Add a "known issues & limitations" page

[leofang]

• cuda.core does not allow any side calls to cudaDeviceReset or alike that tear down the primary contexts
• avoid multiple frees of the same buffer in the child process (https://github.com/NVIDIA/cuda-python/issues/1074#issuecomment-3367798948, #1118)
• when going out of scope without an explicit .close() call, a buffer is deallocated on the allocation stream (https://github.com/NVIDIA/cuda-python/pull/1032)
• compute-sanitizer needs a flag to work with cuda.core (https://github.com/NVIDIA/cuda-python/issues/562#issuecomment-3009049723)
• ...