full diff: https://github.com/containerd/consolve/compare/v1.0.0...v1.0.1
vendor golang.org/x/sys 2334cc1a136f0a5b4384f1a27e3634457be08553
full diff: ed371f2e16...2334cc1a13
- unix: add Darwin support for clonefile syscalls
- Adds openat2 for linux
openat2 is a new syscall added to Linux 5.6. It provides a superset of
openat(2) functionality, extending it with flags telling the kernel how
to resolve the paths.
For more info, see https://lwn.net/Articles/803237/
NOTE that this is a second attempt to add the call; the previous one
(https://golang.org/cl/227280) was reverted
(https://golang.org/cl/227846) due to the test case failure on ARM
(https://golang.org/issue/38357).
This CL has the test case reworked to be less assumptive to the testing
environment. In particular, it first tries if the most simplistic
openat2() call succeeds, and skips the test otherwise. It is done that
way because CI can be under under different kernels and in various
envrionments -- in particular, Docker+seccomp can result in EPERM from a
system call (which is not expected otherwise).
For previous discussions about the test case, see
https://golang.org/cl/227865.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
The shim delete action needs bundle information to cleanup resources
created by shim. If the cleanup dead shim is called after delete bundle,
the part of resources maybe leaky.
The ttrpc client UserOnCloseWait() can make sure that resources are
cleanup before delete bundle, which synchronizes task deletion and
cleanup deadshim. It might slow down the task deletion, but it can make
sure that resources can be cleanup and avoid EBUSY umount case. For
example, the sandbox container like Kata/Firecracker might have mount
points over the rootfs. If containerd handles task deletion and cleanup
deadshim parallelly, the task deletion will meet EBUSY during umount and
fail to cleanup bundle, which makes case worse.
And also update cleanupAfterDeadshim, which makes sure that
cleanupAfterDeadshim must be called after shim disconnected. In some
case, shim fails to call runc-create for some reason, but the runc-create
already makes runc-init into ready state. If containerd doesn't call shim
deletion, the runc-init process will be leaky and hold the cgroup, which
makes pod terminating :(.
Signed-off-by: Wei Fu <fuweid89@gmail.com>
full diff: d3ef23f19f...efbc4488d8
Fix sameFile() to recognize empty files as the same
- fixes "Empty files can diff as "modified" even when they're not"
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
This includes an update of the selinux package and the ability in the CRI
configuration to set the upper bounds for the selinux category labels that are
generated.
Signed-off-by: Michael Crosby <michael@thepasture.io>
full diff: 0dbf7f05ba...318312a373
relevant changes:
- cpuset: typo fix for function name
- Retry file writes on EINTR errors to work with Go 1.14 asynchronous preemption
- Various linting issues and cleanup
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
full diff: 0ec596719c...d3ef23f19f
- fs: support for OpenBSD
- sysx/xattr: fix and improve
- remove Windows' Readlink fork
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
- Remove dependency on libcontainer/system
- Get rid of socat for port forwarding
- Roll docker/distribution back to latest (v2.7.1) release
Now that 901bcb2231 was merged in containerd,
we no longer depend on the ParseDockerRef utility from docker/distribution,
so we can safely roll back to the latest release for this dependency.
- vendor: kubernetes v1.18.2
Fix client watch reestablishment handling of client-side timeouts
- Add config flag to default empty seccomp profile
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
