hexagon: optimization for HMX mat_mul (#21554)

* hexagon: add async HMX worker

Introduce hmx-worker (dedicated thread for HMX compute) to overlap HMX
matmul with HVX dequant/DMA stages in the pipeline path, replacing the
previous synchronous HMX calls that blocked the main thread.

* hexagon: cost-based VTCM chunk search for out-stationary matmul

* hexagon: fix futex race in hmx_worker_drain
Store the boolean to local variable avoid atomic load twice

* hex-mm: hmx optimize scatter/transpose and use HMX intrinsics

* hex-vmem: drop vmem limit a touch under 3GB on v73

* hexagon: add fwd declaration of htp_context

* hex-hmx: replace hmx-worker with hmx-queue that mimics dma-queue interface

Simplifies the overall implemantion, reduces thread wakeup roundtrips.

* hex-mm: add debug log to hmx work func called from hmx-queue

* Update hmx-queue.h

Co-authored-by: Max Krasnyansky <max.krasnyansky@gmail.com>

---------

Co-authored-by: Kim-Chyan Gan <kgan@qti.qualcomm.com>
Co-authored-by: Max Krasnyansky <maxk@qti.qualcomm.com>
Co-authored-by: Max Krasnyansky <max.krasnyansky@gmail.com>

ggml/src/ggml-hexagon/htp/htp-ctx.h

@@ -2,6 +2,7 @@
 #define HTP_CTX_H
 
 #include "hex-dma.h"
+#include "hmx-queue.h"
 #include "htp-ops.h"
 #include "worker-pool.h"
 
@@ -30,6 +31,8 @@ struct htp_spad {
     uint32_t                  size_per_thread; // size per thread
 };
 
+struct htp_context;
+
 // Context while processing an Op
 // TODO: fold this into the main context
 struct htp_ops_context {
@@ -72,6 +75,10 @@ struct htp_context {
     atomic_bool            vtcm_needs_release;
 
     struct htp_ops_context octx;
+
+#ifdef HTP_HAS_HMX
+    struct hmx_queue *     hmx_queue; // Async HMX queue for pipeline overlap
+#endif
 };
 
 int op_matmul(struct htp_ops_context * octx);