get_dist_addr() -> u64 {
layout::MAPPED_IO_START - KvmGICv3::KVM_VGIC_V3_DIST_SIZE
}
/// Get the size of the GIC distributor.
pub fn get_dist_size() -> u64 {
KvmGICv3::KVM_VGIC_V3_DIST_SIZE
}
/// Get the address of the GIC redistributors.
pub fn get_redists_addr(vcpu_count: u64) -> u64 {
KvmGICv3::get_dist_addr() - KvmGICv3::get_redists_size(vcpu_count)
}
/// Get the size of the GIC redistributors.
pub fn get_redists_size(vcpu_count: u64) -> u64 {
vcpu_count * KvmGICv3::KVM_VGIC_V3_REDIST_SIZE
}
}
impl GICDevice for KvmGICv3 {
fn device(&self) -> &Arc<dyn hypervisor::Device> {
&self.device
}
fn fdt_compatibility(&self) -> &str {
"arm,gic-v3"
}
fn fdt_maint_irq(&self) -> u32 {
KvmGICv3::ARCH_GIC_V3_MAINT_IRQ
}
fn device_properties(&self) -> &[u64] {
&self.properties
}
fn vcpu_count(&self) -> u64 {
self.vcpu_count
}
}
impl KvmGICDevice for KvmGICv3 {
fn version() -> u32 {
kvm_bindings::kvm_device_type_KVM_DEV_TYPE_ARM_VGIC_V3
}
fn create_device(
device: Arc<dyn hypervisor::Device>,
vcpu_count: u64,
) -> Box<dyn GICDevice> {
Box::new(KvmGICv3 {
device,
properties: [
KvmGICv3::get_dist_addr(),
KvmGICv3::get_dist_size(),
KvmGICv3::get_redists_addr(vcpu_count),
KvmGICv3::get_redists_size(vcpu_count),
],
vcpu_count,
})
}
fn init_device_attributes(
_vm: &Arc<dyn hypervisor::Vm>,
gic_device: &dyn GICDevice,
) -> Result<()> {
/* Setting up the distributor attribute.
We are placing the GIC below 1GB so we need to substract the size of the distributor.
*/
Self::set_device_attribute(
gic_device.device(),
kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ADDR,
u64::from(kvm_bindings::KVM_VGIC_V3_ADDR_TYPE_DIST),
&KvmGICv3::get_dist_addr() as *const u64 as u64,
0,
)?;
/* Setting up the redistributors' attribute.
We are calculating here the start of the redistributors address. We have one per CPU.
*/
Self::set_device_attribute(
gic_device.device(),
kvm_bindings::KVM_DEV_ARM_VGIC_GRP_ADDR,
u64::from(kvm_bindings::KVM_VGIC_V3_ADDR_TYPE_REDIST),
&KvmGICv3::get_redists_addr(gic_device.vcpu_count()) as *const u64 as u64,
0,
)?;
Ok(())
}
}
}

View File

@@ -37,7 +37,7 @@
// | |
// 144 M +---------------------------------------------------------------|
// | |
// | Reserverd (now GIC is here) |
// | Reserved (now GIC is here) |
// | |
// 0GB +---------------------------------------------------------------+
//

View File

@@ -6,9 +6,6 @@
pub mod fdt;
/// Module for the global interrupt controller configuration.
pub mod gic;
mod gicv2;
mod gicv3;
mod gicv3_its;
/// Layout for this aarch64 system.
pub mod layout;
/// Logic for configuring aarch64 registers.
@@ -17,7 +14,7 @@ pub mod regs;
pub use self::fdt::DeviceInfoForFDT;
use crate::DeviceType;
use crate::RegionType;
use hypervisor::kvm::kvm_bindings;
use aarch64::gic::GICDevice;
use std::collections::HashMap;
use std::ffi::CStr;
use std::fmt::Debug;
@@ -42,11 +39,8 @@ pub enum Error {
/// Error configuring the general purpose registers
REGSConfiguration(regs::Error),
/// Error fetching prefered target
VcpuArmPreferredTarget(hypervisor::HypervisorVmError),
/// Error doing Vcpu Init on Arm.
VcpuArmInit(hypervisor::HypervisorCpuError),
/// Error configuring the MPIDR register
VcpuRegMPIDR(hypervisor::HypervisorCpuError),
}
impl From<Error> for super::Error {
@@ -67,23 +61,10 @@ pub struct EntryPoint {
pub fn configure_vcpu(
fd: &Arc<dyn hypervisor::Vcpu>,
id: u8,
vm: &Arc<dyn hypervisor::Vm>,
kernel_entry_point: Option<EntryPoint>,
vm_memory: &GuestMemoryAtomic<GuestMemoryMmap>,
_phys_bits: u8,
) -> super::Result<u64> {
let mut kvi: kvm_bindings::kvm_vcpu_init = kvm_bindings::kvm_vcpu_init::default();
// This reads back the kernel's preferred target type.
vm.get_preferred_target(&mut kvi)
.map_err(Error::VcpuArmPreferredTarget)?;
// We already checked that the capability is supported.
kvi.features[0] |= 1 << kvm_bindings::KVM_ARM_VCPU_PSCI_0_2;
// Non-boot cpus are powered off initially.
if id > 0 {
kvi.features[0] |= 1 << kvm_bindings::KVM_ARM_VCPU_POWER_OFF;
}
fd.vcpu_init(&kvi).map_err(Error::VcpuArmInit)?;
if let Some(kernel_entry_point) = kernel_entry_point {
regs::setup_regs(
fd,
@@ -94,7 +75,7 @@ pub fn configure_vcpu(
.map_err(Error::REGSConfiguration)?;
}
let mpidr = regs::read_mpidr(fd).map_err(Error::REGSConfiguration)?;
let mpidr = fd.read_mpidr().map_err(Error::VcpuRegMPIDR)?;
Ok(mpidr)
}
@@ -145,13 +126,9 @@ pub fn configure_system<T: DeviceInfoForFDT + Clone + Debug, S: ::std::hash::Bui
vcpu_mpidr: Vec<u64>,
device_info: &HashMap<(DeviceType, String), T, S>,
initrd: &Option<super::InitramfsConfig>,
pci_space_address: &Option<(u64, u64)>,
) -> super::Result<()> {
// If pci_space_address is present, it means PCI devices are used ("pci" feature enabled).
// Then GITv3-ITS is required for MSI messaging.
// Otherwise ("mmio" feature enabled), any version of GIC is OK.
let gic_device = gic::kvm::create_gic(vm, vcpu_count, pci_space_address.is_some())
.map_err(Error::SetupGIC)?;
pci_space_address: &(u64, u64),
) -> super::Result<Box<dyn GICDevice>> {
let gic_device = gic::kvm::create_gic(vm, vcpu_count).map_err(Error::SetupGIC)?;
fdt::create_fdt(
guest_mem,
@@ -164,7 +141,7 @@ pub fn configure_system<T: DeviceInfoForFDT + Clone + Debug, S: ::std::hash::Bui
)
.map_err(Error::SetupFDT)?;
Ok(())
Ok(gic_device)
}
/// Returns the memory address where the initramfs could be loaded.
@@ -210,10 +187,10 @@ fn get_fdt_addr(mem: &GuestMemoryMmap) -> u64 {
pub fn get_host_cpu_phys_bits() -> u8 {
// The value returned here is used to determine the physical address space size
// for a VM (IPA size).
// In recent kernel versions, the maxium IPA size supported by the host can be
// In recent kernel versions, the maximum IPA size supported by the host can be
// known by querying cap KVM_CAP_ARM_VM_IPA_SIZE. And the IPA size for a
// guest can be configured smaller.
// But in Cloud-Hypervisor we simply use the maxium value for the VM.
// But in Cloud-Hypervisor we simply use the maximum value for the VM.
// Reference https://lwn.net/Articles/766767/.
//
// The correct way to query KVM_CAP_ARM_VM_IPA_SIZE is via rust-vmm/kvm-ioctls,

View File

@@ -7,12 +7,9 @@
use super::get_fdt_addr;
use hypervisor::kvm::kvm_bindings::{
user_pt_regs, KVM_REG_ARM64, KVM_REG_ARM64_SYSREG, KVM_REG_ARM64_SYSREG_CRM_MASK,
KVM_REG_ARM64_SYSREG_CRM_SHIFT, KVM_REG_ARM64_SYSREG_CRN_MASK, KVM_REG_ARM64_SYSREG_CRN_SHIFT,
KVM_REG_ARM64_SYSREG_OP0_MASK, KVM_REG_ARM64_SYSREG_OP0_SHIFT, KVM_REG_ARM64_SYSREG_OP1_MASK,
KVM_REG_ARM64_SYSREG_OP1_SHIFT, KVM_REG_ARM64_SYSREG_OP2_MASK, KVM_REG_ARM64_SYSREG_OP2_SHIFT,
KVM_REG_ARM_CORE, KVM_REG_SIZE_U64,
kvm_regs, user_pt_regs, KVM_REG_ARM64, KVM_REG_ARM_CORE, KVM_REG_SIZE_U64,
};
use hypervisor::{arm64_core_reg_id, offset__of};
use std::sync::Arc;
use std::{mem, result};
use vm_memory::GuestMemoryMmap;
@@ -38,80 +35,6 @@ const PSR_D_BIT: u64 = 0x0000_0200;
// Taken from arch/arm64/kvm/inject_fault.c.
const PSTATE_FAULT_BITS_64: u64 = PSR_MODE_EL1h | PSR_A_BIT | PSR_F_BIT | PSR_I_BIT | PSR_D_BIT;
// Following are macros that help with getting the ID of a aarch64 core register.
// The core register are represented by the user_pt_regs structure. Look for it in
// arch/arm64/include/uapi/asm/ptrace.h.
// This macro gets the offset of a structure (i.e `str`) member (i.e `field`) without having
// an instance of that structure.
// It uses a null pointer to retrieve the offset to the field.
// Inspired by C solution: `#define offsetof(str, f) ((size_t)(&((str *)0)->f))`.
// Doing `offset__of!(user_pt_regs, pstate)` in our rust code will trigger the following:
// unsafe { &(*(0 as *const user_pt_regs)).pstate as *const _ as usize }
// The dereference expression produces an lvalue, but that lvalue is not actually read from,
// we're just doing pointer math on it, so in theory, it should safe.
macro_rules! offset__of {
($str:ty, $field:ident) => {
unsafe { &(*std::ptr::null::<user_pt_regs>()).$field as *const _ as usize }
};
}
macro_rules! arm64_core_reg {
($reg: tt) => {
// As per `kvm_arm_copy_reg_indices`, the id of a core register can be obtained like this:
// `const u64 core_reg = KVM_REG_ARM64 | KVM_REG_SIZE_U64 | KVM_REG_ARM_CORE | i`, where i is obtained with:
// `for (i = 0; i < sizeof(struct kvm_regs) / sizeof(__u32); i++) {`
// We are using here `user_pt_regs` since this structure contains the core register and it is at
// the start of `kvm_regs`.
// struct kvm_regs {
// struct user_pt_regs regs; /* sp = sp_el0 */
//
// __u64 sp_el1;
// __u64 elr_el1;
//
// __u64 spsr[KVM_NR_SPSR];
//
// struct user_fpsimd_state fp_regs;
//};
// struct user_pt_regs {
// __u64 regs[31];
// __u64 sp;
// __u64 pc;
// __u64 pstate;
//};
// In our implementation we need: pc, pstate and user_pt_regs->regs[0].
KVM_REG_ARM64 as u64
| KVM_REG_SIZE_U64 as u64
| u64::from(KVM_REG_ARM_CORE)
| ((offset__of!(user_pt_regs, $reg) / mem::size_of::<u32>()) as u64)
};
}
// This macro computes the ID of a specific ARM64 system register similar to how
// the kernel C macro does.
// https://elixir.bootlin.com/linux/v4.20.17/source/arch/arm64/include/uapi/asm/kvm.h#L203
macro_rules! arm64_sys_reg {
($name: tt, $op0: tt, $op1: tt, $crn: tt, $crm: tt, $op2: tt) => {
const $name: u64 = KVM_REG_ARM64 as u64
| KVM_REG_SIZE_U64 as u64
| KVM_REG_ARM64_SYSREG as u64
| ((($op0 as u64) << KVM_REG_ARM64_SYSREG_OP0_SHIFT)
& KVM_REG_ARM64_SYSREG_OP0_MASK as u64)
| ((($op1 as u64) << KVM_REG_ARM64_SYSREG_OP1_SHIFT)
& KVM_REG_ARM64_SYSREG_OP1_MASK as u64)
| ((($crn as u64) << KVM_REG_ARM64_SYSREG_CRN_SHIFT)
& KVM_REG_ARM64_SYSREG_CRN_MASK as u64)
| ((($crm as u64) << KVM_REG_ARM64_SYSREG_CRM_SHIFT)
& KVM_REG_ARM64_SYSREG_CRM_MASK as u64)
| ((($op2 as u64) << KVM_REG_ARM64_SYSREG_OP2_SHIFT)
& KVM_REG_ARM64_SYSREG_OP2_MASK as u64);
};
}
// Constant imported from the Linux kernel:
// https://elixir.bootlin.com/linux/v4.20.17/source/arch/arm64/include/asm/sysreg.h#L135
arm64_sys_reg!(MPIDR_EL1, 3, 0, 0, 0, 5);
/// Configure core registers for a given CPU.
///
/// # Arguments
@@ -126,31 +49,33 @@ pub fn setup_regs(
boot_ip: u64,
mem: &GuestMemoryMmap,
) -> Result<()> {
let kreg_off = offset__of!(kvm_regs, regs);
// Get the register index of the PSTATE (Processor State) register.
vcpu.set_one_reg(arm64_core_reg!(pstate), PSTATE_FAULT_BITS_64)
.map_err(Error::SetCoreRegister)?;
let pstate = offset__of!(user_pt_regs, pstate) + kreg_off;
vcpu.set_reg(
arm64_core_reg_id!(KVM_REG_SIZE_U64, pstate),
PSTATE_FAULT_BITS_64,
)
.map_err(Error::SetCoreRegister)?;
// Other vCPUs are powered off initially awaiting PSCI wakeup.
if cpu_id == 0 {
// Setting the PC (Processor Counter) to the current program address (kernel address).
vcpu.set_one_reg(arm64_core_reg!(pc), boot_ip)
let pc = offset__of!(user_pt_regs, pc) + kreg_off;
vcpu.set_reg(arm64_core_reg_id!(KVM_REG_SIZE_U64, pc), boot_ip as u64)
.map_err(Error::SetCoreRegister)?;
// Last mandatory thing to set -> the address pointing to the FDT (also called DTB).
// "The device tree blob (dtb) must be placed on an 8-byte boundary and must
// not exceed 2 megabytes in size." -> https://www.kernel.org/doc/Documentation/arm64/booting.txt.
// We are choosing to place it the end of DRAM. See `get_fdt_addr`.
vcpu.set_one_reg(arm64_core_reg!(regs), get_fdt_addr(mem) as u64)
.map_err(Error::SetCoreRegister)?;
let regs0 = offset__of!(user_pt_regs, regs) + kreg_off;
vcpu.set_reg(
arm64_core_reg_id!(KVM_REG_SIZE_U64, regs0),
get_fdt_addr(mem) as u64,
)
.map_err(Error::SetCoreRegister)?;
}
Ok(())
}
/// Read the MPIDR - Multiprocessor Affinity Register.
///
/// # Arguments
///
/// * `vcpu` - Structure for the VCPU that holds the VCPU's fd.
pub fn read_mpidr(vcpu: &Arc<dyn hypervisor::Vcpu>) -> Result<u64> {
vcpu.get_one_reg(MPIDR_EL1).map_err(Error::GetSysRegister)
}

View File

@@ -14,6 +14,7 @@
clippy::cast_ptr_alignment
)]
extern crate anyhow;
extern crate byteorder;
extern crate hypervisor;
extern crate libc;
@@ -23,7 +24,13 @@ extern crate log;
extern crate acpi_tables;
extern crate arch_gen;
extern crate linux_loader;
extern crate serde;
extern crate vm_memory;
extern crate vm_migration;
#[cfg(target_arch = "aarch64")]
#[macro_use]
extern crate serde_derive;
extern crate serde_json;
use std::fmt;
use std::result;

View File

@@ -10,7 +10,6 @@ use std::sync::Arc;
mod gdt;
pub mod interrupts;
pub mod layout;
#[cfg(not(feature = "acpi"))]
mod mptable;
pub mod regs;
use crate::InitramfsConfig;
@@ -26,6 +25,7 @@ use vm_memory::{
GuestMemoryMmap, GuestMemoryRegion, GuestUsize,
};
mod smbios;
use std::arch::x86_64;
#[derive(Debug, Copy, Clone)]
pub enum BootProtocol {
@@ -141,7 +141,7 @@ unsafe impl ByteValued for BootParamsWrapper {}
pub enum Error {
/// Invalid e820 setup params.
E820Configuration,
#[cfg(not(feature = "acpi"))]
/// Error writing MP table to memory.
MpTableSetup(mptable::Error),
@@ -337,6 +337,7 @@ pub fn configure_vcpu(
vm_memory: &GuestMemoryAtomic<GuestMemoryMmap>,
cpuid: CpuId,
kvm_hyperv: bool,
phys_bits: u8,
) -> super::Result<()> {
let mut cpuid = cpuid;
CpuidPatch::set_cpuid_reg(&mut cpuid, 0xb, None, CpuidReg::EDX, u32::from(id));
@@ -394,6 +395,29 @@ pub fn configure_vcpu(
}
}
// Copy CPU identification string
for i in 0x8000_0002..=0x8000_0004 {
cpuid.retain(|c| c.function != i);
let leaf = unsafe { x86_64::__cpuid(i) };
cpuid
.push(CpuIdEntry {
function: i,
eax: leaf.eax,
ebx: leaf.ebx,
ecx: leaf.ecx,
edx: leaf.edx,
..Default::default()
})
.map_err(|_| Error::PopulatingCpuid)?;
}
// Set CPU physical bits
for entry in cpuid.as_mut_slice().iter_mut() {
if entry.function == 0x8000_0008 {
entry.eax = (entry.eax & 0xffff_ff00) | (phys_bits as u32 & 0xff);
}
}
fd.set_cpuid2(&cpuid)
.map_err(|e| Error::SetSupportedCpusFailed(e.into()))?;
@@ -487,11 +511,12 @@ pub fn configure_system(
boot_prot: BootProtocol,
sgx_epc_region: Option<SgxEpcRegion>,
) -> super::Result<()> {
smbios::setup_smbios(guest_mem).map_err(Error::SmbiosSetup)?;
let size = smbios::setup_smbios(guest_mem).map_err(Error::SmbiosSetup)?;
// Note that this puts the mptable at the last 1k of Linux's 640k base RAM
#[cfg(not(feature = "acpi"))]
mptable::setup_mptable(guest_mem, _num_cpus).map_err(Error::MpTableSetup)?;
// Place the MP table after the SMIOS table aligned to 16 bytes
let offset = GuestAddress(layout::SMBIOS_START).unchecked_add(size);
let offset = GuestAddress((offset.0 + 16) & !0xf);
mptable::setup_mptable(offset, guest_mem, _num_cpus).map_err(Error::MpTableSetup)?;
// Check that the RAM is not smaller than the RSDP start address
if let Some(rsdp_addr) = rsdp_addr {
@@ -808,7 +833,6 @@ pub fn initramfs_load_addr(
}
pub fn get_host_cpu_phys_bits() -> u8 {
use std::arch::x86_64;
unsafe {
let leaf = x86_64::__cpuid(0x8000_0000);

View File

@@ -13,8 +13,10 @@ use std::slice;
use libc::c_char;
use arch_gen::x86::mpspec;
use layout::{APIC_START, IOAPIC_START, MPTABLE_START};
use vm_memory::{Address, ByteValued, Bytes, GuestMemory, GuestMemoryError, GuestMemoryMmap};
use layout::{APIC_START, HIGH_RAM_START, IOAPIC_START};
use vm_memory::{
Address, ByteValued, Bytes, GuestAddress, GuestMemory, GuestMemoryError, GuestMemoryMmap,
};
// This is a workaround to the Rust enforcement specifying that any implementation of a foreign
// trait (in this case `ByteValued`) where:
@@ -121,16 +123,21 @@ fn compute_mp_size(num_cpus: u8) -> usize {
}
/// Performs setup of the MP table for the given `num_cpus`.
pub fn setup_mptable(mem: &GuestMemoryMmap, num_cpus: u8) -> Result<()> {
pub fn setup_mptable(offset: GuestAddress, mem: &GuestMemoryMmap, num_cpus: u8) -> Result<()> {
if num_cpus as u32 > MAX_SUPPORTED_CPUS {
return Err(Error::TooManyCpus);
}
// Used to keep track of the next base pointer into the MP table.
let mut base_mp = MPTABLE_START;
let mut base_mp = offset;
let mp_size = compute_mp_size(num_cpus);
if offset.unchecked_add(mp_size as u64) >= HIGH_RAM_START {
warn!("Skipping mptable creation due to insufficient space");
return Ok(());
}
let mut checksum: u8 = 0;
let ioapicid: u8 = num_cpus + 1;
@@ -280,6 +287,7 @@ pub fn setup_mptable(mem: &GuestMemoryMmap, num_cpus: u8) -> Result<()> {
#[cfg(test)]
mod tests {
use super::*;
use layout::MPTABLE_START;
use vm_memory::{GuestAddress, GuestUsize};
fn table_entry_size(type_: u8) -> usize {
@@ -299,7 +307,7 @@ mod tests {
let mem =
GuestMemoryMmap::from_ranges(&[(MPTABLE_START, compute_mp_size(num_cpus))]).unwrap();
setup_mptable(&mem, num_cpus).unwrap();
setup_mptable(MPTABLE_START, &mem, num_cpus).unwrap();
}
#[test]
@@ -308,7 +316,7 @@ mod tests {
let mem = GuestMemoryMmap::from_ranges(&[(MPTABLE_START, compute_mp_size(num_cpus) - 1)])
.unwrap();
assert!(setup_mptable(&mem, num_cpus).is_err());
assert!(setup_mptable(MPTABLE_START, &mem, num_cpus).is_err());
}
#[test]
@@ -317,7 +325,7 @@ mod tests {
let mem =
GuestMemoryMmap::from_ranges(&[(MPTABLE_START, compute_mp_size(num_cpus))]).unwrap();
setup_mptable(&mem, num_cpus).unwrap();
setup_mptable(MPTABLE_START, &mem, num_cpus).unwrap();
let mpf_intel: MpfIntelWrapper = mem.read_obj(MPTABLE_START).unwrap();
@@ -333,7 +341,7 @@ mod tests {
let mem =
GuestMemoryMmap::from_ranges(&[(MPTABLE_START, compute_mp_size(num_cpus))]).unwrap();
setup_mptable(&mem, num_cpus).unwrap();
setup_mptable(MPTABLE_START, &mem, num_cpus).unwrap();
let mpf_intel: MpfIntelWrapper = mem.read_obj(MPTABLE_START).unwrap();
let mpc_offset = GuestAddress(mpf_intel.0.physptr as GuestUsize);
@@ -367,7 +375,7 @@ mod tests {
.unwrap();
for i in 0..MAX_SUPPORTED_CPUS as u8 {
setup_mptable(&mem, i).unwrap();
setup_mptable(MPTABLE_START, &mem, i).unwrap();
let mpf_intel: MpfIntelWrapper = mem.read_obj(MPTABLE_START).unwrap();
let mpc_offset = GuestAddress(mpf_intel.0.physptr as GuestUsize);
@@ -400,7 +408,7 @@ mod tests {
let mem =
GuestMemoryMmap::from_ranges(&[(MPTABLE_START, compute_mp_size(cpus as u8))]).unwrap();
let result = setup_mptable(&mem, cpus as u8);
let result = setup_mptable(MPTABLE_START, &mem, cpus as u8);
assert!(result.is_err());
}
}

View File

@@ -161,7 +161,7 @@ fn write_string(
Ok(curptr)
}
pub fn setup_smbios(mem: &GuestMemoryMmap) -> Result<()> {
pub fn setup_smbios(mem: &GuestMemoryMmap) -> Result<u64> {
let physptr = GuestAddress(SMBIOS_START)
.checked_add(mem::size_of::<Smbios30Entrypoint>() as u64)
.ok_or(Error::NotEnoughMemory)?;
@@ -224,7 +224,7 @@ pub fn setup_smbios(mem: &GuestMemoryMmap) -> Result<()> {
.map_err(|_| Error::WriteSmbiosEp)?;
}
Ok(())
Ok(curptr.unchecked_offset_from(physptr))
}
#[cfg(test)]

View File

@@ -10,12 +10,12 @@ io_uring = []
[dependencies]
io-uring = ">=0.4.0"
libc = "0.2.77"
libc = "0.2.80"
log = "0.4.11"
serde = ">=1.0.27"
serde_derive = ">=1.0.27"
serde_json = ">=1.0.9"
virtio-bindings = { version = "0.1", features = ["virtio-v5_0_0"]}
vm-memory = { version = "0.2.1", features = ["backend-mmap", "backend-atomic"] }
vm-memory = { version = "0.3.0", features = ["backend-mmap", "backend-atomic"] }
vm-virtio = { path = "../vm-virtio" }
vmm-sys-util = ">=0.3.1"

View File

@@ -14,13 +14,15 @@ extern crate log;
extern crate serde_derive;
#[cfg(feature = "io_uring")]
use io_uring::{opcode, squeue, IoUring, Probe};
use io_uring::Probe;
use io_uring::{opcode, squeue, IoUring};
use serde::ser::{Serialize, SerializeStruct, Serializer};
use std::cmp;
use std::io::{self, Read, Seek, SeekFrom, Write};
use std::os::linux::fs::MetadataExt;
#[cfg(feature = "io_uring")]
use std::os::unix::io::{AsRawFd, RawFd};
use std::os::unix::io::AsRawFd;
use std::os::unix::io::RawFd;
use std::path::PathBuf;
use std::result;
use virtio_bindings::bindings::virtio_blk::*;
@@ -266,7 +268,6 @@ impl Request {
Ok(len)
}
#[cfg(feature = "io_uring")]
pub fn execute_io_uring(
&self,
mem: &GuestMemoryMmap,
@@ -565,3 +566,8 @@ pub fn block_io_uring_is_supported() -> bool {
true
}
#[cfg(not(feature = "io_uring"))]
pub fn block_io_uring_is_supported() -> bool {
false
}

View File

@@ -9,7 +9,7 @@ extern crate clap;
use std::process::Command;
fn main() {
let mut version = crate_version!().to_string();
let mut version = "v".to_owned() + crate_version!();
if let Ok(git_out) = Command::new("git").args(&["describe", "--dirty"]).output() {
if git_out.status.success() {

View File

@@ -8,14 +8,14 @@ anyhow = "1.0"
bitflags = ">=1.2.1"
byteorder = "1.3.4"
epoll = ">=4.0.1"
libc = "0.2.77"
libc = "0.2.80"
log = "0.4.11"
serde = {version = ">=1.0.27", features = ["rc"] }
serde_derive = ">=1.0.27"
serde_json = ">=1.0.9"
vm-device = { path = "../vm-device" }
acpi_tables = { path = "../acpi_tables", optional = true }
vm-memory = "0.2.1"
vm-memory = "0.3.0"
vm-migration = { path = "../vm-migration" }
vmm-sys-util = ">=0.3.1"

View File

@@ -8,7 +8,6 @@ use std::sync::Arc;
use vm_device::interrupt::{
InterruptIndex, InterruptManager, InterruptSourceGroup, MsiIrqGroupConfig,
};
use vm_migration::{Migratable, MigratableError, Pausable, Snapshot, Snapshottable, Transportable};
type Result<T> = result::Result<T, Error>;
@@ -63,22 +62,3 @@ impl InterruptController for Gic {
Ok(())
}
}
const GIC_SNAPSHOT_ID: &str = "gic";
impl Snapshottable for Gic {
fn id(&self) -> String {
GIC_SNAPSHOT_ID.to_string()
}
fn snapshot(&mut self) -> std::result::Result<Snapshot, MigratableError> {
unimplemented!();
}
fn restore(&mut self, _snapshot: Snapshot) -> std::result::Result<(), MigratableError> {
unimplemented!();
}
}
impl Pausable for Gic {}
impl Transportable for Gic {}
impl Migratable for Gic {}

View File

@@ -7,8 +7,6 @@ use std::result;
#[derive(Debug)]
pub enum Error {
/// Invalid destination mode.
InvalidDestinationMode,
/// Invalid trigger mode.
InvalidTriggerMode,
/// Invalid delivery mode.

View File

@@ -67,10 +67,12 @@ fn trigger_mode(entry: RedirectionTableEntry) -> u8 {
fn interrupt_mask(entry: RedirectionTableEntry) -> u8 {
((entry >> 16) & 0x1u64) as u8
}
fn destination_field_physical(entry: RedirectionTableEntry) -> u8 {
((entry >> 56) & 0xfu64) as u8
}
fn destination_field_logical(entry: RedirectionTableEntry) -> u8 {
fn destination_field(entry: RedirectionTableEntry) -> u8 {
// When the destination mode is physical, the destination field should only
// be defined through bits 56-59, as defined in the IOAPIC specification.
// But from the APIC specification, the APIC ID is always defined on 8 bits
// no matter which destination mode is selected. That's why we always
// retrieve the destination field based on bits 56-63.
((entry >> 56) & 0xffu64) as u8
}
fn set_delivery_status(entry: &mut RedirectionTableEntry, val: u8) {
@@ -100,12 +102,6 @@ const IOWIN_OFF: u8 = 0x10;
const IOWIN_SCALE: u8 = 0x2;
const REG_MAX_OFFSET: u8 = IOWIN_OFF + (NUM_IOAPIC_PINS as u8 * 2) - 1;
#[repr(u8)]
enum DestinationMode {
Physical = 0,
Logical = 1,
}
#[repr(u8)]
enum TriggerMode {
Edge = 0,
@@ -201,15 +197,13 @@ impl Ioapic {
})
.map_err(Error::CreateInterruptSourceGroup)?;
interrupt_source_group
.enable()
.map_err(Error::EnableInterrupt)?;
// The IOAPIC is created with entries already masked. The guest will be
// in charge of unmasking them if/when necessary.
Ok(Ioapic {
id,
id_reg: 0,
reg_sel: 0,
reg_entries: [0; NUM_IOAPIC_PINS],
reg_entries: [0x10000; NUM_IOAPIC_PINS],
used_entries: [false; NUM_IOAPIC_PINS],
apic_address,
interrupt_source_group,
@@ -220,6 +214,16 @@ impl Ioapic {
debug!("IOAPIC_W reg 0x{:x}, val 0x{:x}", self.reg_sel, val);
match self.reg_sel as u8 {
IOAPIC_REG_VERSION => {
if val == 0 {
// Windows writes zero here (see #1791)
} else {
error!(
"IOAPIC: invalid write to version register (0x{:x}): 0x{:x}",
self.reg_sel, val
);
}
}
IOAPIC_REG_ID => self.id_reg = (val >> 24) & 0xf,
IOWIN_OFF..=REG_MAX_OFFSET => {
let (index, is_high_bits) = decode_irq_from_selector(self.reg_sel as u8);
@@ -241,7 +245,10 @@ impl Ioapic {
// Store the information this IRQ is now being used.
self.used_entries[index] = true;
}
_ => error!("IOAPIC: invalid write to register offset"),
_ => error!(
"IOAPIC: invalid write to register offset 0x{:x}",
self.reg_sel
),
}
}
@@ -260,7 +267,10 @@ impl Ioapic {
}
}
_ => {
error!("IOAPIC: invalid read from register offset");
error!(
"IOAPIC: invalid read from register offset 0x{:x}",
self.reg_sel
);
0
}
}
@@ -296,11 +306,7 @@ impl Ioapic {
// Validate Destination Mode value, and retrieve Destination ID
let destination_mode = destination_mode(entry);
let destination_id: u8 = match destination_mode {
x if x == DestinationMode::Physical as u8 => destination_field_physical(entry),
x if x == DestinationMode::Logical as u8 => destination_field_logical(entry),
_ => return Err(Error::InvalidDestinationMode),
};
let destination_id = destination_field(entry);
// When this bit is set, the message is directed to the processor with
// the lowest interrupt priority among processors that can receive the

View File

@@ -20,7 +20,7 @@ use vm_device::BusDevice;
// From 0x0 to 0x1C we have following registers:
const RTCDR: u64 = 0x0; // Data Register.
const RTCMR: u64 = 0x4; // Match Register.
const RTCLR: u64 = 0x8; // Load Regiser.
const RTCLR: u64 = 0x8; // Load Register.
const RTCCR: u64 = 0xc; // Control Register.
const RTCIMSC: u64 = 0x10; // Interrupt Mask Set or Clear Register.
const RTCRIS: u64 = 0x14; // Raw Interrupt Status.

View File

@@ -12,7 +12,7 @@
* [Reboot a Virtual Machine](#reboot-a-virtual-machine)
* [Shut a Virtual Machine Down](#shut-a-virtual-machine-down)
+ [Command Line Interface](#command-line-interface)
+ [REST API and CLI Architecural Relationship](#rest-api-and-cli-architectural-relationship)
+ [REST API and CLI Architectural Relationship](#rest-api-and-cli-architectural-relationship)
* [Internal API](#internal-api)
+ [Goals and Design](#goals-and-design)
* [End to End Example](#end-to-end-example)

View File

@@ -16,20 +16,10 @@ sudo apt-get install libfdt-dev
## Build
For Virtio devices, you can choose MMIO or PCI as transport option.
### MMIO
Using PCI devices requires GICv3-ITS for MSI messaging. GICv3-ITS is very common in modern servers.
```bash
cargo build --no-default-features --features mmio,kvm
```
### PCI
Using PCI devices requires GICv3-ITS for MSI messaging. GICv3-ITS is very common in modern servers, but your machine happen to be old ones with GICv2(M) (like Raspberry Pi 4) or GICv3 without ITS, MMIO can still work.
```bash
cargo build --no-default-features --features pci,kvm
cargo build --no-default-features --features kvm
```
## Image
@@ -51,7 +41,7 @@ To build the development container:
./scripts/dev_cli.sh build-container
```
To build Cloud-hypervisor in the container: (The default option for Virtio transport is MMIO.)
To build Cloud-hypervisor in the container:
```bash
./scripts/dev_cli.sh build
@@ -62,7 +52,7 @@ To build Cloud-hypervisor in the container: (The default option for Virtio trans
Assuming you have built Cloud-hypervisor with the development container, a VM can be started with command:
```bash
sudo build/cargo_target/aarch64-unknown-linux-gnu/debug/cloud-hypervisor --kernel kernel.bin --disk path=rootfs.ext4 --cmdline "keep_bootcon console=hvc0 reboot=k panic=1 pci=off root=/dev/vda rw" --cpus boot=4 --memory size=512M --serial file=serial.log --log-file log.log -vvv
sudo build/cargo_target/aarch64-unknown-linux-gnu/debug/cloud-hypervisor --kernel kernel.bin --disk path=rootfs.ext4 --cmdline "keep_bootcon console=hvc0 reboot=k panic=1 root=/dev/vda rw" --cpus boot=4 --memory size=512M --serial file=serial.log --log-file log.log -vvv
```
If the build was done out of the container, replace the binary path with `target/debug/cloud-hypervisor`.

View File

@@ -75,12 +75,8 @@ feature is enabled by default.
## Virtio devices
For all virtio devices listed below, both `virtio-mmio` and `virtio-pci`
transport layers are supported, `virtio-pci` being the default.
Both `virtio-mmio` and `virtio-pci` can be compiled out. `virtio-pci` is
built-in by default, and enabled by default. If both transport layers were
built at the same time, `virtio-pci` would be the default transport layer.
For all virtio devices listed below, only `virtio-pci` transport layer is
supported.
### virtio-block

View File

@@ -63,7 +63,7 @@ As per adding CPUs to the guest, after a reboot the VM will be running with the
## Memory Hot Plug
Extra memory can be added from a runing Cloud Hypervisor instance. This is controlled by two mechanisms:
Extra memory can be added from a running Cloud Hypervisor instance. This is controlled by two mechanisms:
1. Allocating some of the guest physical address space for hotplug memory.
2. Making a HTTP API request to the VMM to ask for a new amount of RAM to be assigned to the VM. In the case of expanding the memory for the VM the new memory will be hotplugged into the running VM, if reducing the size of the memory then change will take effect after the next reboot.

View File

@@ -8,7 +8,7 @@ the host kernel. The required Linux and KVM changes can be found in the
[KVM SGX Tree](https://github.com/intel/kvm-sgx).
Utilizing SGX in the guest requires a kernel/OS with SGX support, e.g. a kernel
buit using the [SGX Linux Development Tree](https://github.com/jsakkine-intel/linux-sgx.git)
built using the [SGX Linux Development Tree](https://github.com/jsakkine-intel/linux-sgx.git)
or the [KVM SGX Tree](https://github.com/intel/kvm-sgx). Running KVM SGX as the
guest kernel allows nested virtualization of SGX.

43
docs/logging.md Normal file
View File

@@ -0,0 +1,43 @@
# Logging
The target audience of this document is both:
* Developers who want to understand what log level to use and when,
* Users who want to debug issues with running their workloads in Cloud Hypervisor
## Control
The number of `-v` parameters passed to the `cloud-hypervisor` binary will determine the log level. Currenly the default is log messages up to `WARN:` (`warn!`) are included by default. The `--log-file` allows the log to be sent to a location other than `stderr`.
## Levels
### `error!()`
For immediate, unrecoverable errors where it does not make sense for the execution to continue as the behaviour of the VM is considerablely impacted.
Cloud Hypervisor should exit shortly after reporting this error (with a non-zero exit code). Generally this should be used during initial construction of the VM state before the virtual CPUs have begun running code.
A typical situation where this might occur is when the user is using command line options that conflict with each other or is trying to use a file that is not present on the filesystem.
Users should react to this error by checking their initial VM configuration.
### `warn!()`
A serious problem has occured but the execution of the VM can continue although some functionality might be impacted.
A typical example of where this level of message should be generated is during an API call request that cannot be fulfilled.
The user should investigate the meaning of this warning and take steps to ensure the correct functionality.
### `info!()`
Use `-v` to enable.
This level is for the benefit of developers. It should be used for sporadic and infrequent messages. The same message should not "spam" the logs. The VM should be usable when this level of debugging is enabled and trying to use `stdin/stdout` and the logs are going to `stderr`.
### `debug!()`
Use `-vv` to enable.
For the most verbose of logging messages. It is acceptable to "spam" the log with repeated invocations of the same message. This level of logging would be combined with `--log-file`.

View File

@@ -18,14 +18,12 @@ struct MemoryConfig {
hotplug_method: HotplugMethod,
hotplug_size: Option<u64>,
hotplugged_size: Option<u64>,
balloon: bool,
balloon_size: u64,
zones: Option<Vec<MemoryZoneConfig>>,
}
```
```
--memory <memory> Memory parameters "size=<guest_memory_size>,mergeable=on|off,shared=on|off,hugepages=on|off,hotplug_method=acpi|virtio-mem,hotplug_size=<hotpluggable_memory_size>,hotplugged_size=<hotplugged_memory_size>,balloon=on|off"
--memory <memory> Memory parameters "size=<guest_memory_size>,mergeable=on|off,shared=on|off,hugepages=on|off,hotplug_method=acpi|virtio-mem,hotplug_size=<hotpluggable_memory_size>,hotplugged_size=<hotplugged_memory_size>"
```
### `size`
@@ -142,20 +140,6 @@ _Example_
--memory size=1G,hotplug_method=virtio-mem,hotplug_size=1G,hotplugged_size=512M
```
### `balloon`
Specifies if the `virtio-balloon` device must be activated. This creates a
dedicated virtio device for managing the balloon in the guest, which allows
guest to access more or less memory depending on the balloon size.
By default this option is turned off.
_Example_
```
--memory size=1G,balloon=on
```
## Advanced Parameters
`MemoryZoneConfig` or what is known as `--memory-zone` from the CLI perspective

View File

@@ -1,6 +1,6 @@
# How to use networking
cloud-hypervisor can emulate one or more virtual network interfaces, represented at the hypervisor host by [tap devices](https://www.kernel.org/doc/Documentation/networking/tuntap.txt). This guide briefly describes, in a manual and distribution neutral way, how to setup and use networking with cloud-hypevisor.
cloud-hypervisor can emulate one or more virtual network interfaces, represented at the hypervisor host by [tap devices](https://www.kernel.org/doc/Documentation/networking/tuntap.txt). This guide briefly describes, in a manual and distribution neutral way, how to setup and use networking with cloud-hypervisor.
## Multiple queue support for net devices ##
@@ -14,7 +14,7 @@ Note:
- Currently, it does not support to use ethtool to change the combined queue numbers in guest.
- Multiple queue is enabled for vhost-user-net backend in cloud-hypervisor, however, multiple thread is not added to handle mq, thus, the performance for vhost-user-net backend is not supposed to be improved. The multiple thread will be added for backend later.
- Performance test for vhost-user-net will be covered once vhost-user-net backend has mulitple thread supported.
- Performance test for vhost-user-net will be covered once vhost-user-net backend has multiple thread supported.
- Performance test for virtio-net is done by comparing 2 queue pairs with 1 queue pairs, that to run 2 iperf3 sessions in the same test environments, throughput is improved about 37%.
## Start cloud-hypervisor with net devices

View File

@@ -54,7 +54,7 @@ drwxr-xr-x 47 foo bar 4096 Jul 22 11:47 ../
In this particular example, we can observe that 2 memory region files were
created. That is explained by the size of the guest RAM, which is 4GiB in this
case. Because it exceeds 3GiB (which is where we can find a ~1GiB memory hole),
Cloud-Hypervisor needs 2 distincts memory regions to be created. Each memory
Cloud-Hypervisor needs 2 distinct memory regions to be created. Each memory
region's content is stored through a dedicated file, which explains why we end
up with 2 different files, the first one containing the guest RAM range 0-3GiB
and the second one containing the guest RAM range 3-4GiB.
@@ -63,7 +63,7 @@ and the second one containing the guest RAM range 3-4GiB.
and state. The configuration bits are used to create a similar virtual machine
with the correct amount of CPUs, RAM, and other expected devices. The state
bits are used to restore each component in the state it was left before the
snapshot occured.
snapshot occurred.
## Restore a Cloud-Hypervisor VM

42
docs/uefi.md Normal file
View File

@@ -0,0 +1,42 @@
# UEFI Boot
Cloud Hypervisor supports UEFI boot through the utilization of the EDK II based UEFI firmware.
## Building UEFI Firmware
To avoid any unnecessary issues, it is recommended to use Ubuntu 18.04 and its default toolset. Any other compatible Linux distribution is otherwise suitable, however it is suggested to use a temporary Docker container with Ubuntu 18.04 for a quick build on an existing Linux machine.
The commands below will compile an OVMF firmware suitable for Cloud Hypervisor.
```shell
sudo apt-get update
sudo apt-get install uuid-dev nasm iasl build-essential python3-distutils git
git clone https://github.com/cloud-hypervisor/edk2 -b ch
cd edk2
. edksetup.sh
git submodule update --init
echo "ACTIVE_PLATFORM=OvmfPkg/OvmfCh.dsc" >> Conf/target.txt
echo "TARGET_ARCH=X64" >> Conf/target.txt
echo "TOOL_CHAIN_TAG=GCC5" >> Conf/target.txt
make -C ./BaseTools
build
```
After the successful build, the resulting firmware binaries are available under `Build/OvmfCh/DEBUG_GCC5/FV` underneath the edk2 checkout.
## Using OVMF Binaries
Any UEFI capable image can be booted using the Cloud Hypervisor specific firmware. Windows guests under Cloud Hypervisor only support UEFI boot, therefore OVMF is mandatory there.
To make Cloud Hypervisor use UEFI boot, pass the `OVMF.fd` file path as an argument to the `--kernel` option. The firmware file will be opened in read only mode.
The same firmware can be used with Cloud Hypervisor or with QEMU. This is particularly useful if using QEMU for the preparation phase.
# Links
- [OVMF wiki](https://github.com/tianocore/tianocore.github.io/wiki/OVMF)
- [Cloud Hypervisor specific tree](https://github.com/cloud-hypervisor/edk2/tree/ch)

326
docs/windows.md Normal file
View File

@@ -0,0 +1,326 @@
# Windows Support
Starting with the release version [0.10.0](https://github.com/cloud-hypervisor/cloud-hypervisor/releases/tag/v0.10.0), Cloud Hypervisor supports Windows guests.
__Requirements__
- Host with KVM enabled
- [UEFI](uefi.md) capable Windows guest image with Virtio drivers integrated
Any modern Windows Server version is compatible. Cloud Hypervisor has been successfully tested with Windows Server 2019 and Windows Server Core 2004.
At the current stage, only UEFI capable Windows images are supported. This implies the presence of the OVMF firmware during the Windows installation and in any subsequent usage. BIOS boot is not supported.
The subsequent sections will tell, in detail, how to prepare an appropriate Windows image.
## Image Preparation
### Installation using the stock Windows ISO
__Prerequisites__
- QEMU, version >=5.0.0 is recommended.
- Windows installation ISO. Obtained through MSDN, Visual Studio subscription, evaluation center, etc.
- [VirtIO driver ISO](https://fedorapeople.org/groups/virt/virtio-win/direct-downloads/stable-virtio/)
- Suitable [OVMF](uefi.md) firmware
- With the suggested image size of 30G, there should be enough free disk space to hold the installation ISO and any other necessary files
This step currently requires QEMU to install Windows onto the guest. QEMU is only used at the preparation stage, the resulting image is then fully functional with Cloud Hypervisor.
Preparing several command parts as these will be used in the follow up sections as well.
```shell
IMG_FILE=windows-disk.qcow
WIN_ISO_FILE=en_windows_server_version_2004_updated_may_2020_x64_dvd_1e7f1cfa.iso
VIRTIO_ISO_FILE=virtio-win-0.1.185.iso
OVMF_DIR=./FV
```
Create an empty image file, `qcow` or `raw` is supported.
```shell
qemu-img create -f qcow2 $IMG_FILE 30G
```
Begin the Windows installation process under QEMU
```shell
qemu-system-x86_64 \
-machine q35,accel=kvm \
-cpu host \
-m 4G \
-bios ./$OVMF_DIR/OVMF_CODE.fd \
-cdrom ./$WIN_ISO_FILE \
-drive file=./$VIRTIO_ISO_FILE,index=0,media=cdrom
-drive if=none,id=root,file=./$IMG_FILE \
-device virtio-blk-pci,drive=root,disable-legacy=on \
-device virtio-net-pci,netdev=mynet0,disable-legacy=on \
-netdev user,id=mynet0 \
-vga std
```
Before the installation can proceed, point the Windows installation program to the VirtIO disk and install the necessary storage controller drivers. After that, the attached hard drive will become visible and the actual installation can commence.
After the installation has completed, proceed further to the configuration section. QEMU will be needed at least once more to enable the Windows Special Administration Console (SAC) and to possibly install extra device drivers.
## Image Usage
The basic command to boot a Windows image. The configuration section should be checked before executing it for the first time.
```shell
cloud-hypervisor \
--kernel ./$OVMF_DIR/OVMF.fd \
--disk path=./$IMG_FILE \
--cpus boot=1,kvm_hyperv=on \
--memory size=4G \
--serial tty \
--console off \
--net tap=
```
It is necessary to always:
- Carry the OVMF firmware in the `--kernel` option
- Add `kvm_hyperv=on` to the `--cpus` option
In cases where the host processor supports address space > 39 bits, it might be necessary to limit the address space. It can be done by appending the option `max_phys_bits=X` to the `--cpus` parameter, where `X` is the number of bits to be supported. Windows was tested to support at least 39-bit address space.
To daemonize the Cloud Hypervisor process, `nohup` can be used. Some STDIO redirections might need to be done. In a simple case it is sufficient to just redirect all the output to `/dev/null`.
## Image Configuration
### Device Drivers
After the Windows installation has finished under QEMU, there might be still devices with no drivers installed. This might happen for example, when a device was not used during the installation. In particular it is important to ensure that the VirtIO network device is setup correctly because further steps for the configuration and the usage require network in most case.
Boot once more under QEMU and use the [Device Manager](https://support.microsoft.com/en-in/help/4028443/windows-10-update-drivers), to ensure all the device drivers, and especially the network card, are installed correctly. Also, as Cloud Hypervisor can introduce new devices, it is advisable to repeat the procedure while booted under Cloud Hypervisor, when the RDP access to the image is functional.
### Windows Special Administration Console (SAC) enablement
SAC provides a text based console access to the Windows guest. As Cloud Hypervisor doesn't implement a VGA adaptor, SAC is an important instrument for the Windows guest management.
Boot the Windows image under QEMU and execute the below commands to permanently enable SAC
```cmd
bcdedit /emssettings emsport:1 emsbaudrate:115200
bcdedit /ems on
bcdedit /bootems on
```
Once SAC is enabled, the image can be booted under Cloud Hypervisor. The SAC prompt will show up
<pre>
Computer is booting, SAC started and initialized.
Use the "ch -?" command for information about using channels.
Use the "?" command for general help.
SAC>
</pre>
To open a console on the guest, the command sequence below can be used
<pre>
SAC>cmd
The Command Prompt session was successfully launched.
SAC>
EVENT: A new channel has been created. Use "ch -?" for channel help.
Channel: Cmd0001
SAC>ch -si 1
</pre>
See also the [links](#Links) section for a more extended SAC documentation.
## Network
This section illustrates the Windows specific corner points for the VM network configuration. For the extended networking guide, including bridging for multiple VMs, follow [networking.md](networking.md).
### Basic Networking
As the simplest option, using `--net tap=` in the Cloud Hypervisor command line will create a `vmtapX` device on the host with the default IPv4 adress `192.168.249.1`. After SAC becomes available, the guest configuration can be set with
<pre>
SAC>i 10 192.168.249.2 255.255.255.0 192.168.249.1
</pre>
Where `10` is the device index as shown by the `i` command.
### Guest Internet Connectivity
Additional steps are necessary to provide the guest with internet access.
- On the guest, add the DNS server either by using `netsh` or by opening `Network and Connectivity Center` and editing the adapter properties.
- On the host, configure the traffic forwarding. Replace the `NET_DEV` with the name of your network device.
```shell
NET_DEV=wlp3s0
sysctl -w net.ipv4.ip_forward=1
iptables -t nat -A POSTROUTING -o $NET_DEV -j MASQUERADE
```
### Remote Desktop Protocol (RDP) enablement
#### Using QEMU
- Execute `SystemPropertiesRemote`
- In the properties window, choose "Allow remote connections to this computer"
- Click "Select Users" and add some user to the allow list
#### Using powershell
```powershell
Set-ItemProperty "HKLM:\SYSTEM\CurrentControlSet\Control\Terminal Server\" -Name "fDenyTSConnections" -Value 0
Enable-NetFirewallRule -DisplayGroup "Remote Desktop"
Add-LocalGroupMember -Group "Remote Desktop Users" -Member someuser
```
Administrators can always RDP, non administrator users have to be explicitly enabled.
Once the configuration is set, RDP clients can connect to `192.168.249.2`.
### SSH
#### Enable using powershell
```powershell
Add-WindowsCapability -Online -Name OpenSSH.Server~~~~0.0.1.0
Start-Service sshd
Set-Service -Name sshd -StartupType Automatic
```
This allows for SSH login from a remote machine, for example through the `administrator` user: `ssh administrator@192.168.249.2`. For a more detailed OpenSSH guide, please follow the MSDN article from the [links](#links) section.
## Debugging
The Windows guest debugging process relies heavily on QEMU and [socat](http://www.dest-unreach.org/socat/). The procedure requires two Windows VMs:
- A debugger VM running under QEMU.
- A debuggee, a Windows VM that has been created in the previous steps, running under Cloud Hypervisor or QEMU.
The connection between both guests happens over TCP, whereby on the guest side it is automatically translated to a COM port. Because the VMs are connected through TCP, the debugging infrastructure can be distributed over the network. The serial port, while slowly transferring data, is common enough to support a wide range of cases and tools.
In this excercise, [WinDbg](https://docs.microsoft.com/en-us/windows-hardware/drivers/debugger/) is used. Any other debugger of choice with the ability to use serial connection can be used instead.
### Debugger and Debuggee
#### WinDbg VM
For simplicity, the debugger VM is supposed to be only running under QEMU. It will require VGA and doesn't neccessarily depend on UEFI. As an OS, it can carry any supported Windows OS where the debugger of choice can be installed. The simplest way is to follow the image preparation instructions from the previous chapter, but avoid using the OVMF firmware. It is also not required to use VirtIO drivers, whereby it might be useful in some case. Though, while creating the image file for the debugger VM, be sure to choose a sufficient disk size that counts in the need to save the corresponding debug symbols and sources.
To create the debugger Windows VM, the following command can be used:
```shell
qemu-system-x86_64 \
-machine q35,accel=kvm \
-cpu host \
-smp 1 \
-m 4G \
-cdrom ./$WIN_ISO_FILE \
-drive file=./$VIRTIO_ISO_FILE,index=0,media=cdrom
-drive if=none,id=root,file=./windbg-disk.qcow \
-device virtio-blk-pci,drive=root,disable-legacy=on \
-device virtio-net-pci,netdev=mynet0,disable-legacy=on \
-netdev user,id=mynet0,net=192.168.178.0/24,host=192.168.178.1,dhcpstart=192.168.178.64,hostname=windbg-host \
-vga std
```
A non server Windows OS like Windows 10 can be used to carry the debugging tools in the debugger VM.
#### Debuggee VM
The debuggee VM is the one that we've learned to configure and run in the first section. There might be various reasons to debug. For example, there could be an issue in the Windows guest with an emulated device or an included driver. Or, we might want to develop a custom feature like a kernel driver to be available in the guest.
Note, that there are several ways to debug Windows, not all of them need to be enabled at the same time. For example, if developing a kernel module, the only useful options would be to configure for the serial debugging and enable the kernel debug. In that case, any crash or misbehavior in the boot loader or kernel would be ignored. The commands below must be run as administrator on the debuggee guest VM.
##### Turn On Serial Debugging
This will configure the debugging to be enabled and instruct to use the serial port for it.
```cmd
bcdedit /dbgsettings serial debugport:1 baudrate:115200
```
##### Turn On Kernel Debuging
```cmd
bcdedit /debug on
```
##### Turn On Boot Loader Debug
```cmd
bcdedit /bootdebug on
```
##### Turn on boot manager debug
```cmd
bcdedit /set {bootmgr} bootdebug on
```
##### Disable Recovery Screen On Boot Failure
There could be a situation, where a crash is debugged. In such cases, the guest could be left in an inconsistent state. The default Windows behavior would be to boot into the recovery screen, however in some cases it might be not desired. To make Windows ignore failures and always proceed to booting the OS, use the command below:
```cmd
bcdedit /set {default} bootstatuspolicy ignoreallfailures
```
### Debugging Process
#### Invoke the WinDbg VM
```shell
qemu-system-x86_64 \
-machine q35,accel=kvm \
-cpu host \
-smp 1 \
-m 4G \
-drive if=none,id=root,file=./windbg-disk.qcow \
-device virtio-blk-pci,drive=root,disable-legacy=on \
-serial tcp::4445,server,nowait \
-device virtio-net-pci,netdev=mynet0,disable-legacy=on \
-netdev user,id=mynet0,net=192.168.178.0/24,host=192.168.178.1,dhcpstart=192.168.178.64,hostname=windbg-host \
-vga std
```
Note, this VM has the networking enabled. It is needed, because symbols and sources might need to be fetched from a network location.
Also, notice the `-serial` parameter - that's what does the magic on exposing the serial port to the guest while connecting the debugger VM with a client VM through the network. SAC/EMS needs to be disabled in the debugger VM, as otherwise the COM device might be blocked.
#### Invoke the Debuggee VM
##### Under QEMU
Essentially it would be the command like depicted in the guest preparation sections, with a few modifications:
```shell
qemu-system-x86_64 \
-machine q35,accel=kvm \
-cpu host \
-m 4G \
-bios ./$OVMF_DIR/OVMF_CODE.fd \
-cdrom ./$WIN_ISO_FILE \
-drive file=./$VIRTIO_ISO_FILE,index=0,media=cdrom
-drive if=none,id=root,file=./$IMG_FILE \
-device virtio-blk-pci,drive=root,disable-legacy=on \
-device virtio-net-pci,netdev=mynet0,disable-legacy=on \
-netdev user,id=mynet0 \
-serial tcp:127.0.0.1:4445 \
-vga std
```
It is to see, that `-serial` parameter is used here, to establish the connection with the debugger VM.
To disable HPET, attach `--no-hpet`. To enable hypervisor reference timer, use `-cpu host,hv-time`. These and other options can be used to achieve better [Hyper-V compatibility](https://archive.fosdem.org/2019/schedule/event/vai_enlightening_kvm/attachments/slides/2860/export/events/attachments/vai_enlightening_kvm/slides/2860/vkuznets_fosdem2019_enlightening_kvm.pdf).
##### Cloud Hypervisor
The `socat` tool is used to establish the QEMU compatible behavior. Here as well, the Cloud Hypervisor command used to run the Windows guest is to be used. Put the command into a shell script:
`socat SYSTEM:"./ch-script",openpty,raw,echo=0 TCP:localhost:4445`
The reason to pack the command into the shell script is that the command might contain a comma. When using SYSTEM, the shell command can't contain `,` or `!!`.
## Links
- [Fedora VirtIO guide for Windows](https://docs.fedoraproject.org/en-US/quick-docs/creating-windows-virtual-machines-using-virtio-drivers/)
- [VirtIO driver binaries](https://fedorapeople.org/groups/virt/virtio-win/direct-downloads/stable-virtio/)
- [VirtIO driver sources](https://github.com/virtio-win/kvm-guest-drivers-windows)
- [Emergency Management Services](https://docs.microsoft.com/en-us/previous-versions/windows/it-pro/windows-server-2003/cc787940(v=ws.10))
- [OpenSSH server/client configuration](https://docs.microsoft.com/en-us/windows-server/administration/openssh/openssh_install_firstuse)
- [Windows guest debugging under KVM](https://www.linux-kvm.org/page/WindowsGuestDrivers/GuestDebugging)
- ["ENLIGHTENING" KVM](https://archive.fosdem.org/2019/schedule/event/vai_enlightening_kvm/attachments/slides/2860/export/events/attachments/vai_enlightening_kvm/slides/2860/vkuznets_fosdem2019_enlightening_kvm.pdf)

View File

@@ -10,19 +10,25 @@ edition = "2018"
cargo-fuzz = true
[dependencies]
kvm-ioctls = { git = "https://github.com/rust-vmm/kvm-ioctls", branch = "master" }
libc = "0.2.72"
libfuzzer-sys = "0.3"
qcow = { path = "../qcow" }
seccomp = { git = "https://github.com/firecracker-microvm/firecracker", tag = "v0.22.0" }
virtio-devices = { path = "../virtio-devices" }
vmm-sys-util = ">=0.3.1"
vm-virtio = { path = "../vm-virtio" }
vm-memory = "0.2.1"
vm-memory = "0.3.0"
[dependencies.cloud-hypervisor]
path = ".."
[patch.crates-io]
vm-memory = { git = "https://github.com/cloud-hypervisor/vm-memory", branch = "ch" }
[patch.'https://github.com/rust-vmm/kvm-ioctls']
kvm-ioctls = { git = "https://github.com/cloud-hypervisor/kvm-ioctls", branch = "ch" }
[patch.'https://github.com/rust-vmm/kvm-bindings']
kvm-bindings = { git = "https://github.com/cloud-hypervisor/kvm-bindings", branch = "ch", features = ["with-serde", "fam-wrappers"] }
# Prevent this from interfering with workspaces
[workspace]

View File

@@ -16,6 +16,7 @@ use virtio_devices::{Block, VirtioDevice, VirtioInterrupt, VirtioInterruptType};
use vm_memory::{Bytes, GuestAddress, GuestMemoryAtomic, GuestMemoryMmap};
use vm_virtio::Queue;
use vmm_sys_util::eventfd::EventFd;
use seccomp::SeccompAction;
const MEM_SIZE: u64 = 256 * 1024 * 1024;
const DESC_SIZE: u64 = 16; // Bytes in one virtio descriptor.
@@ -92,6 +93,7 @@ fuzz_target!(|bytes| {
false,
2,
256,
SeccompAction::Allow,
)
.unwrap();

View File

@@ -1,23 +1,25 @@
[package]
name = "hypervisor"
version = "0.1.0"
authors = ["Muminul Islam <muislam@microsoft.com>"]
authors = ["Microsoft Authors"]
edition = "2018"
license = "Apache-2.0 OR BSD-3-Clause"
[features]
kvm = []
[dependencies]
anyhow = "1.0"
arc-swap = ">=0.4.4"
thiserror = "1.0"
libc = "0.2.77"
libc = "0.2.80"
log = "0.4.11"
kvm-ioctls = { git = "https://github.com/cloud-hypervisor/kvm-ioctls", branch = "ch" }
kvm-bindings = { git = "https://github.com/cloud-hypervisor/kvm-bindings", branch = "ch", features = ["with-serde", "fam-wrappers"] }
kvm-ioctls = { git = "https://github.com/rust-vmm/kvm-ioctls", branch = "master" }
kvm-bindings = { git = "https://github.com/rust-vmm/kvm-bindings", branch = "master", features = ["with-serde", "fam-wrappers"] }
serde = {version = ">=1.0.27", features = ["rc"] }
serde_derive = ">=1.0.27"
serde_json = ">=1.0.9"
vm-memory = { version = "0.2.0", features = ["backend-mmap", "backend-atomic"] }
vm-memory = { version = "0.3.0", features = ["backend-mmap", "backend-atomic"] }
vmm-sys-util = { version = ">=0.5.0", features = ["with-serde"] }
[dependencies.linux-loader]

View File

@@ -10,6 +10,8 @@
#[cfg(target_arch = "aarch64")]
use crate::aarch64::VcpuInit;
#[cfg(target_arch = "aarch64")]
use crate::aarch64::{RegList, Register, StandardRegisters};
use crate::{CpuState, MpState};
#[cfg(target_arch = "x86_64")]
@@ -137,12 +139,12 @@ pub enum HypervisorCpuError {
/// Setting one reg error
///
#[error("Failed to init vcpu: {0}")]
SetOneReg(#[source] anyhow::Error),
SetRegister(#[source] anyhow::Error),
///
/// Getting one reg error
///
#[error("Failed to init vcpu: {0}")]
GetOneReg(#[source] anyhow::Error),
GetRegister(#[source] anyhow::Error),
///
/// Getting guest clock paused error
///
@@ -153,6 +155,31 @@ pub enum HypervisorCpuError {
///
#[error("Failed to enable HyperV SynIC")]
EnableHyperVSynIC(#[source] anyhow::Error),
///
/// Getting AArch64 core register error
///
#[error("Failed to get core register: {0}")]
GetCoreRegister(#[source] anyhow::Error),
///
/// Setting AArch64 core register error
///
#[error("Failed to set core register: {0}")]
SetCoreRegister(#[source] anyhow::Error),
///
/// Getting AArch64 registers list error
///
#[error("Failed to retrieve list of registers: {0}")]
GetRegList(#[source] anyhow::Error),
///
/// Getting AArch64 system register error
///
#[error("Failed to get system register: {0}")]
GetSysRegister(#[source] anyhow::Error),
///
/// Setting AArch64 system register error
///
#[error("Failed to set system register: {0}")]
SetSysRegister(#[source] anyhow::Error),
}
#[derive(Debug)]
@@ -299,12 +326,43 @@ pub trait Vcpu: Send + Sync {
/// Sets the value of one register for this vCPU.
///
#[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
fn set_one_reg(&self, reg_id: u64, data: u64) -> Result<()>;
fn set_reg(&self, reg_id: u64, data: u64) -> Result<()>;
///
/// Sets the value of one register for this vCPU.
///
#[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
fn get_one_reg(&self, reg_id: u64) -> Result<u64>;
fn get_reg(&self, reg_id: u64) -> Result<u64>;
///
/// Gets a list of the guest registers that are supported for the
/// KVM_GET_ONE_REG/KVM_SET_ONE_REG calls.
///
#[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
fn get_reg_list(&self, reg_list: &mut RegList) -> Result<()>;
///
/// Save the state of the core registers.
///
#[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
fn core_registers(&self, state: &mut StandardRegisters) -> Result<()>;
///
/// Restore the state of the core registers.
///
#[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
fn set_core_registers(&self, state: &StandardRegisters) -> Result<()>;
///
/// Save the state of the system registers.
///
#[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
fn system_registers(&self, state: &mut Vec<Register>) -> Result<()>;
///
/// Restore the state of the system registers.
///
#[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
fn set_system_registers(&self, state: &[Register]) -> Result<()>;
///
/// Read the MPIDR - Multiprocessor Affinity Register.
///
#[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
fn read_mpidr(&self) -> Result<u64>;
///
/// Retrieve the vCPU state.
/// This function is necessary to snapshot the VM
@@ -315,7 +373,6 @@ pub trait Vcpu: Send + Sync {
/// This function is required when restoring the VM
///
fn set_state(&self, state: &CpuState) -> Result<()>;
///
/// Triggers the running of the current virtual CPU returning an exit reason.
///

View File

@@ -22,6 +22,11 @@ pub enum HypervisorDeviceError {
///
#[error("Failed to set device attribute: {0}")]
SetDeviceAttribute(#[source] anyhow::Error),
///
/// Get device attribute error
///
#[error("Failed to get device attribute: {0}")]
GetDeviceAttribute(#[source] anyhow::Error),
}
///
@@ -37,4 +42,6 @@ pub type Result<T> = std::result::Result<T, HypervisorDeviceError>;
pub trait Device: Send + Sync + AsRawFd {
/// Set device attribute.
fn set_device_attr(&self, attr: &DeviceAttr) -> Result<()>;
/// Get device attribute.
fn get_device_attr(&self, attr: &mut DeviceAttr) -> Result<()>;
}

View File

@@ -12,16 +12,134 @@
/// Export generically-named wrappers of kvm-bindings for Unix-based platforms
///
use crate::kvm::{KvmError, KvmResult};
pub use kvm_bindings::kvm_vcpu_init as VcpuInit;
use kvm_bindings::{
kvm_mp_state, kvm_one_reg, kvm_regs, KVM_REG_ARM64, KVM_REG_ARM64_SYSREG,
KVM_REG_ARM64_SYSREG_CRM_MASK, KVM_REG_ARM64_SYSREG_CRM_SHIFT, KVM_REG_ARM64_SYSREG_CRN_MASK,
KVM_REG_ARM64_SYSREG_CRN_SHIFT, KVM_REG_ARM64_SYSREG_OP0_MASK, KVM_REG_ARM64_SYSREG_OP0_SHIFT,
KVM_REG_ARM64_SYSREG_OP1_MASK, KVM_REG_ARM64_SYSREG_OP1_SHIFT, KVM_REG_ARM64_SYSREG_OP2_MASK,
KVM_REG_ARM64_SYSREG_OP2_SHIFT, KVM_REG_ARM_COPROC_MASK, KVM_REG_ARM_CORE, KVM_REG_SIZE_MASK,
KVM_REG_SIZE_U32, KVM_REG_SIZE_U64,
};
pub use kvm_bindings::{
kvm_one_reg as Register, kvm_regs as StandardRegisters, kvm_vcpu_init as VcpuInit, RegList,
};
use serde_derive::{Deserialize, Serialize};
pub use {kvm_ioctls::Cap, kvm_ioctls::Kvm};
// Following are macros that help with getting the ID of a aarch64 core register.
// The core register are represented by the user_pt_regs structure. Look for it in
// arch/arm64/include/uapi/asm/ptrace.h.
// This macro gets the offset of a structure (i.e `str`) member (i.e `field`) without having
// an instance of that structure.
// It uses a null pointer to retrieve the offset to the field.
// Inspired by C solution: `#define offsetof(str, f) ((size_t)(&((str *)0)->f))`.
// Doing `offset__of!(user_pt_regs, pstate)` in our rust code will trigger the following:
// unsafe { &(*(0 as *const user_pt_regs)).pstate as *const _ as usize }
// The dereference expression produces an lvalue, but that lvalue is not actually read from,
// we're just doing pointer math on it, so in theory, it should safe.
#[macro_export]
macro_rules! offset__of {
($str:ty, $field:ident) => {
unsafe { &(*std::ptr::null::<$str>()).$field as *const _ as usize }
};
}
// Get the ID of a core register
#[macro_export]
macro_rules! arm64_core_reg_id {
($size: tt, $offset: tt) => {
// The core registers of an arm64 machine are represented
// in kernel by the `kvm_regs` structure. This structure is a
// mix of 32, 64 and 128 bit fields:
// struct kvm_regs {
// struct user_pt_regs regs;
//
// __u64 sp_el1;
// __u64 elr_el1;
//
// __u64 spsr[KVM_NR_SPSR];
//
// struct user_fpsimd_state fp_regs;
// };
// struct user_pt_regs {
// __u64 regs[31];
// __u64 sp;
// __u64 pc;
// __u64 pstate;
// };
// The id of a core register can be obtained like this:
// offset = id & ~(KVM_REG_ARCH_MASK | KVM_REG_SIZE_MASK | KVM_REG_ARM_CORE). Thus,
// id = KVM_REG_ARM64 | KVM_REG_SIZE_U64/KVM_REG_SIZE_U32/KVM_REG_SIZE_U128 | KVM_REG_ARM_CORE | offset
KVM_REG_ARM64 as u64
| u64::from(KVM_REG_ARM_CORE)
| $size
| (($offset / mem::size_of::<u32>()) as u64)
};
}
// This macro computes the ID of a specific ARM64 system register similar to how
// the kernel C macro does.
// https://elixir.bootlin.com/linux/v4.20.17/source/arch/arm64/include/uapi/asm/kvm.h#L203
#[macro_export]
macro_rules! arm64_sys_reg {
($name: tt, $op0: tt, $op1: tt, $crn: tt, $crm: tt, $op2: tt) => {
pub const $name: u64 = KVM_REG_ARM64 as u64
| KVM_REG_SIZE_U64 as u64
| KVM_REG_ARM64_SYSREG as u64
| ((($op0 as u64) << KVM_REG_ARM64_SYSREG_OP0_SHIFT)
& KVM_REG_ARM64_SYSREG_OP0_MASK as u64)
| ((($op1 as u64) << KVM_REG_ARM64_SYSREG_OP1_SHIFT)
& KVM_REG_ARM64_SYSREG_OP1_MASK as u64)
| ((($crn as u64) << KVM_REG_ARM64_SYSREG_CRN_SHIFT)
& KVM_REG_ARM64_SYSREG_CRN_MASK as u64)
| ((($crm as u64) << KVM_REG_ARM64_SYSREG_CRM_SHIFT)
& KVM_REG_ARM64_SYSREG_CRM_MASK as u64)
| ((($op2 as u64) << KVM_REG_ARM64_SYSREG_OP2_SHIFT)
& KVM_REG_ARM64_SYSREG_OP2_MASK as u64);
};
}
// Constant imported from the Linux kernel:
// https://elixir.bootlin.com/linux/v4.20.17/source/arch/arm64/include/asm/sysreg.h#L135
arm64_sys_reg!(MPIDR_EL1, 3, 0, 0, 0, 5);
/// Specifies whether a particular register is a system register or not.
/// The kernel splits the registers on aarch64 in core registers and system registers.
/// So, below we get the system registers by checking that they are not core registers.
///
/// # Arguments
///
/// * `regid` - The index of the register we are checking.
pub fn is_system_register(regid: u64) -> bool {
if (regid & KVM_REG_ARM_COPROC_MASK as u64) == KVM_REG_ARM_CORE as u64 {
return false;
}
let size = regid & KVM_REG_SIZE_MASK;
if size != KVM_REG_SIZE_U32 && size != KVM_REG_SIZE_U64 {
panic!("Unexpected register size for system register {}", size);
}
true
}
pub fn check_required_kvm_extensions(kvm: &Kvm) -> KvmResult<()> {
if !kvm.check_extension(Cap::SignalMsi) {
return Err(KvmError::CapabilityMissing(Cap::SignalMsi));
}
if !kvm.check_extension(Cap::OneReg) {
return Err(KvmError::CapabilityMissing(Cap::OneReg));
}
Ok(())
}
#[derive(Clone, Serialize, Deserialize)]
pub struct VcpuKvmState {}
#[derive(Clone, Default, Serialize, Deserialize)]
pub struct VcpuKvmState {
pub mp_state: kvm_mp_state,
pub core_regs: kvm_regs,
pub sys_regs: Vec<kvm_one_reg>,
// We will be using the mpidr for passing it to the VmState.
// The VmState will give this away for saving restoring the icc and redistributor
// registers.
pub mpidr: u64,
}

View File

@@ -9,15 +9,23 @@
//
#[cfg(target_arch = "aarch64")]
pub use crate::aarch64::{check_required_kvm_extensions, VcpuInit, VcpuKvmState as CpuState};
pub use crate::aarch64::{
check_required_kvm_extensions, is_system_register, VcpuInit, VcpuKvmState as CpuState,
MPIDR_EL1,
};
use crate::cpu;
use crate::device;
use crate::hypervisor;
use crate::vm;
use crate::vm::{self, VmmOps};
#[cfg(target_arch = "aarch64")]
use crate::{arm64_core_reg_id, offset__of};
use arc_swap::ArcSwapOption;
use kvm_ioctls::{NoDatamatch, VcpuFd, VmFd};
use serde_derive::{Deserialize, Serialize};
use std::os::unix::io::{AsRawFd, RawFd};
use std::result;
#[cfg(target_arch = "x86_64")]
use std::sync::atomic::{AtomicBool, Ordering};
use std::sync::Arc;
#[cfg(target_arch = "x86_64")]
use vm_memory::Address;
@@ -31,6 +39,9 @@ use x86_64::{
check_required_kvm_extensions, FpuState, SpecialRegisters, StandardRegisters, KVM_TSS_ADDRESS,
};
#[cfg(target_arch = "aarch64")]
use aarch64::{RegList, Register, StandardRegisters};
#[cfg(target_arch = "x86_64")]
pub use x86_64::{
CpuId, CpuIdEntry, ExtendedControlRegisters, LapicState, MsrEntries, VcpuKvmState as CpuState,
@@ -38,7 +49,9 @@ pub use x86_64::{
};
#[cfg(target_arch = "x86_64")]
use kvm_bindings::{kvm_enable_cap, MsrList, KVM_CAP_HYPERV_SYNIC, KVM_CAP_SPLIT_IRQCHIP};
use kvm_bindings::{
kvm_enable_cap, kvm_msr_entry, MsrList, KVM_CAP_HYPERV_SYNIC, KVM_CAP_SPLIT_IRQCHIP,
};
#[cfg(target_arch = "x86_64")]
use crate::arch::x86::NUM_IOAPIC_PINS;
@@ -46,6 +59,13 @@ use crate::arch::x86::NUM_IOAPIC_PINS;
// aarch64 dependencies
#[cfg(target_arch = "aarch64")]
pub mod aarch64;
#[cfg(target_arch = "aarch64")]
use kvm_bindings::{
kvm_regs, user_fpsimd_state, user_pt_regs, KVM_NR_SPSR, KVM_REG_ARM64, KVM_REG_ARM_CORE,
KVM_REG_SIZE_U128, KVM_REG_SIZE_U32, KVM_REG_SIZE_U64,
};
#[cfg(target_arch = "aarch64")]
use std::mem;
pub use kvm_bindings;
pub use kvm_bindings::{
@@ -76,6 +96,7 @@ pub struct KvmVm {
#[cfg(target_arch = "x86_64")]
msrs: MsrEntries,
state: KvmVmState,
vmmops: ArcSwapOption<Box<dyn vm::VmmOps>>,
}
// Returns a `Vec<T>` with a size in bytes at least as large as `size_in_bytes`.
@@ -164,6 +185,9 @@ impl vm::Vm for KvmVm {
fd: vc,
#[cfg(target_arch = "x86_64")]
msrs: self.msrs.clone(),
vmmops: self.vmmops.clone(),
#[cfg(target_arch = "x86_64")]
hyperv_synic: AtomicBool::new(false),
};
Ok(Arc::new(vcpu))
}
@@ -327,7 +351,15 @@ impl vm::Vm for KvmVm {
///
/// Set the VM state
///
fn set_state(&self, _state: &VmState) -> vm::Result<()> {
fn set_state(&self, _state: VmState) -> vm::Result<()> {
Ok(())
}
///
/// Set the VmmOps interface
///
fn set_vmmops(&self, vmmops: Box<dyn VmmOps>) -> vm::Result<()> {
self.vmmops.store(Some(Arc::new(vmmops)));
Ok(())
}
}
@@ -407,6 +439,7 @@ impl hypervisor::Hypervisor for KvmHypervisor {
fd: vm_fd,
msrs,
state: VmState {},
vmmops: ArcSwapOption::from(None),
}))
}
@@ -415,6 +448,7 @@ impl hypervisor::Hypervisor for KvmHypervisor {
Ok(Arc::new(KvmVm {
fd: vm_fd,
state: VmState {},
vmmops: ArcSwapOption::from(None),
}))
}
}
@@ -475,6 +509,9 @@ pub struct KvmVcpu {
fd: VcpuFd,
#[cfg(target_arch = "x86_64")]
msrs: MsrEntries,
vmmops: ArcSwapOption<Box<dyn vm::VmmOps>>,
#[cfg(target_arch = "x86_64")]
hyperv_synic: AtomicBool,
}
/// Implementation of Vcpu trait for KVM
/// Example:
@@ -555,6 +592,10 @@ impl cpu::Vcpu for KvmVcpu {
/// X86 specific call to enable HyperV SynIC
///
fn enable_hyperv_synic(&self) -> cpu::Result<()> {
// Update the information about Hyper-V SynIC being enabled and
// emulated as it will influence later which MSRs should be saved.
self.hyperv_synic.store(true, Ordering::SeqCst);
let mut cap: kvm_enable_cap = Default::default();
cap.cap = KVM_CAP_HYPERV_SYNIC;
self.fd
@@ -666,9 +707,27 @@ impl cpu::Vcpu for KvmVcpu {
match self.fd.run() {
Ok(run) => match run {
#[cfg(target_arch = "x86_64")]
VcpuExit::IoIn(addr, data) => Ok(cpu::VmExit::IoIn(addr, data)),
VcpuExit::IoIn(addr, data) => {
if let Some(vmmops) = self.vmmops.load_full() {
return vmmops
.pio_read(addr.into(), data)
.map(|_| cpu::VmExit::Ignore)
.map_err(|e| cpu::HypervisorCpuError::RunVcpu(e.into()));
}
Ok(cpu::VmExit::IoIn(addr, data))
}
#[cfg(target_arch = "x86_64")]
VcpuExit::IoOut(addr, data) => Ok(cpu::VmExit::IoOut(addr, data)),
VcpuExit::IoOut(addr, data) => {
if let Some(vmmops) = self.vmmops.load_full() {
return vmmops
.pio_write(addr.into(), data)
.map(|_| cpu::VmExit::Ignore)
.map_err(|e| cpu::HypervisorCpuError::RunVcpu(e.into()));
}
Ok(cpu::VmExit::IoOut(addr, data))
}
#[cfg(target_arch = "x86_64")]
VcpuExit::IoapicEoi(vector) => Ok(cpu::VmExit::IoapicEoi(vector)),
#[cfg(target_arch = "x86_64")]
@@ -676,10 +735,12 @@ impl cpu::Vcpu for KvmVcpu {
#[cfg(target_arch = "aarch64")]
VcpuExit::SystemEvent(event_type, flags) => {
use kvm_bindings::KVM_SYSTEM_EVENT_SHUTDOWN;
use kvm_bindings::{KVM_SYSTEM_EVENT_RESET, KVM_SYSTEM_EVENT_SHUTDOWN};
// On Aarch64, when the VM is shutdown, run() returns
// VcpuExit::SystemEvent with reason KVM_SYSTEM_EVENT_SHUTDOWN
if event_type == KVM_SYSTEM_EVENT_SHUTDOWN {
if event_type == KVM_SYSTEM_EVENT_SHUTDOWN
|| event_type == KVM_SYSTEM_EVENT_RESET
{
Ok(cpu::VmExit::Reset)
} else {
Err(cpu::HypervisorCpuError::RunVcpu(anyhow!(
@@ -690,8 +751,26 @@ impl cpu::Vcpu for KvmVcpu {
}
}
VcpuExit::MmioRead(addr, data) => Ok(cpu::VmExit::MmioRead(addr, data)),
VcpuExit::MmioWrite(addr, data) => Ok(cpu::VmExit::MmioWrite(addr, data)),
VcpuExit::MmioRead(addr, data) => {
if let Some(vmmops) = self.vmmops.load_full() {
return vmmops
.mmio_read(addr, data)
.map(|_| cpu::VmExit::Ignore)
.map_err(|e| cpu::HypervisorCpuError::RunVcpu(e.into()));
}
Ok(cpu::VmExit::MmioRead(addr, data))
}
VcpuExit::MmioWrite(addr, data) => {
if let Some(vmmops) = self.vmmops.load_full() {
return vmmops
.mmio_write(addr, data)
.map(|_| cpu::VmExit::Ignore)
.map_err(|e| cpu::HypervisorCpuError::RunVcpu(e.into()));
}
Ok(cpu::VmExit::MmioWrite(addr, data))
}
VcpuExit::Hyperv => Ok(cpu::VmExit::Hyperv),
r => Err(cpu::HypervisorCpuError::RunVcpu(anyhow!(
@@ -749,19 +828,263 @@ impl cpu::Vcpu for KvmVcpu {
/// Sets the value of one register for this vCPU.
///
#[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
fn set_one_reg(&self, reg_id: u64, data: u64) -> cpu::Result<()> {
fn set_reg(&self, reg_id: u64, data: u64) -> cpu::Result<()> {
self.fd
.set_one_reg(reg_id, data)
.map_err(|e| cpu::HypervisorCpuError::SetOneReg(e.into()))
.map_err(|e| cpu::HypervisorCpuError::SetRegister(e.into()))
}
///
/// Gets the value of one register for this vCPU.
///
#[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
fn get_one_reg(&self, reg_id: u64) -> cpu::Result<u64> {
fn get_reg(&self, reg_id: u64) -> cpu::Result<u64> {
self.fd
.get_one_reg(reg_id)
.map_err(|e| cpu::HypervisorCpuError::GetOneReg(e.into()))
.map_err(|e| cpu::HypervisorCpuError::GetRegister(e.into()))
}
///
/// Gets a list of the guest registers that are supported for the
/// KVM_GET_ONE_REG/KVM_SET_ONE_REG calls.
///
#[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
fn get_reg_list(&self, reg_list: &mut RegList) -> cpu::Result<()> {
self.fd
.get_reg_list(reg_list)
.map_err(|e| cpu::HypervisorCpuError::GetRegList(e.into()))
}
///
/// Save the state of the core registers.
///
#[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
fn core_registers(&self, state: &mut StandardRegisters) -> cpu::Result<()> {
let mut off = offset__of!(user_pt_regs, regs);
// There are 31 user_pt_regs:
// https://elixir.free-electrons.com/linux/v4.14.174/source/arch/arm64/include/uapi/asm/ptrace.h#L72
// These actually are the general-purpose registers of the Armv8-a
// architecture (i.e x0-x30 if used as a 64bit register or w0-30 when used as a 32bit register).
for i in 0..31 {
state.regs.regs[i] = self
.fd
.get_one_reg(arm64_core_reg_id!(KVM_REG_SIZE_U64, off))
.map_err(|e| cpu::HypervisorCpuError::GetCoreRegister(e.into()))?;
off += std::mem::size_of::<u64>();
}
// We are now entering the "Other register" section of the ARMv8-a architecture.
// First one, stack pointer.
let off = offset__of!(user_pt_regs, sp);
state.regs.sp = self
.fd
.get_one_reg(arm64_core_reg_id!(KVM_REG_SIZE_U64, off))
.map_err(|e| cpu::HypervisorCpuError::GetCoreRegister(e.into()))?;
// Second one, the program counter.
let off = offset__of!(user_pt_regs, pc);
state.regs.pc = self
.fd
.get_one_reg(arm64_core_reg_id!(KVM_REG_SIZE_U64, off))
.map_err(|e| cpu::HypervisorCpuError::GetCoreRegister(e.into()))?;
// Next is the processor state.
let off = offset__of!(user_pt_regs, pstate);
state.regs.pstate = self
.fd
.get_one_reg(arm64_core_reg_id!(KVM_REG_SIZE_U64, off))
.map_err(|e| cpu::HypervisorCpuError::GetCoreRegister(e.into()))?;
// The stack pointer associated with EL1
let off = offset__of!(kvm_regs, sp_el1);
state.sp_el1 = self
.fd
.get_one_reg(arm64_core_reg_id!(KVM_REG_SIZE_U64, off))
.map_err(|e| cpu::HypervisorCpuError::GetCoreRegister(e.into()))?;
// Exception Link Register for EL1, when taking an exception to EL1, this register
// holds the address to which to return afterwards.
let off = offset__of!(kvm_regs, elr_el1);
state.elr_el1 = self
.fd
.get_one_reg(arm64_core_reg_id!(KVM_REG_SIZE_U64, off))
.map_err(|e| cpu::HypervisorCpuError::GetCoreRegister(e.into()))?;
// Saved Program Status Registers, there are 5 of them used in the kernel.
let mut off = offset__of!(kvm_regs, spsr);
for i in 0..KVM_NR_SPSR as usize {
state.spsr[i] = self
.fd
.get_one_reg(arm64_core_reg_id!(KVM_REG_SIZE_U64, off))
.map_err(|e| cpu::HypervisorCpuError::GetCoreRegister(e.into()))?;
off += std::mem::size_of::<u64>();
}
// Now moving on to floting point registers which are stored in the user_fpsimd_state in the kernel:
// https://elixir.free-electrons.com/linux/v4.9.62/source/arch/arm64/include/uapi/asm/kvm.h#L53
let mut off = offset__of!(kvm_regs, fp_regs) + offset__of!(user_fpsimd_state, vregs);
for i in 0..32 {
state.fp_regs.vregs[i][0] = self
.fd
.get_one_reg(arm64_core_reg_id!(KVM_REG_SIZE_U128, off))
.map_err(|e| cpu::HypervisorCpuError::GetCoreRegister(e.into()))?;
off += mem::size_of::<u128>();
}
// Floating-point Status Register
let off = offset__of!(kvm_regs, fp_regs) + offset__of!(user_fpsimd_state, fpsr);
state.fp_regs.fpsr = self
.fd
.get_one_reg(arm64_core_reg_id!(KVM_REG_SIZE_U32, off))
.map_err(|e| cpu::HypervisorCpuError::GetCoreRegister(e.into()))?
as u32;
// Floating-point Control Register
let off = offset__of!(kvm_regs, fp_regs) + offset__of!(user_fpsimd_state, fpcr);
state.fp_regs.fpcr = self
.fd
.get_one_reg(arm64_core_reg_id!(KVM_REG_SIZE_U32, off))
.map_err(|e| cpu::HypervisorCpuError::GetCoreRegister(e.into()))?
as u32;
Ok(())
}
///
/// Restore the state of the core registers.
///
#[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
fn set_core_registers(&self, state: &StandardRegisters) -> cpu::Result<()> {
// The function follows the exact identical order from `state`. Look there
// for some additional info on registers.
let mut off = offset__of!(user_pt_regs, regs);
for i in 0..31 {
self.fd
.set_one_reg(
arm64_core_reg_id!(KVM_REG_SIZE_U64, off),
state.regs.regs[i],
)
.map_err(|e| cpu::HypervisorCpuError::SetCoreRegister(e.into()))?;
off += std::mem::size_of::<u64>();
}
let off = offset__of!(user_pt_regs, sp);
self.fd
.set_one_reg(arm64_core_reg_id!(KVM_REG_SIZE_U64, off), state.regs.sp)
.map_err(|e| cpu::HypervisorCpuError::SetCoreRegister(e.into()))?;
let off = offset__of!(user_pt_regs, pc);
self.fd
.set_one_reg(arm64_core_reg_id!(KVM_REG_SIZE_U64, off), state.regs.pc)
.map_err(|e| cpu::HypervisorCpuError::SetCoreRegister(e.into()))?;
let off = offset__of!(user_pt_regs, pstate);
self.fd
.set_one_reg(arm64_core_reg_id!(KVM_REG_SIZE_U64, off), state.regs.pstate)
.map_err(|e| cpu::HypervisorCpuError::SetCoreRegister(e.into()))?;
let off = offset__of!(kvm_regs, sp_el1);
self.fd
.set_one_reg(arm64_core_reg_id!(KVM_REG_SIZE_U64, off), state.sp_el1)
.map_err(|e| cpu::HypervisorCpuError::SetCoreRegister(e.into()))?;
let off = offset__of!(kvm_regs, elr_el1);
self.fd
.set_one_reg(arm64_core_reg_id!(KVM_REG_SIZE_U64, off), state.elr_el1)
.map_err(|e| cpu::HypervisorCpuError::SetCoreRegister(e.into()))?;
let mut off = offset__of!(kvm_regs, spsr);
for i in 0..KVM_NR_SPSR as usize {
self.fd
.set_one_reg(arm64_core_reg_id!(KVM_REG_SIZE_U64, off), state.spsr[i])
.map_err(|e| cpu::HypervisorCpuError::SetCoreRegister(e.into()))?;
off += std::mem::size_of::<u64>();
}
let mut off = offset__of!(kvm_regs, fp_regs) + offset__of!(user_fpsimd_state, vregs);
for i in 0..32 {
self.fd
.set_one_reg(
arm64_core_reg_id!(KVM_REG_SIZE_U128, off),
state.fp_regs.vregs[i][0],
)
.map_err(|e| cpu::HypervisorCpuError::SetCoreRegister(e.into()))?;
off += mem::size_of::<u128>();
}
let off = offset__of!(kvm_regs, fp_regs) + offset__of!(user_fpsimd_state, fpsr);
self.fd
.set_one_reg(
arm64_core_reg_id!(KVM_REG_SIZE_U32, off),
state.fp_regs.fpsr as u64,
)
.map_err(|e| cpu::HypervisorCpuError::SetCoreRegister(e.into()))?;
let off = offset__of!(kvm_regs, fp_regs) + offset__of!(user_fpsimd_state, fpcr);
self.fd
.set_one_reg(
arm64_core_reg_id!(KVM_REG_SIZE_U32, off),
state.fp_regs.fpcr as u64,
)
.map_err(|e| cpu::HypervisorCpuError::SetCoreRegister(e.into()))?;
Ok(())
}
///
/// Save the state of the system registers.
///
#[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
fn system_registers(&self, state: &mut Vec<Register>) -> cpu::Result<()> {
// Call KVM_GET_REG_LIST to get all registers available to the guest. For ArmV8 there are
// around 500 registers.
let mut reg_list = RegList::new(512);
self.fd
.get_reg_list(&mut reg_list)
.map_err(|e| cpu::HypervisorCpuError::GetRegList(e.into()))?;
// At this point reg_list should contain: core registers and system registers.
// The register list contains the number of registers and their ids. We will be needing to
// call KVM_GET_ONE_REG on each id in order to save all of them. We carve out from the list
// the core registers which are represented in the kernel by kvm_regs structure and for which
// we can calculate the id based on the offset in the structure.
reg_list.retain(|regid| *regid != 0);
reg_list.as_slice().to_vec().sort_unstable();
reg_list.retain(|regid| is_system_register(*regid));
// Now, for the rest of the registers left in the previously fetched register list, we are
// simply calling KVM_GET_ONE_REG.
let indices = reg_list.as_slice();
for (_pos, index) in indices.iter().enumerate() {
if _pos > 230 {
break;
}
state.push(kvm_bindings::kvm_one_reg {
id: *index,
addr: self
.fd
.get_one_reg(*index)
.map_err(|e| cpu::HypervisorCpuError::GetSysRegister(e.into()))?,
});
}
Ok(())
}
///
/// Restore the state of the system registers.
///
#[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
fn set_system_registers(&self, state: &[Register]) -> cpu::Result<()> {
for reg in state {
self.fd
.set_one_reg(reg.id, reg.addr)
.map_err(|e| cpu::HypervisorCpuError::SetSysRegister(e.into()))?;
}
Ok(())
}
///
/// Read the MPIDR - Multiprocessor Affinity Register.
///
#[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
fn read_mpidr(&self) -> cpu::Result<u64> {
self.fd
.get_one_reg(MPIDR_EL1)
.map_err(|e| cpu::HypervisorCpuError::GetSysRegister(e.into()))
}
#[cfg(target_arch = "x86_64")]
///
@@ -802,6 +1125,7 @@ impl cpu::Vcpu for KvmVcpu {
/// let state = vcpu.state().unwrap();
/// ```
fn state(&self) -> cpu::Result<CpuState> {
let cpuid = self.get_cpuid2(kvm_bindings::KVM_MAX_CPUID_ENTRIES)?;
let mp_state = self.get_mp_state()?;
let regs = self.get_regs()?;
let sregs = self.get_sregs()?;
@@ -816,6 +1140,26 @@ impl cpu::Vcpu for KvmVcpu {
// by chunks. This is the only way to make sure we try to get as many
// MSRs as possible, even if some MSRs are not supported.
let mut msr_entries = self.msrs.clone();
// Save extra MSRs if the Hyper-V synthetic interrupt controller is
// emulated.
if self.hyperv_synic.load(Ordering::SeqCst) {
let hyperv_synic_msrs = vec![
0x40000020, 0x40000021, 0x40000080, 0x40000081, 0x40000082, 0x40000083, 0x40000084,
0x40000090, 0x40000091, 0x40000092, 0x40000093, 0x40000094, 0x40000095, 0x40000096,
0x40000097, 0x40000098, 0x40000099, 0x4000009a, 0x4000009b, 0x4000009c, 0x4000009d,
0x4000009f, 0x400000b0, 0x400000b1, 0x400000b2, 0x400000b3, 0x400000b4, 0x400000b5,
0x400000b6, 0x400000b7,
];
for index in hyperv_synic_msrs {
let msr = kvm_msr_entry {
index,
..Default::default()
};
msr_entries.push(msr).unwrap();
}
}
let expected_num_msrs = msr_entries.as_fam_struct_ref().nmsrs as usize;
let num_msrs = self.get_msrs(&mut msr_entries)?;
let msrs = if num_msrs != expected_num_msrs {
@@ -861,6 +1205,7 @@ impl cpu::Vcpu for KvmVcpu {
let vcpu_events = self.get_vcpu_events()?;
Ok(CpuState {
cpuid,
msrs,
vcpu_events,
regs,
@@ -872,9 +1217,19 @@ impl cpu::Vcpu for KvmVcpu {
mp_state,
})
}
///
/// Get the current AArch64 CPU state
///
#[cfg(target_arch = "aarch64")]
fn state(&self) -> cpu::Result<CpuState> {
unimplemented!();
let mut state = CpuState::default();
// Get this vCPUs multiprocessing state.
state.mp_state = self.get_mp_state()?;
self.core_registers(&mut state.core_regs)?;
self.system_registers(&mut state.sys_regs)?;
state.mpidr = self.read_mpidr()?;
Ok(state)
}
#[cfg(target_arch = "x86_64")]
///
@@ -917,6 +1272,7 @@ impl cpu::Vcpu for KvmVcpu {
/// vcpu.set_state(&state).unwrap();
/// ```
fn set_state(&self, state: &CpuState) -> cpu::Result<()> {
self.set_cpuid2(&state.cpuid)?;
self.set_mp_state(state.mp_state)?;
self.set_regs(&state.regs)?;
self.set_sregs(&state.sregs)?;
@@ -958,10 +1314,15 @@ impl cpu::Vcpu for KvmVcpu {
Ok(())
}
#[allow(unused_variables)]
///
/// Restore the previously saved AArch64 CPU state
///
#[cfg(target_arch = "aarch64")]
fn set_state(&self, state: &CpuState) -> cpu::Result<()> {
warn!("CPU state was not restored");
self.set_core_registers(&state.core_regs)?;
self.set_system_registers(&state.sys_regs)?;
self.set_mp_state(state.mp_state)?;
Ok(())
}
}
@@ -980,6 +1341,14 @@ impl device::Device for KvmDevice {
.set_device_attr(attr)
.map_err(|e| device::HypervisorDeviceError::SetDeviceAttribute(e.into()))
}
///
/// Get device attribute
///
fn get_device_attr(&self, attr: &mut DeviceAttr) -> device::Result<()> {
self.fd
.get_device_attr(attr)
.map_err(|e| device::HypervisorDeviceError::GetDeviceAttribute(e.into()))
}
}
impl AsRawFd for KvmDevice {

View File

@@ -85,6 +85,7 @@ pub fn check_required_kvm_extensions(kvm: &Kvm) -> KvmResult<()> {
}
#[derive(Clone, Serialize, Deserialize)]
pub struct VcpuKvmState {
pub cpuid: CpuId,
pub msrs: MsrEntries,
pub vcpu_events: VcpuEvents,
pub regs: StandardRegisters,

View File

@@ -20,6 +20,7 @@
#[macro_use]
extern crate anyhow;
#[cfg(target_arch = "x86_64")]
#[macro_use]
extern crate log;
extern crate serde;

View File

@@ -117,6 +117,36 @@ pub enum HypervisorVmError {
///
#[error("Failed to create passthrough device: {0}")]
CreatePassthroughDevice(#[source] anyhow::Error),
///
/// Write to Guest memory
///
#[error("Failed to write to guest memory: {0}")]
GuestMemWrite(#[source] anyhow::Error),
///
/// Read Guest memory
///
#[error("Failed to read guest memory: {0}")]
GuestMemRead(#[source] anyhow::Error),
///
/// Read from MMIO Bus
///
#[error("Failed to read from MMIO Bus: {0}")]
MmioBusRead(#[source] anyhow::Error),
///
/// Write to MMIO Bus
///
#[error("Failed to write to MMIO Bus: {0}")]
MmioBusWrite(#[source] anyhow::Error),
///
/// Read from IO Bus
///
#[error("Failed to read from IO Bus: {0}")]
IoBusRead(#[source] anyhow::Error),
///
/// Write to IO Bus
///
#[error("Failed to write to IO Bus: {0}")]
IoBusWrite(#[source] anyhow::Error),
}
///
/// Result type for returning from a function
@@ -183,5 +213,18 @@ pub trait Vm: Send + Sync {
/// Get the Vm state. Return VM specific data
fn state(&self) -> Result<VmState>;
/// Set the VM state
fn set_state(&self, state: &VmState) -> Result<()>;
fn set_state(&self, state: VmState) -> Result<()>;
/// Set VmmOps interface
fn set_vmmops(&self, vmmops: Box<dyn VmmOps>) -> Result<()>;
}
pub trait VmmOps: Send + Sync {
fn guest_mem_write(&self, buf: &[u8], gpa: u64) -> Result<usize>;
fn guest_mem_read(&self, buf: &mut [u8], gpa: u64) -> Result<usize>;
fn mmio_read(&self, addr: u64, data: &mut [u8]) -> Result<()>;
fn mmio_write(&self, addr: u64, data: &[u8]) -> Result<()>;
#[cfg(target_arch = "x86_64")]
fn pio_read(&self, addr: u64, data: &mut [u8]) -> Result<()>;
#[cfg(target_arch = "x86_64")]
fn pio_write(&self, addr: u64, data: &[u8]) -> Result<()>;
}

View File

@@ -5,17 +5,17 @@ authors = ["The Chromium OS Authors"]
[dependencies]
epoll = ">=4.0.1"
libc = "0.2.77"
libc = "0.2.80"
log = "0.4.11"
net_gen = { path = "../net_gen" }
rand = "0.7.3"
serde = "1.0.116"
serde = "1.0.117"
virtio-bindings = "0.1.0"
vm-memory = { version = "0.2.1", features = ["backend-mmap", "backend-atomic"] }
vm-memory = { version = "0.3.0", features = ["backend-mmap", "backend-atomic"] }
vm-virtio = { path = "../vm-virtio" }
vmm-sys-util = ">=0.3.1"
[dev-dependencies]
lazy_static = "1.3.0"
pnet = "0.26.0"
serde_json = "1.0.57"
serde_json = "1.0.59"

View File

@@ -189,7 +189,7 @@ impl FromStr for IntegerList {
let items: Vec<&str> = range.split('-').collect();
if items.len() > 2 {
return Err(IntegerListParseError::InvalidValue(range.to_string()));
return Err(IntegerListParseError::InvalidValue((*range).to_string()));
}
let start_range = items[0]
@@ -203,7 +203,7 @@ impl FromStr for IntegerList {
.parse::<u64>()
.map_err(|_| IntegerListParseError::InvalidValue(items[1].to_owned()))?;
if start_range >= end_range {
return Err(IntegerListParseError::InvalidValue(range.to_string()));
return Err(IntegerListParseError::InvalidValue((*range).to_string()));
}
for i in start_range..end_range {
@@ -233,7 +233,9 @@ impl FromStr for TupleTwoIntegers {
let items: Vec<&str> = tuple.split('@').collect();
if items.len() != 2 {
return Err(TupleTwoIntegersParseError::InvalidValue(tuple.to_string()));
return Err(TupleTwoIntegersParseError::InvalidValue(
(*tuple).to_string(),
));
}
let item1 = items[0]

View File

@@ -10,14 +10,14 @@ byteorder = "1.3.4"
hypervisor = { path = "../hypervisor" }
vfio-ioctls = { git = "https://github.com/cloud-hypervisor/vfio-ioctls", branch = "ch" }
vmm-sys-util = ">=0.3.1"
libc = "0.2.77"
libc = "0.2.80"
log = "0.4.11"
serde = {version = ">=1.0.27", features = ["rc"] }
serde_derive = ">=1.0.27"
serde_json = ">=1.0.9"
vm-allocator = { path = "../vm-allocator" }
vm-device = { path = "../vm-device" }
vm-memory = "0.2.1"
vm-memory = "0.3.0"
vm-migration = { path = "../vm-migration" }
[dependencies.vfio-bindings]

View File

@@ -64,7 +64,7 @@ impl PciClassCode {
}
}
/// A PCI sublcass. Each class in `PciClassCode` can specify a unique set of subclasses. This trait
/// A PCI subclass. Each class in `PciClassCode` can specify a unique set of subclasses. This trait
/// is implemented by each subclass. It allows use of a trait object to generate configurations.
pub trait PciSubclass {
/// Convert this subclass to the value used in the PCI specification.

View File

@@ -56,7 +56,7 @@ impl Default for MsixTableEntry {
msg_addr_lo: 0,
msg_addr_hi: 0,
msg_data: 0,
vector_ctl: 0,
vector_ctl: 0x1,
}
}
}
@@ -97,7 +97,7 @@ impl MsixConfig {
pba_entries,
devid,
interrupt_source_group,
masked: false,
masked: true,
enabled: false,
}
}
@@ -185,12 +185,6 @@ impl MsixConfig {
error!("Failed unmasking vector: {:?}", e);
}
}
if !old_enabled || old_masked {
if let Err(e) = self.interrupt_source_group.enable() {
error!("Failed enabling irq_fd: {:?}", e);
}
}
} else if old_enabled || !old_masked {
if let Err(e) = self.interrupt_source_group.disable() {
error!("Failed disabling irq_fd: {:?}", e);

View File

@@ -714,20 +714,17 @@ impl PciDevice for VfioPciDevice {
// Is this an IO BAR?
let io_bar = if bar_id != VFIO_PCI_ROM_REGION_INDEX {
match lsb_flag & PCI_CONFIG_IO_BAR {
PCI_CONFIG_IO_BAR => true,
_ => false,
}
matches!(lsb_flag & PCI_CONFIG_IO_BAR, PCI_CONFIG_IO_BAR)
} else {
false
};
// Is this a 64-bit BAR?
let is_64bit_bar = if bar_id != VFIO_PCI_ROM_REGION_INDEX {
match lsb_flag & PCI_CONFIG_MEMORY_BAR_64BIT {
PCI_CONFIG_MEMORY_BAR_64BIT => true,
_ => false,
}
matches!(
lsb_flag & PCI_CONFIG_MEMORY_BAR_64BIT,
PCI_CONFIG_MEMORY_BAR_64BIT
)
} else {
false
};
@@ -785,7 +782,7 @@ impl PciDevice for VfioPciDevice {
// We need to allocate a guest MMIO address range for that BAR.
// In case the BAR is mappable directly, this means it might be
// set as user memory region, which expects to deal with 4K
// pages. Therefore, the aligment has to be set accordingly.
// pages. Therefore, the alignment has to be set accordingly.
let bar_alignment = if (bar_id == VFIO_PCI_ROM_REGION_INDEX)
|| (self.device.get_region_flags(bar_id) & VFIO_REGION_INFO_FLAG_MMAP != 0)
{

View File

@@ -10,7 +10,7 @@ path = "src/qcow.rs"
[dependencies]
byteorder = "1.3.4"
libc = "0.2.77"
libc = "0.2.80"
log = "0.4.11"
remain = "0.2.2"
vmm-sys-util = ">=0.3.1"

View File

@@ -769,11 +769,14 @@ impl QcowFile {
let mut ref_table = vec![0; refcount_table_entries as usize];
let mut first_free_cluster: u64 = 0;
for refblock_addr in &mut ref_table {
while refcounts[first_free_cluster as usize] != 0 {
first_free_cluster += 1;
loop {
if first_free_cluster >= refcounts.len() as u64 {
return Err(Error::NotEnoughSpaceForRefcounts);
}
if refcounts[first_free_cluster as usize] == 0 {
break;
}
first_free_cluster += 1;
}
*refblock_addr = first_free_cluster * cluster_size;
@@ -1056,7 +1059,7 @@ impl QcowFile {
// Free the previously used cluster if one exists. Modified tables are always
// witten to new clusters so the L1 table can be committed to disk after they
// are and L1 never points at an invalid table.
// The index must be valid from when it was insterted.
// The index must be valid from when it was inserted.
let addr = self.l1_table[l1_index];
if addr != 0 {
self.unref_clusters.push(addr);
@@ -1354,7 +1357,7 @@ impl QcowFile {
fn sync_caches(&mut self) -> std::io::Result<()> {
// Write out all dirty L2 tables.
for (l1_index, l2_table) in self.l2_cache.iter_mut().filter(|(_k, v)| v.dirty()) {
// The index must be valid from when we insterted it.
// The index must be valid from when we inserted it.
let addr = self.l1_table[*l1_index];
if addr != 0 {
self.raw_file.write_pointer_table(

View File

@@ -1,10 +1,23 @@
- [v0.11.0](#v0110)
- [`io_uring` support by default for `virtio-block`](#io_uring-support-by-default-for-virtio-block)
- [Windows Guest Support](#windows-guest-support)
- [`vhost-user` "Self Spawning" Deprecation](#vhost-user-self-spawning-deprecation)
- [`virtio-mmmio` Removal](#virtio-mmmio-removal)
- [Snapshot/Restore support for ARM64](#snapshotrestore-support-for-arm64)
- [Improved Linux Boot Time](#improved-linux-boot-time)
- [`SIGTERM/SIGINT` Interrupt Signal Handling](#sigtermsigint-interrupt-signal-handling)
- [Default Log Level Changed](#default-log-level-changed)
- [New `--balloon` Parameter Added](#new---balloon-parameter-added)
- [Experimental `virtio-watchdog` Support](#experimental-virtio-watchdog-support)
- [Notable Bug Fixes](#notable-bug-fixes)
- [Contributors](#contributors)
- [v0.10.0](#v0100)
- [`virtio-block` Support for Multiple Descriptors](#virtio-block-support-for-multiple-descriptors)
- [Memory Zones](#memory-zones)
- [`Seccomp` Sandbox Improvements](#seccomp-sandbox-improvements)
- [Preliminary KVM HyperV Emulation Control](#preliminary-kvm-hyperv-emulation-control)
- [Notable Bug Fixes](#notable-bug-fixes)
- [Contributors](#contributors)
- [Notable Bug Fixes](#notable-bug-fixes-1)
- [Contributors](#contributors-1)
- [v0.9.0](#v090)
- [`io_uring` Based Block Device Support](#io_uring-based-block-device-support)
- [Block and Network Device Statistics](#block-and-network-device-statistics)
@@ -17,17 +30,17 @@
- [Enhancements to ARM64 Support](#enhancements-to-arm64-support)
- [Intel SGX Support](#intel-sgx-support)
- [`Seccomp` Sandbox Improvements](#seccomp-sandbox-improvements-1)
- [Notable Bug Fixes](#notable-bug-fixes-1)
- [Contributors](#contributors-1)
- [Notable Bug Fixes](#notable-bug-fixes-2)
- [Contributors](#contributors-2)
- [v0.8.0](#v080)
- [Experimental Snapshot and Restore Support](#experimental-snapshot-and-restore-support)
- [Experimental ARM64 Support](#experimental-arm64-support)
- [Support for Using 5-level Paging in Guests](#support-for-using-5-level-paging-in-guests)
- [Virtio Device Interrupt Suppression for Network Devices](#virtio-device-interrupt-suppression-for-network-devices)
- [`vhost_user_fs` Improvements](#vhost_user_fs-improvements)
- [Notable Bug Fixes](#notable-bug-fixes-2)
- [Notable Bug Fixes](#notable-bug-fixes-3)
- [Command Line and API Changes](#command-line-and-api-changes)
- [Contributors](#contributors-2)
- [Contributors](#contributors-3)
- [v0.7.0](#v070)
- [Block, Network, Persistent Memory (PMEM), VirtioFS and Vsock hotplug](#block-network-persistent-memory-pmem-virtiofs-and-vsock-hotplug)
- [Alternative `libc` Support](#alternative-libc-support)
@@ -37,14 +50,14 @@
- [`Seccomp` Sandboxing](#seccomp-sandboxing)
- [Updated Distribution Support](#updated-distribution-support)
- [Command Line and API Changes](#command-line-and-api-changes-1)
- [Contributors](#contributors-3)
- [Contributors](#contributors-4)
- [v0.6.0](#v060)
- [Directly Assigned Devices Hotplug](#directly-assigned-devices-hotplug)
- [Shared Filesystem Improvements](#shared-filesystem-improvements)
- [Block and Networking IO Self Offloading](#block-and-networking-io-self-offloading)
- [Command Line Interface](#command-line-interface)
- [PVH Boot](#pvh-boot)
- [Contributors](#contributors-4)
- [Contributors](#contributors-5)
- [v0.5.1](#v051)
- [v0.5.0](#v050)
- [Virtual Machine Dynamic Resizing](#virtual-machine-dynamic-resizing)
@@ -52,7 +65,7 @@
- [New Interrupt Management Framework](#new-interrupt-management-framework)
- [Development Tools](#development-tools)
- [Kata Containers Integration](#kata-containers-integration)
- [Contributors](#contributors-5)
- [Contributors](#contributors-6)
- [v0.4.0](#v040)
- [Dynamic virtual CPUs addition](#dynamic-virtual-cpus-addition)
- [Programmatic firmware tables generation](#programmatic-firmware-tables-generation)
@@ -61,7 +74,7 @@
- [Userspace IOAPIC by default](#userspace-ioapic-by-default)
- [PCI BAR reprogramming](#pci-bar-reprogramming)
- [New `cloud-hypervisor` organization](#new-cloud-hypervisor-organization)
- [Contributors](#contributors-6)
- [Contributors](#contributors-7)
- [v0.3.0](#v030)
- [Block device offloading](#block-device-offloading)
- [Network device backend](#network-device-backend)
@@ -88,6 +101,108 @@
- [Unit testing](#unit-testing)
- [Integration tests parallelization](#integration-tests-parallelization)
# v0.11.0
This release has been tracked through the [0.11.0 project](https://github.com/cloud-hypervisor/cloud-hypervisor/projects/14).
Highlights for `cloud-hypervisor` version 0.11.0 include:
### `io_uring` support by default for `virtio-block`
Provided that the host OS supports it (Linux kernel 5.8+) then `io_uring` will
be used for a significantly higher performance block device.
### Windows Guest Support
This is the first release where we officially support Windows running as a
guest. Full details of how to setup the image and run Cloud Hypervisor with a
Windows guest can be found in the dedicated [Windows
documentation](docs/windows.md).
### `vhost-user` "Self Spawning" Deprecation
Automatically spawning a `vhost-user-net` or `vhost-user-block` backend is now
deprecated. Users of this functionality will receive a warning and should make
adjustments. The functionality will be removed in the next release.
### `virtio-mmmio` Removal
Support for using the `virtio-mmio` transport, rather than using PCI, has been
removed. This has been to simplify the code and significantly
reduce the testing burden of the project.
### Snapshot/Restore support for ARM64
When running on the ARM64 architecture snapshot and restore has now been
implemented.
### Improved Linux Boot Time
The time to boot the Linux kernel has been significantly improved by the
identifying some areas of delays around PCI bus probing, IOAPIC programming and
MPTABLE issues. Full details can be seen in #1728.
### `SIGTERM/SIGINT` Interrupt Signal Handling
When the VMM process receives the `SIGTERM` or `SIGINT` signals then it will
trigger the VMM process to cleanly deallocate resources before exiting. The
guest VM will not be cleanly shutdown but the VMM process will clean up its
resources.
### Default Log Level Changed
The default logging level was changed to include warnings which should make it
easier to see potential issues. New [logging
documentation](docs/logging) was also added.
### New `--balloon` Parameter Added
Control of the setup of `virtio-balloon` has been moved from `--memory` to its
own dedicated parameter. This makes it easier to add more balloon specific
controls without overloading `--memory`.
### Experimental `virtio-watchdog` Support
Support for using a new `virtio-watchdog` has been added which can be used to
have the VMM reboot the guest if the guest userspace fails to ping the
watchdog. This is enabled with `--watchdog` and requires kernel support.
### Notable Bug Fixes
* MTRR bit was missing from CPUID advertised to guest
* "Return" key could not be used under `CMD.EXE` under Windows SAC (#1170)
* CPU identification string is now exposed to the guest
* `virtio-pmem` with`discard_writes=on` no longer marks the guest memory as
read only so avoids excessive VM exits (#1795)
* PCI device hotplug after an unplug was fixed (#1802)
* When using the ACPI method to resize the guest memory the full reserved size
can be used (#1803)
* Snapshot and restore followed by a second snapshot and restore now works
correctly
* Snapshot and restore of VMs with more than 2GiB in one region now work
correctly
### Contributors
Many thanks to everyone who has contributed to our 0.11.0 release including some new faces.
* Anatol Belski <anbelski@linux.microsoft.com>
* Bo Chen <chen.bo@intel.com>
* Daniel Verkamp <dverkamp@chromium.org>
* Henry Wang <Henry.Wang@arm.com>
* Hui Zhu <teawater@antfin.com>
* Jiangbo Wu <jiangbo.wu@intel.com>
* Josh Soref <jsoref@users.noreply.github.com>
* Julio Montes <julio.montes@intel.com>
* Michael Zhao <michael.zhao@arm.com>
* Muminul Islam <muislam@microsoft.com>
* pierwill <19642016+pierwill@users.noreply.github.com>
* Praveen Paladugu <prapal@microsoft.com>
* Rob Bradford <robert.bradford@intel.com>
* Sebastien Boeuf <sebastien.boeuf@intel.com>
* Wei Liu <liuwe@microsoft.com>
# v0.10.0
This release has been tracked through the [0.10.0 project](https://github.com/cloud-hypervisor/cloud-hypervisor/projects/13).
@@ -318,7 +433,7 @@ Highlights for `cloud-hypervisor` version 0.7.0 include:
Further to our effort to support modifying a running guest we now support
hotplug and unplug of the following virtio backed devices: block, network,
pmem, virtio-fs and vsock. This functionality is available on the (default) PCI
based tranport and is exposed through the HTTP API. The `ch-remote` utility
based transport and is exposed through the HTTP API. The `ch-remote` utility
provides a CLI for adding or removing these device types after the VM has
booted. User can use the `id` parameter on the devices to choose names for
devices to ease their removal.
@@ -434,7 +549,7 @@ When choosing to offload the paravirtualized block and networking I/O to an
external process (through the `vhost-user` protocol), Cloud Hypervisor now
automatically spawns its default `vhost-user-blk` and `vhost-user-net` backends
into their own, separate processes.
This provides a seamless parvirtualized I/O user experience for those who want
This provides a seamless paravirtualized I/O user experience for those who want
to run their guest I/O into separate executions contexts.
### Command Line Interface
@@ -580,7 +695,7 @@ default Cloud Hypervisor I/O architecture.
### Guest pause and resume
As an initial requiremnt for enabling live migration, we added support for
As an initial requirement for enabling live migration, we added support for
pausing and resuming any VMM components. As an intermediate step towards live
migration, the upcoming guest snapshotting feature will be based on the pause
and resume capabilities.

View File

@@ -1,7 +1,7 @@
FROM ubuntu:18.04 as dev
ARG TARGETARCH="x86_64"
ARG RUST_TOOLCHAIN="1.43.0"
ARG TARGETARCH
ARG RUST_TOOLCHAIN="1.46.0"
ARG CLH_SRC_DIR="/cloud-hypervisor"
ARG CLH_BUILD_DIR="$CLH_SRC_DIR/build"
ARG CARGO_REGISTRY_DIR="$CLH_BUILD_DIR/cargo_registry"
@@ -39,17 +39,18 @@ RUN apt-get update \
cpio \
bsdtar \
libfdt-dev \
python3-setuptools \
&& apt-get clean \
&& rm -rf /var/lib/apt/lists/*
RUN if [ "$TARGETARCH" = "x86_64" ]; then \
RUN if [ "$TARGETARCH" = "amd64" ]; then \
apt-get update \
&& apt-get -yq upgrade \
&& DEBIAN_FRONTEND=noninteractive apt-get install -yq gcc-multilib \
&& apt-get clean \
&& rm -rf /var/lib/apt/lists/*; fi
RUN if [ "$TARGETARCH" = "aarch64" ]; then \
RUN if [ "$TARGETARCH" = "arm64" ]; then \
# On AArch64, `setcap` binary should be installed via `libcap2-bin`.
# The `setcap` binary is used in integration tests.
apt-get update \
@@ -66,17 +67,21 @@ RUN if [ "$TARGETARCH" = "aarch64" ]; then \
&& rm -rf /dtc; fi
# Fix the libssl-dev install
RUN cp /usr/include/"$TARGETARCH"-linux-gnu/openssl/opensslconf.h /usr/include/openssl/
ENV OPENSSL_DIR=/usr/lib/"$TARGETARCH"-linux-gnu/
ENV OPENSSL_LIB_DIR=/usr/lib/"$TARGETARCH"-linux-gnu/
RUN export ARCH="$(uname -m)" \
&& cp /usr/include/$ARCH-linux-gnu/openssl/opensslconf.h /usr/include/openssl/
ENV X86_64_UNKNOWN_LINUX_GNU_OPENSSL_LIB_DIR=/usr/lib/x86_64-linux-gnu/
ENV X86_64_UNKNOWN_LINUX_MUSL_OPENSSL_LIB_DIR=/usr/lib/x86_64-linux-gnu/
ENV AARCH64_UNKNOWN_LINUX_GNU_OPENSSL_LIB_DIR=/usr/lib/aarch64-linux-gnu/
ENV AARCH64_UNKNOWN_LINUX_MUSL_OPENSSL_LIB_DIR=/usr/lib/aarch64-linux-gnu/
ENV OPENSSL_INCLUDE_DIR=/usr/include/
# Install the rust toolchain
RUN nohup curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y --default-toolchain "$RUST_TOOLCHAIN" \
&& rustup target add $TARGETARCH-unknown-linux-musl --toolchain "$RUST_TOOLCHAIN" \
&& if [ "$TARGETARCH" = "x86_64" ]; then rustup toolchain add $RUST_TOOLCHAIN-x86_64-unknown-linux-musl; fi \
&& if [ "$TARGETARCH" = "x86_64" ]; then rustup component add rustfmt; fi \
&& if [ "$TARGETARCH" = "x86_64" ]; then rustup component add clippy; fi \
RUN export ARCH="$(uname -m)" \
&& nohup curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y --default-toolchain "$RUST_TOOLCHAIN" \
&& rustup target add $ARCH-unknown-linux-musl --toolchain "$RUST_TOOLCHAIN" \
&& if [ "$TARGETARCH" = "amd64" ]; then rustup toolchain add $RUST_TOOLCHAIN-x86_64-unknown-linux-musl; fi \
&& if [ "$TARGETARCH" = "amd64" ]; then rustup component add rustfmt; fi \
&& if [ "$TARGETARCH" = "amd64" ]; then rustup component add clippy; fi \
&& cargo install cargo-audit \
&& rm -rf "$CARGO_HOME/registry" \
&& ln -s "$CARGO_REGISTRY_DIR" "$CARGO_HOME/registry" \

View File

@@ -266,10 +266,8 @@ CONFIG_ARCH_MMAP_RND_BITS_MIN=28
CONFIG_ARCH_MMAP_RND_BITS_MAX=32
CONFIG_ARCH_MMAP_RND_COMPAT_BITS_MIN=8
CONFIG_ARCH_MMAP_RND_COMPAT_BITS_MAX=16
CONFIG_GENERIC_ISA_DMA=y
CONFIG_GENERIC_BUG=y
CONFIG_GENERIC_BUG_RELATIVE_POINTERS=y
CONFIG_ARCH_MAY_HAVE_PC_FDC=y
CONFIG_GENERIC_CALIBRATE_DELAY=y
CONFIG_ARCH_HAS_CPU_RELAX=y
CONFIG_ARCH_HAS_CACHE_LINE_SIZE=y
@@ -571,7 +569,7 @@ CONFIG_PCI_MMCONFIG=y
CONFIG_MMCONF_FAM10H=y
# CONFIG_PCI_CNB20LE_QUIRK is not set
# CONFIG_ISA_BUS is not set
CONFIG_ISA_DMA_API=y
# CONFIG_ISA_DMA_API is not set
CONFIG_AMD_NB=y
# CONFIG_X86_SYSFB is not set
# end of Bus options (PCI etc.)
@@ -1150,7 +1148,6 @@ CONFIG_PNP=y
CONFIG_PNPACPI=y
CONFIG_BLK_DEV=y
CONFIG_BLK_DEV_NULL_BLK=y
# CONFIG_BLK_DEV_FD is not set
# CONFIG_BLK_DEV_PCIESSD_MTIP32XX is not set
CONFIG_ZRAM=y
# CONFIG_ZRAM_WRITEBACK is not set
@@ -1447,7 +1444,7 @@ CONFIG_DEVPORT=y
# CONFIG_HPET is not set
CONFIG_HANGCHECK_TIMER=y
# CONFIG_TCG_TPM is not set
CONFIG_TELCLOCK=y
# CONFIG_TELCLOCK is not set
# CONFIG_XILLYBUS is not set
# end of Character devices
@@ -1465,14 +1462,14 @@ CONFIG_TELCLOCK=y
# CONFIG_SPMI is not set
# CONFIG_HSI is not set
CONFIG_PPS=y
CONFIG_PPS_DEBUG=y
# CONFIG_PPS_DEBUG is not set
#
# PPS clients support
#
CONFIG_PPS_CLIENT_KTIMER=y
CONFIG_PPS_CLIENT_LDISC=y
CONFIG_PPS_CLIENT_GPIO=y
# CONFIG_PPS_CLIENT_KTIMER is not set
# CONFIG_PPS_CLIENT_LDISC is not set
# CONFIG_PPS_CLIENT_GPIO is not set
#
# PPS generators support
@@ -1525,7 +1522,68 @@ CONFIG_THERMAL_GOV_USER_SPACE=y
# CONFIG_INTEL_PCH_THERMAL is not set
# end of Intel thermal drivers
# CONFIG_WATCHDOG is not set
CONFIG_WATCHDOG=y
CONFIG_WATCHDOG_CORE=y
# CONFIG_WATCHDOG_NOWAYOUT is not set
CONFIG_WATCHDOG_HANDLE_BOOT_ENABLED=y
CONFIG_WATCHDOG_OPEN_TIMEOUT=0
# CONFIG_WATCHDOG_SYSFS is not set
#
# Watchdog Pretimeout Governors
#
# CONFIG_WATCHDOG_PRETIMEOUT_GOV is not set
#
# Watchdog Device Drivers
#
# CONFIG_SOFT_WATCHDOG is not set
# CONFIG_WDAT_WDT is not set
# CONFIG_XILINX_WATCHDOG is not set
# CONFIG_CADENCE_WATCHDOG is not set
# CONFIG_DW_WATCHDOG is not set
# CONFIG_MAX63XX_WATCHDOG is not set
# CONFIG_ACQUIRE_WDT is not set
# CONFIG_ADVANTECH_WDT is not set
# CONFIG_ALIM1535_WDT is not set
# CONFIG_ALIM7101_WDT is not set
# CONFIG_EBC_C384_WDT is not set
# CONFIG_F71808E_WDT is not set
# CONFIG_SP5100_TCO is not set
# CONFIG_SBC_FITPC2_WATCHDOG is not set
# CONFIG_EUROTECH_WDT is not set
# CONFIG_IB700_WDT is not set
# CONFIG_IBMASR is not set
# CONFIG_WAFER_WDT is not set
# CONFIG_I6300ESB_WDT is not set
# CONFIG_IE6XX_WDT is not set
# CONFIG_ITCO_WDT is not set
# CONFIG_IT8712F_WDT is not set
# CONFIG_IT87_WDT is not set
# CONFIG_HP_WATCHDOG is not set
# CONFIG_SC1200_WDT is not set
# CONFIG_PC87413_WDT is not set
# CONFIG_NV_TCO is not set
# CONFIG_60XX_WDT is not set
# CONFIG_CPU5_WDT is not set
# CONFIG_SMSC_SCH311X_WDT is not set
# CONFIG_SMSC37B787_WDT is not set
# CONFIG_TQMX86_WDT is not set
# CONFIG_VIA_WDT is not set
# CONFIG_W83627HF_WDT is not set
# CONFIG_W83877F_WDT is not set
# CONFIG_W83977F_WDT is not set
# CONFIG_MACHZ_WDT is not set
# CONFIG_SBC_EPX_C3_WATCHDOG is not set
# CONFIG_NI903X_WDT is not set
# CONFIG_NIC7018_WDT is not set
CONFIG_VIRTIO_WDT=y
#
# PCI-based Watchdog Cards
#
# CONFIG_PCIPCWATCHDOG is not set
# CONFIG_WDTPCI is not set
CONFIG_SSB_POSSIBLE=y
# CONFIG_SSB is not set
CONFIG_BCMA_POSSIBLE=y

View File

@@ -7,7 +7,7 @@
CLI_NAME="Cloud Hypervisor"
CTR_IMAGE_TAG="cloudhypervisor/dev"
CTR_IMAGE_VERSION="v4"
CTR_IMAGE_VERSION="latest"
CTR_IMAGE="${CTR_IMAGE_TAG}:${CTR_IMAGE_VERSION}"
DOCKER_RUNTIME="docker"
@@ -136,13 +136,32 @@ fix_dir_perms() {
--workdir "$CTR_CLH_ROOT_DIR" \
--rm \
--volume /dev:/dev \
--volume "$CLH_ROOT_DIR:$CTR_CLH_ROOT_DIR" \
--volume "$CLH_ROOT_DIR:$CTR_CLH_ROOT_DIR" $exported_volumes \
"$CTR_IMAGE" \
chown -R "$(id -u):$(id -g)" "$CTR_CLH_ROOT_DIR"
return $1
}
# Process exported volumes argument, separate the volumes and make docker compatible
# Sample input: --volumes /a:/a#/b:/b
# Sample output: --volume /a:/a --volume /b:/b
#
process_volumes_args() {
if [ -z "$arg_vols" ]; then
return
fi
exported_volumes=""
arr_vols=(${arg_vols//#/ })
for var in "${arr_vols[@]}"
do
parts=(${var//:/ })
if [[ ! -e "${parts[0]}" ]]; then
echo "The volume ${parts[0]} does not exist."
exit 1
fi
exported_volumes="$exported_volumes --volume $var"
done
}
cmd_help() {
echo ""
echo "Cloud Hypervisor $(basename $0)"
@@ -155,15 +174,18 @@ cmd_help() {
echo " --debug Build the debug binaries. This is the default."
echo " --release Build the release binaries."
echo " --libc Select the C library Cloud Hypervisor will be built against. Default is gnu"
echo " --volumes Hash separated volumes to be exported. Example --volumes /mnt:/mnt#/myvol:/myvol"
echo ""
echo " tests [--unit|--cargo|--all] [--libc musl|gnu] [-- [<cargo test args>]]"
echo " Run the Cloud Hypervisor tests."
echo " --unit Run the unit tests."
echo " --cargo Run the cargo tests."
echo " --integration Run the integration tests."
echo " --integration-sgx Run the SGX integration tests."
echo " --libc Select the C library Cloud Hypervisor will be built against. Default is gnu"
echo " --all Run all tests."
echo " --unit Run the unit tests."
echo " --cargo Run the cargo tests."
echo " --integration Run the integration tests."
echo " --integration-sgx Run the SGX integration tests."
echo " --integration-windows Run the Windows guest integration tests."
echo " --libc Select the C library Cloud Hypervisor will be built against. Default is gnu"
echo " --volumes Hash separated volumes to be exported. Example --volumes /mnt:/mnt#/myvol:/myvol"
echo " --all Run all tests."
echo ""
echo " build-container [--type]"
echo " Build the Cloud Hypervisor container."
@@ -195,6 +217,10 @@ cmd_build() {
die "Invalid libc: $1. Valid options are \"musl\" and \"gnu\"."
libc="$1"
;;
"--volumes")
shift
arg_vols="$1"
;;
"--") { shift; break; } ;;
*)
die "Unknown build argument: $1. Please use --help for help."
@@ -202,6 +228,7 @@ cmd_build() {
esac
shift
done
process_volumes_args
target="$(uname -m)-unknown-linux-${libc}"
@@ -209,7 +236,7 @@ cmd_build() {
[ $build = "release" ] && cargo_args+=("--release")
cargo_args+=(--target "$target")
[ $(uname -m) = "aarch64" ] && cargo_args+=("--no-default-features")
[ $(uname -m) = "aarch64" ] && cargo_args+=(--features "mmio,kvm")
[ $(uname -m) = "aarch64" ] && cargo_args+=(--features "kvm")
rustflags=""
if [ $(uname -m) = "aarch64" ] && [ $libc = "musl" ] ; then
@@ -221,7 +248,7 @@ cmd_build() {
--workdir "$CTR_CLH_ROOT_DIR" \
--rm \
--volume /dev:/dev \
--volume "$CLH_ROOT_DIR:$CTR_CLH_ROOT_DIR" \
--volume "$CLH_ROOT_DIR:$CTR_CLH_ROOT_DIR" $exported_volumes \
--env RUSTFLAGS="$rustflags" \
"$CTR_IMAGE" \
cargo build --all \
@@ -236,7 +263,7 @@ cmd_clean() {
--user "$(id -u):$(id -g)" \
--workdir "$CTR_CLH_ROOT_DIR" \
--rm \
--volume "$CLH_ROOT_DIR:$CTR_CLH_ROOT_DIR" \
--volume "$CLH_ROOT_DIR:$CTR_CLH_ROOT_DIR" $exported_volumes \
"$CTR_IMAGE" \
cargo clean \
--target-dir "$CTR_CLH_CARGO_TARGET" \
@@ -248,7 +275,9 @@ cmd_tests() {
cargo=false
integration=false
integration_sgx=false
integration_windows=false
libc="gnu"
arg_vols=""
while [ $# -gt 0 ]; do
case "$1" in
@@ -257,12 +286,17 @@ cmd_tests() {
"--cargo") { cargo=true; } ;;
"--integration") { integration=true; } ;;
"--integration-sgx") { integration_sgx=true; } ;;
"--integration-windows") { integration_windows=true; } ;;
"--libc")
shift
[[ "$1" =~ ^(musl|gnu)$ ]] || \
die "Invalid libc: $1. Valid options are \"musl\" and \"gnu\"."
libc="$1"
;;
"--volumes")
shift
arg_vols="$1"
;;
"--all") { cargo=true; unit=true; integration=true; } ;;
"--") { shift; break; } ;;
*)
@@ -272,6 +306,7 @@ cmd_tests() {
shift
done
process_volumes_args
target="$(uname -m)-unknown-linux-${libc}"
cflags=""
target_cc=""
@@ -288,7 +323,7 @@ cmd_tests() {
--device /dev/kvm \
--device /dev/net/tun \
--cap-add net_admin \
--volume "$CLH_ROOT_DIR:$CTR_CLH_ROOT_DIR" \
--volume "$CLH_ROOT_DIR:$CTR_CLH_ROOT_DIR" $exported_volumes \
--env BUILD_TARGET="$target" \
--env CFLAGS="$cflags" \
--env TARGET_CC="$target_cc" \
@@ -301,7 +336,7 @@ cmd_tests() {
$DOCKER_RUNTIME run \
--workdir "$CTR_CLH_ROOT_DIR" \
--rm \
--volume "$CLH_ROOT_DIR:$CTR_CLH_ROOT_DIR" \
--volume "$CLH_ROOT_DIR:$CTR_CLH_ROOT_DIR" $exported_volumes \
"$CTR_IMAGE" \
./scripts/run_cargo_tests.sh || fix_dir_perms $? || exit $?
fi
@@ -317,7 +352,7 @@ cmd_tests() {
--net="$CTR_CLH_NET" \
--mount type=tmpfs,destination=/tmp \
--volume /dev:/dev \
--volume "$CLH_ROOT_DIR:$CTR_CLH_ROOT_DIR" \
--volume "$CLH_ROOT_DIR:$CTR_CLH_ROOT_DIR" $exported_volumes \
--volume "$CLH_INTEGRATION_WORKLOADS:$CTR_CLH_INTEGRATION_WORKLOADS" \
--env USER="root" \
--env CH_LIBC="${libc}" \
@@ -326,7 +361,26 @@ cmd_tests() {
fi
if [ "$integration_sgx" = true ] ; then
say "Running integration tests for $target..."
say "Running SGX integration tests for $target..."
$DOCKER_RUNTIME run \
--workdir "$CTR_CLH_ROOT_DIR" \
--rm \
--privileged \
--security-opt seccomp=unconfined \
--ipc=host \
--net="$CTR_CLH_NET" \
--mount type=tmpfs,destination=/tmp \
--volume /dev:/dev \
--volume "$CLH_ROOT_DIR:$CTR_CLH_ROOT_DIR" $exported_volumes \
--volume "$CLH_INTEGRATION_WORKLOADS:$CTR_CLH_INTEGRATION_WORKLOADS" \
--env USER="root" \
--env CH_LIBC="${libc}" \
"$CTR_IMAGE" \
./scripts/run_integration_tests_sgx.sh "$@" || fix_dir_perms $? || exit $?
fi
if [ "$integration_windows" = true ] ; then
say "Running Windows integration tests for $target..."
$DOCKER_RUNTIME run \
--workdir "$CTR_CLH_ROOT_DIR" \
--rm \
@@ -341,9 +395,8 @@ cmd_tests() {
--env USER="root" \
--env CH_LIBC="${libc}" \
"$CTR_IMAGE" \
./scripts/run_integration_tests_sgx.sh "$@" || fix_dir_perms $? || exit $?
./scripts/run_integration_tests_windows.sh "$@" || fix_dir_perms $? || exit $?
fi
fix_dir_perms $?
}
@@ -367,11 +420,14 @@ cmd_build-container() {
mkdir -p $BUILD_DIR
cp $CLH_DOCKERFILE $BUILD_DIR
[ $(uname -m) = "aarch64" ] && TARGETARCH="arm64"
[ $(uname -m) = "x86_64" ] && TARGETARCH="amd64"
$DOCKER_RUNTIME build \
--target $container_type \
-t $CTR_IMAGE \
-f $BUILD_DIR/Dockerfile \
--build-arg TARGETARCH="$(uname -m)" \
--build-arg TARGETARCH=$TARGETARCH \
$BUILD_DIR
}
@@ -388,7 +444,7 @@ cmd_shell() {
--net="$CTR_CLH_NET" \
--tmpfs /tmp:exec \
--volume /dev:/dev \
--volume "$CLH_ROOT_DIR:$CTR_CLH_ROOT_DIR" \
--volume "$CLH_ROOT_DIR:$CTR_CLH_ROOT_DIR" $exported_volumes \
--volume "$CLH_INTEGRATION_WORKLOADS:$CTR_CLH_INTEGRATION_WORKLOADS" \
--env USER="root" \
--entrypoint bash \
@@ -427,9 +483,4 @@ if [ $(uname -m) = "x86_64" ]; then
ensure_latest_ctr
fi
# Before a public image for AArch64 ready, we build the container if needed.
if [ $(uname -m) = "aarch64" ]; then
cmd_build-container
fi
$cmd "$@"

View File

@@ -15,11 +15,9 @@ time cargo rustc --bin cloud-hypervisor -- -D warnings
time cargo rustc -p vhost_user_net --bin vhost_user_net -- -D warnings
time cargo test
time cargo audit
time cargo clippy --all-targets --no-default-features --features "pci,acpi,kvm" -- -D warnings
time cargo rustc --bin cloud-hypervisor --no-default-features --features "pci,acpi,kvm" -- -D warnings
time cargo clippy --all-targets --no-default-features --features "pci,kvm" -- -D warnings
time cargo rustc --bin cloud-hypervisor --no-default-features --features "pci,kvm" -- -D warnings
time cargo clippy --all-targets --no-default-features --features "mmio,kvm" -- -D warnings
time cargo rustc --bin cloud-hypervisor --no-default-features --features "mmio,kvm" -- -D warnings
time cargo clippy --all-targets --no-default-features --features "acpi,kvm" -- -D warnings
time cargo rustc --bin cloud-hypervisor --no-default-features --features "acpi,kvm" -- -D warnings
time cargo clippy --all-targets --no-default-features --features "kvm" -- -D warnings
time cargo rustc --bin cloud-hypervisor --no-default-features --features "kvm" -- -D warnings
time cargo fmt -- --check
time cargo build --all --release

View File

@@ -126,14 +126,13 @@ update_workloads() {
if [ ! -f "$VIRTIOFSD" ]; then
pushd $WORKLOADS_DIR
git clone --depth 1 "https://gitlab.com/virtio-fs/qemu.git" -b "virtio-fs-dev" $QEMU_DIR
git clone --depth 1 "https://gitlab.com/virtio-fs/qemu.git" -b "qemu5.0-virtiofs-dax" $QEMU_DIR
pushd $QEMU_DIR
time ./configure --prefix=$PWD --target-list=aarch64-softmmu
time make virtiofsd -j `nproc`
cp virtiofsd $VIRTIOFSD || exit 1
popd
rm -rf $QEMU_DIR
sudo setcap cap_chown,cap_dac_override,cap_dac_read_search,cap_fowner,cap_fsetid,cap_setgid,cap_setuid,cap_mknod,cap_setfcap,cap_sys_admin+epi "virtiofsd" || exit 1
popd
fi
@@ -173,9 +172,9 @@ if [ $RES -ne 0 ]; then
exit 1
fi
# Create tap interface without multipe queues support for vhost_user_net test.
# Create tap interface without multiple queues support for vhost_user_net test.
sudo ip tuntap add name vunet-tap0 mode tap
# Create tap interface with multipe queues support for vhost_user_net test.
# Create tap interface with multiple queues support for vhost_user_net test.
sudo ip tuntap add name vunet-tap1 mode tap multi_queue
BUILD_TARGET="aarch64-unknown-linux-${CH_LIBC}"
@@ -186,52 +185,21 @@ TARGET_CC="musl-gcc"
CFLAGS="-I /usr/include/aarch64-linux-musl/ -idirafter /usr/include/"
fi
cargo build --all --release --no-default-features --features pci,kvm --target $BUILD_TARGET
cargo build --all --release --no-default-features --features kvm --target $BUILD_TARGET
strip target/$BUILD_TARGET/release/cloud-hypervisor
strip target/$BUILD_TARGET/release/vhost_user_net
strip target/$BUILD_TARGET/release/ch-remote
# Copy for non-privileged net test
cp target/$BUILD_TARGET/release/cloud-hypervisor target/$BUILD_TARGET/release/cloud-hypervisor-unprivileged
sudo setcap cap_net_admin+ep target/$BUILD_TARGET/release/cloud-hypervisor
sudo setcap cap_net_admin+ep target/$BUILD_TARGET/release/vhost_user_net
# Enable KSM with some reasonable parameters so that it won't take too long
# for the memory to be merged between two processes.
sudo bash -c "echo 1000000 > /sys/kernel/mm/ksm/pages_to_scan"
sudo bash -c "echo 10 > /sys/kernel/mm/ksm/sleep_millisecs"
sudo bash -c "echo 1 > /sys/kernel/mm/ksm/run"
sudo adduser $USER kvm
newgrp kvm << EOF
export RUST_BACKTRACE=1
time cargo test --no-default-features --features "integration_tests,pci,kvm" "tests::parallel::$@"
EOF
time cargo test --no-default-features --features "integration_tests,kvm" "tests::parallel::$@" -- --skip test_snapshot_restore
RES=$?
if [ $RES -eq 0 ]; then
# virtio-mmio based testing
cargo build --release --target $BUILD_TARGET --no-default-features --features "mmio,kvm"
strip target/$BUILD_TARGET/release/cloud-hypervisor
strip target/$BUILD_TARGET/release/vhost_user_net
strip target/$BUILD_TARGET/release/ch-remote
sudo setcap cap_net_admin+ep target/$BUILD_TARGET/release/cloud-hypervisor
sudo setcap cap_net_admin+ep target/$BUILD_TARGET/release/vhost_user_net
# Ensure test binary has the same caps as the cloud-hypervisor one
time cargo test --no-run --no-default-features --features "integration_tests,mmio,kvm" || exit 1
ls target/debug/deps/cloud_hypervisor-* | xargs -n 1 sudo setcap cap_net_admin+ep
newgrp kvm << EOF
export RUST_BACKTRACE=1
time cargo test --no-default-features --features "integration_tests,mmio,kvm" "tests::parallel::$@"
EOF
RES=$?
fi
# Tear vhost_user_net test network down
sudo ip link del vunet-tap0
sudo ip link del vunet-tap1

View File

@@ -19,13 +19,4 @@ export RUST_BACKTRACE=1
time cargo test --features "integration_tests" "tests::sgx::$@"
RES=$?
if [ $RES -eq 0 ]; then
# virtio-mmio based testing
cargo build --all --release --target $BUILD_TARGET --no-default-features --features "mmio,kvm"
strip target/$BUILD_TARGET/release/cloud-hypervisor
time cargo test --features "integration_tests,mmio" "tests::sgx::$@"
RES=$?
fi
exit $RES

View File

@@ -0,0 +1,24 @@
#!/bin/bash
set -x
source $HOME/.cargo/env
BUILD_TARGET="$(uname -m)-unknown-linux-${CH_LIBC}"
CFLAGS=""
TARGET_CC=""
if [[ "${BUILD_TARGET}" == "x86_64-unknown-linux-musl" ]]; then
TARGET_CC="musl-gcc"
CFLAGS="-I /usr/include/x86_64-linux-musl/ -idirafter /usr/include/"
fi
cargo build --all --release --target $BUILD_TARGET
strip target/$BUILD_TARGET/release/cloud-hypervisor
export RUST_BACKTRACE=1
# Only run with 1 thread to avoid tests interfering with one another because
# Windows has a static IP configured
time cargo test --features "integration_tests" "tests::windows::$@" -- --test-threads=1
RES=$?
exit $RES

View File

@@ -127,14 +127,13 @@ VIRTIOFSD="$WORKLOADS_DIR/virtiofsd"
QEMU_DIR="qemu_build"
if [ ! -f "$VIRTIOFSD" ]; then
pushd $WORKLOADS_DIR
git clone --depth 1 "https://gitlab.com/virtio-fs/qemu.git" -b "virtio-fs-dev" $QEMU_DIR
git clone --depth 1 "https://gitlab.com/virtio-fs/qemu.git" -b "qemu5.0-virtiofs-dax" $QEMU_DIR
pushd $QEMU_DIR
time ./configure --prefix=$PWD --target-list=x86_64-softmmu
time make virtiofsd -j `nproc`
cp virtiofsd $VIRTIOFSD || exit 1
popd
rm -rf $QEMU_DIR
sudo setcap cap_chown,cap_dac_override,cap_dac_read_search,cap_fowner,cap_fsetid,cap_setgid,cap_setuid,cap_mknod,cap_setfcap,cap_sys_admin+epi "virtiofsd" || exit 1
popd
fi
@@ -189,9 +188,9 @@ sudo ip tuntap add vfio-tap3 mode tap
sudo ip link set vfio-tap3 master vfio-br0
sudo ip link set vfio-tap3 up
# Create tap interface without multipe queues support for vhost_user_net test.
# Create tap interface without multiple queues support for vhost_user_net test.
sudo ip tuntap add name vunet-tap0 mode tap
# Create tap interface with multipe queues support for vhost_user_net test.
# Create tap interface with multiple queues support for vhost_user_net test.
sudo ip tuntap add name vunet-tap1 mode tap multi_queue
BUILD_TARGET="$(uname -m)-unknown-linux-${CH_LIBC}"
@@ -207,12 +206,6 @@ strip target/$BUILD_TARGET/release/cloud-hypervisor
strip target/$BUILD_TARGET/release/vhost_user_net
strip target/$BUILD_TARGET/release/ch-remote
# Copy for non-privileged net test
cp target/$BUILD_TARGET/release/cloud-hypervisor target/$BUILD_TARGET/release/cloud-hypervisor-unprivileged
sudo setcap cap_net_admin+ep target/$BUILD_TARGET/release/cloud-hypervisor
sudo setcap cap_net_admin+ep target/$BUILD_TARGET/release/vhost_user_net
# We always copy a fresh version of our binary for our L2 guest.
cp target/$BUILD_TARGET/release/cloud-hypervisor $VFIO_DIR
cp target/$BUILD_TARGET/release/ch-remote $VFIO_DIR
@@ -227,58 +220,18 @@ sudo bash -c "echo 1 > /sys/kernel/mm/ksm/run"
echo 4096 | sudo tee /proc/sys/vm/nr_hugepages
sudo chmod a+rwX /dev/hugepages
# Ensure test binary has the same caps as the cloud-hypervisor one
time cargo test --no-run --features "integration_tests" || exit 1
ls target/debug/deps/cloud_hypervisor-* | xargs -n 1 sudo setcap cap_net_admin+ep
sudo adduser $USER kvm
newgrp kvm << EOF
export RUST_BACKTRACE=1
time cargo test --features "integration_tests" "tests::parallel::$@"
EOF
RES=$?
# Run some tests in sequence since the result could be affected by other tests
# running in parallel.
if [ $RES -eq 0 ]; then
newgrp kvm << EOF
export RUST_BACKTRACE=1
time cargo test --features "integration_tests" "tests::sequential::$@" -- --test-threads=1
EOF
export RUST_BACKTRACE=1
time cargo test --features "integration_tests" "tests::sequential::$@" -- --test-threads=1
RES=$?
fi
if [ $RES -eq 0 ]; then
# virtio-mmio based testing
cargo build --all --release --target $BUILD_TARGET --no-default-features --features "mmio,kvm"
strip target/$BUILD_TARGET/release/cloud-hypervisor
strip target/$BUILD_TARGET/release/vhost_user_net
strip target/$BUILD_TARGET/release/ch-remote
sudo setcap cap_net_admin+ep target/$BUILD_TARGET/release/cloud-hypervisor
# Ensure test binary has the same caps as the cloud-hypervisor one
time cargo test --no-run --features "integration_tests,mmio" || exit 1
ls target/debug/deps/cloud_hypervisor-* | xargs -n 1 sudo setcap cap_net_admin+ep
newgrp kvm << EOF
export RUST_BACKTRACE=1
time cargo test --features "integration_tests,mmio" "tests::parallel::$@"
EOF
RES=$?
# Run some tests in sequence since the result could be affected by other tests
# running in parallel.
if [ $RES -eq 0 ]; then
newgrp kvm << EOF
export RUST_BACKTRACE=1
time cargo test --features "integration_tests,mmio" "tests::sequential::$@" -- --test-threads=1
EOF
RES=$?
fi
fi
# Tear VFIO test network down
sudo ip link del vfio-br0
sudo ip link del vfio-tap0

View File

@@ -5,15 +5,7 @@ source $HOME/.cargo/env
BUILD_TARGET=${BUILD_TARGET-x86_64-unknown-linux-gnu}
cargo_args=("$@")
[ $(uname -m) = "aarch64" ] && cargo_args+=("--no-default-features")
[ $(uname -m) = "aarch64" ] && cargo_args+=("--features mmio,kvm")
[ $(uname -m) = "aarch64" ] && cargo_args+=("--features kvm")
cargo test --target $BUILD_TARGET --workspace --no-run ${cargo_args[@]}
pushd target/$BUILD_TARGET/debug
ls | grep net_util | grep -v "\.d" | xargs -n 1 sudo setcap cap_net_admin,cap_net_raw+ep
popd
sudo adduser $USER kvm
newgrp kvm << EOF || exit 1
export RUST_BACKTRACE=1
cargo test --target $BUILD_TARGET --workspace ${cargo_args[@]} || exit 1;
EOF
export RUST_BACKTRACE=1
cargo test --target $BUILD_TARGET --workspace ${cargo_args[@]} || exit 1;

View File

@@ -5,23 +5,22 @@
#[macro_use(crate_authors)]
extern crate clap;
extern crate api_client;
extern crate serde_json;
extern crate vmm;
use api_client::simple_api_command;
use api_client::Error as ApiClientError;
use clap::{App, AppSettings, Arg, ArgMatches, SubCommand};
use option_parser::{ByteSized, ByteSizedParseError};
use std::fmt;
use std::io::{Read, Write};
use std::os::unix::net::UnixStream;
use std::process;
#[derive(Debug)]
enum Error {
Socket(std::io::Error),
StatusCodeParsing(std::num::ParseIntError),
MissingProtocol,
ContentLengthParsing(std::num::ParseIntError),
ServerResponse(StatusCode, Option<String>),
Connect(std::io::Error),
ApiClient(ApiClientError),
InvalidCPUCount(std::num::ParseIntError),
InvalidMemorySize(ByteSizedParseError),
InvalidBalloonSize(ByteSizedParseError),
@@ -38,17 +37,8 @@ impl fmt::Display for Error {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
use Error::*;
match self {
Socket(e) => write!(f, "Error writing to HTTP socket: {}", e),
StatusCodeParsing(e) => write!(f, "Error parsing HTTP status code: {}", e),
MissingProtocol => write!(f, "HTTP output is missing protocol statement"),
ContentLengthParsing(e) => write!(f, "Error parsing HTTP Content-Length field: {}", e),
ServerResponse(s, o) => {
if let Some(o) = o {
write!(f, "Server responded with an error: {:?}: {}", s, o)
} else {
write!(f, "Server responded with an error: {:?}", s)
}
}
ApiClient(e) => e.fmt(f),
Connect(e) => write!(f, "Error openning HTTP socket: {}", e),
InvalidCPUCount(e) => write!(f, "Error parsing CPU count: {}", e),
InvalidMemorySize(e) => write!(f, "Error parsing memory size: {:?}", e),
InvalidBalloonSize(e) => write!(f, "Error parsing balloon size: {:?}", e),
@@ -63,146 +53,6 @@ impl fmt::Display for Error {
}
}
#[derive(Clone, Copy, Debug)]
pub enum StatusCode {
Continue,
OK,
NoContent,
BadRequest,
NotFound,
InternalServerError,
NotImplemented,
Unknown,
}
impl StatusCode {
fn from_raw(code: usize) -> StatusCode {
match code {
100 => StatusCode::Continue,
200 => StatusCode::OK,
204 => StatusCode::NoContent,
400 => StatusCode::BadRequest,
404 => StatusCode::NotFound,
500 => StatusCode::InternalServerError,
501 => StatusCode::NotImplemented,
_ => StatusCode::Unknown,
}
}
fn parse(code: &str) -> Result<StatusCode, Error> {
Ok(StatusCode::from_raw(
code.trim().parse().map_err(Error::StatusCodeParsing)?,
))
}
fn is_server_error(self) -> bool {
match self {
StatusCode::OK | StatusCode::Continue | StatusCode::NoContent => false,
_ => true,
}
}
}
fn get_header<'a>(res: &'a str, header: &'a str) -> Option<&'a str> {
let header_str = format!("{}: ", header);
if let Some(o) = res.find(&header_str) {
Some(&res[o + header_str.len()..o + res[o..].find('\r').unwrap()])
} else {
None
}
}
fn get_status_code(res: &str) -> Result<StatusCode, Error> {
if let Some(o) = res.find("HTTP/1.1") {
Ok(StatusCode::parse(
&res[o + "HTTP/1.1 ".len()..res[o..].find('\r').unwrap()],
)?)
} else {
Err(Error::MissingProtocol)
}
}
fn parse_http_response(socket: &mut UnixStream) -> Result<Option<String>, Error> {
let mut res = String::new();
let mut body_offset = None;
let mut content_length: Option<usize> = None;
loop {
let mut bytes = vec![0; 256];
let count = socket.read(&mut bytes).map_err(Error::Socket)?;
res.push_str(std::str::from_utf8(&bytes[0..count]).unwrap());
// End of headers
if let Some(o) = res.find("\r\n\r\n") {
body_offset = Some(o + "\r\n\r\n".len());
// With all headers available we can see if there is any body
content_length = if let Some(length) = get_header(&res, "Content-Length") {
Some(length.trim().parse().map_err(Error::ContentLengthParsing)?)
} else {
None
};
if content_length.is_none() {
break;
}
}
if let Some(body_offset) = body_offset {
if let Some(content_length) = content_length {
if res.len() >= content_length + body_offset {
break;
}
}
}
}
let body_string = content_length.and(Some(String::from(&res[body_offset.unwrap()..])));
let status_code = get_status_code(&res)?;
if status_code.is_server_error() {
Err(Error::ServerResponse(status_code, body_string))
} else {
Ok(body_string)
}
}
fn simple_api_command(
socket: &mut UnixStream,
method: &str,
c: &str,
request_body: Option<&str>,
) -> Result<(), Error> {
socket
.write_all(
format!(
"{} /api/v1/vm.{} HTTP/1.1\r\nHost: localhost\r\nAccept: */*\r\n",
method, c
)
.as_bytes(),
)
.map_err(Error::Socket)?;
if let Some(request_body) = request_body {
socket
.write_all(format!("Content-Length: {}\r\n", request_body.len()).as_bytes())
.map_err(Error::Socket)?;
}
socket.write_all(b"\r\n").map_err(Error::Socket)?;
if let Some(request_body) = request_body {
socket
.write_all(request_body.as_bytes())
.map_err(Error::Socket)?;
}
socket.flush().map_err(Error::Socket)?;
if let Some(body) = parse_http_response(socket)? {
println!("{}", body);
}
Ok(())
}
fn resize_api_command(
socket: &mut UnixStream,
cpus: Option<&str>,
@@ -226,7 +76,7 @@ fn resize_api_command(
None
};
let desired_ram_w_balloon: Option<u64> = if let Some(balloon) = balloon {
let desired_balloon: Option<u64> = if let Some(balloon) = balloon {
Some(
balloon
.parse::<ByteSized>()
@@ -240,7 +90,7 @@ fn resize_api_command(
let resize = vmm::api::VmResizeData {
desired_vcpus,
desired_ram,
desired_ram_w_balloon,
desired_balloon,
};
simple_api_command(
@@ -249,6 +99,7 @@ fn resize_api_command(
"resize",
Some(&serde_json::to_string(&resize).unwrap()),
)
.map_err(Error::ApiClient)
}
fn resize_zone_api_command(socket: &mut UnixStream, id: &str, size: &str) -> Result<(), Error> {
@@ -266,6 +117,7 @@ fn resize_zone_api_command(socket: &mut UnixStream, id: &str, size: &str) -> Res
"resize-zone",
Some(&serde_json::to_string(&resize_zone).unwrap()),
)
.map_err(Error::ApiClient)
}
fn add_device_api_command(socket: &mut UnixStream, config: &str) -> Result<(), Error> {
@@ -277,6 +129,7 @@ fn add_device_api_command(socket: &mut UnixStream, config: &str) -> Result<(), E
"add-device",
Some(&serde_json::to_string(&device_config).unwrap()),
)
.map_err(Error::ApiClient)
}
fn remove_device_api_command(socket: &mut UnixStream, id: &str) -> Result<(), Error> {
@@ -288,6 +141,7 @@ fn remove_device_api_command(socket: &mut UnixStream, id: &str) -> Result<(), Er
"remove-device",
Some(&serde_json::to_string(&remove_device_data).unwrap()),
)
.map_err(Error::ApiClient)
}
fn add_disk_api_command(socket: &mut UnixStream, config: &str) -> Result<(), Error> {
@@ -299,6 +153,7 @@ fn add_disk_api_command(socket: &mut UnixStream, config: &str) -> Result<(), Err
"add-disk",
Some(&serde_json::to_string(&disk_config).unwrap()),
)
.map_err(Error::ApiClient)
}
fn add_fs_api_command(socket: &mut UnixStream, config: &str) -> Result<(), Error> {
@@ -310,6 +165,7 @@ fn add_fs_api_command(socket: &mut UnixStream, config: &str) -> Result<(), Error
"add-fs",
Some(&serde_json::to_string(&fs_config).unwrap()),
)
.map_err(Error::ApiClient)
}
fn add_pmem_api_command(socket: &mut UnixStream, config: &str) -> Result<(), Error> {
@@ -321,6 +177,7 @@ fn add_pmem_api_command(socket: &mut UnixStream, config: &str) -> Result<(), Err
"add-pmem",
Some(&serde_json::to_string(&pmem_config).unwrap()),
)
.map_err(Error::ApiClient)
}
fn add_net_api_command(socket: &mut UnixStream, config: &str) -> Result<(), Error> {
@@ -332,6 +189,7 @@ fn add_net_api_command(socket: &mut UnixStream, config: &str) -> Result<(), Erro
"add-net",
Some(&serde_json::to_string(&net_config).unwrap()),
)
.map_err(Error::ApiClient)
}
fn add_vsock_api_command(socket: &mut UnixStream, config: &str) -> Result<(), Error> {
@@ -343,6 +201,7 @@ fn add_vsock_api_command(socket: &mut UnixStream, config: &str) -> Result<(), Er
"add-vsock",
Some(&serde_json::to_string(&vsock_config).unwrap()),
)
.map_err(Error::ApiClient)
}
fn snapshot_api_command(socket: &mut UnixStream, url: &str) -> Result<(), Error> {
@@ -356,6 +215,7 @@ fn snapshot_api_command(socket: &mut UnixStream, url: &str) -> Result<(), Error>
"snapshot",
Some(&serde_json::to_string(&snapshot_config).unwrap()),
)
.map_err(Error::ApiClient)
}
fn restore_api_command(socket: &mut UnixStream, config: &str) -> Result<(), Error> {
@@ -367,15 +227,20 @@ fn restore_api_command(socket: &mut UnixStream, config: &str) -> Result<(), Erro
"restore",
Some(&serde_json::to_string(&restore_config).unwrap()),
)
.map_err(Error::ApiClient)
}
fn do_command(matches: &ArgMatches) -> Result<(), Error> {
let mut socket =
UnixStream::connect(matches.value_of("api-socket").unwrap()).map_err(Error::Socket)?;
UnixStream::connect(matches.value_of("api-socket").unwrap()).map_err(Error::Connect)?;
match matches.subcommand_name() {
Some("info") => simple_api_command(&mut socket, "GET", "info", None),
Some("counters") => simple_api_command(&mut socket, "GET", "counters", None),
Some("info") => {
simple_api_command(&mut socket, "GET", "info", None).map_err(Error::ApiClient)
}
Some("counters") => {
simple_api_command(&mut socket, "GET", "counters", None).map_err(Error::ApiClient)
}
Some("resize") => resize_api_command(
&mut socket,
matches
@@ -476,7 +341,7 @@ fn do_command(matches: &ArgMatches) -> Result<(), Error> {
.value_of("restore_config")
.unwrap(),
),
Some(c) => simple_api_command(&mut socket, "PUT", c, None),
Some(c) => simple_api_command(&mut socket, "PUT", c, None).map_err(Error::ApiClient),
None => unreachable!(),
}
}
@@ -577,7 +442,7 @@ fn main() {
.arg(
Arg::with_name("balloon")
.long("balloon")
.help("New memory with balloon size in bytes (supports K/M/G suffix)")
.help("New balloon size in bytes (supports K/M/G suffix)")
.takes_value(true)
.number_of_values(1),
),

View File

@@ -3,6 +3,7 @@
// SPDX-License-Identifier: Apache-2.0
//
extern crate anyhow;
extern crate vmm;
extern crate vmm_sys_util;
@@ -13,14 +14,39 @@ use clap::{App, Arg, ArgGroup, ArgMatches};
use libc::EFD_NONBLOCK;
use log::LevelFilter;
use seccomp::SeccompAction;
use std::env;
use std::sync::mpsc::channel;
use std::sync::{Arc, Mutex};
use std::{env, process};
use thiserror::Error;
use vhost_user_block::start_block_backend;
use vhost_user_net::start_net_backend;
use vmm::config;
use vmm_sys_util::eventfd::EventFd;
#[derive(Error, Debug)]
enum Error {
#[error("Failed to create API EventFd: {0}")]
CreateAPIEventFd(#[source] std::io::Error),
#[error("Failed to open hypervisor interface (is /dev/kvm available?): {0}")]
CreateHypervisor(#[source] hypervisor::HypervisorError),
#[error("Failed to start the VMM thread: {0}")]
StartVMMThread(#[source] vmm::Error),
#[error("Error parsing config: {0}")]
ParsingConfig(vmm::config::Error),
#[error("Error creating VM: {0:?}")]
VmCreate(vmm::api::ApiError),
#[error("Error booting VM: {0:?}")]
VmBoot(vmm::api::ApiError),
#[error("Error restoring VM: {0:?}")]
VmRestore(vmm::api::ApiError),
#[error("Error parsing restore: {0}")]
ParsingRestore(vmm::config::Error),
#[error("Failed to join on VMM thread: {0:?}")]
ThreadJoin(std::boxed::Box<dyn std::any::Any + std::marker::Send>),
#[error("VMM thread exited with error: {0}")]
VmmThread(#[source] vmm::Error),
}
struct Logger {
output: Mutex<Box<dyn std::io::Write + Send>>,
start: std::time::Instant,
@@ -97,7 +123,8 @@ fn create_app<'a, 'b>(
.long("cpus")
.help(
"boot=<boot_vcpus>,max=<max_vcpus>,\
topology=<threads_per_core>:<cores_per_die>:<dies_per_package>:<packages>,kvm_hyperv=on|off",
topology=<threads_per_core>:<cores_per_die>:<dies_per_package>:<packages>,\
kvm_hyperv=on|off,max_phys_bits=<maximum_number_of_physical_bits>",
)
.default_value(&default_vcpus)
.group("vm-config"),
@@ -132,7 +159,10 @@ fn create_app<'a, 'b>(
.arg(
Arg::with_name("kernel")
.long("kernel")
.help("Path to kernel image (vmlinux)")
.help(
"Path to loaded kernel. This may be a kernel or firmware that supports a PVH \
entry point, a vmlinux ELF file or a Linux bzImage or achitecture equivalent",
)
.takes_value(true)
.group("vm-config"),
)
@@ -175,6 +205,13 @@ fn create_app<'a, 'b>(
.default_value(&default_rng)
.group("vm-config"),
)
.arg(
Arg::with_name("balloon")
.long("balloon")
.help(config::BalloonConfig::SYNTAX)
.takes_value(true)
.group("vm-config"),
)
.arg(
Arg::with_name("fs")
.long("fs")
@@ -231,6 +268,13 @@ fn create_app<'a, 'b>(
.min_values(1)
.group("vm-config"),
)
.arg(
Arg::with_name("watchdog")
.long("watchdog")
.help("Enable virtio-watchdog")
.takes_value(false)
.group("vm-config"),
)
.arg(
Arg::with_name("v")
.short("v")
@@ -302,13 +346,9 @@ fn create_app<'a, 'b>(
app
}
fn start_vmm(cmd_arguments: ArgMatches) {
let api_socket_path = cmd_arguments
.value_of("api-socket")
.expect("Missing argument: api-socket");
fn start_vmm(cmd_arguments: ArgMatches, api_socket_path: &str) -> Result<(), Error> {
let (api_request_sender, api_request_receiver) = channel();
let api_evt = EventFd::new(EFD_NONBLOCK).expect("Cannot create API EventFd");
let api_evt = EventFd::new(EFD_NONBLOCK).map_err(Error::CreateAPIEventFd)?;
let http_sender = api_request_sender.clone();
let seccomp_action = if let Some(seccomp_value) = cmd_arguments.value_of("seccomp") {
@@ -317,15 +357,15 @@ fn start_vmm(cmd_arguments: ArgMatches) {
"false" => SeccompAction::Allow,
"log" => SeccompAction::Log,
_ => {
eprintln!("Invalid parameter {} for \"--seccomp\" flag", seccomp_value);
process::exit(1);
// The user providing an invalid value will be rejected by clap
panic!("Invalid parameter {} for \"--seccomp\" flag", seccomp_value);
}
}
} else {
SeccompAction::Trap
};
let hypervisor = hypervisor::new().unwrap();
let vmm_thread = match vmm::start_vmm_thread(
let hypervisor = hypervisor::new().map_err(Error::CreateHypervisor)?;
let vmm_thread = vmm::start_vmm_thread(
env!("CARGO_PKG_VERSION").to_string(),
api_socket_path,
api_evt.try_clone().unwrap(),
@@ -333,25 +373,14 @@ fn start_vmm(cmd_arguments: ArgMatches) {
api_request_receiver,
&seccomp_action,
hypervisor,
) {
Ok(t) => t,
Err(e) => {
eprintln!("Failed spawning the VMM thread {:?}", e);
process::exit(1);
}
};
)
.map_err(Error::StartVMMThread)?;
// Can't test for "vm-config" group as some have default values. The kernel
// is the only required option for booting the VM.
if cmd_arguments.is_present("kernel") {
let vm_params = config::VmParams::from_arg_matches(&cmd_arguments);
let vm_config = match config::VmConfig::parse(vm_params) {
Ok(config) => config,
Err(e) => {
eprintln!("{}", e);
process::exit(1);
}
};
let vm_config = config::VmConfig::parse(vm_params).map_err(Error::ParsingConfig)?;
println!(
"Cloud Hypervisor Guest\n\tAPI server: {}\n\tvCPUs: {}\n\tMemory: {} MB\n\tKernel: \
@@ -372,36 +401,21 @@ fn start_vmm(cmd_arguments: ArgMatches) {
api_request_sender,
Arc::new(Mutex::new(vm_config)),
)
.expect("Could not create the VM");
vmm::api::vm_boot(api_evt.try_clone().unwrap(), sender).expect("Could not boot the VM");
.map_err(Error::VmCreate)?;
vmm::api::vm_boot(api_evt.try_clone().unwrap(), sender).map_err(Error::VmBoot)?;
} else if let Some(restore_params) = cmd_arguments.value_of("restore") {
vmm::api::vm_restore(
api_evt.try_clone().unwrap(),
api_request_sender,
Arc::new(match config::RestoreConfig::parse(restore_params) {
Ok(config) => config,
Err(e) => {
println!("{}", e);
process::exit(1);
}
}),
Arc::new(config::RestoreConfig::parse(restore_params).map_err(Error::ParsingRestore)?),
)
.expect("Could not restore the VM");
.map_err(Error::VmRestore)?;
}
match vmm_thread.join() {
Ok(res) => match res {
Ok(_) => (),
Err(e) => {
eprintln!("VMM thread failed {:?}", e);
process::exit(1);
}
},
Err(e) => {
eprintln!("Could not join VMM thread {:?}", e);
process::exit(1);
}
}
vmm_thread
.join()
.map_err(Error::ThreadJoin)?
.map_err(Error::VmmThread)
}
fn main() {
@@ -435,10 +449,9 @@ fn main() {
.get_matches();
let log_level = match cmd_arguments.occurrences_of("v") {
0 => LevelFilter::Error,
1 => LevelFilter::Warn,
2 => LevelFilter::Info,
3 => LevelFilter::Debug,
0 => LevelFilter::Warn,
1 => LevelFilter::Info,
2 => LevelFilter::Debug,
_ => LevelFilter::Trace,
};
@@ -463,7 +476,17 @@ fn main() {
} else if let Some(backend_command) = cmd_arguments.value_of("block-backend") {
start_block_backend(backend_command);
} else {
start_vmm(cmd_arguments);
let api_socket_path = cmd_arguments
.value_of("api-socket")
.expect("Missing argument: api-socket")
.to_string();
if let Err(e) = start_vmm(cmd_arguments, &api_socket_path) {
eprintln!("{}", e);
std::fs::remove_file(api_socket_path).ok();
std::process::exit(1);
}
std::fs::remove_file(api_socket_path).ok();
}
}
@@ -535,6 +558,7 @@ mod unit_tests {
max_vcpus: 1,
topology: None,
kvm_hyperv: false,
max_phys_bits: None,
},
memory: MemoryConfig {
size: 536_870_912,
@@ -544,8 +568,6 @@ mod unit_tests {
hotplugged_size: None,
shared: false,
hugepages: false,
balloon: false,
balloon_size: 0,
zones: None,
},
kernel: Some(KernelConfig {
@@ -561,6 +583,7 @@ mod unit_tests {
src: PathBuf::from("/dev/urandom"),
iommu: false,
},
balloon: None,
fs: None,
pmem: None,
serial: ConsoleConfig {
@@ -579,6 +602,7 @@ mod unit_tests {
#[cfg(target_arch = "x86_64")]
sgx_epc: None,
numa: None,
watchdog: false,
};
aver_eq!(tb, expected_vm_config, result_vm_config);

View File

@@ -9,6 +9,11 @@ users:
ssh_pwauth: True
runcmd:
- [ systemctl, daemon-reload]
- [ systemctl, enable, notify-booted.service]
- [ systemctl, start, --no-block, notify-booted.service ]
write_files:
-
path: /etc/systemd/system/vfio.service
@@ -39,3 +44,26 @@ write_files:
echo 512 | sudo tee /proc/sys/vm/nr_hugepages
sudo chmod a+rwX /dev/hugepages
/mnt/cloud-hypervisor --kernel /mnt/vmlinux --cmdline "console=hvc0 reboot=k panic=1 nomodules i8042.noaux i8042.nomux i8042.nopnp i8042.dumbkbd root=/dev/vda1 VFIOTAG" --disk path=/mnt/focal-server-cloudimg-amd64-custom.qcow2 path=/mnt/cloudinit.img --cpus boot=1 --memory size=512M,hotplug_size=1G,hugepages=on --device path=/sys/bus/pci/devices/0000:00:06.0/ path=/sys/bus/pci/devices/0000:00:07.0/ --api-socket /tmp/ch_api.sock
-
path: /etc/systemd/system/notify-booted.service
permissions: 0644
content: |
[Unit]
Description=Notify the tcp listener on the host that the guest is booted
After=sshd.service
[Service]
Type=simple
ExecStart=/usr/bin/cloud-hypervisor-notify-booted.sh
[Install]
WantedBy=multi-user.target
-
path: /usr/bin/cloud-hypervisor-notify-booted.sh
permissions: 0755
content: |
#!/bin/bash
echo -n "@DEFAULT_TCP_LISTENER_MESSAGE" | nc -w0 @HOST_IP @TCP_LISTENER_PORT

File diff suppressed because it is too large Load Diff

View File

@@ -9,10 +9,10 @@ default = []
[dependencies]
epoll = ">=4.0.1"
libc = "0.2.77"
libc = "0.2.80"
log = "0.4.11"
virtio-bindings = "0.1.0"
vm-memory = "0.2.1"
vm-memory = "0.3.0"
vm-virtio = { path = "../vm-virtio" }
vmm-sys-util = ">=0.3.1"
vhost_rs = { git = "https://github.com/rust-vmm/vhost", branch = "master", package = "vhost", features = ["vhost-user-slave"] }

View File

@@ -8,12 +8,12 @@ edition = "2018"
block_util = { path = "../block_util" }
clap = { version = "2.33.3", features=["wrap_help"] }
epoll = ">=4.0.1"
libc = "0.2.77"
libc = "0.2.80"
log = "0.4.11"
option_parser = { path = "../option_parser" }
qcow = { path = "../qcow" }
vhost_user_backend = { path = "../vhost_user_backend" }
vhost_rs = { git = "https://github.com/rust-vmm/vhost", branch = "master", package = "vhost", features = ["vhost-user-slave"] }
virtio-bindings = "0.1.0"
vm-memory = "0.2.1"
vm-memory = "0.3.0"
vmm-sys-util = ">=0.3.1"

View File

@@ -8,8 +8,8 @@ edition = "2018"
bitflags = "1.1.0"
clap = { version = "2.33.3", features=["wrap_help"] }
epoll = ">=4.0.1"
futures = { version = "0.3.5", features = ["thread-pool"] }
libc = "0.2.77"
futures = { version = "0.3.7", features = ["thread-pool"] }
libc = "0.2.80"
log = "0.4.11"
# Match the version in vmm
seccomp = { git = "https://github.com/firecracker-microvm/firecracker", tag = "v0.22.0" }
@@ -18,6 +18,6 @@ virtio-bindings = { version = "0.1", features = ["virtio-v5_0_0"]}
virtio-devices = { path = "../virtio-devices" }
vhost_rs = { git = "https://github.com/rust-vmm/vhost", branch = "master", package = "vhost", features = ["vhost-user-slave"] }
vhost_user_backend = { path = "../vhost_user_backend"}
vm-memory = "0.2.1"
vm-memory = "0.3.0"
vm-virtio = { path = "../vm-virtio" }
vmm-sys-util = "0.6.1"

View File

@@ -763,7 +763,7 @@ pub trait FileSystem {
/// Allocate requested space for file data.
///
/// If this function returns success, then the file sytem must guarantee that it is possible to
/// If this function returns success, then the file system must guarantee that it is possible to
/// write up to `length` bytes of data starting at `offset` without failing due to a lack of
/// free space on the disk.
///

View File

@@ -260,11 +260,8 @@ impl Sandbox {
// /proc/sys/fs/nr_open is a sysctl file that shows the maximum number
// of file-handles a process can allocate.
let path = "/proc/sys/fs/nr_open";
let max_str = fs::read_to_string(path).map_err(|e| Error::ReadProc(e))?;
let max = max_str
.trim()
.parse()
.map_err(|e| Error::InvalidNrOpen(e))?;
let max_str = fs::read_to_string(path).map_err(Error::ReadProc)?;
let max = max_str.trim().parse().map_err(Error::InvalidNrOpen)?;
let limit = libc::rlimit {
rlim_cur: max,

View File

@@ -137,6 +137,5 @@ fn vuf_filter(action: SeccompAction) -> Result<SeccompFilter, Error> {
pub fn enable_seccomp(action: SeccompAction) -> Result<(), Error> {
let scfilter = vuf_filter(action)?;
let bpfprog: BpfProgram = scfilter.try_into()?;
SeccompFilter::apply(bpfprog.try_into().unwrap()).unwrap();
Ok(())
SeccompFilter::apply(bpfprog).map_err(Error::ApplySeccompFilter)
}

View File

@@ -7,12 +7,12 @@ edition = "2018"
[dependencies]
clap = { version = "2.33.3", features=["wrap_help"] }
epoll = ">=4.0.1"
libc = "0.2.77"
libc = "0.2.80"
log = "0.4.11"
net_util = { path = "../net_util" }
option_parser = { path = "../option_parser" }
vhost_user_backend = { path = "../vhost_user_backend" }
vhost_rs = { git = "https://github.com/rust-vmm/vhost", branch = "master", package = "vhost", features = ["vhost-user-slave"] }
virtio-bindings = "0.1.0"
vm-memory = "0.2.1"
vm-memory = "0.3.0"
vmm-sys-util = ">=0.3.1"

View File

@@ -6,8 +6,6 @@ edition = "2018"
[features]
default = []
pci_support = ["pci"]
mmio_support = []
io_uring = ["block_util/io_uring"]
[dependencies]
@@ -17,11 +15,11 @@ block_util = { path = "../block_util" }
byteorder = "1.3.4"
epoll = ">=4.0.1"
io-uring = ">=0.4.0"
libc = "0.2.77"
libc = "0.2.80"
log = "0.4.11"
net_gen = { path = "../net_gen" }
net_util = { path = "../net_util" }
pci = { path = "../pci", optional = true }
pci = { path = "../pci" }
seccomp = { git = "https://github.com/firecracker-microvm/firecracker", tag = "v0.22.0" }
serde = ">=1.0.27"
serde_derive = ">=1.0.27"
@@ -32,7 +30,7 @@ vhost_rs = { git = "https://github.com/rust-vmm/vhost", branch = "master", packa
virtio-bindings = { version = "0.1", features = ["virtio-v5_0_0"]}
vm-allocator = { path = "../vm-allocator" }
vm-device = { path = "../vm-device" }
vm-memory = { version = "0.2.1", features = ["backend-mmap", "backend-atomic"] }
vm-memory = { version = "0.3.0", features = ["backend-mmap", "backend-atomic"] }
vm-migration = { path = "../vm-migration" }
vm-virtio = { path = "../vm-virtio" }
vmm-sys-util = ">=0.3.1"

View File

@@ -47,9 +47,6 @@ const INFLATE_QUEUE_EVENT: u16 = EPOLL_HELPER_EVENT_LAST + 2;
// New descriptors are pending on the virtio queue.
const DEFLATE_QUEUE_EVENT: u16 = EPOLL_HELPER_EVENT_LAST + 3;
// Page shift in the host.
const PAGE_SHIFT: u32 = 12;
// Size of a PFN in the balloon interface.
const VIRTIO_BALLOON_PFN_SHIFT: u64 = 12;
@@ -78,7 +75,7 @@ pub enum Error {
}
// Got from include/uapi/linux/virtio_balloon.h
#[repr(C, packed)]
#[repr(C)]
#[derive(Copy, Clone, Debug, Default)]
struct VirtioBalloonConfig {
// Number of pages host wants Guest to give up.
@@ -87,6 +84,9 @@ struct VirtioBalloonConfig {
actual: u32,
}
const CONFIG_ACTUAL_OFFSET: u64 = 4;
const CONFIG_ACTUAL_SIZE: usize = 4;
// Safe because it only has data and has no implicit padding.
unsafe impl ByteValued for VirtioBalloonConfig {}
@@ -207,7 +207,7 @@ impl BalloonEpollHandler {
let res = unsafe {
libc::madvise(
hva as *mut libc::c_void,
(1 << PAGE_SHIFT) as libc::size_t,
(1 << VIRTIO_BALLOON_PFN_SHIFT) as libc::size_t,
advice,
)
};
@@ -258,7 +258,8 @@ impl EpollHelperHandler for BalloonEpollHandler {
let mut signal_error = false;
let r = {
let mut config = self.config.lock().unwrap();
config.num_pages = (self.resize_receiver.get_size() >> PAGE_SHIFT) as u32;
config.num_pages =
(self.resize_receiver.get_size() >> VIRTIO_BALLOON_PFN_SHIFT) as u32;
if let Err(e) = self.signal(&VirtioInterruptType::Config, None) {
signal_error = true;
Err(e)
@@ -297,7 +298,7 @@ impl EpollHelperHandler for BalloonEpollHandler {
}
}
_ => {
error!("Unknown event for virtio-mem");
error!("Unknown event for virtio-balloon");
return true;
}
}
@@ -321,7 +322,7 @@ impl Balloon {
let avail_features = 1u64 << VIRTIO_F_VERSION_1;
let mut config = VirtioBalloonConfig::default();
config.num_pages = (size >> PAGE_SHIFT) as u32;
config.num_pages = (size >> VIRTIO_BALLOON_PFN_SHIFT) as u32;
Ok(Balloon {
common: VirtioCommon {
@@ -341,6 +342,11 @@ impl Balloon {
pub fn resize(&self, size: u64) -> Result<(), Error> {
self.resize.work(size)
}
// Get the actual size of the virtio-balloon.
pub fn get_actual(&self) -> u64 {
(self.config.lock().unwrap().actual as u64) << VIRTIO_BALLOON_PFN_SHIFT
}
}
impl Drop for Balloon {
@@ -373,6 +379,20 @@ impl VirtioDevice for Balloon {
self.read_config_from_slice(self.config.lock().unwrap().as_slice(), offset, data);
}
fn write_config(&mut self, offset: u64, data: &[u8]) {
// The "actual" field is the only mutable field
if offset != CONFIG_ACTUAL_OFFSET || data.len() != CONFIG_ACTUAL_SIZE {
error!(
"Attempt to write to read-only field: offset {:x} length {}",
offset,
data.len()
);
return;
}
self.write_config_helper(self.config.lock().unwrap().as_mut_slice(), offset, data);
}
fn activate(
&mut self,
mem: GuestMemoryAtomic<GuestMemoryMmap>,

View File

@@ -176,6 +176,28 @@ pub trait VirtioDevice: Send {
.unwrap();
}
}
/// Helper to allow common implementation of write_config
fn write_config_helper(&self, config: &mut [u8], offset: u64, data: &[u8]) {
let config_len = config.len() as u64;
let data_len = data.len() as u64;
if offset + data_len > config_len {
error!(
"Out-of-bound access to configuration: config_len = {} offset = {:x} length = {} for {}",
config_len,
offset,
data_len,
self.device_type()
);
return;
}
if let Some(end) = offset.checked_add(config.len() as u64) {
let mut offset_config =
&mut config[offset as usize..std::cmp::min(end, config_len) as usize];
offset_config.write_all(data).unwrap();
}
}
}
/// Trait providing address translation the same way a physical DMA remapping

View File

@@ -809,7 +809,7 @@ impl Iommu {
// virtual IOMMU. This list is translated into a virtio-iommu configuration
// topology, so that it can be understood by the guest driver.
//
// The topology is overriden everytime this function is being invoked.
// The topology is overridden everytime this function is being invoked.
//
// This function is dedicated to PCI, which means it will exclusively
// create VIRTIO_IOMMU_TOPO_PCI_RANGE entries.

View File

@@ -14,7 +14,6 @@ extern crate arc_swap;
extern crate epoll;
#[macro_use]
extern crate log;
#[cfg(feature = "pci_support")]
extern crate pci;
extern crate serde;
#[macro_use]
@@ -31,7 +30,6 @@ use std::io;
mod device;
pub mod balloon;
pub mod block;
#[cfg(feature = "io_uring")]
pub mod block_io_uring;
mod console;
pub mod epoll_helper;
@@ -45,10 +43,10 @@ pub mod seccomp_filters;
pub mod transport;
pub mod vhost_user;
pub mod vsock;
pub mod watchdog;
pub use self::balloon::*;
pub use self::block::*;
#[cfg(feature = "io_uring")]
pub use self::block_io_uring::*;
pub use self::console::*;
pub use self::device::*;
@@ -60,6 +58,7 @@ pub use self::net_util::*;
pub use self::pmem::*;
pub use self::rng::*;
pub use self::vsock::*;
pub use self::watchdog::*;
use vm_virtio::{queue::*, VirtioDeviceType};
const DEVICE_INIT: u32 = 0x00;
@@ -73,11 +72,6 @@ const VIRTIO_F_VERSION_1: u32 = 32;
const VIRTIO_F_IOMMU_PLATFORM: u32 = 33;
const VIRTIO_F_IN_ORDER: u32 = 35;
#[allow(dead_code)]
const INTERRUPT_STATUS_USED_RING: u32 = 0x1;
#[allow(dead_code)]
const INTERRUPT_STATUS_CONFIG_CHANGED: u32 = 0x2;
#[cfg(feature = "pci_support")]
const VIRTIO_MSI_NO_VECTOR: u16 = 0xffff;
#[derive(Debug)]

View File

@@ -66,7 +66,7 @@ const VIRTIO_MEM_RESP_NACK: u16 = 1;
// - VIRTIO_MEM_REQ_UNPLUG_ALL
// VIRTIO_MEM_RESP_BUSY: u16 = 2;
// Error in request (e.g. addresses/alignemnt), applicable for
// Error in request (e.g. addresses/alignment), applicable for
// - VIRTIO_MEM_REQ_PLUG
// - VIRTIO_MEM_REQ_UNPLUG
// - VIRTIO_MEM_REQ_STATE
@@ -159,7 +159,7 @@ struct VirtioMemResp {
unsafe impl ByteValued for VirtioMemResp {}
// Got from qemu/include/standard-headers/linux/virtio_mem.h
#[repr(C, packed)]
#[repr(C)]
#[derive(Copy, Clone, Debug, Default)]
struct VirtioMemConfig {
// Block size and alignment. Cannot change.
@@ -664,7 +664,7 @@ impl EpollHelperHandler for MemEpollHandler {
},
};
if let Err(e) = self.resize.send(r) {
error!("Sending \"resize\" reponse: {:?}", e);
error!("Sending \"resize\" response: {:?}", e);
return true;
}
if signal_error {

View File

@@ -82,7 +82,7 @@ pub enum Error {
InvalidQueuePairsNum,
/// No memory passed in.
NoMemory,
/// No ueue pairs nummber.
/// No ueue pairs number.
NoQueuePairsNum,
}

View File

@@ -16,7 +16,6 @@ use crate::seccomp_filters::{get_seccomp_filter, Thread};
use crate::{VirtioInterrupt, VirtioInterruptType};
use anyhow::anyhow;
use seccomp::{SeccompAction, SeccompFilter};
use serde::ser::{Serialize, SerializeStruct, Serializer};
use std::fmt::{self, Display};
use std::fs::File;
use std::io;
@@ -46,33 +45,13 @@ const VIRTIO_PMEM_RESP_TYPE_EIO: u32 = 1;
// New descriptors are pending on the virtio queue.
const QUEUE_AVAIL_EVENT: u16 = EPOLL_HELPER_EVENT_LAST + 1;
#[derive(Copy, Clone, Debug, Default, Deserialize)]
#[repr(C, packed)]
#[derive(Copy, Clone, Debug, Default, Serialize, Deserialize)]
#[repr(C)]
struct VirtioPmemConfig {
start: u64,
size: u64,
}
// We must explicitly implement Serialize since the structure is packed and
// it's unsafe to borrow from a packed structure. And by default, if we derive
// Serialize from serde, it will borrow the values from the structure.
// That's why this implementation copies each field separately before it
// serializes the entire structure field by field.
impl Serialize for VirtioPmemConfig {
fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
where
S: Serializer,
{
let start = self.start;
let size = self.size;
let mut virtio_pmem_config = serializer.serialize_struct("VirtioPmemConfig", 16)?;
virtio_pmem_config.serialize_field("start", &start)?;
virtio_pmem_config.serialize_field("size", &size)?;
virtio_pmem_config.end()
}
}
// Safe because it only has data and has no implicit padding.
unsafe impl ByteValued for VirtioPmemConfig {}

View File

@@ -27,6 +27,7 @@ pub enum Thread {
VirtioVhostNet,
VirtioVhostNetCtl,
VirtioVsock,
VirtioWatchdog,
}
/// Shorthand for chaining `SeccompCondition`s with the `and` operator in a `SeccompRule`.
@@ -126,6 +127,7 @@ fn virtio_blk_io_uring_thread_rules() -> Result<Vec<SyscallRuleSet>, Error> {
#[cfg(target_arch = "x86_64")]
allow_syscall(libc::SYS_epoll_wait),
allow_syscall(libc::SYS_exit),
allow_syscall(libc::SYS_fsync),
allow_syscall(libc::SYS_futex),
allow_syscall(SYS_IO_URING_ENTER),
allow_syscall(libc::SYS_madvise),
@@ -166,14 +168,21 @@ fn virtio_console_thread_rules() -> Result<Vec<SyscallRuleSet>, Error> {
fn virtio_iommu_thread_rules() -> Result<Vec<SyscallRuleSet>, Error> {
Ok(vec![
allow_syscall(libc::SYS_brk),
allow_syscall(libc::SYS_close),
allow_syscall(libc::SYS_dup),
allow_syscall(libc::SYS_epoll_create1),
allow_syscall(libc::SYS_epoll_ctl),
allow_syscall(libc::SYS_epoll_pwait),
#[cfg(target_arch = "x86_64")]
allow_syscall(libc::SYS_epoll_wait),
allow_syscall(libc::SYS_exit),
allow_syscall(libc::SYS_futex),
allow_syscall(libc::SYS_madvise),
allow_syscall(libc::SYS_mmap),
allow_syscall(libc::SYS_mprotect),
allow_syscall(libc::SYS_munmap),
allow_syscall(libc::SYS_read),
allow_syscall(libc::SYS_sigaltstack),
allow_syscall(libc::SYS_write),
])
}
@@ -408,6 +417,33 @@ fn virtio_vsock_thread_rules() -> Result<Vec<SyscallRuleSet>, Error> {
])
}
fn virtio_watchdog_thread_rules() -> Result<Vec<SyscallRuleSet>, Error> {
Ok(vec![
allow_syscall(libc::SYS_brk),
allow_syscall(libc::SYS_close),
allow_syscall(libc::SYS_dup),
allow_syscall(libc::SYS_epoll_create1),
allow_syscall(libc::SYS_epoll_ctl),
allow_syscall(libc::SYS_epoll_pwait),
#[cfg(target_arch = "x86_64")]
allow_syscall(libc::SYS_epoll_wait),
allow_syscall(libc::SYS_exit),
allow_syscall(libc::SYS_futex),
allow_syscall(libc::SYS_madvise),
allow_syscall(libc::SYS_mmap),
allow_syscall(libc::SYS_mprotect),
allow_syscall(libc::SYS_munmap),
allow_syscall(libc::SYS_prctl),
allow_syscall(libc::SYS_read),
allow_syscall(libc::SYS_rt_sigprocmask),
allow_syscall(libc::SYS_sched_getaffinity),
allow_syscall(libc::SYS_set_robust_list),
allow_syscall(libc::SYS_sigaltstack),
allow_syscall(libc::SYS_timerfd_settime),
allow_syscall(libc::SYS_write),
])
}
fn get_seccomp_filter_trap(thread_type: Thread) -> Result<SeccompFilter, Error> {
let rules = match thread_type {
Thread::VirtioBalloon => virtio_balloon_thread_rules()?,
@@ -425,6 +461,7 @@ fn get_seccomp_filter_trap(thread_type: Thread) -> Result<SeccompFilter, Error>
Thread::VirtioVhostNet => virtio_vhost_net_thread_rules()?,
Thread::VirtioVhostNetCtl => virtio_vhost_net_ctl_thread_rules()?,
Thread::VirtioVsock => virtio_vsock_thread_rules()?,
Thread::VirtioWatchdog => virtio_watchdog_thread_rules()?,
};
Ok(SeccompFilter::new(
@@ -450,6 +487,7 @@ fn get_seccomp_filter_log(thread_type: Thread) -> Result<SeccompFilter, Error> {
Thread::VirtioVhostNet => virtio_vhost_net_thread_rules()?,
Thread::VirtioVhostNetCtl => virtio_vhost_net_ctl_thread_rules()?,
Thread::VirtioVsock => virtio_vsock_thread_rules()?,
Thread::VirtioWatchdog => virtio_watchdog_thread_rules()?,
};
Ok(SeccompFilter::new(

View File

@@ -1,537 +0,0 @@
// Copyright 2017 The Chromium OS Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
use crate::transport::{VirtioTransport, NOTIFY_REG_OFFSET};
use crate::{
Queue, VirtioDevice, VirtioInterrupt, VirtioInterruptType, DEVICE_ACKNOWLEDGE, DEVICE_DRIVER,
DEVICE_DRIVER_OK, DEVICE_FAILED, DEVICE_FEATURES_OK, DEVICE_INIT,
INTERRUPT_STATUS_CONFIG_CHANGED, INTERRUPT_STATUS_USED_RING,
};
use anyhow::anyhow;
use byteorder::{ByteOrder, LittleEndian};
use libc::EFD_NONBLOCK;
use std::num::Wrapping;
use std::result;
use std::sync::atomic::{AtomicUsize, Ordering};
use std::sync::{Arc, Mutex};
use vm_device::interrupt::InterruptSourceGroup;
use vm_device::BusDevice;
use vm_memory::{GuestAddress, GuestAddressSpace, GuestMemoryAtomic, GuestMemoryMmap};
use vm_migration::{
Migratable, MigratableError, Pausable, Snapshot, SnapshotDataSection, Snapshottable,
Transportable,
};
use vm_virtio::queue;
use vmm_sys_util::{errno::Result, eventfd::EventFd};
const VENDOR_ID: u32 = 0;
const MMIO_MAGIC_VALUE: u32 = 0x7472_6976;
const MMIO_VERSION: u32 = 2;
#[derive(Debug)]
enum Error {
/// Failed to retrieve queue ring's index.
QueueRingIndex(queue::Error),
}
pub struct VirtioInterruptIntx {
interrupt_status: Arc<AtomicUsize>,
interrupt: Arc<Box<dyn InterruptSourceGroup>>,
}
impl VirtioInterruptIntx {
pub fn new(
interrupt_status: Arc<AtomicUsize>,
interrupt: Arc<Box<dyn InterruptSourceGroup>>,
) -> Self {
VirtioInterruptIntx {
interrupt_status,
interrupt,
}
}
}
impl VirtioInterrupt for VirtioInterruptIntx {
fn trigger(
&self,
int_type: &VirtioInterruptType,
_queue: Option<&Queue>,
) -> std::result::Result<(), std::io::Error> {
let status = match int_type {
VirtioInterruptType::Config => INTERRUPT_STATUS_CONFIG_CHANGED,
VirtioInterruptType::Queue => INTERRUPT_STATUS_USED_RING,
};
self.interrupt_status
.fetch_or(status as usize, Ordering::SeqCst);
self.interrupt.trigger(0)
}
}
#[derive(Serialize, Deserialize)]
struct VirtioMmioDeviceState {
device_activated: bool,
features_select: u32,
acked_features_select: u32,
queue_select: u32,
interrupt_status: usize,
driver_status: u32,
queues: Vec<Queue>,
shm_region_select: u32,
}
/// Implements the
/// [MMIO](http://docs.oasis-open.org/virtio/virtio/v1.0/cs04/virtio-v1.0-cs04.html#x1-1090002)
/// transport for virtio devices.
///
/// This requires 3 points of installation to work with a VM:
///
/// 1. Mmio reads and writes must be sent to this device at what is referred to here as MMIO base.
/// 1. `Mmio::queue_evts` must be installed at `virtio::NOTIFY_REG_OFFSET` offset from the MMIO
/// base. Each event in the array must be signaled if the index is written at that offset.
/// 1. `Mmio::interrupt_evt` must signal an interrupt that the guest driver is listening to when it
/// is written to.
///
/// Typically one page (4096 bytes) of MMIO address space is sufficient to handle this transport
/// and inner virtio device.
pub struct MmioDevice {
id: String,
device: Arc<Mutex<dyn VirtioDevice>>,
device_activated: bool,
features_select: u32,
acked_features_select: u32,
queue_select: u32,
interrupt_status: Arc<AtomicUsize>,
interrupt_cb: Option<Arc<dyn VirtioInterrupt>>,
driver_status: u32,
config_generation: u32,
queues: Vec<Queue>,
queue_evts: Vec<EventFd>,
mem: Option<GuestMemoryAtomic<GuestMemoryMmap>>,
shm_region_select: u32,
}
impl MmioDevice {
/// Constructs a new MMIO transport for the given virtio device.
pub fn new(
id: String,
mem: GuestMemoryAtomic<GuestMemoryMmap>,
device: Arc<Mutex<dyn VirtioDevice>>,
) -> Result<MmioDevice> {
let device_clone = device.clone();
let locked_device = device_clone.lock().unwrap();
let mut queue_evts = Vec::new();
for _ in locked_device.queue_max_sizes().iter() {
queue_evts.push(EventFd::new(EFD_NONBLOCK)?)
}
let queues = locked_device
.queue_max_sizes()
.iter()
.map(|&s| Queue::new(s))
.collect();
Ok(MmioDevice {
id,
device,
device_activated: false,
features_select: 0,
acked_features_select: 0,
queue_select: 0,
interrupt_status: Arc::new(AtomicUsize::new(0)),
interrupt_cb: None,
driver_status: DEVICE_INIT,
config_generation: 0,
queues,
queue_evts,
mem: Some(mem),
shm_region_select: 0,
})
}
fn state(&self) -> VirtioMmioDeviceState {
VirtioMmioDeviceState {
device_activated: self.device_activated,
features_select: self.features_select,
acked_features_select: self.acked_features_select,
queue_select: self.queue_select,
interrupt_status: self.interrupt_status.load(Ordering::SeqCst),
driver_status: self.driver_status,
queues: self.queues.clone(),
shm_region_select: self.shm_region_select,
}
}
fn set_state(&mut self, state: &VirtioMmioDeviceState) -> std::result::Result<(), Error> {
self.device_activated = state.device_activated;
self.features_select = state.features_select;
self.acked_features_select = state.acked_features_select;
self.queue_select = state.queue_select;
self.interrupt_status
.store(state.interrupt_status, Ordering::SeqCst);
self.driver_status = state.driver_status;
// Update virtqueues indexes for both available and used rings.
if let Some(mem) = self.mem.as_ref() {
let mem = mem.memory();
for (i, queue) in self.queues.iter_mut().enumerate() {
queue.max_size = state.queues[i].max_size;
queue.size = state.queues[i].size;
queue.ready = state.queues[i].ready;
queue.vector = state.queues[i].vector;
queue.desc_table = state.queues[i].desc_table;
queue.avail_ring = state.queues[i].avail_ring;
queue.used_ring = state.queues[i].used_ring;
queue.next_avail = Wrapping(
queue
.used_index_from_memory(&mem)
.map_err(Error::QueueRingIndex)?,
);
queue.next_used = Wrapping(
queue
.used_index_from_memory(&mem)
.map_err(Error::QueueRingIndex)?,
);
}
}
self.shm_region_select = state.shm_region_select;
Ok(())
}
/// Gets the list of queue events that must be triggered whenever the VM writes to
/// `virtio::NOTIFY_REG_OFFSET` past the MMIO base. Each event must be triggered when the
/// value being written equals the index of the event in this list.
fn queue_evts(&self) -> &[EventFd] {
self.queue_evts.as_slice()
}
fn is_driver_ready(&self) -> bool {
let ready_bits = DEVICE_ACKNOWLEDGE | DEVICE_DRIVER | DEVICE_DRIVER_OK | DEVICE_FEATURES_OK;
self.driver_status == ready_bits && self.driver_status & DEVICE_FAILED == 0
}
/// Determines if the driver has requested the device (re)init / reset itself
fn is_driver_init(&self) -> bool {
self.driver_status == DEVICE_INIT
}
fn are_queues_valid(&self) -> bool {
if let Some(mem) = self.mem.as_ref() {
self.queues.iter().all(|q| q.is_valid(&mem.memory()))
} else {
false
}
}
fn with_queue<U, F>(&self, d: U, f: F) -> U
where
F: FnOnce(&Queue) -> U,
{
match self.queues.get(self.queue_select as usize) {
Some(queue) => f(queue),
None => d,
}
}
fn with_queue_mut<F: FnOnce(&mut Queue)>(&mut self, f: F) -> bool {
if let Some(queue) = self.queues.get_mut(self.queue_select as usize) {
f(queue);
true
} else {
false
}
}
pub fn assign_interrupt(&mut self, interrupt: Arc<Box<dyn InterruptSourceGroup>>) {
self.interrupt_cb = Some(Arc::new(VirtioInterruptIntx::new(
self.interrupt_status.clone(),
interrupt,
)));
}
}
impl VirtioTransport for MmioDevice {
fn ioeventfds(&self, base_addr: u64) -> Vec<(&EventFd, u64)> {
let notify_base = base_addr + u64::from(NOTIFY_REG_OFFSET);
self.queue_evts()
.iter()
.map(|event| (event, notify_base))
.collect()
}
}
impl BusDevice for MmioDevice {
fn read(&mut self, _base: u64, offset: u64, data: &mut [u8]) {
match offset {
0x00..=0xff if data.len() == 4 => {
let v = match offset {
0x0 => MMIO_MAGIC_VALUE,
0x04 => MMIO_VERSION,
0x08 => self.device.lock().unwrap().device_type(),
0x0c => VENDOR_ID, // vendor id
0x10 => {
if self.features_select < 2 {
(self.device.lock().unwrap().features() >> (self.features_select * 32))
as u32
} else {
0
}
}
0x34 => self.with_queue(0, |q| u32::from(q.get_max_size())),
0x44 => self.with_queue(0, |q| q.ready as u32),
0x60 => self.interrupt_status.load(Ordering::SeqCst) as u32,
0x70 => self.driver_status,
0xfc => self.config_generation,
0xb0..=0xbc => {
// For no SHM region or invalid region the kernel looks for length of -1
let (shm_offset, shm_len) = if let Some(shm_regions) =
self.device.lock().unwrap().get_shm_regions()
{
if self.shm_region_select as usize > shm_regions.region_list.len() {
(0, !0 as u64)
} else {
(
shm_regions.region_list[self.shm_region_select as usize].offset
+ shm_regions.addr.0,
shm_regions.region_list[self.shm_region_select as usize].len,
)
}
} else {
(0, !0 as u64)
};
match offset {
0xb0 => shm_len as u32,
0xb4 => (shm_len >> 32) as u32,
0xb8 => shm_offset as u32,
0xbc => (shm_offset >> 32) as u32,
_ => {
error!("invalid shm region offset");
0
}
}
}
_ => {
warn!("unknown virtio mmio register read: 0x{:x}", offset);
return;
}
};
LittleEndian::write_u32(data, v);
}
0x100..=0xfff => self
.device
.lock()
.unwrap()
.read_config(offset - 0x100, data),
_ => {
warn!(
"invalid virtio mmio read: 0x{:x}:0x{:x}",
offset,
data.len()
);
}
};
}
fn write(&mut self, _base: u64, offset: u64, data: &[u8]) {
fn hi(v: &mut GuestAddress, x: u32) {
*v = (*v & 0xffff_ffff) | (u64::from(x) << 32)
}
fn lo(v: &mut GuestAddress, x: u32) {
*v = (*v & !0xffff_ffff) | u64::from(x)
}
let mut mut_q = false;
match offset {
0x00..=0xff if data.len() == 4 => {
let v = LittleEndian::read_u32(data);
match offset {
0x14 => self.features_select = v,
0x20 => {
if self.acked_features_select < 2 {
self.device
.lock()
.unwrap()
.ack_features(u64::from(v) << (self.acked_features_select * 32));
} else {
warn!(
"invalid ack_features (page {}, value 0x{:x})",
self.acked_features_select, v
);
}
}
0x24 => self.acked_features_select = v,
0x30 => self.queue_select = v,
0x38 => mut_q = self.with_queue_mut(|q| q.size = v as u16),
0x44 => mut_q = self.with_queue_mut(|q| q.ready = v == 1),
0x64 => {
self.interrupt_status
.fetch_and(!(v as usize), Ordering::SeqCst);
}
0x70 => self.driver_status = v,
0x80 => mut_q = self.with_queue_mut(|q| lo(&mut q.desc_table, v)),
0x84 => mut_q = self.with_queue_mut(|q| hi(&mut q.desc_table, v)),
0x90 => mut_q = self.with_queue_mut(|q| lo(&mut q.avail_ring, v)),
0x94 => mut_q = self.with_queue_mut(|q| hi(&mut q.avail_ring, v)),
0xa0 => mut_q = self.with_queue_mut(|q| lo(&mut q.used_ring, v)),
0xa4 => mut_q = self.with_queue_mut(|q| hi(&mut q.used_ring, v)),
0xac => self.shm_region_select = v,
_ => {
warn!("unknown virtio mmio register write: 0x{:x}", offset);
return;
}
}
}
0x100..=0xfff => {
return self
.device
.lock()
.unwrap()
.write_config(offset - 0x100, data)
}
_ => {
warn!(
"invalid virtio mmio write: 0x{:x}:0x{:x}",
offset,
data.len()
);
return;
}
}
if self.device_activated && mut_q {
warn!("virtio queue was changed after device was activated");
}
if !self.device_activated && self.is_driver_ready() && self.are_queues_valid() {
if let Some(interrupt_cb) = self.interrupt_cb.take() {
if self.mem.is_some() {
let mem = self.mem.as_ref().unwrap().clone();
self.device
.lock()
.unwrap()
.activate(
mem,
interrupt_cb,
self.queues.clone(),
self.queue_evts.split_off(0),
)
.expect("Failed to activate device");
self.device_activated = true;
}
}
}
// Device has been reset by the driver
if self.device_activated && self.is_driver_init() {
let mut device = self.device.lock().unwrap();
if let Some((interrupt_cb, mut queue_evts)) = device.reset() {
// Upon reset the device returns its interrupt EventFD and it's queue EventFDs
self.interrupt_cb = Some(interrupt_cb);
self.queue_evts.append(&mut queue_evts);
self.device_activated = false;
// Reset queue readiness (changes queue_enable), queue sizes
// and selected_queue as per spec for reset
self.queues.iter_mut().for_each(Queue::reset);
self.queue_select = 0;
} else {
error!("Attempt to reset device when not implemented in underlying device");
self.driver_status = DEVICE_FAILED;
}
}
}
}
impl Pausable for MmioDevice {
fn pause(&mut self) -> result::Result<(), MigratableError> {
Ok(())
}
fn resume(&mut self) -> result::Result<(), MigratableError> {
Ok(())
}
}
impl Snapshottable for MmioDevice {
fn id(&self) -> String {
self.id.clone()
}
fn snapshot(&mut self) -> std::result::Result<Snapshot, MigratableError> {
let snapshot =
serde_json::to_vec(&self.state()).map_err(|e| MigratableError::Snapshot(e.into()))?;
let mut virtio_mmio_dev_snapshot = Snapshot::new(self.id.as_str());
virtio_mmio_dev_snapshot.add_data_section(SnapshotDataSection {
id: format!("{}-section", self.id),
snapshot,
});
Ok(virtio_mmio_dev_snapshot)
}
fn restore(&mut self, snapshot: Snapshot) -> std::result::Result<(), MigratableError> {
if let Some(virtio_mmio_dev_section) =
snapshot.snapshot_data.get(&format!("{}-section", self.id))
{
let virtio_mmio_dev_state =
match serde_json::from_slice(&virtio_mmio_dev_section.snapshot) {
Ok(state) => state,
Err(error) => {
return Err(MigratableError::Restore(anyhow!(
"Could not deserialize VIRTIO_MMIO_DEVICE {}",
error
)))
}
};
// First restore the status of the virtqueues.
self.set_state(&virtio_mmio_dev_state).map_err(|e| {
MigratableError::Restore(anyhow!(
"Could not restore VIRTIO_MMIO_DEVICE state {:?}",
e
))
})?;
// Then we can activate the device, as we know at this point that
// the virtqueues are in the right state and the device is ready
// to be activated, which will spawn each virtio worker thread.
if self.device_activated && self.is_driver_ready() && self.are_queues_valid() {
if let Some(interrupt_cb) = self.interrupt_cb.take() {
if self.mem.is_some() {
let mem = self.mem.as_ref().unwrap().clone();
self.device
.lock()
.unwrap()
.activate(
mem,
interrupt_cb,
self.queues.clone(),
self.queue_evts.split_off(0),
)
.map_err(|e| {
MigratableError::Restore(anyhow!(
"Failed activating the device: {:?}",
e
))
})?;
}
}
}
return Ok(());
}
Err(MigratableError::Restore(anyhow!(
"Could not find VIRTIO_MMIO_DEVICE snapshot section"
)))
}
}
impl Transportable for MmioDevice {}
impl Migratable for MmioDevice {}

View File

@@ -3,22 +3,11 @@
// SPDX-License-Identifier: Apache-2.0
use vmm_sys_util::eventfd::EventFd;
#[cfg(feature = "pci_support")]
mod pci_common_config;
#[cfg(feature = "pci_support")]
mod pci_device;
#[cfg(feature = "pci_support")]
pub use pci_common_config::VirtioPciCommonConfig;
#[cfg(feature = "pci_support")]
pub use pci_device::VirtioPciDevice;
#[cfg(feature = "mmio_support")]
mod mmio;
#[cfg(feature = "mmio_support")]
pub use mmio::MmioDevice;
#[cfg(feature = "mmio_support")]
pub const NOTIFY_REG_OFFSET: u32 = 0x50;
pub trait VirtioTransport {
fn ioeventfds(&self, base_addr: u64) -> Vec<(&EventFd, u64)>;
}

Some files were not shown because too many files have changed in this diff Show More