Improve efficiency and formatting of ITM output

SerialWireOutput was outputting 1 character per 32-bit write to the
ITM stimulus port. This is inefficient, and causes processing problems
with some viewers due to them receiving 3 NUL bytes between each
desired character.

Rework to allow us to be more efficient, and eliminate those NUL bytes:

* Retain existing mbed_itm_send() and clarify it's a single 32-bit write.
* Add new mbed_itm_send_block() that is appropriate for sending
  character data, and modify SerialWireOutput to use it.
* Move "wait for FIFO ready" check to before the write, rather than
  after.

One minor correction - FIFOREADY is a single bit of the register read.
Don't interpret reserved bits.

drivers/SerialWireOutput.h

@@ -33,12 +33,8 @@ class SerialWireOutput : public FileHandle {
 
     virtual ssize_t write(const void *buffer, size_t size)
     {
-        const unsigned char *buf = static_cast<const unsigned char *>(buffer);
+        mbed_itm_send_block(ITM_PORT_SWO, buffer, size);
 
-        /* Send buffer one character at a time over the ITM SWO port */
-        for (size_t i = 0; i < size; i++) {
-            mbed_itm_send(ITM_PORT_SWO, buf[i]);
-        }
         return size;
     }
 

hal/itm_api.h

@@ -22,6 +22,7 @@
 #if defined(DEVICE_ITM)
 
 #include <stdint.h>
+#include <stddef.h>
 
 #ifdef __cplusplus
 extern "C" {
@@ -68,12 +69,26 @@ void mbed_itm_init(void);
  * @brief      Send data over ITM stimulus port.
  *
  * @param[in]  port  The stimulus port to send data over.
- * @param[in]  data  The data to send.
+ * @param[in]  data  The 32-bit data to send.
+ *
+ * The data is written as a single 32-bit write to the port.
  *
  * @return     value of data sent.
  */
 uint32_t mbed_itm_send(uint32_t port, uint32_t data);
 
+/**
+ * @brief      Send a block of data over ITM stimulus port.
+ *
+ * @param[in]  port  The stimulus port to send data over.
+ * @param[in]  data  The block of data to send.
+ * @param[in]  len   The number of bytes of data to send.
+ *
+ * The data is written using multiple appropriately-sized port accesses for
+ * efficient transfer.
+ */
+void mbed_itm_send_block(uint32_t port, const void *data, size_t len);
+
 /**@}*/
 
 #ifdef __cplusplus

hal/mbed_itm_api.c

@@ -21,6 +21,10 @@
 
 #include <stdbool.h>
 
+#ifndef ITM_STIM_FIFOREADY_Msk
+#define ITM_STIM_FIFOREADY_Msk 1
+#endif
+
 #define ITM_ENABLE_WRITE 0xC5ACCE55
 
 #define SWO_NRZ 0x02
@@ -66,21 +70,64 @@ void mbed_itm_init(void)
     }
 }
 
+static void itm_out8(uint32_t port, uint8_t data)
+{
+    /* Wait until port is available */
+    while ((ITM->PORT[port].u32 & ITM_STIM_FIFOREADY_Msk) == 0) {
+        __NOP();
+    }
+
+    /* write data to port */
+    ITM->PORT[port].u8 = data;
+}
+
+static void itm_out32(uint32_t port, uint32_t data)
+{
+    /* Wait until port is available */
+    while ((ITM->PORT[port].u32 & ITM_STIM_FIFOREADY_Msk) == 0) {
+        __NOP();
+    }
+
+    /* write data to port */
+    ITM->PORT[port].u32 = data;
+}
+
 uint32_t mbed_itm_send(uint32_t port, uint32_t data)
 {
     /* Check if ITM and port is enabled */
     if (((ITM->TCR & ITM_TCR_ITMENA_Msk) != 0UL) &&      /* ITM enabled */
-            ((ITM->TER & (1UL << port)) != 0UL)) {           /* ITM Port enabled */
-        /* write data to port */
-        ITM->PORT[port].u32 = data;
-
-        /* Wait until data has been clocked out */
-        while (ITM->PORT[port].u32 == 0UL) {
-            __NOP();
-        }
+            ((ITM->TER & (1UL << port)) != 0UL)) {       /* ITM Port enabled */
+        itm_out32(port, data);
     }
 
     return data;
 }
 
+void mbed_itm_send_block(uint32_t port, const void *data, size_t len)
+{
+    const char *ptr = data;
+
+    /* Check if ITM and port is enabled */
+    if (((ITM->TCR & ITM_TCR_ITMENA_Msk) != 0UL) &&      /* ITM enabled */
+            ((ITM->TER & (1UL << port)) != 0UL)) {       /* ITM Port enabled */
+        /* Output single byte at a time until data is aligned */
+        while ((((uintptr_t) ptr) & 3) && len != 0) {
+            itm_out8(port, *ptr++);
+            len--;
+        }
+
+        /* Output bulk of data one word at a time */
+        while (len >= 4) {
+            itm_out32(port, *(const uint32_t *) ptr);
+            ptr += 4;
+            len -= 4;
+        }
+
+        /* Output any trailing bytes */
+        while (len != 0) {
+            itm_out8(port, *ptr++);
+            len--;
+        }
+    }
+}
 #endif // defined(DEVICE_ITM)