Merge branch 'master' into rp2040-hcd-bulk

7 files changed, 77 insertions, 42 deletions

diff --git a/src/common/tusb_common.h b/src/common/tusb_common.h
index 58591f0be..b1ee40a1a 100644
--- a/src/common/tusb_common.h
+++ b/src/common/tusb_common.h
@@ -134,12 +134,6 @@ TU_ATTR_ALWAYS_INLINE static inline uint32_t tu_offset4k(uint32_t value) { retur
 //------------- Mathematics -------------//
 TU_ATTR_ALWAYS_INLINE static inline uint32_t tu_div_ceil(uint32_t v, uint32_t d) { return (v + d -1)/d; }
 
-/// inclusive range checking TODO remove
-TU_ATTR_ALWAYS_INLINE static inline bool tu_within(uint32_t lower, uint32_t value, uint32_t upper)
-{
-  return (lower <= value) && (value <= upper);
-}
-
 // log2 of a value is its MSB's position
 // TODO use clz TODO remove
 static inline uint8_t tu_log2(uint32_t value)
@@ -149,6 +143,16 @@ static inline uint8_t tu_log2(uint32_t value)
   return result;
 }
 
+//static inline uint8_t tu_log2(uint32_t value)
+//{
+//   return sizeof(uint32_t) * CHAR_BIT - __builtin_clz(x) - 1;
+//}
+
+static inline bool tu_is_power_of_two(uint32_t value)
+{
+   return (value != 0) && ((value & (value - 1)) == 0);
+}
+
 //------------- Unaligned Access -------------//
 #if TUP_ARCH_STRICT_ALIGN
 
diff --git a/src/common/tusb_compiler.h b/src/common/tusb_compiler.h
index ae00f4e75..2c30daf6f 100644
--- a/src/common/tusb_compiler.h
+++ b/src/common/tusb_compiler.h
@@ -131,6 +131,12 @@
   #define TU_ATTR_BIT_FIELD_ORDER_BEGIN
   #define TU_ATTR_BIT_FIELD_ORDER_END
 
+  #if __has_attribute(__fallthrough__)
+    #define TU_ATTR_FALLTHROUGH         __attribute__((fallthrough))
+  #else
+    #define TU_ATTR_FALLTHROUGH         do {} while (0)  /* fallthrough */
+  #endif
+
   // Endian conversion use well-known host to network (big endian) naming
   #if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
     #define TU_BYTE_ORDER TU_LITTLE_ENDIAN
@@ -156,6 +162,7 @@
   #define TU_ATTR_DEPRECATED(mess)      __attribute__ ((deprecated(mess))) // warn if function with this attribute is used
   #define TU_ATTR_UNUSED                __attribute__ ((unused))           // Function/Variable is meant to be possibly unused
   #define TU_ATTR_USED                  __attribute__ ((used))
+  #define TU_ATTR_FALLTHROUGH           __attribute__((fallthrough))
 
   #define TU_ATTR_PACKED_BEGIN
   #define TU_ATTR_PACKED_END
@@ -182,6 +189,7 @@
   #define TU_ATTR_DEPRECATED(mess)      __attribute__ ((deprecated(mess))) // warn if function with this attribute is used
   #define TU_ATTR_UNUSED                __attribute__ ((unused))           // Function/Variable is meant to be possibly unused
   #define TU_ATTR_USED                  __attribute__ ((used))             // Function/Variable is meant to be used
+  #define TU_ATTR_FALLTHROUGH           __attribute__((fallthrough))
 
   #define TU_ATTR_PACKED_BEGIN
   #define TU_ATTR_PACKED_END
@@ -207,6 +215,7 @@
   #define TU_ATTR_DEPRECATED(mess)
   #define TU_ATTR_UNUSED
   #define TU_ATTR_USED
+  #define TU_ATTR_FALLTHROUGH           do {} while (0)  /* fallthrough */
 
   #define TU_ATTR_PACKED_BEGIN          _Pragma("pack")
   #define TU_ATTR_PACKED_END            _Pragma("packoption")
@@ -227,6 +236,7 @@
   #error "Compiler attribute porting is required"
 #endif
 
+
 #if (TU_BYTE_ORDER == TU_LITTLE_ENDIAN)
 
   #define tu_htons(u16)  (TU_BSWAP16(u16))
diff --git a/src/common/tusb_debug.h b/src/common/tusb_debug.h
index 647e8a8db..ac5bee6ec 100644
--- a/src/common/tusb_debug.h
+++ b/src/common/tusb_debug.h
@@ -57,7 +57,7 @@ void tu_print_mem(void const *buf, uint32_t count, uint8_t indent);
   #define tu_printf    printf
 #endif
 
-static inline void tu_print_var(uint8_t const* buf, uint32_t bufsize)
+static inline void tu_print_arr(uint8_t const* buf, uint32_t bufsize)
 {
   for(uint32_t i=0; i<bufsize; i++) tu_printf("%02X ", buf[i]);
 }
@@ -65,6 +65,7 @@ static inline void tu_print_var(uint8_t const* buf, uint32_t bufsize)
 // Log with Level
 #define TU_LOG(n, ...)        TU_XSTRCAT(TU_LOG, n)(__VA_ARGS__)
 #define TU_LOG_MEM(n, ...)    TU_XSTRCAT3(TU_LOG, n, _MEM)(__VA_ARGS__)
+#define TU_LOG_ARR(n, ...)    TU_XSTRCAT3(TU_LOG, n, _ARR)(__VA_ARGS__)
 #define TU_LOG_VAR(n, ...)    TU_XSTRCAT3(TU_LOG, n, _VAR)(__VA_ARGS__)
 #define TU_LOG_INT(n, ...)    TU_XSTRCAT3(TU_LOG, n, _INT)(__VA_ARGS__)
 #define TU_LOG_HEX(n, ...)    TU_XSTRCAT3(TU_LOG, n, _HEX)(__VA_ARGS__)
@@ -74,7 +75,8 @@ static inline void tu_print_var(uint8_t const* buf, uint32_t bufsize)
 // Log Level 1: Error
 #define TU_LOG1               tu_printf
 #define TU_LOG1_MEM           tu_print_mem
-#define TU_LOG1_VAR(_x)       tu_print_var((uint8_t const*)(_x), sizeof(*(_x)))
+#define TU_LOG1_ARR(_x, _n)   tu_print_arr((uint8_t const*)(_x), _n)
+#define TU_LOG1_VAR(_x)       tu_print_arr((uint8_t const*)(_x), sizeof(*(_x)))
 #define TU_LOG1_INT(_x)       tu_printf(#_x " = %ld\r\n", (unsigned long) (_x) )
 #define TU_LOG1_HEX(_x)       tu_printf(#_x " = %lX\r\n", (unsigned long) (_x) )
 
@@ -82,6 +84,7 @@ static inline void tu_print_var(uint8_t const* buf, uint32_t bufsize)
 #if CFG_TUSB_DEBUG >= 2
   #define TU_LOG2             TU_LOG1
   #define TU_LOG2_MEM         TU_LOG1_MEM
+  #define TU_LOG2_ARR         TU_LOG1_ARR
   #define TU_LOG2_VAR         TU_LOG1_VAR
   #define TU_LOG2_INT         TU_LOG1_INT
   #define TU_LOG2_HEX         TU_LOG1_HEX
@@ -91,6 +94,7 @@ static inline void tu_print_var(uint8_t const* buf, uint32_t bufsize)
 #if CFG_TUSB_DEBUG >= 3
   #define TU_LOG3             TU_LOG1
   #define TU_LOG3_MEM         TU_LOG1_MEM
+  #define TU_LOG3_ARR         TU_LOG1_ARR
   #define TU_LOG3_VAR         TU_LOG1_VAR
   #define TU_LOG3_INT         TU_LOG1_INT
   #define TU_LOG3_HEX         TU_LOG1_HEX
diff --git a/src/common/tusb_fifo.c b/src/common/tusb_fifo.c
index 183c9c6fc..895b9208b 100644
--- a/src/common/tusb_fifo.c
+++ b/src/common/tusb_fifo.c
@@ -79,7 +79,7 @@ bool tu_fifo_config(tu_fifo_t *f, void* buffer, uint16_t depth, uint16_t item_si
   // Limit index space to 2*depth - this allows for a fast "modulo" calculation
   // but limits the maximum depth to 2^16/2 = 2^15 and buffer overflows are detectable
   // only if overflow happens once (important for unsupervised DMA applications)
-  f->max_pointer_idx = 2*depth - 1;
+  f->max_pointer_idx = (uint16_t) (2*depth - 1);
   f->non_used_index_space = UINT16_MAX - f->max_pointer_idx;
 
   f->rd_idx = f->wr_idx = 0;
@@ -205,7 +205,7 @@ static void _ff_push_n(tu_fifo_t* f, void const * app_buf, uint16_t n, uint16_t
         uint8_t rem = nLin_bytes & 0x03;
         if (rem > 0)
         {
-          uint8_t remrem = tu_min16(nWrap_bytes, 4-rem);
+          uint8_t remrem = (uint8_t) tu_min16(nWrap_bytes, 4-rem);
           nWrap_bytes -= remrem;
 
           uint32_t tmp32 = *rx_fifo;
@@ -288,7 +288,7 @@ static void _ff_pull_n(tu_fifo_t* f, void* app_buf, uint16_t n, uint16_t rel, tu
         uint8_t rem = nLin_bytes & 0x03;
         if (rem > 0)
         {
-          uint8_t remrem = tu_min16(nWrap_bytes, 4-rem);
+          uint8_t remrem = (uint8_t) tu_min16(nWrap_bytes, 4-rem);
           nWrap_bytes -= remrem;
 
           uint32_t tmp32=0;
@@ -325,7 +325,7 @@ static uint16_t advance_pointer(tu_fifo_t* f, uint16_t p, uint16_t offset)
   // We are exploiting the wrap around to the correct index
   if ((p > (uint16_t)(p + offset)) || ((uint16_t)(p + offset) > f->max_pointer_idx))
   {
-    p = (p + offset) + f->non_used_index_space;
+    p = (uint16_t) ((p + offset) + f->non_used_index_space);
   }
   else
   {
@@ -342,7 +342,7 @@ static uint16_t backward_pointer(tu_fifo_t* f, uint16_t p, uint16_t offset)
   // We are exploiting the wrap around to the correct index
   if ((p < (uint16_t)(p - offset)) || ((uint16_t)(p - offset) > f->max_pointer_idx))
   {
-    p = (p - offset) - f->non_used_index_space;
+    p = (uint16_t) ((p - offset) - f->non_used_index_space);
   }
   else
   {
@@ -818,7 +818,7 @@ bool tu_fifo_clear(tu_fifo_t *f)
   _ff_lock(f->mutex_rd);
 
   f->rd_idx = f->wr_idx = 0;
-  f->max_pointer_idx = 2*f->depth-1;
+  f->max_pointer_idx = (uint16_t) (2*f->depth-1);
   f->non_used_index_space = UINT16_MAX - f->max_pointer_idx;
 
   _ff_unlock(f->mutex_wr);
diff --git a/src/common/tusb_mcu.h b/src/common/tusb_mcu.h
index 8eb4ad475..86c68baf8 100644
--- a/src/common/tusb_mcu.h
+++ b/src/common/tusb_mcu.h
@@ -45,8 +45,7 @@
  * - ENDPOINT_MAX: max (logical) number of endpoint
  * - ENDPOINT_EXCLUSIVE_NUMBER: endpoint number with different direction IN and OUT aren't allowed,
  *                              e.g EP1 OUT & EP1 IN cannot exist together
- * - RHPORT_HIGHSPEED: mask to indicate which port support highspeed mode (without external PHY)
- *                     bit0 for port0 and so on.
+ * - RHPORT_HIGHSPEED: support highspeed with on-chip PHY
  */
 
 //------------- NXP -------------//
@@ -63,8 +62,7 @@
   #define TUP_USBIP_EHCI
 
   #define TUP_DCD_ENDPOINT_MAX    6
-  #define TUP_RHPORT_HIGHSPEED    0x01 // Port0 HS, Port1 FS
-
+  #define TUP_RHPORT_HIGHSPEED    1 // Port0 HS, Port1 FS
 
 #elif TU_CHECK_MCU(OPT_MCU_LPC51UXX)
    #define TUP_DCD_ENDPOINT_MAX   5
@@ -77,13 +75,12 @@
   // TODO USB0 has 5, USB1 has 6
   #define TUP_DCD_ENDPOINT_MAX    6
 
-#elif TU_CHECK_MCU(OPT_MCU_MIMXRT10XX)
+#elif TU_CHECK_MCU(OPT_MCU_MIMXRT)
   #define TUP_USBIP_CHIPIDEA_HS
   #define TUP_USBIP_EHCI
 
   #define TUP_DCD_ENDPOINT_MAX    8
-  #define TUP_RHPORT_HIGHSPEED    0x03 // Port0 HS, Port1 HS
-
+  #define TUP_RHPORT_HIGHSPEED    1 // Port0 HS, Port1 HS
 
 #elif TU_CHECK_MCU(OPT_MCU_MKL25ZXX, OPT_MCU_K32L2BXX)
   #define TUP_DCD_ENDPOINT_MAX    16
@@ -107,13 +104,18 @@
 
 #elif TU_CHECK_MCU(OPT_MCU_SAMX7X)
   #define TUP_DCD_ENDPOINT_MAX    10
-  #define TUP_RHPORT_HIGHSPEED    0x01
+  #define TUP_RHPORT_HIGHSPEED    1
   #define TUP_DCD_ENDPOINT_EXCLUSIVE_NUMBER
 
 #elif TU_CHECK_MCU(OPT_MCU_PIC32MZ)
   #define TUP_DCD_ENDPOINT_MAX    8
   #define TUP_DCD_ENDPOINT_EXCLUSIVE_NUMBER
 
+#elif TU_CHECK_MCU(OPT_MCU_PIC32MX, OPT_MCU_PIC32MM, OPT_MCU_PIC32MK) || \
+      TU_CHECK_MCU(OPT_MCU_PIC24, OPT_MCU_DSPIC33)
+  #define TUP_DCD_ENDPOINT_MAX    16
+  #define TUP_DCD_ENDPOINT_EXCLUSIVE_NUMBER
+
 //------------- ST -------------//
 #elif TU_CHECK_MCU(OPT_MCU_STM32F0)
   #define TUP_DCD_ENDPOINT_MAX    8
@@ -155,7 +157,7 @@
 
   // MCU with on-chip HS Phy
   #if defined(STM32F723xx) || defined(STM32F730xx) || defined(STM32F733xx)
-    #define TUP_RHPORT_HIGHSPEED  0x02 // Port 0: FS, Port 1: HS
+    #define TUP_RHPORT_HIGHSPEED  1 // Port0: FS, Port1: HS
   #endif
 
 #elif TU_CHECK_MCU(OPT_MCU_STM32H7)
@@ -185,12 +187,17 @@
   #endif
 
 #elif TU_CHECK_MCU(OPT_MCU_STM32WB)
-#define TUP_DCD_ENDPOINT_MAX    8
+  #define TUP_DCD_ENDPOINT_MAX    8
+
+#elif TU_CHECK_MCU(OPT_MCU_STM32U5)
+  #define TUP_USBIP_DWC2
+  #define TUP_USBIP_DWC2_STM32
+  #define TUP_DCD_ENDPOINT_MAX    6
 
 //------------- Sony -------------//
 #elif TU_CHECK_MCU(OPT_MCU_CXD56)
   #define TUP_DCD_ENDPOINT_MAX    7
-  #define TUP_RHPORT_HIGHSPEED    0x01
+  #define TUP_RHPORT_HIGHSPEED    1
   #define TUP_DCD_ENDPOINT_EXCLUSIVE_NUMBER
 
 //------------- TI -------------//
@@ -213,7 +220,7 @@
 
 #elif TU_CHECK_MCU(OPT_MCU_NUC505)
   #define TUP_DCD_ENDPOINT_MAX    12
-  #define TUP_RHPORT_HIGHSPEED    0x01
+  #define TUP_RHPORT_HIGHSPEED    1
 
 //------------- Espressif -------------//
 #elif TU_CHECK_MCU(OPT_MCU_ESP32S2, OPT_MCU_ESP32S3)
@@ -228,6 +235,8 @@
 #elif TU_CHECK_MCU(OPT_MCU_RP2040)
   #define TUP_DCD_ENDPOINT_MAX    16
 
+  #define TU_ATTR_FAST_FUNC       __attribute__((section(".time_critical.tinyusb")))
+
 //------------- Silabs -------------//
 #elif TU_CHECK_MCU(OPT_MCU_EFM32GG)
   #define TUP_USBIP_DWC2
@@ -246,7 +255,7 @@
 #elif TU_CHECK_MCU(OPT_MCU_BCM2711, OPT_MCU_BCM2835, OPT_MCU_BCM2837)
   #define TUP_USBIP_DWC2
   #define TUP_DCD_ENDPOINT_MAX    8
-  #define TUP_RHPORT_HIGHSPEED    0x01
+  #define TUP_RHPORT_HIGHSPEED    1
 
 //------------- Broadcom -------------//
 #elif TU_CHECK_MCU(OPT_MCU_XMC4000)
@@ -256,11 +265,11 @@
 //------------- BridgeTek -------------//
 #elif TU_CHECK_MCU(OPT_MCU_FT90X)
   #define TUP_DCD_ENDPOINT_MAX    8
-  #define TUP_RHPORT_HIGHSPEED    0x01
+  #define TUP_RHPORT_HIGHSPEED    1
 
 #elif TU_CHECK_MCU(OPT_MCU_FT93X)
   #define TUP_DCD_ENDPOINT_MAX    16
-  #define TUP_RHPORT_HIGHSPEED    0x01
+  #define TUP_RHPORT_HIGHSPEED    1
 
 //------------ Allwinner -------------//
 #elif TU_CHECK_MCU(OPT_MCU_F1C100S)
@@ -279,7 +288,12 @@
 
 // Default to fullspeed if not defined
 #ifndef TUP_RHPORT_HIGHSPEED
-  #define TUP_RHPORT_HIGHSPEED    0x00
+  #define TUP_RHPORT_HIGHSPEED    0
+#endif
+
+// fast function, normally mean placing function in SRAM
+#ifndef TU_ATTR_FAST_FUNC
+  #define TU_ATTR_FAST_FUNC
 #endif
 
 #endif
diff --git a/src/common/tusb_types.h b/src/common/tusb_types.h
index 7eab396ef..cfa24bd82 100644
--- a/src/common/tusb_types.h
+++ b/src/common/tusb_types.h
@@ -492,23 +492,23 @@ TU_ATTR_BIT_FIELD_ORDER_END
 //--------------------------------------------------------------------+
 
 // Get direction from Endpoint address
-static inline tusb_dir_t tu_edpt_dir(uint8_t addr)
+TU_ATTR_ALWAYS_INLINE static inline tusb_dir_t tu_edpt_dir(uint8_t addr)
 {
   return (addr & TUSB_DIR_IN_MASK) ? TUSB_DIR_IN : TUSB_DIR_OUT;
 }
 
 // Get Endpoint number from address
-static inline uint8_t tu_edpt_number(uint8_t addr)
+TU_ATTR_ALWAYS_INLINE static inline uint8_t tu_edpt_number(uint8_t addr)
 {
   return (uint8_t)(addr & (~TUSB_DIR_IN_MASK));
 }
 
-static inline uint8_t tu_edpt_addr(uint8_t num, uint8_t dir)
+TU_ATTR_ALWAYS_INLINE static inline uint8_t tu_edpt_addr(uint8_t num, uint8_t dir)
 {
   return (uint8_t)(num | (dir ? TUSB_DIR_IN_MASK : 0));
 }
 
-static inline uint16_t tu_edpt_packet_size(tusb_desc_endpoint_t const* desc_ep)
+TU_ATTR_ALWAYS_INLINE static inline uint16_t tu_edpt_packet_size(tusb_desc_endpoint_t const* desc_ep)
 {
   return tu_le16toh(desc_ep->wMaxPacketSize) & TU_GENMASK(10, 0);
 }
@@ -526,18 +526,18 @@ static inline const char *tu_edpt_type_str(tusb_xfer_type_t t) {
 //--------------------------------------------------------------------+
 // Descriptor helper
 //--------------------------------------------------------------------+
-static inline uint8_t const * tu_desc_next(void const* desc)
+TU_ATTR_ALWAYS_INLINE static inline uint8_t const * tu_desc_next(void const* desc)
 {
   uint8_t const* desc8 = (uint8_t const*) desc;
   return desc8 + desc8[DESC_OFFSET_LEN];
 }
 
-static inline uint8_t tu_desc_type(void const* desc)
+TU_ATTR_ALWAYS_INLINE static inline uint8_t tu_desc_type(void const* desc)
 {
   return ((uint8_t const*) desc)[DESC_OFFSET_TYPE];
 }
 
-static inline uint8_t tu_desc_len(void const* desc)
+TU_ATTR_ALWAYS_INLINE static inline uint8_t tu_desc_len(void const* desc)
 {
   return ((uint8_t const*) desc)[DESC_OFFSET_LEN];
 }
diff --git a/src/common/tusb_verify.h b/src/common/tusb_verify.h
index f4a08ce2f..a52a6d269 100644
--- a/src/common/tusb_verify.h
+++ b/src/common/tusb_verify.h
@@ -90,6 +90,9 @@
 #elif defined(__riscv)
   #define TU_BREAKPOINT() do { __asm("ebreak\n"); } while(0)
 
+#elif defined(_mips)
+  #define TU_BREAKPOINT() do { __asm("sdbbp 0"); } while (0)
+
 #else
   #define TU_BREAKPOINT() do {} while (0)
 #endif
@@ -99,8 +102,8 @@
  *------------------------------------------------------------------*/
 
 // Helper to implement optional parameter for TU_VERIFY Macro family
-#define GET_3RD_ARG(arg1, arg2, arg3, ...)        arg3
-#define GET_4TH_ARG(arg1, arg2, arg3, arg4, ...)  arg4
+#define _GET_3RD_ARG(arg1, arg2, arg3, ...)        arg3
+#define _GET_4TH_ARG(arg1, arg2, arg3, arg4, ...)  arg4
 
 /*------------- Generator for TU_VERIFY and TU_VERIFY_HDLR -------------*/
 #define TU_VERIFY_DEFINE(_cond, _handler, _ret)  do            \
@@ -116,7 +119,7 @@
 #define TU_VERIFY_1ARGS(_cond)                         TU_VERIFY_DEFINE(_cond, , false)
 #define TU_VERIFY_2ARGS(_cond, _ret)                   TU_VERIFY_DEFINE(_cond, , _ret)
 
-#define TU_VERIFY(...)                   GET_3RD_ARG(__VA_ARGS__, TU_VERIFY_2ARGS, TU_VERIFY_1ARGS, UNUSED)(__VA_ARGS__)
+#define TU_VERIFY(...)                   _GET_3RD_ARG(__VA_ARGS__, TU_VERIFY_2ARGS, TU_VERIFY_1ARGS, UNUSED)(__VA_ARGS__)
 
 
 /*------------------------------------------------------------------*/
@@ -127,7 +130,7 @@
 #define TU_VERIFY_HDLR_2ARGS(_cond, _handler)           TU_VERIFY_DEFINE(_cond, _handler, false)
 #define TU_VERIFY_HDLR_3ARGS(_cond, _handler, _ret)     TU_VERIFY_DEFINE(_cond, _handler, _ret)
 
-#define TU_VERIFY_HDLR(...)              GET_4TH_ARG(__VA_ARGS__, TU_VERIFY_HDLR_3ARGS, TU_VERIFY_HDLR_2ARGS,UNUSED)(__VA_ARGS__)
+#define TU_VERIFY_HDLR(...)              _GET_4TH_ARG(__VA_ARGS__, TU_VERIFY_HDLR_3ARGS, TU_VERIFY_HDLR_2ARGS,UNUSED)(__VA_ARGS__)
 
 /*------------------------------------------------------------------*/
 /* ASSERT
@@ -139,7 +142,7 @@
 #define ASSERT_2ARGS(_cond, _ret)      TU_VERIFY_DEFINE(_cond, _MESS_FAILED(); TU_BREAKPOINT(), _ret)
 
 #ifndef TU_ASSERT
-#define TU_ASSERT(...)             GET_3RD_ARG(__VA_ARGS__, ASSERT_2ARGS, ASSERT_1ARGS,UNUSED)(__VA_ARGS__)
+#define TU_ASSERT(...)             _GET_3RD_ARG(__VA_ARGS__, ASSERT_2ARGS, ASSERT_1ARGS,UNUSED)(__VA_ARGS__)
 #endif
 
 /*------------------------------------------------------------------*/