All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
- #4907: Bump snapshot version to 4.0.0.
- #4834: Add
VIRTIO_NET_F_RX_MRGBUF
support to thevirtio-net
device. When this feature is negotiated, guestvirtio-net
driver can perform more efficient memory management which in turn improves RX and TX performance. - #4460: Add a
call to
KVM_KVMCLOCK_CTRL
after pausing vCPUs on x86_64 architectures. This ioctl sets a flag in the KVM state of the vCPU indicating that it has been paused by the host userspace. In guests that use kvmclock, the soft lockup watchdog checks this flag. If it is set, it won't trigger the lockup condition. Calling the ioctl for guests that don't use kvmclock will fail. These failures are not fatal. We log the failure and increase thevcpu.kvmclock_ctrl_fails
metric. - #4869: Added
support for Aarch64 systems which feature CPU caches with a number of sets
higher than
u16::MAX
. - #4797, #4854: Added GDB debugging support for a microVM guest kernel. Please see our GDB debugging documentation for more information.
- #4844: Upgrade
virtio-net
device to usereadv
syscall to avoid unnecessary memory copies on RX path, increasing the RX performance.
- #4804: Drop Support for guest kernel 4.14. Linux 4.14 reached end-of-life in January 2024 The minimum supported guest kernel now is 5.10.
- #4796: Fixed
Vsock not notifying guest about
TRANSPORT_RESET_EVENT
event after snapshot restore. This resulted in guest waiting indefinitely on a connection which was reset during snapshot creation. - #4790: v1.9.0 was missing most of the debugging information in the debuginfo file, due to a change in the Cargo defaults. This has been corrected.
- #4826: Add missing configuration of tap offload features when restoring from a snapshot. Setting the features was previously moved from net device creation to device activation time, but it was not reflected in the restore path. This was leading to inability to connect to the restored VM if the offload features were used.
- #4687: Added VMGenID support for microVMs running on ARM hosts with 6.1 guest kernels. Support for VMGenID via DeviceTree bindings exists only on mainline 6.10 Linux onwards. Users of Firecracker will need to backport the relevant patches on top of their 6.1 kernels to make use of the feature.
- #4732, #4733, #4741, #4746: Added official support for 6.1 microVM guest kernels.
- #4743: Added
support for
-h
help flag to the Jailer. The Jailer will now print the help message with either--help
or-h
.
- Support for guest kernel 4.14 is now deprecated. We will completely remove 4.14 support with Firecracker version v1.10
- #4689: Drop support for host kernel 4.14. Linux 4.14 reached end-of-life in January 2024. The minimum supported kernel now is 5.10. Guest kernel 4.14 is still supported.
- 4680: Fixed an issue (#4659) where the virtio-net device implementation would always assume the guest accepts all VirtIO features the device offers. This is always true with the Linux guest kernels we are testing but other kernels, like FreeBSD make different assumptions. This PR fixes the emulation code to set the TAP features based on the features accepted by the guest.
- #4428: Added ACPI support to Firecracker for x86_64 microVMs. Currently, we pass ACPI tables with information about the available vCPUs, interrupt controllers, VirtIO and legacy x86 devices to the guest. This allows booting kernels without MPTable support. Please see our kernel policy documentation for more information regarding relevant kernel configurations.
- #4487: Added support for the Virtual Machine Generation Identifier (VMGenID) device on x86_64 platforms. VMGenID is a virtual device that allows VMMs to notify guests when they are resumed from a snapshot. Linux includes VMGenID support since version 5.18. It uses notifications from the device to reseed its internal CSPRNG. Please refer to snapshot support and random for clones documention for more info on VMGenID. VMGenID state is part of the snapshot format of Firecracker. As a result, Firecracker snapshot version is now 2.0.0.
- #4492: Changed
--config
parameter ofcpu-template-helper
optional. Users no longer need to prepare kernel, rootfs and Firecracker configuration files to usecpu-template-helper
. - #4537 Changed
T2CL template to pass through bit 27 and 28 of
MSR_IA32_ARCH_CAPABILITIES
(RFDS_NO
andRFDS_CLEAR
) since KVM consider they are able to be passed through and T2CL isn't designed for secure snapshot migration between different processors. - #4537 Changed
T2S template to set bit 27 of
MSR_IA32_ARCH_CAPABILITIES
(RFDS_NO
) to 1 since it assumes that the fleet only consists of processors that are not affected by RFDS. - #4388: Avoid
setting
kvm_immediate_exit
to 1 if are already handling an exit, or if the vCPU is stopped. This avoids a spurious KVM exit upon restoring snapshots. - #4567: Do not initialize vCPUs in powered-off state upon snapshot restore. No functional change, as vCPU initialization is only relevant for the booted case (where the guest expects CPUs to be powered off).
- Firecracker's
--start-time-cpu-us
and--start-time-us
parameters are deprecated and will be removed in v2.0 or later. They are used by the jailer to pass the value that should be subtracted from the (CPU) time, when emitting thestart_time_us
andstart_time_cpu_us
metrics. These parameters were never meant to be used by end customers, and we recommend doing any such time adjustments outside Firecracker. - Booting with microVM kernels that rely on MPTable on x86_64 is deprecated and support will be removed in v2.0 or later. We suggest to users of Firecracker to use guest kernels with ACPI support. For x86_64 microVMs, ACPI will be the only way Firecracker passes hardware information to the guest once MPTable support is removed.
- #4526: Added a
check in the network TX path that the size of the network frames the guest
passes to us is not bigger than the maximum frame the device expects to
handle. On the TX path, we copy frames destined to MMDS from guest memory to
Firecracker memory. Without the check, a mis-behaving virtio-net driver could
cause an increase in the memory footprint of the Firecracker process. Now, if
we receive such a frame, we ignore it and increase
Net::tx_malformed_frames
metric. - #4536: Make the first differential snapshot taken after a full snapshot contain only the set of memory pages changed since the full snapshot. Previously, these differential snapshots would contain all memory pages. This will result in potentially much smaller differential snapshots after a full snapshot.
- #4578: Fix UFFD support not being forward-compatible with new ioctl options introduced in Linux 6.6. See also bytecodealliance/userfaultfd-rs#61.
- #4618: On x86_64, when taking a snapshot, if a vCPU has MSR_IA32_TSC_DEADLINE set to 0, Firecracker will replace it with the MSR_IA32_TSC value from the same vCPU. This is to guarantee that the vCPU will continue receiving TSC interrupts after restoring from the snapshot even if an interrupt is lost when taking a snapshot.
- #4666: Fixed
Firecracker sometimes restoring
MSR_IA32_TSC_DEADLINE
beforeMSR_IA32_TSC
. Now it always restoresMSR_IA32_TSC_DEADLINE
MSR afterMSR_IA32_TSC
, as KVM relies on the guest TSC for correct restoration ofMSR_IA32_TSC_DEADLINE
. This fixed guests using theTSC_DEADLINE
hardware feature receiving incorrect timer interrupts after snapshot restoration, which could lead to them seemingly getting stuck in sleep-related syscalls (see also #4099).
- #4346: Added
support to emit aggregate (minimum/maximum/sum) latency for
VcpuExit::MmioRead
,VcpuExit::MmioWrite
,VcpuExit::IoIn
andVcpuExit::IoOut
. The average for these VM exits is not emitted since it can be deduced from the available emitted metrics. - #4360: Added dev-preview support for backing a VM's guest memory by 2M hugetlbfs pages. Please see the documentation for more information
- #4486: Added block and net device metrics for file/tap access latencies and queue backlog lengths, which can be used to analyse saturation of the Firecracker VMM thread and underlying layers. Queue backlog length metrics are flushed periodically. They can be used to esimtate an average queue length by request by dividing its value by the number of requests served.
-
#4230: Changed microVM snapshot format version strategy. Firecracker snapshot format now has a version that is independent of Firecracker version. The current version of the snapshot format is v1.0.0. From now on, the Firecracker binary will define the snapshot format version it supports and it will only be able to load snapshots with format that is backwards compatible with that version. Users can pass the
--snapshot-version
flag to the Firecracker binary to see its supported snapshot version format. This change renders all previous Firecracker snapshots (up to Firecracker version v1.6.0) incompatible with the current Firecracker version. -
#4449: Added information about page size to the payload Firecracker sends to the UFFD handler. Each memory region object now contains a
page_size_kib
field. See also the hugepages documentation. -
#4498: Only use memfd to back guest memory if a vhost-user-blk device is configured, otherwise use anonymous private memory. This is because serving page faults of shared memory used by memfd is slower and may impact workloads.
- #4409: Fixed a bug in the cpu-template-helper that made it panic during conversion of cpu configuration with SVE registers to the cpu template on aarch64 platform. Now cpu-template-helper will print warnings if it encounters SVE registers during the conversion process. This is because cpu templates are limited to only modify registers less than 128 bits.
- #4413: Fixed a bug in the Firecracker that prevented it to restore snapshots of VMs that had SVE enabled.
- #4414: Made
PATCH
requests to the/machine-config
endpoint transactional, meaning Firecracker's configuration will be unchanged if the request returns an error. This fixes a bug where a microVM with incompatible balloon and guest memory size could be booted, due to the check for this condition happening after Firecracker's configuration was updated. - #4259: Added a double fork mechanism in the Jailer to avoid setsid() failures occurred while running Jailer as the process group leader. However, this changed the behaviour of Jailer and now the Firecracker process will always have a different PID than the Jailer process. #4436: Added a "Known Limitations" section in the Jailer docs to highlight the above change in behaviour introduced in PR#4259. #4442: As a solution to the change in behaviour introduced in PR#4259, provided a mechanism to reliably fetch Firecracker PID. With this change, Firecracker process's PID will always be available in the Jailer's root directory regardless of whether new_pid_ns was set.
- #4468: Fixed a bug where a client would hang or timeout when querying for an MMDS path whose content is empty, because the 'Content-Length' header field was missing in a response.
- #4145: Added
support for per net device metrics. In addition to aggregate metrics
net
, each individual net device will emit metrics under the label"net_{iface_id}"
. E.g. the associated metrics for the endpoint"/network-interfaces/eth0"
will be available under"net_eth0"
in the metrics json object. - #4202: Added
support for per block device metrics. In addition to aggregate metrics
block
, each individual block device will emit metrics under the label"block_{drive_id}"
. E.g. the associated metrics for the endpoint"/drives/{drive_id}"
will be available under"block_drive_id"
in the metrics json object. - #4205: Added a
new
vm-state
subcommand toinfo-vmstate
command in thesnapshot-editor
tool to print MicrovmState of vmstate snapshot file in a readable format. Also made thevcpu-states
subcommand available on x86_64. - #4063: Added source-level instrumentation based tracing. See tracing for more details.
- #4138,
#4170,
#4223,
#4247,
#4226: Added
developer preview only (NOT for production use) support for vhost-user
block devices. Firecracker implements a vhost-user frontend. Users are free to
choose from existing open source backend solutions or their own
implementation. Known limitation: snapshotting is not currently supported for
microVMs containing vhost-user block devices. See the
related doc page for details. The
device emits metrics under the label
"vhost_user_{device}_{drive_id}"
.
- #4309: The
jailer's option
--parent-cgroup
will move the process to that cgroup if nocgroup
options are provided. - Simplified and clarified the removal policy of deprecated API elements to follow semantic versioning 2.0.0. For more information, please refer to this GitHub discussion.
- #4180: Refactored error propagation to avoid logging and printing an error on exits with a zero exit code. Now, on successful exit "Firecracker exited successfully" is logged.
- #4194: Removed support for creating Firecracker snapshots targeting older versions of Firecracker. With this change, running 'firecracker --version' will not print the supported snapshot versions.
- #4301: Allow
merging of diff snapshots into base snapshots by directly writing the diff
snapshot on top of the base snapshot's memory file. This can be done by
setting the
mem_file_path
to the path of the pre-existing full snapshot.
- #4209:
rebase-snap
tool is now deprecated. Users should usesnapshot-editor
for rebasing diff snapshots.
- #4171: Fixed a
bug that ignored the
--show-log-origin
option, preventing it from printing the source code file of the log messages. - #4178: Fixed a
bug reporting a non-zero exit code on successful shutdown when starting
Firecracker with
--no-api
. - #4261: Fixed a bug where Firecracker would log "RunWithApiError error: MicroVMStopped without an error: GenericError" when exiting after encountering an emulation error. It now correctly prints "RunWithApiError error: MicroVMStopped with an error: GenericError".
- #4242: Fixed a
bug introduced in #4047 that limited the
--level
option of logger to Pascal-cased values (e.g. accepting "Info", but not "info"). It now ignores case again. - #4286: Fixed a bug in the asynchronous virtio-block engine that rendered the device non-functional after a PATCH request was issued to Firecracker for updating the path to the host-side backing file of the device.
- #4301: Fixed a
bug where if Firecracker was instructed to take a snapshot of a microvm which
itself was restored from a snapshot, specifying
mem_file_path
to be the path of the memory file from which the microvm was restored would result in both the microvm and the snapshot being corrupted. It now instead performs a "write-back" of all memory that was updated since the snapshot was originally loaded.
- #3837: Added official support for Linux 6.1. See prod-host-setup for some security and performance considerations.
- #4045 and
#4075: Added
snapshot-editor
tool for modifications of snapshot files. It allows for rebasing of memory snapshot files, printing and removing aarch64 registers from the vmstate and obtaining snapshot version. - #3967: Added
new fields to the custom CPU templates. (aarch64 only)
vcpu_features
field allows modifications of vCPU features enabled during vCPU initialization.kvm_capabilities
field allows modifications of KVM capability checks that Firecracker performs during boot. If any of these fields are in use, minimal target snapshot version is restricted to 1.5.
- Updated deserialization of
bitmap
for custom CPU templates to allow usage of '_' as a separator. - Changed the strip feature of
cpu-template-helper
tool to operate bitwise. - Better logs during validation of CPU ID in snapshot restoration path. Also Firecracker now does not fail if it can't get CPU ID from the host or can't find CPU ID in the snapshot.
- Changed the serial device to only try to initialize itself if stdin is a terminal or a FIFO pipe. This fixes logged warnings about the serial device failing to initialize if the process is daemonized (in which case stdin is /dev/null instead of a terminal).
- Changed to show a warning message when launching a microVM with C3 template on a processor prior to Intel Cascade Lake, because the guest kernel does not apply the mitigation against MMIO stale data vulnerability when it is running on a processor that does not enumerate FBSDP_NO, PSDP_NO and SBDR_SSDP_NO on IA32_ARCH_CAPABILITIES MSR.
- Made Firecracker resize its file descriptor table on process start. It now
preallocates the in-kernel fdtable to hold
RLIMIT_NOFILE
many fds (or 2048 if no limit is set). This avoids the kernel reallocating the fdtable during Firecracker operations, resulting in a 30ms to 70ms reduction of snapshot restore times for medium to large microVMs with many devices attached. - Changed the dump feature of
cpu-template-helper
tool not to enumerate program counter (PC) on ARM because it is determined by the given kernel image and it is useless in the custom CPU template context. - The ability to create snapshots for an older version of Firecracker is now
deprecated. As a result, the
version
body field inPUT
on/snapshot/create
request in deprecated. - Added support for the /dev/userfaultfd device available on linux kernels >= 6.1. This is the default for creating UFFD handlers on these kernel versions. If it is unavailable, Firecracker falls back to the userfaultfd syscall.
- Deprecated
cpu_template
field inPUT
andPATCH
requests on/machine-config
API, which is used to set a static CPU template. Custom CPU templates added in v1.4.0 are available as an improved iteration of the static CPU templates. For more information about the transition from static CPU templates to custom CPU templates, please refer to this GitHub discussion. - Changed default log level from
Warn
toInfo
. This results in more logs being output by default.
- Fixed a change in behavior of normalize host brand string that breaks Firecracker on external instances.
- Fixed the T2A CPU template not to unset the MMX bit (CPUID.80000001h:EDX[23]) and the FXSR bit (CPUID.80000001h:EDX[24]).
- Fixed the T2A CPU template to set the RstrFpErrPtrs bit (CPUID.80000008h:EBX[2]).
- Fixed a bug where Firecracker would crash during boot if a guest set up a virtio queue that partially overlapped with the MMIO gap. Now Firecracker instead correctly refuses to activate the corresponding virtio device.
- Fixed the T2CL CPU template to pass through security mitigation bits that are listed by KVM as bits able to be passed through. By making the most use of the available hardware security mitigations on a processor that a guest is running on, the guest might be able to benefit from performance improvements.
- Fixed the T2S CPU template to set the GDS_NO bit of the IA32_ARCH_CAPABILITIES MSR to 1 in accordance with an Intel microcode update. To use the template securely, users should apply the latest microcode update on the host.
- Fixed the spelling of the
nomodule
param passed in the default kernel command line parameters. This is a breaking change for setups that use the default kernel command line which also depend on being able to load kernel modules at runtime. This may also break setups which use the default kernel command line and which use an init binary that inadvertently depends on the misspelled param ("nomodules") being present at the command line, since this param will no longer be passed.
- Added support for custom CPU templates allowing users to adjust vCPU features exposed to the guest via CPUID, MSRs and ARM registers.
- Introduced V1N1 static CPU template for ARM to represent Neoverse V1 CPU as Neoverse N1.
- Added support for the
virtio-rng
entropy device. The device is optional. A single device can be enabled per VM using the/entropy
endpoint. - Added a
cpu-template-helper
tool for assisting with creating and managing custom CPU templates.
- Set FDP_EXCPTN_ONLY bit (CPUID.7h.0:EBX[6]) and ZERO_FCS_FDS bit (CPUID.7h.0:EBX[13]) in Intel's CPUID normalization process.
- Fixed feature flags in T2S CPU template on Intel Ice Lake.
- Fixed CPUID leaf 0xb to be exposed to guests running on AMD host.
- Fixed a performance regression in the jailer logic for closing open file descriptors. Related to: #3542.
- A race condition that has been identified between the API thread and the VMM
thread due to a misconfiguration of the
api_event_fd
. - Fixed CPUID leaf 0x1 to disable perfmon and debug feature on x86 host.
- Fixed passing through cache information from host in CPUID leaf 0x80000006.
- Fixed the T2S CPU template to set the RRSBA bit of the IA32_ARCH_CAPABILITIES MSR to 1 in accordance with an Intel microcode update.
- Fixed the T2CL CPU template to pass through the RSBA and RRSBA bits of the IA32_ARCH_CAPABILITIES MSR from the host in accordance with an Intel microcode update.
- Fixed passing through cache information from host in CPUID leaf 0x80000005.
- Fixed the T2A CPU template to disable SVM (nested virtualization).
- Fixed the T2A CPU template to set EferLmsleUnsupported bit (CPUID.80000008h:EBX[20]), which indicates that EFER[LMSLE] is not supported.
- Introduced T2CL (Intel) and T2A (AMD) CPU templates to provide instruction set feature parity between Intel and AMD CPUs when using these templates.
- Added Graviton3 support (c7g instance type).
- Improved error message when invalid network backend provided.
- Improved TCP throughput by between 5% and 15% (depending on CPU) by using scatter-gather I/O in the net device's TX path.
- Upgraded Rust toolchain from 1.64.0 to 1.66.0.
- Made seccompiler output bit-reproducible.
- Fixed feature flags in T2 CPU template on Intel Ice Lake.
- Added a new CPU template called
T2S
. This exposes the same CPUID asT2
to the Guest and also overwrites theARCH_CAPABILITIES
MSR to expose a reduced set of capabilities. With regards to hardware vulnerabilities and mitigations, the Guest vCPU will apear to look like a Skylake CPU, making it safe to snapshot uVMs running on a newer host CPU (Cascade Lake) and restore on a host that has a Skylake CPU. - Added a new CLI option
--metrics-path PATH
. It accepts a file parameter where metrics will be sent to. - Added baselines for m6i.metal and m6a.metal for all long running performance tests.
- Releases now include debuginfo files.
- Changed the jailer option
--exec-file
to fail if the filename does not contain the stringfirecracker
to prevent from running non-firecracker binaries. - Upgraded Rust toolchain from 1.52.1 to 1.64.0.
- Switched to specifying our dependencies using caret requirements instead of comparison requirements.
- Updated all dependencies to their respective newest versions.
- Made the
T2
template more robust by explicitly disabling additional CPUID flags that should be off but were missed initially or that were not available in the spec when the template was created. - Now MAC address is correctly displayed when queried with GET
/vm/config
if left unspecified in both pre and post snapshot states. - Fixed a self-DoS scenario in the virtio-queue code by reporting and terminating execution when the number of available descriptors reported by the driver is higher than the queue size.
- Fixed the bad handling of kernel cmdline parameters when init arguments were
provided in the
boot_args
field of the JSON body of the PUT/boot-source
request. - Fixed a bug on ARM64 hosts where the upper 64bits of the V0-V31 FL/SIMD registers were not saved correctly when taking a snapshot, potentially leading to data loss. This change invalidates all ARM64 snapshots taken with versions of Firecracker <= 1.1.3.
- Improved stability and security when saving CPU MSRs in snapshots.
- The API
PATCH
methods formachine-config
can now be used to reset thecpu_template
to"None"
. Until this change there was no way to reset thecpu_template
once it was set. - Added a
rebase-snap
tool for rebasing a diff snapshot over a base snapshot. - Mmds version is persisted across snapshot-restore. Snapshot compatibility is preserved bidirectionally, to and from a Firecracker version that does not support persisting the Mmds version. In such cases, the default V1 option is used.
- Added
--mmds-size-limit
for limiting the mmds data store size instead of piggy-backing on--http-api-max-payload-size
. If left unconfigured it defaults to the value of--http-api-max-payload-size
, to provide backwards compatibility. - Added optional
mem_backend
body field inPUT
requests on/snapshot/load
. This new parameter is an object that defines the configuration of the backend responsible for handling memory loading during snapshot restore. Themem_backend
parameter containsbackend_type
andbackend_path
required fields.backend_type
is an enum that can take eitherFile
orUffd
as value. Interpretation ofbackend_path
field depends on the value ofbackend_type
. IfFile
, then the user must provide the path to file that contains the guest memory to be loaded. Otherwise, ifbackend_type
isUffd
, thenbackend_path
is the path to a unix domain socket where a custom page fault handler process is listening and expecting a UFFD to be sent by Firecracker. The UFFD is used to handle the guest memory page faults in the separate process. - Added logging for the snapshot/restore and async block device IO engine features to indicate they are in development preview.
- The API
PATCH
method for/machine-config
can be now used to changetrack_dirty_pages
on aarch64. - MmdsV2 is now Generally Available.
- MmdsV1 is now deprecated and will be removed in Firecracker v2.0.0. Use MmdsV2 instead.
- Deprecated
mem_file_path
body field inPUT
on/snapshot/load
request.
- Fixed inconsistency that allowed the start of a microVM from a JSON file
without specifying the
vcpu_count
andmem_size_mib
parameters formachine-config
although they are mandatory when configuring via the API. Now these fields are mandatory when specifyingmachine-config
in the JSON file and when using thePUT
request on/machine-config
. - Fixed inconsistency that allowed a user to specify the
cpu_template
parameter and setsmt
toTrue
inmachine-config
when starting from a JSON file on aarch64 even though they are not permitted when usingPUT
orPATCH
in the API. Now Firecracker will return an error on aarch64 ifsmt
is set toTrue
or ifcpu_template
is specified. - Fixed inconsistent behaviour of the
PUT
method for/machine-config
that would reset thetrack_dirty_pages
parameter tofalse
if it was not specified in the JSON body of the request, but left thecpu_template
parameter intact if it was not present in the request. Now aPUT
request for/machine-config
will reset all optional parameters (smt
,cpu_template
,track_dirty_pages
) to their default values if they are not specified in thePUT
request. - Fixed incosistency in the swagger definition with the current state of the
/vm/config
endpoint.
- Added jailer option
--parent-cgroup <relative_path>
to allow the placement of microvm cgroups in custom cgroup nested hierarchies. The default value is<exec-file>
which is backwards compatible to the behavior before this change. - Added jailer option
--cgroup-version <1|2>
to support running the jailer on systems that have cgroup-v2. Default value is1
which means that if--cgroup-version
is not specified, the jailer will try to create cgroups on cgroup-v1 hierarchies only. - Added
--http-api-max-payload-size
parameter to configure the maximum payload size for PUT and PATCH requests. - Limit MMDS data store size to
--http-api-max-payload-size
. - Cleanup all environment variables in Jailer.
- Added metrics for accesses to deprecated HTTP and command line API endpoints.
- Added permanent HTTP endpoint for
GET
on/version
for getting the Firecracker version. - Added
--metadata
parameter to enable MMDS content to be supplied from a file allowing the MMDS to be used when using--no-api
to disable the API server. - Checksum file for the release assets.
- Added support for custom headers to MMDS requests. Accepted headers are:
X-metadata-token
, which accepts a string value that provides a session token for MMDS requests; andX-metadata-token-ttl-seconds
, which specifies the lifetime of the session token in seconds. - Support and validation for host and guest kernel 5.10.
- A kernel support policy.
- Added
io_engine
to the pre-boot block device configuration. Possible values:Sync
(the default option) orAsync
(only available for kernels newer than 5.10.51). TheAsync
variant introduces a block device engine that uses io_uring for executing requests asynchronously, which is in developer preview (NOT for production use). Seedocs/api_requests/block-io-engine.md
. - Added
block.io_engine_throttled_events
metric for measuring the number of virtio events throttled because of the IO engine. - New optional
version
field to PUT requests towards/mmds/config
to configure MMDS version. Accepted values areV1
andV2
and default isV1
. MMDSV2
is developer preview only (NOT for production use) and it does not currently work after snapshot load. - Mandatory
network_interfaces
field to PUT requests towards/mmds/config
which contains a list of network interface IDs capable of forwarding packets to MMDS.
- Removed the
--node
jailer parameter. - Deprecated
vsock_id
body field inPUT
s on/vsock
. - Removed the deprecated the
--seccomp-level parameter
. GET
requests to MMDS require a session token to be provided throughX-metadata-token
header when using V2.- Allow
PUT
requests to MMDS in order to generate a session token to be used for futureGET
requests when version 2 is used. - Remove
allow_mmds_requests
field from the request body that attaches network interfaces. Specifying interfaces that allow forwarding requests to MMDS is done by adding the network interface's ID to thenetwork_interfaces
field of PUT/mmds/config
request's body. - Renamed
/machine-config
ht_enabled
tosmt
. smt
field is now optional on PUT/machine-config
, defaulting tofalse
.- Configuring
smt: true
on aarch64 via the API is forbidden.
- GET
/vm/config
was returning a default config object after restoring from a snapshot. It now correctly returns the config of the original microVM, except for boot_config and the cpu_template and smt fields of the machine config, which are currently lost. - Fixed incorrect propagation of init parameters in kernel commandline. Related to: #2709.
- Adapt T2 and C3 CPU templates for kernel 5.10. Firecracker was not previously
masking some CPU features of the host or emulated by KVM, introduced in more
recent kernels:
umip
,vmx
,avx512_vnni
. - Fix jailer's cgroup implementation to accept properties that contain multiple dots.
- Added devtool build
--ssh-keys
flag to support fetching from private git repositories. - Added option to configure block device flush.
- Added
--new-pid-ns
flag to the Jailer in order to spawn the Firecracker process in a new PID namespace. - Added API metrics for
GET
,PUT
andPATCH
requests on/mmds
endpoint. - Added
--describe-snapshot
flag to Firecracker to fetch the data format version of a snapshot state file provided as argument. - Added
--no-seccomp
parameter for disabling the default seccomp filters. - Added
--seccomp-filter
parameter for supplying user-provided, custom filters. - Added the
seccompiler-bin
binary that is used to compile JSON seccomp filters into serialized BPF for Firecracker consumption. - Snapshotting support for GICv2 enabled guests.
- Added
devtool install
to deploy built binaries in/usr/local/bin
or a given path. - Added code logic to send
VIRTIO_VSOCK_EVENT_TRANSPORT_RESET
on snapshot creation, when the Vsock device is active. The event will close active connections on the guest. - Added
GET
request on/vm/config
that provides full microVM configuration as a JSON HTTP response. - Added
--resource-limit
flag to jailer to limit resources such as: number of file descriptors allowed at a time (with a default value of 2048) and maximum size of files created by the process.
- Changed Docker images repository from DockerHub to Amazon ECR.
- Fixed off-by-one error in virtio-block descriptor address validation.
- Changed the
PATCH
request on/balloon/statistics
to schedule the first statistics update immediately after processing the request. - Deprecated the
--seccomp-level parameter
. It will be removed in a future release. Using it logs a runtime warning. - Experimental gnu libc builds use empty default seccomp filters, allowing all system calls.
- Fixed non-compliant check for the RTC device ensuring a fixed 4-sized data buffer.
- Unnecessary interrupt assertion was removed from the RTC. However, a dummy interrupt is still allocated for snapshot compatibility reasons.
- Fixed the SIGPIPE signal handler so Firecracker no longer exits. The signal is still recorded in metrics and logs.
- Fixed ballooning API definitions by renaming all fields which mentioned "MB" to use "MiB" instead.
- Snapshot related host files (vm-state, memory, block backing files) are now flushed to their backing mediums as part of the CreateSnapshot operation.
- Fixed the SSBD mitigation not being enabled on
aarch64
with the providedprod-host-setup.md
. - Fixed the balloon statistics not working after a snapshot restore event.
- The
utc_timestamp_ms
now reports the timestamp in ms from the UTC UNIX Epoch, as the name suggests. It was previously using a monotonic clock with an undefined starting point.
- Added optional
resume_vm
field to/snapshot/load
API call. - Added support for block rate limiter PATCH.
- Added devtool test
-c|--cpuset-cpus
flag for cpus confinement when tests run. - Added devtool test
-m|--cpuset-mems
flag for memory confinement when tests run. - Added the virtio traditional memory ballooning device.
- Added a mechanism to handle vCPU/VMM errors that result in process termination.
- Added incremental guest memory snapshot support.
- Added aarch64 snapshot support.
- Change the information provided in
DescribeInstance
command to provide microVM state information (Not started/Running/Paused) instead of whether it's started or not. - Removed the jailer
--extra-args
parameter. It was a noop, having been replaced by the--
separator for extra arguments. - Changed the output of the
--version
command line parameter to include a list of supported snapshot data format versions for the firecracker binary. - Increased the maximum number of virtio devices from 11 to 19.
- Added a new check that prevents creating v0.23 snapshots when more than 11 devices are attached.
- If the stdout buffer is full and non-blocking, the serial writes no longer
block. Any new bytes will be lost, until the buffer is freed. The device also
logs these errors and increments the
uart.error_count
metric for each lost byte.
- Fixed inconsistency in YAML file InstanceInfo definition
- Added metric for throttled block device events.
- Added metrics for counting rate limiter throttling events.
- Added metric for counting MAC address updates.
- Added metrics for counting TAP read and write errors.
- Added metrics for counting RX and TX partial writes.
- Added metrics that measure the duration of pausing and resuming the microVM, from the VMM perspective.
- Added metric for measuring the duration of the last full/diff snapshot created, from the VMM perspective.
- Added metric for measuring the duration of loading a snapshot, from the VMM perspective.
- Added metrics that measure the duration of pausing and resuming the microVM, from the API (user) perspective.
- Added metric for measuring the duration of the last full/diff snapshot created, from the API (user) perspective.
- Added metric for measuring the duration of loading a snapshot, from the API (user) perspective.
- Added
track_dirty_pages
field tomachine-config
. If enabled, Firecracker can create incremental guest memory snapshots by saving the dirty guest pages in a sparse file. - Added a new API call,
PATCH /vm
, for changing the microVM state (toPaused
orResumed
). - Added a new API call,
PUT /snapshot/create
, for creating a full or diff snapshot. - Added a new API call,
PUT /snapshot/load
, for loading a snapshot. - Added new jailer command line argument
--cgroup
which allow the user to specify the cgroups that are going to be set by the Jailer. - Added full support for AMD CPUs (General Availability). More details here.
- Boot time on AMD achieves the desired performance (i.e under 150ms).
- The logger
level
field is now case-insensitive. - Disabled boot timer device after restoring a snapshot.
- Enabled boot timer device only when specifically requested, by using the
--boot-timer
dedicated cmdline parameter. - firecracker and jailer
--version
now gets updated on each devtool build to the output ofgit describe --dirty
, if the git repo is available. - MicroVM process is only attached to the cgroups defined by using
--cgroups
or the ones defined indirectly by using--node
. - Changed
devtool build
to build jailer binary formusl
only targets. Building jailer binary fornon-musl
targets have been removed.
- Added a new API call,
PUT /metrics
, for configuring the metrics system. - Added
app_name
field in InstanceInfo struct for storing the application name. - New command-line parameters for
firecracker
, named--log-path
,--level
,--show-level
and--show-log-origin
that can be used for configuring the Logger when starting the process. When using this method for configuration, only--log-path
is mandatory. - Added a guide for updating the dev container image.
- Added a new API call,
PUT /mmds/config
, for configuring theMMDS
with a custom valid link-local IPv4 address. - Added experimental JSON response format support for MMDS guest applications requests.
- Added metrics for the vsock device.
- Added
devtool strip
command which removes debug symbols from the release binaries. - Added the
tx_malformed_frames
metric for the virtio net device, emitted when a TX frame missing the VNET header is encountered.
- Added
--version
flag to both Firecracker and Jailer. - Return
405 Method Not Allowed
MMDS response for non HTTPGET
MMDS requests originating from guest. - Fixed folder permissions in the jail (#1802).
- Any number of whitespace characters are accepted after ":" when parsing HTTP headers.
- Potential panic condition caused by the net device expecting to find a VNET header in every frame.
- Potential crash scenario caused by "Content-Length" HTTP header field accepting negative values.
- Fixed #1754 - net: traffic blocks when running ingress UDP performance tests with very large buffers.
- Updated CVE-2019-3016 mitigation information in Production Host Setup
- In case of using an invalid JSON as a 'config-file' for Firecracker, the process will exit with return code 152.
- Removed the
testrun.sh
wrapper. - Removed
metrics_fifo
field from the logger configuration. - Renamed
log_fifo
field from LoggerConfig tolog_path
andmetrics_fifo
field from MetricsConfig tometrics_path
. PATCH /drives/{id}
only allowed post-boot. UsePUT
for pre-boot updates to existing configurations.PATCH /network-interfaces/{id}
only allowed post-boot. UsePUT
for pre-boot updates to existing configurations.- Changed returned status code from
500 Internal Server Error
to501 Not Implemented
, for queries on the MMDS endpoint in IMDS format, when the requested resource value type is unsupported. - Allowed the MMDS data store to be initialized with all supported JSON types. Retrieval of these values within the guest, besides String, Array, and Dictionary, is only possible in JSON mode.
PATCH
request on/mmds
before the data store is initialized returns403 BadRequest
.- Segregated MMDS documentation in MMDS design documentation and MMDS user guide documentation.
- Support for booting with an initial RAM disk image. This image can be
specified through the new
initrd_path
field of the/boot-source
API request.
- Fixed #1469 - Broken GitHub location for Firecracker release binary.
- The jailer allows changing the default api socket path by using the extra arguments passed to firecracker.
- Fixed #1456 - Occasional KVM_EXIT_SHUTDOWN and bad syscall (14) during VM shutdown.
- Updated the production host setup guide with steps for addressing CVE-2019-18960.
- The HTTP header parsing is now case insensitive.
- The
put_api_requests
andpatch_api_requests
metrics for net devices were un-swapped.
- Removed redundant
--seccomp-level
jailer parameter since it can be simply forwarded to the Firecracker executable using "end of command options" convention. - Removed
memory.dirty_pages
metric. - Removed
options
field from the logger configuration. - Decreased release binary size by ~15%.
- Changed default API socket path to
/run/firecracker.socket
. This path also applies when running with the jailer. - Disabled KVM dirty page tracking by default.
- Removed redundant RescanBlockDevice action from the /actions API. The
functionality is available through the PATCH /drives API. See
docs/api_requests/patch-block.md
.
- Added support for GICv2.
- Fixed CVE-2019-18960 - Fixed a logical error in bounds checking performed on vsock virtio descriptors.
- Fixed #1283 - Can't start a VM in AARCH64 with vcpus number more than 16.
- Fixed #1088 - The backtrace are printed on
panic
, no longer causing a seccomp fault. - Fixed #1375 - Change logger options type from
Value
toVec<LogOption>
to prevent potential unwrap on None panics. - Fixed #1436 - Raise interrupt for TX queue used descriptors
- Fixed #1439 - Prevent achieving 100% cpu load when the net device rx is throttled by the ratelimiter
- Fixed #1437 - Invalid fields in rate limiter related API requests are now failing with a proper error message.
- Fixed #1316 - correctly determine the size of a virtio device backed by a block device.
- Fixed #1383 - Log failed api requests.
- Decreased release binary size by 10%.
- New command-line parameter for
firecracker
, named--no-api
, which will disable the API server thread. If set, the user won't be able to send any API requests, neither before, nor after the vm has booted. It must be paired with--config-file
parameter. Also, when API server is disabled, MMDS is no longer available now. - New command-line parameter for
firecracker
, named--config-file
, which represents the path to a file that contains a JSON which can be used for configuring and starting a microVM without sending any API requests. - The jailer adheres to the "end of command options" convention, meaning all
parameters specified after
--
are forwarded verbatim to Firecracker. - Added
KVM_PTP
support to the recommended guest kernel config. - Added entry in FAQ.md for Firecracker Guest timekeeping.
- Vsock API call:
PUT /vsocks/{id}
changed toPUT /vsock
and no longer appear to support multiple vsock devices. Any subsequent calls to this API endpoint will override the previous vsock device configuration. - Removed unused 'Halting' and 'Halted' instance states.
- Vsock host-initiated connections now implement a trivial handshake protocol. See the vsock doc for details. Related to: #1253, #1432, #1443
- Fixed serial console on aarch64 (GitHub issue #1147).
- Upon panic, the terminal is now reset to canonical mode.
- Explicit error upon failure of vsock device creation.
- The failure message returned by an API call is flushed in the log FIFOs.
- Insert virtio devices in the FDT in order of their addresses sorted from low to high.
- Enforce the maximum length of the network interface name to be 16 chars as specified in the Linux Kernel.
- Changed the vsock property
id
tovsock_id
so that the API client can be successfully generated from the swagger definition.
- New device: virtio-vsock, backed by Unix domain sockets (GitHub issue #650).
See
docs/vsock.md
.
- No error is thrown upon a flush metrics intent if logger has not been configured.
- Updated the documentation for integration tests.
- Fixed high CPU usage before guest network interface is brought up (GitHub issue #1049).
- Fixed an issue that caused the wrong date (month) to appear in the log.
- Fixed a bug that caused the seccomp filter to reject legit syscalls in some rare cases (GitHub issue #1206).
- Docs: updated the production host setup guide.
- Docs: updated the rootfs and kernel creation guide.
- Removed experimental support for vhost-based vsock devices.
- New API call:
PATCH /machine-config/
, used to update VM configuration, before the microVM boots. - Added an experimental swagger definition that includes the specification for the vsock API call.
- Added a signal handler for
SIGBUS
andSIGSEGV
that immediately terminates the process upon intercepting the signal. - Added documentation for signal handling utilities.
- Added [alpha] aarch64 support.
- Added metrics for successful read and write operations of MMDS, Net and Block devices.
vcpu_count
,mem_size_mib
andht_enabled
have been changed to be mandatory forPUT
requests on/machine-config/
.- Disallow invalid seccomp levels by exiting with error.
- Incorrect handling of bind mounts within the jailed rootfs.
- Corrected the guide for
Alpine
guest setup.
- Added [alpha] AMD support.
- New
devtool
command:prepare_release
. This updates the Firecracker version, crate dependencies and credits in preparation for a new release. - New
devtool
command:tag
. This creates a new git tag for the specified release number, based on the changelog contents. - New doc section about building with glibc.
- Dropped the JSON-formatted
context
command-line parameter from Firecracker in favor of individual classic command-line parameters. - When running with
jailer
the location of the API socket has changed to<jail-root-path>/api.socket
(API socket was moved inside the jail). PUT
andPATCH
requests on/mmds
with data containing any value type other thanString
,Array
,Object
will returns status code 400.- Improved multiple error messages.
- Removed all kernel modules from the recommended kernel config.
- Corrected the seccomp filter when building with glibc.
- Removed the
seccomp.bad_syscalls
metric.
- Corrected the conditional compilation of the seccomp rule for
madvise
.
- A
madvise
call issued by themusl
allocator was added to the seccomp allow list to prevent Firecracker from terminating abruptly when allocating memory in certain conditions.
- New API action: SendCtrlAltDel, used to initiate a graceful shutdown, if the guest has driver support for i8042 and AT Keyboard. See the docs for details.
- New metric counting the number of egress packets with a spoofed MAC:
net.tx_spoofed_mac_count
. - New API call:
PATCH /network-interfaces/
, used to update the rate limiters on a network interface, after the start of a microVM.
- Added missing
vmm_version
field to the InstanceInfo API swagger definition, and marked several other mandatory fields as such. - New default command line for guest kernel:
reboot=k panic=1 pci=off nomodules 8250.nr_uarts=0 i8042.noaux i8042.nomux i8042.nopnp i8042.dumbkbd
.
- virtio-blk: VIRTIO_BLK_T_FLUSH now working as expected.
- Vsock devices can be attached when starting Firecracker using the jailer.
- Vsock devices work properly when seccomp filtering is enabled.
- Documentation for development environment setup on AWS in
dev-machine-setup.md
. - Documentation for microVM networking setup in
docs/network-setup.md
. - Limit the maximum supported vCPUs to 32.
- Log the app version when the
Logger
is initialized. - Pretty print panic information.
- Firecracker terminates with exit code 148 when a syscall which is not present in the allow list is intercepted.
- Fixed build with the
vsock
feature.
- Documentation for Logger API Requests in
docs/api_requests/logger.md
. - Documentation for Actions API Requests in
docs/api_requests/actions.md
. - Documentation for MMDS in
docs/mmds.md
. - Flush metrics on request via a PUT
/actions
with theaction_type
field set toFlushMetrics
.
- Updated the swagger definition of the
Logger
to specify the required fields and provide default values for optional fields. - Default
seccomp-level
is2
(was previously 0). - API Resource IDs can only contain alphanumeric characters and underscores.
- Seccomp filters are now applied to all Firecracker threads.
- Enforce minimum length of 1 character for the jailer ID.
- Exit with error code when starting the jailer process fails.
- Removed
InstanceHalt
from the list of possible actions.
- The
/logger
API has a new field calledoptions
. This is an array of strings that specify additional logging configurations. The only supported value isLogDirtyPages
. - When the
LogDirtyPages
option is configured viaPUT /logger
, a new metric calledmemory.dirty_pages
is computed as the number of pages dirtied by the guest since the last time the metric was flushed. - Log messages on both graceful and forceful termination.
- Availability of the list of dependencies for each commit inside the code base.
- Documentation on vsock experimental feature and host setup recommendations.
PUT
requests on/mmds
always return 204 on success.PUT
operations on/network-interfaces
API resources no longer accept the previously requiredstate
parameter.- The jailer starts with
--seccomp-level=2
(was previously 0) by default. - Log messages use
anonymous-instance
as instance id if none is specified.
- Fixed crash upon instance start on hosts without 1GB huge page support.
- Fixed "fault_message" inconsistency between Open API specification and code base.
- Ensure MMDS compatibility with C5's IMDS implementation.
- Corrected the swagger specification to ensure
OpenAPI 2.0
compatibility.
- Apache-2.0 license
- Docs:
- Experimental vhost-based vsock implementation.
- Improved MMDS network stack performance.
- If the logging system is not yet initialized (via
PUT /logger
), log events are now sent to stdout/stderr. - Moved the
instance_info_fails
metric underget_api_requests
- Improved readme and added links to more detailed information, now featured in subject-specific docs.
- Fixed bug in the MMDS network stack, that caused some RST packets to be sent without a destination.
- Fixed bug in
PATCH /drives
, whereby the ID in the path was not checked against the ID in the body.
- The Swagger definition was corrected.
- Each Firecracker process has an associated microVM Metadata Store (MMDS). Its
contents can be configured using the
/mmds
API resource.
- The boot source is specified only with the
kernel_image_path
and the optional parameterboot_args
. All other fields are removed. - The
path_on_host
property in the drive specification is now marked as mandatory. - PATCH drive only allows patching/changing the
path_on_host
property. - All PUT and PATCH requests return the status code 204.
- CPUID brand string (aka model name) now includes the host CPU frequency.
- API requests which add guest network interfaces have an additional parameter,
allow_mmds_requests
which defaults tofalse
. - Stopping the guest (e.g. using the
reboot
command) also terminates the Firecracker process. When the Firecracker process ends for any reason, (other thankill -9
), metrics are flushed at the very end. - On startup
jailer
closes all inherited file descriptors based onsysconf(_SC_OPEN_MAX)
except input, output and error. - The microVM ID prefixes each Firecracker log line. This ID also appears in the
process
cmdline
so it's now possible tops | grep <ID>
for it.
- Seccomp filtering is configured via the
--seccomp-level
jailer parameter. - Firecracker logs the starting addresses of host memory areas provided as guest memory slots to KVM.
- The metric
panic_count
gets incremented to signal that a panic has occurred. - Firecracker logs a backtrace when it crashes following a panic.
- Added basic instrumentation support for measuring boot time.
StartInstance
is a synchronous API request (it used to be an asynchronous request).
- Ensure that fault messages sent by the API have valid JSON bodies.
- Use HTTP response code 500 for internal Firecracker errors, and 400 for user errors on InstanceStart.
- Serialize the machine configuration fields to the correct data types (as specified in the Swagger definition).
- NUMA node assignment is properly enforced by the jailer.
- The
is_root_device
andis_read_only
properties are now marked as required in the Swagger definition ofDrive
object properties.
GET
requests on the/actions
API resource are no longer supported.- The metrics associated with asynchronous actions have been removed.
- Remove the
action_id
parameter forInstanceStart
, both from the URI and the JSON request body.
- The jailer can now be configured to enter a preexisting network namespace, and to run as a daemon.
- Enabled PATCH operations on
/drives
resources.
- The microVM
id
supplied to the jailer may now contain alphanumeric characters and hyphens, up to a maximum length of 64 characters. - Replaced the
permissions
property of/drives
resources with a boolean. - Removed the
state
property of/drives
resources.
- Rate limiting functionality allows specifying an initial one time burst size.
- Firecracker can now boot from an arbitrary boot partition by specifying its unique id in the driver's API call.
- Block device rescan is triggered via a PUT
/actions
with the drive ID in the action body'spayload
field and theaction_type
field set toBlockDeviceRescan
.
- Removed
noapic
from the default guest kernel command line. - The
action_id
parameter is no longer required for synchronous PUT requests to/actions
. - PUT requests are no longer allowed on
/drives
resources after the guest has booted.
- Fixed guest instance kernel loader to accelerate vCPUs launch and consequently guest kernel boot.
- Fixed network emulation to improve IO performance.
- Firecracker uses two different named pipes to record human readable logs and metrics, respectively.
- Seccomp filtering can be enabled via setting the
USE_SECCOMP
environment variable. - It is possible to supply only a partial specification when attaching a rate limiter (i.e. just the bandwidth or ops parameter).
- Errors related to guest network interfaces are now more detailed.
- Fixed a bug that was causing Firecracker to panic whenever a
PUT
request was sent on an existing network interface. - The
id
parameter of thejailer
is required to be an RFC 4122-compliant UUID. - Fixed an issue which caused the network RX rate limiter to be more restrictive than intended.
- API requests which contain unknown fields will generate an error.
- Fixed an issue related to high CPU utilization caused by improper
KVM PIT
configuration. - It is now possible to create more than one network tun/tap interface inside a jailed Firecracker.
- Added metrics for API requests, VCPU and device actions for the serial console
(
UART
), keyboard (i8042
), block and network devices. Metrics are logged every 60 seconds. - A CPU features template for C3 is available, in addition to the one for T2.
- Seccomp filters restrict Firecracker from calling any other system calls than
the minimum set it needs to function properly. The filters are enabled by
setting the
USE_SECCOMP
environment variable to 1 before running Firecracker. - Firecracker can be started by a new binary called
jailer
. The jailer takes as command line arguments a unique ID, the path to the Firecracker binary, the NUMA node that Firecracker will be assigned to and auid
andgid
for Firecracker to run under. It sets up achroot
environment and acgroup
, and calls exec to morph into Firecracker.
- In case of failure, the metrics and the panic location are logged before aborting.
- Metric values are reset with every flush.
CPUTemplate
is now calledCpuTemplate
in order to work seamlessly with the swagger code generator for Go.firecracker-beta.yaml
is now calledfirecracker.yaml
.
- Handling was added for several untreated KVM exit scenarios, which could have led to panic.
- Fixed a bug that caused Firecracker to crash when attempting to disable the
IA32_DEBUG_INTERFACE MSR
flag in the T2 CPU features.
- Removed a leftover file generated by the logger unit tests.
- Removed
firecracker-v1.0.yaml
.
- The CPU Template can be set with an API call on
PUT /machine-config
. The only available template is T2. - Hyperthreading can be enabled/disabled with an API call on
PUT /machine-config
. By default, hyperthreading is disabled. - Added boot time performance test (
tests/performance/test_boottime.py
). - Added Rate Limiter for VirtIO/net and VirtIO/net devices. The Rate Limiter
uses two token buckets to limit rate on bytes/s and ops/s. The rate limiter
can be (optionally) configured per drive with a
PUT
on/drives/{drive_id}
and per network interface with aPUT
on/network-interface/{iface_id}
. - Implemented pre-boot PUT updates for
/boot-source
,/drives
,/network-interfaces
and/vsock
. - Added integration tests for
PUT
updates.
- Moved the API definition (
swagger/firecracker-beta.yaml
) to theapi_server
crate. - Removed
"console=ttyS0"
and added"8250.nr_uarts=0"
to the default kernel command line to decrease the boot time. - Changed the CPU topology to have all logical CPUs on a single socket.
- Removed the upper bound on CPU count as with musl there is no good way to get the total number of logical processors on a host.
- Build time tests now print the full output of commands.
- Disabled the Performance Monitor Unit and the Turbo Boost.
- Check the expected KVM capabilities before starting the VM.
- Logs now have timestamps.
testrun.sh
can run on platforms with more than one package manager by setting the package manager via a command line parameter (-p
).- Allow correct set up of multiple network-interfaces with auto-generated MAC.
- Fixed sporadic bug in VirtIO which was causing lost packages.
- Don't allow
PUT
requests with empty body on/machine-config
. - Deny
PUT
operations after the microvm boots (exception: the temporarily fix for live resize of block devices).
- Removed examples crate. This used to have a Python example of starting
Firecracker. This is replaced by
test_api.py
integration tests. - Removed helper scripts for getting coverage and coding style errors. These
were replaced by
test_coverage.py
andtest_style.py
test integration tests. - Removed
--vmm-no-api
command line option. Firecracker can only be started via the API.
- Users can interrogate the Machine Configuration (i.e. vcpu count and memory
size) using a
GET
request on/machine-config
. - The logging system can be configured through the API using a
PUT
on/logger
. - Block devices support live resize by calling
PUT
with the same parameters as when the block was created. - Release builds have Link Time Optimization (LTO) enabled.
- Firecracker is built with
musl
, resulting in a statically linked binary. - More in-tree integration tests were added as part of the continuous integration system.
- The vcpu count is enforced to
1
or an even number. - The Swagger definition of rate limiters was updated.
- Syslog-enabled logs were replaced with a host-file backed mechanism.
- The host topology of the CPU and the caches is not leaked into the microvm anymore.
- Boot time was improved by advertising the availability of the TSC deadline timer.
- Fixed an issue which prevented Firecracker from working on 4.14 (or newer) host kernels.
- Specifying the MAC address for an interface through the API is optional.
- Removed support for attaching vsock devices.
- Removed support for building Firecracker with glibc.
- Users can now interrogate Instance Information (currently just instance state) through the API.
- Renamed
api/swagger/all.yaml
toapi/swagger/firecracker-v1.0.yaml
which specifies targeted API support for Firecracker v1.0. - Renamed
api/swagger/firecracker-v0.1.yaml
toapi/swagger/firecracker-beta.yaml
which specifies the currently supported API. - Users can now enforce that an emulated block device is read-only via the API.
To specify whether a block device is read-only or read-write, an extra
"permissions" field was added to the Drive definition in the API. The root
filesystem is automatically mounted in the guest OS as
ro
/rw
according to the specified "permissions". It's the responsibility of the user to mount any other read-only block device as such within the guest OS. - Users can now stop the guest VM using the API. Actions of type
InstanceHalt
are now supported via the API.
- Added support for
getDeviceID()
invirtIO-block
. Without this, the guest Linux kernel would complain at boot time that the operation is unsupported. stdin
control is returned to the Firecracker process when guest VM is inactive. Raw modestdin
is forwarded to the guest OS when guest VM is running.
- Removed
api/swagger/actions.yaml
. - Removed
api/swagger/devices.yaml
. - Removed
api/swagger/firecracker-mvp.yaml
. - Removed
api/swagger/limiters.yaml
.
- Users can now specify the MAC address of a guest network interface via the
PUT
network interface API request. Previously, the guest MAC address parameter was ignored.
- Fixed a guest memory allocation issue, which previously led to a potentially significant memory chunk being wasted.
- Fixed an issue which caused compilation problems, due to a compatibility breaking transitive dependency in the tokio suite of crates.
- One-process virtual machine manager (one Firecracker per microVM).
- RESTful API running on a unix socket. The API supported by v0.1 can be found
at
api/swagger/firecracker-v0.1.yaml
. - Emulated keyboard (
i8042
) and serial console (UART
). The microVM serial console input and output are connected to those of the Firecracker process (this allows direct console access to the guest OS). - The capability of mapping an existing host tun-tap device as a VirtIO/net device into the microVM.
- The capability of mapping an existing host file as a GirtIO/block device into the microVM.
- The capability of creating a VirtIO/vsock between the host and the microVM.
- Default demand fault paging & CPU oversubscription.