- Added a new option `strict-cpu-reservation` for CPU Manager static policy. When this option is enabled, CPU cores in `reservedSystemCPUs` will be strictly used for system daemons and interrupt processing no longer available for any workload. ([#130290](https://github.com/kubernetes/kubernetes/pull/130290), [@psasnal...

- The CPU Manager will now validate the state of the node,  enabling Kubernetes to maintain the CPU topology even if resources change. ([#66718](https://github.com/kubernetes/kubernetes/pull/66718), [@ipuustin](https://github.com/ipuustin))

  //   have observed 12 micros on 64-bit linux systems to wake up a parked thread). So if the
  //   timeout is small we shouldn't park(). This needs to be traded off with the cpu overhead of
  //   spinning, so we use SPIN_THRESHOLD_NANOS which is what AbstractQueuedSynchronizer uses for
  //   similar purposes.
  // * We want to behave reasonably for timeouts of 0

- github.com/kisielk/errcheck: [v1.2.0](https://github.com/kisielk/errcheck/tree/v1.2.0)
- github.com/kisielk/gotool: [v1.0.0](https://github.com/kisielk/gotool/tree/v1.0.0)
- github.com/klauspost/cpuid: [v1.2.0](https://github.com/klauspost/cpuid/tree/v1.2.0)
- github.com/kr/pty: [v1.1.5](https://github.com/kr/pty/tree/v1.1.5)
- github.com/kylelemons/godebug: [d65d576](https://github.com/kylelemons/godebug/tree/d65d576)

- github.com/jimstudt/http-authentication: [3eca13d](https://github.com/jimstudt/http-authentication/tree/3eca13d)
- github.com/klauspost/cpuid: [v1.2.0](https://github.com/klauspost/cpuid/tree/v1.2.0)
- github.com/kr/logfmt: [b84e30a](https://github.com/kr/logfmt/tree/b84e30a)
- github.com/kylelemons/godebug: [d65d576](https://github.com/kylelemons/godebug/tree/d65d576)

        TF_NewOperation(host_graph_, func_name_, func_node_name_);
    for (auto input : inputs) {
      TF_AddInput(desc, input);
    }
    // Set device to CPU because some ops inside the function might not be
    // available on GPU.
    TF_SetDevice(desc, "/cpu:0");
    *op = TF_FinishOperation(desc, s_);
    ASSERT_EQ(TF_OK, TF_GetCode(s_)) << TF_Message(s_);
    ASSERT_NE(*op, nullptr);
  }

  FunctionDef fdef() {

- Fixes regression in CPUManager that caused freeing of exclusive CPUs at incorrect times ([#90377](https://github.com/kubernetes/kubernetes/pull/90377), [@cbf123](https://github.com/cbf123)) [SIG Cloud Provider and Node]

  //   have observed 12 micros on 64-bit linux systems to wake up a parked thread). So if the
  //   timeout is small we shouldn't park(). This needs to be traded off with the cpu overhead of
  //   spinning, so we use SPIN_THRESHOLD_NANOS which is what AbstractQueuedSynchronizer uses for
  //   similar purposes.
  // * We want to behave reasonably for timeouts of 0

 * either of these extremes, {@code Striped} allows the user to trade between required concurrency
 * and memory footprint. For example, if a set of tasks are CPU-bound, one could easily create a
 * very compact {@code Striped<Lock>} of {@code availableProcessors() * 4} stripes, instead of
 * possibly thousands of locks which could be created in a {@code Map<K, Lock>} structure.
 *

- Add a new `distribute-cpus-across-numa` option to the static `CPUManager` policy. When enabled, this will trigger the `CPUManager` to evenly distribute CPUs across NUMA nodes in cases where more than one NUMA node is required to satisfy the allocation. ([#105631](https://github.com/kubernetes/kubernetes/pull/105631), [@klueska](https://github.com/klueska))...