The situation is not a IPC scene, pipe(2) is too heavy.
We have tokio::sync:⌚:channel after tokio has been introduced.
The channel has better performance and easy to use.
Fixes: #1721
Signed-off-by: Tim Zhang <tim@hyper.sh>
Update:
- Make the type of errnoRet in oci.proto oneof
- Update seccomp_grpc_to_oci that can set errnoRet as EPREM if the
value is empty.
- Update the oci.pb.go based on the above fixes
- Add seccomp errnoRet and flags option to configs in rustjail
Fixes: #1719
Signed-off-by: Manabu Sugimoto <Manabu.Sugimoto@sony.com>
Occassional coredumps and OOMs observed on shutdown path due to improper
BH handling during aio cleanup in QEMU. Thankfully this has been fixed
in upstream already -- let's carry this patch.
Fixes: #1717
Signed-off-by: Eric Ernst <eric_ernst@apple.com>
Remove the prohibition of vhost-user devices on s390x, which are by now
supported (e.g. vhost-user-fs-ccw). As a consequence,
appendVhostUserDevice no longer needs an error in its signature.
This enables virtio-fs support on s390x.
Fixes: #1469
Signed-off-by: Jakob Naucke <jakob.naucke@ibm.com>
Update GoVMM to get memory backend support for non-DIMM setups. This is
necessary for virtio-fs on s390x.
Signed-off-by: Jakob Naucke <jakob.naucke@ibm.com>
register_memory_event_v2() includes a closure spawned as an async task
with tokio. At the end of that closure, there's a test for a closed fd
exiting if so. But this is right at the end of the closure when it was
about to exit anyway, so this does nothing.
This code was originally an explicit thread, converted to a tokio task
by 332fa4c "agent: switch to async runtime". It looks like there was an
error during conversion, where this logic was accidentally moved out of the
while loop above, where it makes a lot more sense.
Put it back into the loop.
fixes#1702
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Rather than relying on the system clock, use a channel timeout to avoid
problems if the system time changed.
Signed-off-by: James O. D. Hunt <james.o.hunt@intel.com>
Fixed logic used to handle static agent tracing.
For a standard (untraced) hypervisor shutdown, the runtime kills the VM
process once the workload has finished. But if static agent tracing is
enabled, the agent running inside the VM is responsible for the
shutdown. The existing code handled this scenario but did not wait for
the hypervisor process to end. The outcome of this being that the
console watcher thread was killed too early.
Although not a problem for an untraced system, if static agent tracing
was enabled, the logs from the hypervisor would be truncated, missing the
crucial final stages of the agents shutdown sequence.
The fix necessitated adding a new parameter to the `stopSandbox()` API,
which if true requests the runtime hypervisor logic simply to wait for
the hypervisor process to exit rather than killing it.
Fixes: #1696.
Signed-off-by: James O. D. Hunt <james.o.hunt@intel.com>
Previously, the hypervisors were sending a signal and then checking to
see if the process had died by sending the magic null signal (`0`). However,
that doesn't work as it was written: the logic was assuming sending the
null signal to a process that was dead would return `ESRCH`, but it
doesn't: you first need to you `wait(2)` for the process before sending
that signal. This means that previously, all affected hypervisors would
appear to take `timeout` seconds to end, even though they had _already_
finished.
Now, the hypervisors true end time will be seen as we wait for the
processes before sending the null signal to ensure the process has
finished.
Signed-off-by: James O. D. Hunt <james.o.hunt@intel.com>
Refactored some of the hypervisors to remove the duplicated code used to
trigger a shutdown.
Also added some unit tests.
Signed-off-by: James O. D. Hunt <james.o.hunt@intel.com>
Currently runtime and agent special case virtio-blk devices under clh,
ostensibly because the PCI address information is not available in that
case.
In fact, cloud-hypervisor's VmAddDiskPut API does return a PciDeviceInfo,
which includes a PCI address. That API is broken, because PCI addressing
depends on guest (firmware or OS) actions that the hypervisor won't know
about. clh only gets away with this because it only uses a single PCI root
and never uses PCI bridges, in which case the guest addresses are
accurately predictable: they always have domain and bus zero.
Until https://github.com/kata-containers/kata-containers/pull/1190, Kata
couldn't handle PCI addressing unless there was exactly one bridge, which
might be why this was actually special-cased for clh.
With #1190 merged, we can handle more general PCI paths, and we can derive
a trivial (one element) PCI path from the information that the clh API
gives us. We can use that to remove this special case.
fixes#1431
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Change the "moo FAILURE" message shown in a couple of the unit tests to
"moo message". This means that searching for unrelated failures in the
test output by looking for "FAIL" won't show these messages as false
positives any more.
fixes#1683
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Instead of giving too many options, let's just mention the script and
rely on it entirely for the release.
This helps to simplify the document and have one well stablished
process.
Fixes: #1680
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
During the 2.1.0-alpha2 / 2.0.3 release, I had a hard time trying to
perform anything related to hub as the app token should be used instead
of the user password. Thankfully Carlos pointed me out to that
direction, but it'd be good to have it explicitly documented.
Fixes: #1680
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Since we stopped building kata-containers packages as part of our
release process, there's no need to have an OBS account to be able to do
the release.
Fixes: #1680
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
With tracing enabled, grpc health check generates a large number of
spans which creates too much data for tasks running longer than a few
minutes. To solve this, remove span creation from kata agent check() and
sendReq() where the majority of the spans come from. Leave contexts in
functions for subsequent calls that create spans.
Fixes#1395
Signed-off-by: Chelsea Mafrica <chelsea.e.mafrica@intel.com>
- release: Do not git add kata-{deploy,cleanup}.yaml for the tests repo
- kata-deploy: add runtimeclass that includes pod overhead
- release: automatically bump the version of the kata-deploy images
- Refine uevent matching conditions
- docs: update dev-guide to include fixes from 1.x
- virtcontainers: replace newStore by store in Sandbox struct
- agent: log the mount point if it is already mounted
- tools/agent-ctl: Update Cargo.lock
- agent: Rework the debug console
- oci: Update seccomp configuration
- kernel: update experimental kernel to 5.10.x
- kata-deploy: Fix `test-kata.sh` and do some small cleanups / improvements in the kata-deploy script
- github: Fix slash-command-action usage
- rustjail: fix the issue of missing default home env
- Make uevent watching mechanism more flexible
- ci/openshift-ci: Prepare to build on CentOS 8
- docs: update configuration for passing annotations in conatinerd
- Revert "github: Remove kata-deploy-test action"
- runtime: increase dial timeout
- qemu experimental: Move to latest tree on virtio-fs-dev (qemu 6.0 + DAX patches).
- github: Remove kata-deploy-test action
- agent: s390x statfs constants
- kernel: upgrade kernel to 5.10.x for arm64.
- Don't do anything in Pipestream::shutdown
- Fix fsgroup
- agent: Remove many "panic message is not string literal" warnings
- osbuilder: Update QAT Dockerfile with new QAT driver version
- osbuilder: update dockerfiles to utilize IMAGE_REGISTRY
- Only keep one VERSION file
- Dechat deruntime
- runtime: Format auto-generated client code for cloud-hypervisor API
- runtime: use concrete KataAgentConfig instead of interface type
- versions: Update cloud-hypervisor to release v0.14.1
- runtime: import runtime/v2/runc/options to decode request from Docker
- virtcontainers/fc: Upgrade Firecracker to v0.23.1
- docs: Remove ubuntu installation guide
- docs: Update snap install guide
- docs: update how-to-use-k8s-with-cri-containerd-and-kata.md
- Update install docs for Fedora and CentOS
- action: fix missing qemu tag
- Remove installation guides for SLE and openSUSE
- kernel: Enable OVERLAY_FS_{METACOPY,XINO_AUTO}
- versions: kernel 5.10.x
- virtcontainers: Fix missing contexts in s390x
- runtime: makefile allow override DAX value
11897248 release: Do not git add kata-{deploy,cleanup}.yaml for the tests repo
2b5f79d6 release: automatically bump the version of the kata-deploy images
8682d6b7 docs: update dev-guide to include fixes from 1.x
f444adb5 kata-cleanup: Explicitly add tag to the container image
12582c2f kata-deploy: add runtimeclass that includes pod overhead
d75fe956 virtcontainers: replace newStore by store in Sandbox struct
342eb765 tools/agent-ctl: Update Cargo.lock
24b0703f agent: fix test for the debug console
79033257 agent: async the debug console
8ea2ce9a agent/device: Remove legacy uevent matching
5d007743 agent/device: Refine uevent matching for pmem devices
9017e110 agent: start to rework the debug console
a59e07c1 agent/define: Refine uevent matching for virtio-scsi devices
484a3647 agent/device: Rework uevent handling for virtio-blk devices
7873b7a1 github: Fix slash-command-action usage
eda8da1e github: Revert "github: Remove kata-deploy-test action"
a938d903 rustjail: fix the issue of missing default home env
b0e4618e docs: update configuration for passing annotations in conatinerd
d43098ec kata-deploy: Adapt regex for testing kata-deploy
107ceca6 kernel: update experimental kernel to 5.10.x
ca4dccf9 release: Get rid of "master"
c2197cbf release: Use sudo to install hub
49eec920 agent: log the tag and mount point if it is already mounted
16f732fc ci/lib: Use git to clone the tests repository
9281e567 ci/openshift-ci: Add build root dockerfile
1cce9300 github: Remove kata-deploy-test action
0828f9ba agent/uevent: Introduce wait_for_uevent() helper
16ed55e4 agent/device: Use consistent matching for past and future uevents
4b16681d agent/uevent: Put matcher object rather than "device address" in watch list
b8b32248 agent/uevent: Consolidate event matching logic
d2caff6c agent: Re-organize uevent processing
55ed2ddd agent: Store uevent watchers in Vec rather than HashMap
91e0ef5c agent/uevent: Report whole Uevents to device watchers
36420054 agent: Store whole Uevent in map, rather than just /dev name
06162025 agent/device: Move GLOBAL_DEVICE_WATCHER into Sandbox
11ae32e3 agent/device: Fix path matching for PCI devices
4f608804 agent/device: Update test_get_device_name()
ee6a590d agent: add test test_pipestream_shutdown
4a2d4370 agent: don't do anything in Pipestream::shutdown
e3e670c5 agent/device: Forward port test for get_device_name() from Kata 1.x
ed08980f agent: Remove many "panic message is not string literal" warnings
f365bdb7 versions: qemu-experimental: 6.0~rc 470dd6
6491b9d7 qemu: Add support to build static qemu for dev tree
13653e7b runtime: increase dial timeout
935460e5 osbuilder: update dockerfiles to utilize IMAGE_REGISTRY
010d57f4 osbuilder: Update QAT Dockerfile with new QAT driver version
adb866ad kata-deploy: Adapt to the correct tag name
60adc7f0 VERSION: Use the correct form
a4c125a8 trace: move gRPC requests from debug to trace
50fff977 trace: move trace span chatter to trace rather than info
28bd8c11 kernel: upgrade kernel to 5.10.x for arm64.
6fe48329 runtime: use concrete KataAgentConfig instead of interface type
64939425 mount: fix the issue of missing set fsGroup
88e58a4f agent: fix the issue of missing pass fsGroup
572aff53 build: Only keep one VERSION file
0c38d9ec runtime: Fix the format of the client code of cloud-hypervisor APIs
52cacf88 runtime: Format auto-generated client code for cloud-hypervisor API
84b62dc3 versions: Update cloud-hypervisor to release v0.14.1
4a38ff41 docs: Update snap install guide
ede1ab86 docs: Remove ubuntu installation guide
6255cc19 virtcontainers/fc: Upgrade Firecracker to v0.23.1
2c47277c docs: update how-to-use-k8s-with-cri-containerd-and-kata.md
317f55f8 docs: Update minimum version for Fedora
1ce29fc9 docs: Update CentOS install docs
3f90561b docs: Update Fedora install docs
8a1c6c3f action: fix missing qemu tag
a9ff9c87 docs: Remove openSUSE installation guide
2888ceb0 docs: Remove SLE installation guide
09d454ac runtime: import runtime/v2/runc/options to decode request from Docker
0b502d15 runtime: makefile allow override DAX value
a65519b9 versions: keep using kernel 5.4.x for ARM
31ced01e virtcontainers: Fix missing contexts in s390x
52a276fb agent: Fix type for PROC_SUPER_MAGIC on s390x
5b7c8b7d agent: Update cgroups-rs to 0.2.5
c035cdb3 versions: kernel 5.10.x
660b0473 oci: Update seccomp configuration
8c1e0d30 kernel: Enable OVERLAY_FS_{METACOPY,XINO_AUTO}
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Let's teach `update-repository-version.sh` to automatically bump the
version of the kata-deploy images to be used within that release, when
running against the `kata-containers` repo.
Fixes: #1665
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
This addresses a few gaps with respect to fixes in 1.x docs:
- Cleanup QEMU information in order to drop references to qemu-lite
- Make sure we include directions for debug console in case of QEMU
Fixes: #574
Signed-off-by: Eric Ernst <eric.g.ernst@gmail.com>
The property name make newcomers confused when reading code.
Since in Kata Containers 2.0 there will only be one type of store,
so it's safe to replace it by `store` simply.
Fixes: #1660
Signed-off-by: bin <bin@hyper.sh>
While performing a `make` in the top directory of the project, I've
noticed that agent-ctl's Cargo.lock file was not up-to-date.
Fixes: #1655
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
DevAddrMatcher existed purely as a transitional step as we refined the
uevent matching logic for each of the different device types we care about.
We've now done that, so it can be removed along with several related
pieces.
fixes#1628
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Use the new uevent matching infrastructure to refine the matching for pmem
devices to something more pinned down to that device type. While we're
there, fix a few anciliary problems with get_pmem_device_name():
- The name is poor - the *input* to this function is the expected device
name, so the result isn't helpful, except that it needs to wait for the
device to be ready in the guest. Change it to wait_for_pmem_device() and
explicitly check that the returned device name matches the one expected.
- Remove an incorrect comment in nvdimm_storage_handler() (the only caller)
which appears to have been copied from the virtio-blk path, but then
become stale.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Current get_scsi_device_name() uses the legacy uevent matching which
isn't very precise. This refines it to use a specific matcher
implementation. While we're at it:
- No longer insist on the SCSI controller being under the PCI root.
It generally will be, but there's no particular reason to require
it.
The matcher still has a problem in that it won't work sensibly if
there are multiple SCSI busses in the guest. Fixing that requires
changes on the runtime side as well, though, so it's beyond scope for
this change.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
