samd5x_e5x: ETM trace support for same54_xplained

The populated 20-pin Cortex Debug+ETM header carries 4-bit trace
(TRACECLK=PC27, D0-3=PC28/PC26/PC25/PC24, mux H). TRACE_ETM builds mux
the pins and enable GCLK channel 47 (GCLK_CM4_TRACE) from GCLK0 - without
that gate the port stays silent with pins and TPIU armed. Chip-max
120 MHz core / 60 MHz TRACECLK validated (3x 280M-fetch captures).

hw/bsp/samd5x_e5x/boards/same54_xplained/ozone/same54.jdebug

@@ -0,0 +1,89 @@
+/*********************************************************************
+*
+*       OnProjectLoad
+*
+* Function description
+*   Project load routine. Required.
+*
+* Notes
+*   SAM E54 Xplained Pro carries a populated 20-pin Cortex Debug+ETM
+*   connector (Table 5-11): TRACECLK=PC27, D0=PC28, D1=PC26, D2=PC25,
+*   D3=PC24 - no rework needed. Firmware must be built with TRACE_ETM=1
+*   (mux function H on those pins in board_init); TPIU/ETM are
+*   ROM-table-discoverable so no J-Link script is required.
+*   Trace clock is CPU/2 = 60 MHz at the stock 120 MHz core.
+*
+**********************************************************************
+*/
+void OnProjectLoad (void) {
+  Project.SetTraceSource ("Trace Pins");
+  Project.SetTracePortWidth (4);
+  Project.SetSWO (0);
+  Edit.SysVar (VAR_TRACE_CORE_CLOCK, 120000000);
+  Project.AddSvdFile ("$(InstallDir)/Config/CPU/Cortex-M4F.svd");
+
+  Project.SetDevice ("ATSAME54P20");
+  Project.SetHostIF ("USB", "");
+  Project.SetTargetIF ("SWD");
+  Project.SetTIFSpeed ("4 MHz");
+
+  File.Open ("../../../../../../examples/cmake-build-same54_xplained/device/cdc_msc/cdc_msc.elf");
+}
+
+/*********************************************************************
+*
+*       AfterTargetReset
+*
+* Function description
+*   Event handler routine.
+*    - Sets the PC register to program reset value.
+*    - Sets the SP register to program reset value on Cortex-M.
+*
+**********************************************************************
+*/
+void AfterTargetReset (void) {
+  unsigned int SP;
+  unsigned int PC;
+  unsigned int VectorTableAddr;
+
+  VectorTableAddr = Elf.GetBaseAddr();
+
+  if (VectorTableAddr == 0xFFFFFFFF) {
+    Util.Log("Project file error: failed to get program base");
+  } else {
+    SP = Target.ReadU32(VectorTableAddr);
+    Target.SetReg("SP", SP);
+
+    PC = Target.ReadU32(VectorTableAddr + 4);
+    Target.SetReg("PC", PC);
+  }
+}
+
+/*********************************************************************
+*
+*       AfterTargetDownload
+*
+* Function description
+*   Event handler routine.
+*    - Sets the PC register to program reset value.
+*    - Sets the SP register to program reset value on Cortex-M.
+*
+**********************************************************************
+*/
+void AfterTargetDownload (void) {
+  unsigned int SP;
+  unsigned int PC;
+  unsigned int VectorTableAddr;
+
+  VectorTableAddr = Elf.GetBaseAddr();
+
+  if (VectorTableAddr == 0xFFFFFFFF) {
+    Util.Log("Project file error: failed to get program base");
+  } else {
+    SP = Target.ReadU32(VectorTableAddr);
+    Target.SetReg("SP", SP);
+
+    PC = Target.ReadU32(VectorTableAddr + 4);
+    Target.SetReg("PC", PC);
+  }
+}

hw/bsp/samd5x_e5x/family.c

@@ -87,6 +87,31 @@ void USB_3_Handler(void) { USB_Any_Handler(); }
 static void max3421_init(void);
 #endif
 
+#if defined(TRACE_ETM)
+// same54_xplained routes 4-bit trace to its 20-pin Cortex Debug+ETM header:
+// TRACECLK=PC27, D0=PC28, D1=PC26, D2=PC25, D3=PC24 - all peripheral
+// function H (CM4 trace). TPIU/ETM are ROM-table-discoverable; the debugger
+// arms them, firmware only muxes the pins.
+static void trace_etm_init(void) {
+  // the CM4 trace unit runs from its own GCLK channel (47): feed it GCLK0
+  // (CPU clock) - without this the pins mux fine but the port stays silent
+  GCLK->PCHCTRL[47].reg = GCLK_PCHCTRL_GEN_GCLK0 | GCLK_PCHCTRL_CHEN;
+  while (!(GCLK->PCHCTRL[47].reg & GCLK_PCHCTRL_CHEN)) {}
+
+  const uint8_t pin[] = {24, 25, 26, 27, 28};
+  for (unsigned i = 0; i < 5; i++) {
+    PORT->Group[2].PINCFG[pin[i]].reg = PORT_PINCFG_PMUXEN | PORT_PINCFG_DRVSTR;
+    if (pin[i] & 1) {
+      PORT->Group[2].PMUX[pin[i] >> 1].bit.PMUXO = 7; // function H
+    } else {
+      PORT->Group[2].PMUX[pin[i] >> 1].bit.PMUXE = 7;
+    }
+  }
+}
+#else
+  #define trace_etm_init()
+#endif
+
 void board_init(void) {
   // Clock init ( follow hpl_init.c )
   hri_nvmctrl_set_CTRLA_RWS_bf(NVMCTRL, 0);
@@ -104,6 +129,8 @@ void board_init(void) {
   // Init 1ms tick timer (samd SystemCoreClock may not correct)
   SystemCoreClock = CONF_CPU_FREQUENCY;
 
+  trace_etm_init();
+
 #if CFG_TUSB_OS == OPT_OS_NONE
   SysTick_Config(CONF_CPU_FREQUENCY / 1000);
 #elif CFG_TUSB_OS == OPT_OS_FREERTOS