llama : refactor kv cache guard

[ggerganov]

Simplify the KV cache guard mechanism. Prepare for separate recurrent cache implementation.

Also, llama_decode now correctly returns 1 when the batch cannot fit in the KV cache and the KV cache state is correctly restored upon failure to process the batch.

[LostRuins]

After this commit, it seems like RNN based models like RWKV don't work anymore, and asset at llama-kv-cache.cpp:594: GGML_ASSERT(empty_cell.is_empty()) failed. Reverting the early return line at

llama.cpp/src/llama-kv-cache.cpp

Line 208 in 626f822

return true;

seems to allow RWKV to work again.

cc: @MollySophia