Fix TARGET_WITH in wrong places.

adler32_avx2.c

@@ -33,9 +33,9 @@
 #include "simd.h"
 
 /**
- * @brief Calculate Adler-32 checksum for a given data using TARGET_WITH_AVX2 instructions.
+ * @brief Calculate Adler-32 checksum for a given data using AVX2 instructions.
  *
- * This function calculates the Adler-32 checksum for a block of data using TARGET_WITH_AVX2 vector instructions.
+ * This function calculates the Adler-32 checksum for a block of data using AVX2 vector instructions.
  *
  * @param sum1 Pointer to the variable where the first 16-bit checksum value is stored.
  * @param sum2 Pointer to the variable where the second 16-bit checksum value is stored.

adler32_ssse3.c

@@ -41,9 +41,9 @@
 
 
 /**
- * @brief Calculate Adler-32 checksum for a given data using TARGET_WITH_SSSE3 instructions.
+ * @brief Calculate Adler-32 checksum for a given data using SSSE3 instructions.
  *
- * This function calculates the Adler-32 checksum for a block of data using TARGET_WITH_SSSE3 vector instructions.
+ * This function calculates the Adler-32 checksum for a block of data using SSSE3 vector instructions.
  *
  * @param sum1 Pointer to the variable where the first 16-bit checksum value is stored.
  * @param sum2 Pointer to the variable where the second 16-bit checksum value is stored.

crc32_clmul.c

@@ -224,7 +224,7 @@ TARGET_WITH_CLMUL static void partial_fold(const size_t len,
 #define XOR_INITIAL(where) ONCE(where = _mm_xor_si128(where, xmm_initial))
 
 /**
- * @brief Calculate the CRC32 checksum using TARGET_WITH_CLMUL instruction extension.
+ * @brief Calculate the CRC32 checksum using CLMUL instruction extension.
  *
  * @param previous_crc The previously calculated CRC32 checksum.
  * @param data Pointer to the input data buffer.
@@ -437,7 +437,7 @@ done:
     /*
      * could just as well write xmm_crc3[2], doing a movaps and truncating, but
      * no real advantage - it's a tiny bit slower per call, while no additional CPUs
-     * would be supported by only requiring TARGET_WITH_SSSE3 and TARGET_WITH_CLMUL instead of SSE4.1 + TARGET_WITH_CLMUL
+     * would be supported by only requiring SSSE3 and CLMUL instead of SSE4.1 + CLMUL
      */
     crc = _mm_extract_epi32(xmm_crc3, 2);
     return ~crc;

crc32_vmull.c

@@ -460,7 +460,7 @@ done:
     /*
      * could just as well write q_crc3[2], doing a movaps and truncating, but
      * no real advantage - it's a tiny bit slower per call, while no additional CPUs
-     * would be supported by only requiring TARGET_WITH_SSSE3 and TARGET_WITH_CLMUL instead of SSE4.1 + TARGET_WITH_CLMUL
+     * would be supported by only requiring SSSE3 and CLMUL instead of SSE4.1 + CLMUL
      */
     crc = vgetq_lane_u32(vreinterpretq_u32_u64(q_crc3), (2));
     return ~crc;

crc64_clmul.c

@@ -75,7 +75,7 @@ TARGET_WITH_CLMUL static __m128i fold(__m128i in, __m128i foldConstants)
 }
 
 /**
- * @brief Calculate the CRC-64 checksum using TARGET_WITH_CLMUL instruction extension.
+ * @brief Calculate the CRC-64 checksum using CLMUL instruction extension.
  *
  * @param previous_crc The previously calculated CRC-64 checksum.
  * @param data Pointer to the input data buffer.

fletcher32_neon.c

@@ -38,8 +38,6 @@
 #include "fletcher32.h"
 #include "simd.h"
 
-TARGET_WITH_NEON /***/
-
 /**
  * @brief Calculate Fletcher-32 checksum for a given data using NEON instructions.
  *
@@ -50,7 +48,7 @@ TARGET_WITH_NEON /***/
  * @param data Pointer to the data buffer.
  * @param len Length of the data buffer in bytes.
  */
-void fletcher32_neon(uint16_t *sum1, uint16_t *sum2, const uint8_t *data, uint32_t len)
+TARGET_WITH_NEON void fletcher32_neon(uint16_t *sum1, uint16_t *sum2, const uint8_t *data, uint32_t len)
 {
     /*
      * Split Fletcher-32 into component sums.

fletcher32_ssse3.c

@@ -40,9 +40,9 @@
 #include "fletcher32.h"
 
 /**
- * @brief Calculate Fletcher-32 checksum for a given data using TARGET_WITH_SSSE3 instructions.
+ * @brief Calculate Fletcher-32 checksum for a given data using SSSE3 instructions.
  *
- * This function calculates the Fletcher-32 checksum for a block of data using TARGET_WITH_SSSE3 vector instructions.
+ * This function calculates the Fletcher-32 checksum for a block of data using SSSE3 vector instructions.
  *
  * @param sum1 Pointer to the variable where the first 16-bit checksum value is stored.
  * @param sum2 Pointer to the variable where the second 16-bit checksum value is stored.

simd.c

@@ -123,15 +123,15 @@ static void cpuidex(int info, int count, unsigned* eax, unsigned* ebx, unsigned*
 }
 
 /**
- * @brief Checks if the hardware supports the TARGET_WITH_CLMUL instruction set.
+ * @brief Checks if the hardware supports the CLMUL instruction set.
  *
- * The function checks if the system's CPU supports the TARGET_WITH_CLMUL (Carry-Less Multiplication) instruction set.
+ * The function checks if the system's CPU supports the CLMUL (Carry-Less Multiplication) instruction set.
- * TARGET_WITH_CLMUL is an extension to the x86 instruction set architecture and provides hardware acceleration for
+ * CLMUL is an extension to the x86 instruction set architecture and provides hardware acceleration for
  * carry-less multiplication operations.
  *
- * @return True if TARGET_WITH_CLMUL instruction set is supported, False otherwise.
+ * @return True if CLMUL instruction set is supported, False otherwise.
  *
- * @see https://software.intel.com/sites/landingpage/IntrinsicsGuide/#techs=TARGET_WITH_CLMUL
+ * @see https://software.intel.com/sites/landingpage/IntrinsicsGuide/#techs=CLMUL
  * @see https://en.wikipedia.org/wiki/Carry-less_multiplication
  */
 int have_clmul(void)
@@ -148,17 +148,17 @@ int have_clmul(void)
 }
 
 /**
- * @brief Checks if the current processor supports TARGET_WITH_SSSE3 instructions.
+ * @brief Checks if the current processor supports SSSE3 instructions.
  *
- * The function detects whether the current processor supports TARGET_WITH_SSSE3 instructions by
+ * The function detects whether the current processor supports SSSE3 instructions by
- * checking the CPU feature flags. TARGET_WITH_SSSE3 (Supplemental Streaming SIMD Extensions 3)
+ * checking the CPU feature flags. SSSE3 (Supplemental Streaming SIMD Extensions 3)
  * is an extension to the x86 instruction set architecture that introduces
  * additional SIMD instructions useful for multimedia and signal processing tasks.
  *
- * @return true if the current processor supports TARGET_WITH_SSSE3 instructions, false otherwise.
+ * @return true if the current processor supports SSSE3 instructions, false otherwise.
  *
- * @see https://software.intel.com/sites/landingpage/IntrinsicsGuide/#techs=TARGET_WITH_SSSE3
+ * @see https://software.intel.com/sites/landingpage/IntrinsicsGuide/#techs=SSSE3
- * @see https://en.wikipedia.org/wiki/TARGET_WITH_SSSE3
+ * @see https://en.wikipedia.org/wiki/SSSE3
  */
 int have_ssse3(void)
 {
@@ -169,16 +169,16 @@ int have_ssse3(void)
 }
 
 /**
- * @brief Checks if the current processor supports TARGET_WITH_AVX2 instructions.
+ * @brief Checks if the current processor supports AVX2 instructions.
  *
- * The function detects whether the current processor supports TARGET_WITH_AVX2 instructions by
+ * The function detects whether the current processor supports AVX2 instructions by
- * checking the CPU feature flags. TARGET_WITH_AVX2 (Advanced Vector Extensions 2) is an extension
+ * checking the CPU feature flags. AVX2 (Advanced Vector Extensions 2) is an extension
  * to the x86 instruction set architecture that introduces additional SIMD instructions
  * useful for multimedia and signal processing tasks.
  *
- * @return true if the current processor supports TARGET_WITH_AVX2 instructions, false otherwise.
+ * @return true if the current processor supports AVX2 instructions, false otherwise.
  *
- * @see https://software.intel.com/sites/landingpage/IntrinsicsGuide/#techs=TARGET_WITH_AVX2
+ * @see https://software.intel.com/sites/landingpage/IntrinsicsGuide/#techs=AVX2
  * @see https://en.wikipedia.org/wiki/Advanced_Vector_Extensions
  */