Build error with GCC : incompatible types

[jeromecoutant]

Hello

I am using GCC 13.3.1
and got errors when building this lib:

| xxx/Source/SupportFunctions/arm_q15_to_float.c:74:16: error: incompatible types when assigning to type 'q15x8_t' {aka 'int16x8_t'} from type 'int32x4_t'
| xxx/Source/SupportFunctions/arm_q7_to_float.c:72:18: error: incompatible types when assigning to type 'q7x16_t' {aka 'int8x16_t'} from type 'int32x4_t'
| xxx/Source/MatrixFunctions/arm_mat_ldlt_f32.c:150:19: error: incompatible types when assigning to type 'int32x4_t' from type 'uint32x4_t'
| xxx/Source/FilteringFunctions/arm_biquad_cascade_df1_q15.c:105:13: error: incompatible types when assigning to type 'q15x8_t' {aka 'int16x8_t'} from type 'int32x4_t'
| xxx/Source/FilteringFunctions/arm_biquad_cascade_df1_q15.c:107:13: error: incompatible types when assigning to type 'q15x8_t' {aka 'int16x8_t'} from type 'int32x4_t'

[christophe0606]

@jeromecoutant Which compile options are you using ?
I don't see the warning and yet from looking at the code I can confirm there is an issue.

[jeromecoutant]

Hi

Sorry, I agree I didn't add many details :-(

Projects with some M33 STM32 are OK.

I got issues only for projects with M85

Here are processor option in the generated cbuild.yml file

dsp: 'on'
endian: little
fpu: dp
mve: fp
trustzone: secure-only
core: Cortex-M85

[christophe0606]

@jeromecoutant It should be fixed in latest commit.

[jeromecoutant]

Hi
I checked the latest commit!
Few patches are missing in PrivateInclude/arm_vec_fft.h :

+    bitRevTabOff = (uint32x4_t)vldrhq_u16(pBitRevTab);
+    bitRevOff0Low = (uint32x4_t)vmullbq_int_u16((uint16x8_t)bitRevTabOff, one);
+    bitRevOff0High = (uint32x4_t)vmulltq_int_u16((uint16x8_t)bitRevTabOff, one);
+    bitRevOff0Low = (uint32x4_t)vshrq_n_u16((uint16x8_t)bitRevOff0Low, 3);
+    bitRevOff0High = (uint32x4_t)vshrq_n_u16((uint16x8_t)bitRevOff0High, 3);
+        bitRevTabOff = (uint32x4_t)vldrhq_u16(pBitRevTab);
+        bitRevOff1Low = (uint32x4_t)vshrq_n_u16((uint16x8_t)bitRevOff1Low, 3);
+        bitRevOff1High = (uint32x4_t)vshrq_n_u16((uint16x8_t)bitRevOff1High, 3);
+        bitRevTabOff = (uint32x4_t)vldrhq_u16(pBitRevTab);
+        bitRevOff0Low = (uint32x4_t)vshrq_n_u16((uint16x8_t)bitRevOff0Low, 3);
+        bitRevOff0High = (uint32x4_t)vshrq_n_u16((uint16x8_t)bitRevOff0High, 3);
+        bitRevTabOff = (uint32x4_t)vldrhq_z_u16(pBitRevTab, p);
+        bitRevOff0Low = (uint32x4_t)vshrq_n_u16((uint16x8_t)bitRevOff0Low, 3);
+        bitRevOff0High = (uint32x4_t)vshrq_n_u16((uint16x8_t)bitRevOff0High, 3);
+        vecIn = (uint32x4_t)vldrwq_gather_offset_s32(pSrc, bitRevOffs0);
+        vecIn = (uint32x4_t)vldrwq_gather_offset_s32(pSrc, bitRevOffs1);