`CUDA_ERROR_ILLEGAL_ADDRESS` failures.

    A fix in CUDA 9.1.121 is expected in late February 2018. We do not expect a
    fix for CUDA 9.0.x. Until the fix is available, the only workaround is to
    [downgrade](https://developer.nvidia.com/cuda-toolkit-archive) to CUDA 8.0.x
    or disable XLA:GPU.

    TensorFlow will print a warning if you use XLA:GPU with a known-bad version

                    }
                    context.success("Model upgrade completed");
                    modifiedPoms.add(pomPath);
                } else {
                    // Treat invalid upgrades (including downgrades) as errors, not warnings
                    context.failure("Cannot upgrade from " + currentVersion + " to " + targetModelVersion);
                    errorPoms.add(pomPath);
                }

        }
        return model;
    }

    static void warnNotDowngraded(MavenProject project) {
        LOGGER.warn("The consumer POM for " + project.getId() + " cannot be downgraded to 4.0.0. "
                + "If you intent your build to be consumed with Maven 3 projects, you need to remove "
                + "the features that request a newer model version.  If you're fine with having the "

## Introduction to 1.9.0

	// Current version being written.
	xlVersionCurrent [4]byte
)

const (
	// Breaking changes.
	// Newer versions cannot be read by older software.
	// This will prevent downgrades to incompatible versions.
	xlVersionMajor = 1

	// Non breaking changes.
	// Bumping this is informational, but should be done
	// if any change is made to the data stored, bumping this

    * To downgrade the cluster to version that does not support scale-to-zero feature:
    * 1. make sure there are no hpa objects with minReplicas=0. Here is a oneliner to update it to 1:

- DRA: The scheduler plugin prevented abnormal filter runtimes by timing out after 10...

        "ReadWriteC -> ReadWriteA",
        "ReadWriteB -> ReadWriteC",
        "ReadWriteA -> ReadWriteB");
  }

  public void testWriteToReadLockDowngrading() {
    writeLockA.lock(); // writeLockA downgrades to readLockA
    readLockA.lock();
    writeLockA.unlock();

    lockB.lock(); // readLockA -> lockB
    readLockA.unlock();

    // lockB -> writeLockA should fail

- Support a smooth upgrade from client-side apply to server-side apply without conflicts, as well as support the corresponding downgrade. ([#90187](https://github.com/kubernetes/kubernetes/pull/90187), [@julianvmodesto](https://github.com/julianvmodesto)) [SIG API Machinery and Testing]

				if !ok {
					ckvs.Set(kv.Key, kv.Value)
				}
			}
			if _, ok := cp[subSys]; !ok {
				rnSubSys, ok := renamedSubsys[subSys]
				if !ok {
					// A config subsystem was removed or server was downgraded.
					continue
				}
				// Copy over settings from previous sub-system
				// to newly renamed sub-system
				for _, kv := range cp[rnSubSys][Default] {
					_, ok := c[subSys][tgt].Lookup(kv.Key)