The memory-backend-ram should also be set to a numa node instead of
being inserted as a new device. Otherwise it becomes additional memory
and requires explicit online to be available, instead of just being a
backend of the memory specified by -m option.
Signed-off-by: Peng Tao <bergwolf@gmail.com>
This will kill the process when the context is cancelled. As using a nil
context is not permitted it is necessary to substitute with a real
context if it is not initialised in the Config struct.
Signed-off-by: Rob Bradford <robert.bradford@intel.com>
If no RTC is specified in the config then do not generate any RTC command line
options. RTC command line options are optional for QEMU so make Valid() return
false when presented with the empty version of the RTC struct containing empty
strings.
Signed-off-by: Rob Bradford <robert.bradford@intel.com>
`vhostfd` is the vhost file descriptor that holds the socket context ID
`disable-modern` prevents qemu from relying on fast MMIO
Signed-off-by: Julio Montes <julio.montes@intel.com>
Implement function to hotplug vsocks, vsocks are needed
to communicate processes are running inside the VM
with processes are running on the host.
Signed-off-by: Julio Montes <julio.montes@intel.com>
This commit enables the gas security checker on govmm builds. The
security checker has signalled 4 issues all of which I've checked
and have determined to be non issues. These issues are disabled
by this commit.
Signed-off-by: Mark Ryan <mark.d.ryan@intel.com>
This commit enables staticcheck in the travis builds and fixes the existing
errors detected by staticcheck. There was one type of error repeated in
qemu.go in which the type of some constants was not explicitly specified.
Signed-off-by: Mark Ryan <mark.d.ryan@intel.com>
It is useful when we want to specify migration incoming source.
Supported source are fd and exec right now.
Signed-off-by: Peng Tao <bergwolf@gmail.com>
It allows a caller to use a local file as the memory backend of the
guest, and it also allows the file backed memory device to be set shared
or not.
Signed-off-by: Peng Tao <bergwolf@gmail.com>
We need to be able to specify the PCI slot for a bridge while
adding it.
Add test to verify bridge is correctly added.
Signed-off-by: Archana Shinde <archana.m.shinde@intel.com>
This enable data-plane for scsi. All drives attached to the
scsi controller will have their IO processed in a single separate
IO thread instead of qemu's main event loop.
Signed-off-by: Archana Shinde <archana.m.shinde@intel.com>
IOthreads also known as x-data-plane allow IO to
be processed in a separate thread rather than the main event
loop. This produces much better IO throughput and latency.
Signed-off-by: Archana Shinde <archana.m.shinde@intel.com>
certain machines types need to have options to enable or disable features
For example the machine type virt in certain hosts must have the gic version
(gic-version=3 or gic-version=host) to start without problems
Signed-off-by: Julio Montes <julio.montes@intel.com>
VSOCK sockets are added through a vhost PCI device.
It takes a device ID and a context ID, the latter being
the endpoint value to be reached from the host.
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
Introduce basic vhost-user-blk-pci support.
In adding this, cleaned up the QemuParams function to use a more
appropriate switch statement. Similarly, cleanup up the Valid() logic.
We still need to look into parameterization of the block parameter
fields as well as introducing multiqueue support for the vhost-user devices.
Signed-off-by: Eric Ernst <eric.ernst@intel.com>
Some comments were network specific for vhost-user devices, which is
incorect. Fixed these.
Renamed the HWAddress field to be Address, so that it could potentially
be used more generically for non-network based vhost-user types.
Signed-off-by: Eric Ernst <eric.ernst@intel.com>
maxcpus is used to specify how many cpus a VM can have.
This attribute must be specified to enable the hotplugging CPUs capability,
otherwise the maximum number of CPU will be defined by the number of CPU
in -smp.
Signed-off-by: Julio Montes <julio.montes@intel.com>
