Document NVLink API lifecycle changes

1. Review finding: the 1.1.0 release notes did not mention the new Device.get_nvlink_count and Device.get_nvlinks APIs, the changed Device.get_nvlink validation behavior, or the NvlinkInfo.max_links deprecation.\n\n2. Suggested fix implemented in this commit: add release-note entries for device-specific NVLink enumeration, the stricter get_nvlink ValueError behavior, and the deprecated NvlinkInfo.max_links replacement path.

Author: rwgk

cuda_core/docs/source/release/1.1.0-notes.rst

@@ -44,6 +44,11 @@ New features
   :attr:`~ManagedBuffer.accessed_by`). Locations are expressed via
   :class:`Device` or :class:`Host`.
 
+- Added :meth:`system.Device.get_nvlink_count` and
+  :meth:`system.Device.get_nvlinks` for device-specific NVLink enumeration.
+  These APIs avoid relying on the static NVML ``NVML_NVLINK_MAX_LINKS`` macro
+  when querying the links available on a particular device.
+
 Bug fixes
 ---------
 
@@ -52,3 +57,12 @@ Bug fixes
 - Calling ``cuda.core.system.get_process_name`` before querying any device's
   ``compute_running_processes`` would raise a ``NvmlNotFoundError``.  Now it will
   correctly return the process name, if it is a GPU-using process.
+- :meth:`system.Device.get_nvlink` now validates link numbers against the
+  device-specific NVLink count and raises ``ValueError`` for unsupported links.
+
+Deprecated APIs
+---------------
+
+- Deprecated :attr:`system.NvlinkInfo.max_links`. Use
+  :meth:`system.Device.get_nvlink_count` or :meth:`system.Device.get_nvlinks`
+  to query NVLink availability for a specific device.