input range must be in [0:2PI]

Source/FastMathFunctions/arm_cos_f32.c

@@ -38,7 +38,7 @@
  * Computes the trigonometric cosine function using a combination of table lookup
  * and linear interpolation.  There are separate functions for
  * Q15, Q31, and floating-point data types.
- * The input to the floating-point version is in radians while the
+ * The input to the floating-point version is in radians and in the range [0 2*pi) while the
  * fixed-point Q15 and Q31 have a scaled input with the range
  * [0 +0.9999] mapping to [0 2*pi).  The fixed-point range is chosen so that a
  * value of 2*pi wraps around to 0.

Source/FilteringFunctions/arm_conv_partial_f32.c

@@ -420,7 +420,14 @@ arm_status arm_conv_partial_f32(
         count += 4U;
 
         /* Update the inputA and inputB pointers for next MAC calculation */
-        px = pIn1 + count;
+        if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0)
+        {
+          px = pIn1 + firstIndex - srcBLen + 1 + count;
+        }
+        else
+        {
+          px = pIn1 + count;
+        }
         py = pSrc2;
 
         /* Decrement the loop counter */
@@ -473,7 +480,14 @@ arm_status arm_conv_partial_f32(
         count++;
 
         /* Update the inputA and inputB pointers for next MAC calculation */
-        px = pIn1 + count;
+        if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0)
+        {
+          px = pIn1 + firstIndex - srcBLen + 1 + count;
+        }
+        else
+        {
+          px = pIn1 + count;
+        }
         py = pSrc2;
 
         /* Decrement the loop counter */
@@ -510,7 +524,14 @@ arm_status arm_conv_partial_f32(
         count++;
 
         /* Update the inputA and inputB pointers for next MAC calculation */
-        px = pIn1 + count;
+        if ((int32_t)firstIndex - (int32_t)srcBLen + 1 > 0)
+        {
+          px = pIn1 + firstIndex - srcBLen + 1 + count;
+        }
+        else
+        {
+          px = pIn1 + count;
+        }
         py = pSrc2;
 
         /* Decrement the loop counter */