🐛 fix stack alignment upon first procedure entry (#1021)

CHANGELOG.md

@@ -29,6 +29,7 @@ mimpid = 0x01040312 -> Version 01.04.03.12 -> v1.4.3.12
 
 | Date | Version | Comment | Ticket |
 |:----:|:-------:|:--------|:------:|
+| 15.09.2024 | 1.10.3.10 | :bug: SW: fix stack-alignment (has to be 128-bit-aligned) before entering the very first procedure (`main`) | [#1021](https://github.com/stnolting/neorv32/pull/1021) |
 | 14.09.2024 | 1.10.3.9 | massive rtl code cleanup | [#1019](https://github.com/stnolting/neorv32/pull/1019) |
 | 14.09.2024 | 1.10.3.8 | :bug: fix `b.ctz` instruction decoding (bug introduced in v1.10.3.6) | [#1018](https://github.com/stnolting/neorv32/pull/1018) |
 | 14.09.2024 | 1.10.3.7 | :warning: rework RTL files / hierarchy | [#1017](https://github.com/stnolting/neorv32/pull/1017) |

docs/datasheet/software.adoc

@@ -444,7 +444,8 @@ contains _explicitly initialized_ global variables. This section is initialized
 like global variables without explicit initialization. This section is cleared by the start-up code `crt0.S`.
 . **Heap (`.heap`)**: The heap is used for dynamic memory that is managed by functions like `malloc()` and `free()`. The heap
 grows upwards. This section is not initialized at all.
-. **Stack**: The stack starts at the very end of the RAM at address `ORIGIN(ram) + LENGTH(ram) - 4`. The stack grows downwards.
+. **Stack**: The stack starts at the very end of the RAM at address `ORIGIN(ram) + LENGTH(ram) - 1`. According to the RISC-V ABI / calling
+convention the stack is 128-bit-aligned before procedure entry. The stack grows downwards.
 
 There is _no explicit limit_ for the maximum stack size as this is hard to check. However, a physical memory protection rule could
 be used to configure a maximum size by adding a "protection area" between stack and heap (a PMP region without any access rights).
@@ -546,7 +547,7 @@ The `crt0.S` start-up performs the following operations:
 . Clear <<_mstatus>>.
 . Clear <<_mie>> disabling all interrupt sources.
 . Install an <<_early_trap_handler>> to <<_mtvec>>.
-. Initialize the global pointer `gp` and the stack pointer `sp` according to the <<_ram_layout>> provided by the linker script.
+. Initialize the global pointer `gp` and the stack pointer `sp` according to the <<_ram_layout>> provided by the linker script. According to the RISC-V ABI the stack pointer gets 128-bit-aligned.
 . Initialize all integer register `x1` - `x31` (only `x1` - `x15` if the `E` CPU extension is enabled).
 . Setup `.data` section to configure initialized variables.
 . Clear the `.bss` section.

rtl/core/neorv32_application_image.vhd

@@ -1,8 +1,8 @@
 -- The NEORV32 RISC-V Processor: https://github.com/stnolting/neorv32
 -- Auto-generated memory initialization file (for APPLICATION) from source file <demo_blink_led/main.bin>
--- Size: 1080 bytes
+-- Size: 1100 bytes
 -- MARCH: default
--- Built: 13.03.2024 16:12:32
+-- Built: 15.09.2024 21:48:24 (dd.mm.yyyy hh:mm:ss)
 
 -- prototype defined in 'neorv32_package.vhd'
 package body neorv32_application_image is
@@ -12,15 +12,15 @@ x"000020b7",
 x"80008093",
 x"30009073",
 x"00000097",
-x"12c08093",
+x"13808093",
 x"30509073",
 x"30401073",
 x"80002217",
 x"fe320213",
-x"ffc27113",
-x"80000297",
-x"7d828293",
-x"ffc2f193",
+x"ff027113",
+x"80000197",
+x"7d818193",
+x"00000293",
 x"00000313",
 x"00000393",
 x"00000413",
@@ -41,7 +41,7 @@ x"00000e13",
 x"00000e93",
 x"00000f13",
 x"00000f93",
-x"43800593",
+x"44c00593",
 x"80000617",
 x"f7860613",
 x"80000697",
@@ -62,23 +62,26 @@ x"00072023",
 x"00470713",
 x"ff5ff06f",
 x"00000417",
-x"36440413",
+x"37840413",
 x"00000497",
-x"35c48493",
+x"37048493",
 x"00945a63",
 x"00042083",
 x"000080e7",
 x"00440413",
 x"ff1ff06f",
 x"00000513",
 x"00000593",
-x"078000ef",
+x"084000ef",
 x"30401073",
 x"34051073",
 x"00000417",
-x"32c40413",
+x"03840413",
+x"30541073",
+x"00000417",
+x"33440413",
 x"00000497",
-x"32448493",
+x"32c48493",
 x"00945a63",
 x"00042083",
 x"000080e7",
@@ -107,40 +110,33 @@ x"00000513",
 x"00000593",
 x"00112623",
 x"00812423",
-x"138000ef",
+x"11c000ef",
 x"00000513",
 x"00150413",
 x"00000593",
 x"0ff57513",
-x"124000ef",
+x"108000ef",
 x"0fa00513",
-x"03c000ef",
+x"020000ef",
 x"00040513",
 x"fe5ff06f",
-x"fffff7b7",
-x"40078793",
-x"0047a583",
-x"0007a503",
-x"0047a703",
-x"fee59ae3",
-x"00008067",
 x"c80027f3",
 x"c0002573",
 x"c80025f3",
 x"fef59ae3",
 x"00008067",
 x"fe010113",
-x"00a12623",
-x"e0002503",
-x"3e800593",
 x"00112e23",
 x"00812c23",
 x"00912a23",
-x"18c000ef",
+x"00a12623",
+x"0f8000ef",
+x"3e800593",
+x"1b0000ef",
 x"00c12603",
 x"00000693",
 x"00000593",
-x"0e8000ef",
+x"10c000ef",
 x"00050413",
 x"00058493",
 x"fc0027f3",
@@ -166,12 +162,12 @@ x"00008067",
 x"e0802783",
 x"00f79713",
 x"02075663",
-x"f3dff0ef",
+x"05c000ef",
 x"00850433",
 x"00a43533",
 x"009585b3",
 x"00b504b3",
-x"f29ff0ef",
+x"048000ef",
 x"fe95eee3",
 x"fcb492e3",
 x"fe856ae3",
@@ -189,6 +185,15 @@ x"c0000793",
 x"00a7a423",
 x"00b7a623",
 x"00008067",
+x"fffff7b7",
+x"40078793",
+x"0047a583",
+x"0007a503",
+x"0047a703",
+x"fee59ae3",
+x"00008067",
+x"e0002503",
+x"00008067",
 x"00050613",
 x"00000513",
 x"0015f693",

rtl/core/neorv32_bootloader_image.vhd

@@ -2,7 +2,7 @@
 -- Auto-generated memory initialization file (for BOOTLOADER) from source file <bootloader/main.bin>
 -- Size: 4068 bytes
 -- MARCH: default
--- Built: 19.07.2024 21:13:24 (dd.mm.yyyy hh:mm:ss)
+-- Built: 15.09.2024 21:48:51 (dd.mm.yyyy hh:mm:ss)
 
 -- prototype defined in 'neorv32_package.vhd'
 package body neorv32_bootloader_image is
@@ -17,10 +17,10 @@ x"30509073",
 x"30401073",
 x"80004217",
 x"1e320213",
-x"ffc27113",
-x"80004297",
-x"7d828293",
-x"ffc2f193",
+x"ff027113",
+x"80004197",
+x"7d818193",
+x"00000293",
 x"00000313",
 x"00000393",
 x"00000413",
@@ -893,9 +893,9 @@ x"6f6c746f",
 x"72656461",
 x"0a3e3e20",
 x"444c420a",
-x"4a203a56",
-x"31206c75",
-x"30322039",
+x"53203a56",
+x"31207065",
+x"30322035",
 x"480a3432",
 x"203a5657",
 x"00000020",

rtl/core/neorv32_package.vhd

@@ -29,7 +29,7 @@ package neorv32_package is
 
   -- Architecture Constants -----------------------------------------------------------------
   -- -------------------------------------------------------------------------------------------
-  constant hw_version_c : std_ulogic_vector(31 downto 0) := x"01100309"; -- hardware version
+  constant hw_version_c : std_ulogic_vector(31 downto 0) := x"01100310"; -- hardware version
   constant archid_c     : natural := 19; -- official RISC-V architecture ID
   constant XLEN         : natural := 32; -- native data path width
 

sw/common/crt0.S

@@ -39,10 +39,9 @@ __crt0_pointer_init:
   .option push
   .option norelax
   // setup pointers using linker script symbols
-  la   x4, __crt0_stack_end  // stack pointer
-  andi sp, x4, 0xfffffffc    // word-aligned
-  la   x5, __global_pointer$ // global pointer
-  andi gp, x5, 0xfffffffc    // word-aligned
+  la   x4, __crt0_ram_last   // last address of RAM, stack starts here
+  andi sp, x4, 0xfffffff0    // align to 16-bytes / 128-bit according to RISC-V ABI (#1021)
+  la   gp, __global_pointer$ // global pointer
   .option pop
 
 __crt0_reg_file_init:
@@ -51,7 +50,7 @@ __crt0_reg_file_init:
 //addi  x2, x0, 0 // stack pointer sp
 //addi  x3, x0, 0 // global pointer gp
 //addi  x4, x0, 0 // implicitly initialized within crt0
-//addi  x5, x0, 0 // implicitly initialized within crt0
+  addi  x5, x0, 0
   addi  x6, x0, 0
   addi  x7, x0, 0
   addi  x8, x0, 0

sw/common/neorv32.ld

@@ -289,9 +289,7 @@ SECTIONS
 /* Export symbols for neorv32 crt0 start-up code                                                     */
 /* ************************************************************************************************* */
   PROVIDE(__crt0_max_heap            = __neorv32_heap_size);
-  PROVIDE(__crt0_imem_begin          = ORIGIN(rom));
-  PROVIDE(__crt0_dmem_begin          = ORIGIN(ram));
-  PROVIDE(__crt0_stack_end           = (ORIGIN(ram) + LENGTH(ram)) - 1);
+  PROVIDE(__crt0_ram_last            = (ORIGIN(ram) + LENGTH(ram)) - 1);
   PROVIDE(__crt0_bss_start           = __BSS_START__);
   PROVIDE(__crt0_bss_end             = __BSS_END__);
   PROVIDE(__crt0_copy_data_src_begin = LOADADDR(.data));

sw/lib/include/neorv32.h

@@ -193,28 +193,13 @@ extern "C" {
  * @name Export linker script symbols
  **************************************************************************/
 /**@{*/
-extern char __heap_start[];               /**< heap start address */
-extern char __heap_end[];                 /**< heap end address */
-extern char __crt0_max_heap[];            /**< heap size in bytes */
-extern char __crt0_imem_begin[];          /**< instruction memory/ROM start address */
-extern char __crt0_dmem_begin[];          /**< data memory/RAM start address */
-extern char __crt0_stack_end[];           /**< last address of stack space */
-extern char __crt0_bss_start[];           /**< bss start address */
-extern char __crt0_bss_end[];             /**< bss end address */
-extern char __crt0_copy_data_dst_begin[]; /**< data start address */
-extern char __crt0_copy_data_dst_end[];   /**< data end address */
-
+extern char __heap_start[];    /**< heap start address */
+extern char __heap_end[];      /**< heap last address */
+extern char __crt0_max_heap[]; /**< heap size in bytes */
 // aliases
 #define neorv32_heap_begin_c ((uint32_t)&__heap_start[0])
 #define neorv32_heap_end_c   ((uint32_t)&__heap_end[0])
 #define neorv32_heap_size_c  ((uint32_t)&__crt0_max_heap[0])
-#define neorv32_imem_begin_c ((uint32_t)&__crt0_imem_begin[0])
-#define neorv32_dmem_begin_c ((uint32_t)&__crt0_dmem_begin[0])
-#define neorv32_stack_end_c  ((uint32_t)&__crt0_stack_end[0])
-#define neorv32_bss_start_c  ((uint32_t)&__crt0_bss_start[0])
-#define neorv32_bss_end_c    ((uint32_t)&__crt0_bss_end[0])
-#define neorv32_data_start_c ((uint32_t)&__crt0_copy_data_dst_begin[0])
-#define neorv32_data_end_c   ((uint32_t)&__crt0_copy_data_dst_end[0])
 /**@}*/