"//tensorflow/core/common_runtime/next_pluggable_device/c:tf_rendezvous_c_api",
        "@local_tsl//tsl/platform",
    ],
)

tf_cuda_library(
    name = "kernels",
    srcs = [
        "kernels.cc",
    ],
    hdrs = [
        "kernels.h",
    ],
    copts = tf_copts(),
    visibility = ["//visibility:public"],
    deps = [
        ":c_api_internal",
        ":c_api_macros_hdrs",

# Only include debug info for files under tensorflow/, excluding kernels, to
# reduce the size of the debug info in the binary. This is because if the debug
# sections in the ELF binary are too large, errors can occur. See
# https://github.com/tensorflow/tensorflow/issues/48919.
# Users can still include debug info for a specific kernel, e.g. with:
#     --config=dbg --per_file_copt=+tensorflow/core/kernels/identity_op.*@-g

```

TensorFlow kernels and TensorFlow's dependencies are still not built with
debugging information with `--config=dbg`, as issues occur on Linux if
there is too much debug info (see [this GitHub
issue](https://github.com/tensorflow/tensorflow/issues/48919) for context). If
you want to debug a kernel, you can compile specific files with `-g` using the

        mean, var = tf.nn.moments(self.kernel, axes=[0, 1, 2], keepdims=True)
        return self.convolution_op(inputs, (self.kernel - mean) / tf.sqrt(var +
        1e-10))` Alternatively, you can override `convolution_op`: `python class
        StandardizedConv2D(tf.keras.Layer): def convolution_op(self, inputs,
        kernel): mean, var = tf.nn.moments(kernel, axes=[0, 1, 2],

//tensorflow/cc/saved_model:saved_model_half_plus_two
//tensorflow:no_tensorflow_py_deps
//tensorflow/tools/pip_package:win_pip_package_marker
//tensorflow/core:image_testdata
//tensorflow/core/kernels/cloud:bigquery_reader_ops
//tensorflow/python:extra_py_tests_deps
//tensorflow/python:mixed_precision
//tensorflow/python:tf_optimizer
//tensorflow/python:compare_test_proto_py

- The (alpha) nftables mode of kube-proxy now requires version 1.0.1 or later
  of the nft command-line, and kernel 5.13 or later. (For testing/development
  purposes, you can use older kernels, as far back as 5.4, if you set the
  `nftables.skipKernelVersionCheck` option in the kube-proxy config, but this is not

## Ports {#ports}

### Linux {#linux}

As [announced](go1.23#linux) in the Go 1.23 release notes, Go 1.24 requires Linux
kernel version 3.2 or later.

### WebAssembly {#wasm}

Go 1.22 changed how the runtime interacts with transparent huge pages on Linux.
In particular, a common default Linux kernel configuration can result in
significant memory overheads, and Go 1.22 no longer works around this default.
To work around this issue without adjusting kernel settings, transparent huge
pages can be disabled for Go memory with the
[`disablethp` setting](/pkg/runtime#hdr-Environment_Variable).

- Allow pods to use the `net.ipv4.tcp_rmem` and `net.ipv4.tcp_wmem` sysctl by default
  when the kernel version is 4.15 or higher. With the kernel 4.15 the sysctl became namespaced.

- `kube-proxy`: Added an option/flag for configuring the `nf_conntrack_tcp_be_liberal` sysctl (in the kernel's netfilter conntrack subsystem).  When enabled, `kube-proxy` will not install the `DROP` rule for invalid conntrack states, which currently breaks users of asymmetric routing. ([#120354](https://github.com/kubernetes/kubernetes/pull/120354), [@ar...