[S32K3] [RTD 7.0.1] Three Issues in the BCTU Module

[chenwilsoft]

I have found three problems with the BCTU module in the RTD code version S32K3_RTD_7_0_1_D2602_ASR_REL_4_9_REV_0000_20260206. The detailed issues are listed below:

Issue 1: In file Adc_TS_T40D34M70I1R0/src/Bctu_Ip.c, line 1558, the current code uses bitwise OR assignment: BctuBasePtr->FIFOERR |= FifoWatermarkMask; It should be modified to direct assignment: BctuBasePtr->FIFOERR = FifoWatermarkMask;

Issue 2: In file Adc_TS_T40D34M70I1R0/src/Bctu_Ip.c, lines 567 and 568, the existing code contains incorrect bit shift operations. The original code is shown below:

ClrMask |= ((u32Mask & (BCTU_IP_STATUS_FIFO1_OVR_ERR << (Index * 2u))) != 0U) ? (BCTU_FIFOERR_OVR_ERR_FIFO1_MASK << Index) : 0U;

ClrMask |= ((u32Mask & (BCTU_IP_STATUS_FIFO1_UNDR_ERR << (Index * 2u))) != 0U) ? (BCTU_FIFOERR_UNDR_ERR_FIFO1_MASK << Index) : 0U;

These two lines should be revised to the correct shift logic as follows:

ClrMask |= ((u32Mask & (BCTU_IP_STATUS_FIFO1_OVR_ERR << Index)) != 0U) ? (BCTU_FIFOERR_OVR_ERR_FIFO1_MASK << (Index * 2u)) : 0U;

ClrMask |= ((u32Mask & (BCTU_IP_STATUS_FIFO1_UNDR_ERR << Index)) != 0U) ? (BCTU_FIFOERR_UNDR_ERR_FIFO1_MASK << (Index * 2u)) : 0U;

Issue 3: There is a type inconsistency alignment issue between the MCAL/SDK code generation tools and the static source code of the BCTU module. In the MCAL generated code (Adc_TS_T40D34M70I1R0/generate_PB/Adc_RegOperations.m, line 2397), the arrays defined in both header and C files uniformly adopt the uint32 type. In the SDK generated code, the type definition is inconsistent between header and C files: the array type in eclipse/mcu_data/components/PlatformSDK_S32K3/Bctu_Ip/Bctu_Ip_PBcfg.h (line 249) switches between uint16 and uint32 depending on whether the BctuFifoDmaRawData option is enabled, while the corresponding array in eclipse/mcu_data/components/PlatformSDK_S32K3/Bctu_Ip/Bctu_Ip_PBcfg.c (line 317) is fixed as uint32 type. In the static code (Adc_TS_T40D34M70I1R0/src/Bctu_Ip.c, line 1629), the transmission length (2-byte or 4-byte) is determined dynamically based on the BctuFifoDmaRawData configuration. All the following abnormal problems occur when theBctuFifoDmaRawData option is unchecked:

1. For SDK projects: The mismatched array types between header and C files cause direct compilation failures.

2. For MCAL projects: Although compilation can succeed, the fixed uint32 array definition leads to half of the memory space being wasted. Besides, each uint32 data unit contains two ADC results, requiring manual splitting of high and low uint16 values during data parsing.

The above three bugs exist in the official version S32K3_RTD_7_0_1_D2602_ASR_REL_4_9_REV_0000_20260206. We hope NXP software engineers can fix these BCTU module defects in the next RTD release.

[Senlent · Accepted solution]

Hi@chenwilsoft

Thank you for your feedback.

The software team and I have confirmed these issues, they are indeed bugs.

We have escalated these issues internally and will fix them in a future update.

[Senlent]

Hi@chenwilsoft

Your question has been escalated to the internal forum and is awaiting confirmation from the design team.