Commit Graph

362 Commits

Author SHA1 Message Date
Sebastien Boeuf
d34f31fe7b vmm: Fix KvmInterruptManager when base is different from 0
When the base InterruptIndex is different from 0, the loop allocating
GSI and HashMap entries won't work as expected. The for loop needs to
start from base, but the limit must be base+count so that we allocate
a number of "count" entries.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-01-21 10:44:48 +01:00
Sebastien Boeuf
e73cb1ff80 vmm: Initialize InterruptManager sooner
In order to let the InterruptManager be shared across both PCI and MMIO
devices, this commit moves the initialization earlier in the code.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-01-21 10:44:48 +01:00
Rob Bradford
3901a1dd7d vmm: Log an error if VM resize fails
As well as returing an error to the API caller.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-01-17 23:44:21 +01:00
Rob Bradford
76d9bf2792 vmm: Start memory slots at zero
After refactoring a common function is used to setup these slots and
that function takes care of allocating a new slot so it is not necessary
to reserve the initial region slots.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-01-17 23:44:21 +01:00
Rob Bradford
0ab22fea2c vmm: Only generate GED event when new DIMM added
Avoid the ACPI scan in the guest OS when no new DIMM is hotplugged.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-01-17 23:44:21 +01:00
Rob Bradford
211786ab42 vmm: Only generate GED interrupt when the number of vCPUs has changed
Avoid activity in the the guest OS if the number of vCPUs has not
changed.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-01-17 23:44:21 +01:00
Sebastien Boeuf
4bb12a2d8d interrupt: Reorganize all interrupt management with InterruptManager
Based on all the previous changes, we can at this point replace the
entire interrupt management with the implementation of InterruptManager
and InterruptSourceGroup traits.

By using KvmInterruptManager from the DeviceManager, we can provide both
VirtioPciDevice and VfioPciDevice a way to pick the kind of
InterruptSourceGroup they want to create. Because they choose the type
of interrupt to be MSI/MSI-X, they will be given a MsiInterruptGroup.

Both MsixConfig and MsiConfig are responsible for the update of the GSI
routes, which is why, by passing the MsiInterruptGroup to them, they can
still perform the GSI route management without knowing implementation
details. That's where the InterruptSourceGroup is powerful, as it
provides a generic way to manage interrupt, no matter the type of
interrupt and no matter which hypervisor might be in use.

Once the full replacement has been achieved, both SystemAllocator and
KVM specific dependencies can be removed.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-01-17 23:43:45 +01:00
Sebastien Boeuf
92082ad439 vmm: Fully implement interrupt traits
After the skeleton of InterruptManager and InterruptSourceGroup traits
have been implemented, this new commit takes care of fully implementing
the content of KvmInterruptManager (InterruptManager trait) and
MsiInterruptGroup (InterruptSourceGroup).

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-01-17 23:43:45 +01:00
Sebastien Boeuf
0f727127d5 vmm: Implement InterruptSourceGroup and InterruptManager skeleton
This commit introduces an empty implementation of both InterruptManager
and InterruptSourceGroup traits, as a proper basis for further
implementation.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-01-17 23:43:45 +01:00
Sebastien Boeuf
c396baca46 vm-virtio: Modify VirtioInterrupt callback into a trait
Callbacks are not the most Rust idiomatic way of programming. The right
way is to use a Trait to provide multiple implementation of the same
interface.

Additionally, a Trait will allow for multiple functions to be defined
while using callbacks means that a new callback must be introduced for
each new function we want to add.

For these two reasons, the current commit modifies the existing
VirtioInterrupt callback into a Trait of the same name.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-01-17 23:43:45 +01:00
Sebastien Boeuf
2381f32ae0 msix: Add gsi_msi_routes to MsixConfig
Because MsixConfig will be responsible for updating KVM GSI routes at
some point, it is necessary that it can access the list of routes
contained by gsi_msi_routes.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-01-17 23:43:45 +01:00
Sebastien Boeuf
9b60fcdc39 msix: Add VmFd to MsixConfig
Because MsixConfig will be responsible for updating the KVM GSI routes
at some point, it must have access to the VmFd to invoke the KVM ioctl
KVM_SET_GSI_ROUTING.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-01-17 23:43:45 +01:00
Sebastien Boeuf
86c760a0d9 msix: Add SystemAllocator to MsixConfig
The point here is to let MsixConfig take care of the GSI allocation,
which means the SystemAllocator must be passed from the vmm crate all
the way down to the pci crate.

Once this is done, the GSI allocation and irq_fd creation is performed
by MsixConfig directly.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-01-17 23:43:45 +01:00
Sebastien Boeuf
f5704d32b3 vmm: Move gsi_msi_routes creation to be shared across all PCI devices
Because we will need to share the same list of GSI routes across
multiple PCI devices (virtio-pci, VFIO), this commit moves the creation
of such list to a higher level location in the code.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-01-17 23:43:45 +01:00
Sergio Lopez
a14aee9213 qcow: Use RawFile as backend instead of File
Use RawFile as backend instead of File. This allows us to abstract
the access to the actual image with a specialized layer, so we have a
place where we can deal with the low-level peculiarities.

Signed-off-by: Sergio Lopez <slp@redhat.com>
2020-01-17 17:28:44 +00:00
Sergio Lopez
c5a656c9dc vm-virtio: block: Add support for alignment restrictions
Doing I/O on an image opened with O_DIRECT requires to adhere to
certain restrictions, requiring the following elements to be aligned:

 - Address of the source/destination memory buffer.
 - File offset.
 - Length of the data to be read/written.

The actual alignment value depends on various elements, and according
to open(2) "(...) there is currently no filesystem-independent
interface for an application to discover these restrictions (...)".

To discover such value, we iterate through a list of alignments
(currently, 512 and 4096) calling pread() with each one and checking
if the operation succeeded.

We also extend RawFile so it can be used as a backend for QcowFile,
so the later can be easily adapted to support O_DIRECT too.

Signed-off-by: Sergio Lopez <slp@redhat.com>
2020-01-17 17:28:44 +00:00
Cathy Zhang
652e7b9b8a vm-virtio: Implement multiple queue support for net devices
Update the common part in net_util.rs under vm-virtio to add mq
support, meanwhile enable mq for virtio-net device, vhost-user-net
device and vhost-user-net backend. Multiple threads will be created,
one thread will be responsible to handle one queue pair separately.
To gain the better performance, it requires to have the same amount
of vcpus as queue pair numbers defined for the net device, due to
the cpu affinity.

Multiple thread support is not added for vhost-user-net backend
currently, it will be added in future.

Signed-off-by: Cathy Zhang <cathy.zhang@intel.com>
2020-01-17 12:06:19 +01:00
Cathy Zhang
404316eea1 vmm: Add multiple queue option and update config for virtio-net device
Add num_queues and queue_size for virtio-net device to make them configurable,
while add the associated options in command line.

Update cloud-hypervisor.yaml with the new options for NetConfig.

Signed-off-by: Cathy Zhang <cathy.zhang@intel.com>
2020-01-17 12:06:19 +01:00
Cathy Zhang
4ab88a8173 net_util: Add multiple queue support for tap
Add support to allow VMMs to open the same tap device many times, it will
create multiple file descriptors meanwhile.

Signed-off-by: Cathy Zhang <cathy.zhang@intel.com>
2020-01-17 12:06:19 +01:00
Cathy Zhang
1ae7deb393 vm-virtio: Implement refactor for net devices and backend
Since the common parts are put into net_util.rs under vm-virtio,
refactoring code for virtio-net device, vhost-user-net device
and backend to shrink the code size and improve readability
meanwhile.

Signed-off-by: Cathy Zhang <cathy.zhang@intel.com>
2020-01-17 12:06:19 +01:00
Rob Bradford
8b500d7873 deps: Bump vm-memory and linux-loader version
The function GuestMemory::end_addr() has been renamed to last_addr()

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-01-15 20:21:22 +01:00
Rob Bradford
7310ab6fa7 devices, vmm: Use a bit field for ACPI GED interrupt type
Use independent bits for storing whether there is a CPU or memory device
changed when reporting changes via ACPI GED interrupt. This prevents a
later notification squashing an earlier one and ensure that hotplugging
both CPU and memory at the same time succeeds.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-01-15 20:21:22 +01:00
Rob Bradford
28c6652e57 vmm: Upon VmResize attempt to hotplug the memory
If a new amount of RAM is requested in the VmResize command try and
hotplug if it an increase (MemoryManager::Resize() silently ignores
decreases.)

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-01-15 20:21:22 +01:00
Rob Bradford
4e414f0d84 vmm: device_manager: Scan memory devices upon GED interrupt
If there is a GED interrupt and the field indicates that the memory
device has changed triggers a scan of the memory devices.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-01-15 20:21:22 +01:00
Rob Bradford
284d5e011a vmm: Add memory hotplug ACPI entries to DSDT
Generate and expose the DSDT table entries required to support memory
hotplug. The AML methods call into the MemoryManager via I/O ports
exposed as fields.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-01-15 20:21:22 +01:00
Rob Bradford
8ecf736982 vmm: device_manager: Add the MemoryManager to the I/O bus
Now that the MemoryManager has I/O port functionality it needs to be
exposed on the I/O bus.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-01-15 20:21:22 +01:00
Rob Bradford
1218765df2 vmm: memory_manager: Expose the slots details via an I/O port
Expose the details of hotplug RAM slots via an I/O port. This will be
consumed by the ACPI DSDT tables to report the hotplug memory details to
the guest.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-01-15 20:21:22 +01:00
Rob Bradford
9880a2aba9 vmm: memory_manger: Add support for adding new memory to the VM
Add a "resize()" method on MemoryManager which will create a new memory
allocation based on the difference between the desired RAM amount and
the amount already in use. After allocating the added RAM using the same
backing method as the boot RAM store the details in a vector and update
the KVM map and create a new GuestMemoryMmap and replace all the users.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-01-15 20:21:22 +01:00
Rob Bradford
82fce5a4e2 vmm: Add support for resizing the memory used by the VM
For now the new memory size is only used after a reboot but support for
hotplugging memory will be added in a later commit.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-01-15 20:21:22 +01:00
Rob Bradford
78dcb1862c vmm: device_manager: Store the type of notification in a local value
When the value is read from the I/O port via the ACPI AML functions to
determine what has been triggered the notifiction value is reset
preventing a second read from exposing the value. If we need support
multiple types of GED notification (such as memory hotplug) then we
should avoid reading the value multiple times.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-01-15 20:21:22 +01:00
Rob Bradford
f5137e84bb vmm, main: Add optional "hotplug_size" to --mem
This specifies how much address space should be reserved for hotplugging
of RAM. This space is reserved by adding move the start of the device
area by the desired amount.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-01-15 20:21:22 +01:00
Rob Bradford
f1b6657833 vmm: Make desired vCPUs optional in resize command
In order to be able to support resizing either vCPUs or memory or both
make the fields in the resize command optional.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-01-15 20:21:22 +01:00
Rob Bradford
72b9e920a3 vmm: memory_manager: Further refactor memory region allocation
This allows the memory regions to be allocated later which is necessary
for hotplug memory.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-01-15 20:21:22 +01:00
Rob Bradford
1af11a7c92 vmm: memory_manager: Refactor GuestMemoryMmap construction
Make the GuestMemoryMmap from a Vec<Arc<GuestRegionMmap>> by using this
method we can persist a set of regions in the MemoryManager and then
extend this set with a newly created region. Ultimately that will allow
the hotplug of memory.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-01-15 20:21:22 +01:00
Samuel Ortiz
5788d36583 vmm: Do not create virtio devices when missing a transport
If neither PCI or MMIO are built in, we should not bother creating any
virtio devices at all.
When building a minimal VMM made of a kernel with an initramfs and a
serial console, the RNG virtio device is still created even though there
is no way it can ever get probed.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
2020-01-14 07:42:09 +01:00
Sebastien Boeuf
ae6f27277b acpi: Introduce VIOT to support latest virtio-iommu implementation
Because virtio-iommu is still evolving (as it's only partly upstream),
some pieces like the ACPI declaration of the different nodes and devices
attached to the virtual IOMMU are changing.

This patch introduces a new ACPI table called VIOT, standing as the high
level table overseeing the IORT table and associated subtables.

Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
2020-01-08 09:27:07 +01:00
Rob Bradford
b2589d4f3f vm-virtio, vmm, vfio: Store GuestMemoryMmap in an Arc<ArcSwap<T>>
This allows us to change the memory map that is being used by the
devices via an atomic swap (by replacing the map with another one). The
ArcSwap provides the mechanism for atomically swapping from to another
whilst still giving good read performace. It is inside an Arc so that we
can use a single ArcSwap for all users.

Not covered by this change is replacing the GuestMemoryMmap itself.

This change also removes some vertical whitespace from use blocks in the
files that this commit also changed. Vertical whitespace was being used
inconsistently and broke rustfmt's behaviour of ordering the imports as
it would only do it within the block.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2020-01-02 13:20:11 +00:00
Rob Bradford
a551398135 vmm: device_manager: Use MemoryManager to create KVM mapping
Use the newly exported funtionality to reduce the amount of duplicated
code.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2019-12-23 10:25:40 +00:00
Rob Bradford
962dec2913 vmm: memory_manager: Refactor KVM userspace mapping creation
This function will be useful for other parts of the VMM that also
estabilish their own mappings.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2019-12-23 10:25:40 +00:00
Rob Bradford
7df88793a0 vmm: device_manager: Get device range from MemoryManager
This removes the duplication of these values.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2019-12-23 10:25:40 +00:00
Rob Bradford
61cfe3e72d vmm: Obtain sequential KVM memory slot numbers from MemoryManager
This removes the need to handle a mutable integer and also centralises
the allocation of these slot numbers.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2019-12-23 10:25:40 +00:00
Rob Bradford
260cebb8cf vmm: Introduce MemoryManager
The memory manager is responsible for setting up the guest memory and in
the long term will also handle addition of guest memory.

In this commit move code for creating the backing memory and populating
the allocator into the new implementation trying to make as minimal
changes to other code as possible.

Follow on commits will further reduce some of the duplicated code.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2019-12-23 10:25:40 +00:00
Rob Bradford
d5682cd306 vmm: device_manager: Rewrite if chain using match
To reflect updated clippy rules:

error: `if` chain can be rewritten with `match`
    --> vmm/src/device_manager.rs:1508:25
     |
1508 | /                         if ret > 0 {
1509 | |                             debug!("MSI message successfully delivered");
1510 | |                         } else if ret == 0 {
1511 | |                             warn!("failed to deliver MSI message, blocked by guest");
1512 | |                         }
     | |_________________________^
     |
     = note: `-D clippy::comparison-chain` implied by `-D warnings`
     = help: Consider rewriting the `if` chain to use `cmp` and `match`.
     = help: for further information visit https://rust-lang.github.io/rust-clippy/master/index.html#comparison_chain

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2019-12-20 00:52:03 +01:00
Rob Bradford
21b88c3ea0 vmm: cpu: Rewrite if chain using match
Address updated clippy error:

error: `if` chain can be rewritten with `match`
   --> vmm/src/cpu.rs:668:9
    |
668 | /         if desired_vcpus > self.present_vcpus() {
669 | |             self.activate_vcpus(desired_vcpus, None)?;
670 | |         } else if desired_vcpus < self.present_vcpus() {
671 | |             self.mark_vcpus_for_removal(desired_vcpus)?;
672 | |         }
    | |_________^
    |
    = note: `-D clippy::comparison-chain` implied by `-D warnings`
    = help: Consider rewriting the `if` chain to use `cmp` and `match`.
    = help: for further information visit https://rust-lang.github.io/rust-clippy/master/index.html#comparison_chain

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2019-12-20 00:52:03 +01:00
Rob Bradford
e25a47b32c vmm: device_manager: Remove redundant clones
Address updated clippy errors:

error: redundant clone
   --> vmm/src/device_manager.rs:699:32
    |
699 |             .insert(acpi_device.clone(), 0x3c0, 0x4)
    |                                ^^^^^^^^ help: remove this
    |
    = note: `-D clippy::redundant-clone` implied by `-D warnings`
note: this value is dropped without further use
   --> vmm/src/device_manager.rs:699:21
    |
699 |             .insert(acpi_device.clone(), 0x3c0, 0x4)
    |                     ^^^^^^^^^^^
    = help: for further information visit https://rust-lang.github.io/rust-clippy/master/index.html#redundant_clone

error: redundant clone
   --> vmm/src/device_manager.rs:737:26
    |
737 |             .insert(i8042.clone(), 0x61, 0x4)
    |                          ^^^^^^^^ help: remove this
    |
note: this value is dropped without further use
   --> vmm/src/device_manager.rs:737:21
    |
737 |             .insert(i8042.clone(), 0x61, 0x4)
    |                     ^^^^^
    = help: for further information visit https://rust-lang.github.io/rust-clippy/master/index.html#redundant_clone

error: redundant clone
   --> vmm/src/device_manager.rs:754:29
    |
754 |                 .insert(cmos.clone(), 0x70, 0x2)
    |                             ^^^^^^^^ help: remove this
    |
note: this value is dropped without further use
   --> vmm/src/device_manager.rs:754:25
    |
754 |                 .insert(cmos.clone(), 0x70, 0x2)
    |                         ^^^^
    = help: for further information visit https://rust-lang.github.io/rust-clippy/master/index.html#redundant_clone

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2019-12-20 00:52:03 +01:00
Rob Bradford
a6878accd5 vmm: cpu: Implement CPU removal
When the running OS has been told that a CPU should be removed it will
shutdown the CPU and then signal to the hypervisor via the "_EJ0" method
on the device that ultimately writes into an I/O port than the vCPU
should be shutdown. Upon notification the hypervisor signals to the
individual thread that it should shutdown and waits for that thread to
end.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2019-12-18 08:23:53 +00:00
Rob Bradford
7b3fc72aea vmm: cpu: Notify guest OS that it should offline vCPUs
Allow the resizing of the number of vCPUs to less than the current
active vCPUs. This does not currently remove them from the system but
the kernel will take them offline.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2019-12-18 08:23:53 +00:00
Rob Bradford
7e81b0ded7 vmm: cpu: Create vCPU state for all possible vCPUs
This will make it more straightforward when we attempt to remove vCPUs.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2019-12-18 08:23:53 +00:00
Rob Bradford
156ea392a2 vmm: cpu: Only do ACPI notify on newly added vCPUs
When we add a vCPU set an "inserting" boolean that is exposed as an ACPI
field that will be checked for and reset when the ACPI GED notification
for CPU devices happens.

This change is a precursor for CPU unplug.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2019-12-16 23:57:14 +01:00
Rob Bradford
e8313e3e69 vmm: acpi: Refactor ACPI CPU notification
Continue to notify on all vCPUs but instead separate the notification
functionality into two methods, CSCN that walks through all the CPUs
and CTFY which notifies based on the numerical CPU id. This is an
interim step towards only notifying on changed CPUs and ultimately CPU
removal.

Signed-off-by: Rob Bradford <robert.bradford@intel.com>
2019-12-16 23:57:14 +01:00