fix: #254 bound receiver queue by aggregate payload bytes

The per-connection memory budget (MAX_CONNECTION_MEMORY, 256 MiB) only
charged wire buffers via receive_data_limited.  Decompression buffers and
per-file chunk copies were not accounted for, and the multithreaded
receiver could enqueue up to 100 files (each up to 64 MiB uncompressed)
ahead of a slow disk writer, retaining ~6.4 GiB per connection.  A client
sending highly compressible chunks with little bandwidth could OOM the
host while the reserve never tripped.

Bound the receive pipeline by aggregate payload bytes instead of item
count alone:
- Export MAX_CONNECTION_MEMORY from protocol.h.
- PipelineContextReceiver tracks queued_bytes (payload bytes received but
  not yet released by the disk writer, i.e. queued or in the writer's
  hand) under the existing mutex.
- receiver enqueue now blocks while the queue is full by count OR when
  adding the file would push queued_bytes over the configured byte limit,
  applying backpressure to the sender instead of failing the transfer.
- The disk writer releases the byte budget after each file is freed and
  signals the not-full condition.
- The server sets the byte ceiling to
  MAX_CONNECTION_MEMORY - 2*MAX_CHUNK_SIZE so that the queued payloads
  plus the transient wire/decompression buffers of the one in-flight
  chunk stay within the per-connection budget.

The single-threaded receive path is already bounded: it writes files to
disk before reading the next chunk, so its transient is at most one
chunk's wire + decompressed + copied payload (~3 * MAX_CHUNK_SIZE, below
the budget).  Wire buffers remain charged exactly once by
receive_data_limited; this change does not double charge them.

Adds a deterministic unit test in test_multiprocessing.c proving that an
enqueue which would exceed the byte budget blocks until the writer
releases bytes.

src/shared/multiprocessing.h

@@ -45,6 +45,14 @@ typedef struct PipelineContextReceiver {
   cnd_t condition_not_empty;
   bool receiver_done;
   atomic_bool cancelled;
+  /* Aggregate payload bytes that have been received but not yet released by
+     the disk writer (queued or in the writer's hand).  Guarded by `mutex`.
+     When `max_queue_bytes` is non-zero the receiver blocks before enqueuing
+     once this total would exceed it, so decompressed/copied file payloads
+     buffered ahead of a slow disk writer respect the per-connection memory
+     budget instead of growing without bound. */
+  size_t queued_bytes;
+  size_t max_queue_bytes;
 } PipelineContextReceiver;
 
 PipelineContextSender* pipeline_context_sender_create(Config* config, Queue* queue_scanner,
@@ -53,6 +61,18 @@ void pipeline_context_sender_destroy(PipelineContextSender* context);
 PipelineContextReceiver* pipeline_context_receiver_create(Config* config, Queue* queue_receiver,
                                                           int file_descriptor, SSL* ssl);
 void pipeline_context_receiver_destroy(PipelineContextReceiver* context);
+/* Bound the bytes buffered ahead of the disk writer (see max_queue_bytes). */
+void pipeline_context_receiver_set_queue_byte_limit(PipelineContextReceiver* context,
+                                                    size_t max_bytes);
+/* Blocking enqueue used by the receive pipeline sink.  Blocks while the queue
+   is full by element count or when adding `file` would push queued_bytes over
+   the configured byte limit; waits until the disk writer releases bytes.
+   Takes ownership of `file` on success and destroys it on failure/cancel. */
+bool pipeline_context_receiver_enqueue_file(PipelineContextReceiver* context, File* file);
+/* Account for `released_bytes` of payload memory that has been freed by the
+   disk writer, unblocking a receiver that is waiting on the byte limit. */
+void pipeline_context_receiver_note_bytes_released(PipelineContextReceiver* context,
+                                                   size_t released_bytes);
 int receive_thread(void* pipeline_context);
 int write_thread(void* pipeline_context);
 #endif