SD_Read bug?

[con3]

Posted on November 29, 2017 at 14:53

Good day everyone,

I'm not sure if this is a bug or I'm looking at this incorrectly, but It seems to be a bug.

While calling the f_mount function there is a point where SD_read is called.

DRESULT SD_read(BYTE lun, BYTE *buff, DWORD sector, UINT count)

{ DRESULT res = RES_ERROR; ReadStatus = 0; uint32_t timeout; if (ENABLE_SD_DMA_CACHE_MAINTENANCE == 1) uint32_t alignedAddr; endif

if(BSP_SD_ReadBlocks_DMA((uint32_t*)buff,

(uint32_t) (sector), count) == MSD_OK) { /* Wait that the reading process is completed or a timeout occurs */ timeout = HAL_GetTick(); while((ReadStatus == 0) && ((HAL_GetTick() - timeout) < SD_TIMEOUT)) { } /* incase of a timeout return error */ if (ReadStatus == 0) { res = RES_ERROR; } else { ReadStatus = 0; timeout = HAL_GetTick();

while((HAL_GetTick() - timeout) < SD_TIMEOUT)

{ if (BSP_SD_GetCardState() == SD_TRANSFER_OK) { res = RES_OK; if (ENABLE_SD_DMA_CACHE_MAINTENANCE == 1) /* the SCB_InvalidateDCache_by_Addr() requires a 32-Bit aligned address, adjust the address and the D-Cache size to invalidate accordingly. */ alignedAddr = (uint32_t)buff & ~3; SCB_InvalidateDCache_by_Addr((uint32_t*)alignedAddr, count*BLOCKSIZE + ((uint32_t)buff - alignedAddr)); endif break; } } } }

return res;

}

After this call my code enters

f(BSP_SD_ReadBlocks_DMA((uint32_t*)buff,(uint32_t) (sector),count) == MSD_OK)

and passes it successfully, after which it gets stuck in the while loop.

while((ReadStatus == 0) && ((HAL_GetTick() - timeout) < SD_TIMEOUT))

{ }

After this the problem occurs and that is:

if (ReadStatus == 0)

{ res = RES_ERROR; }

the function always gets stuck here no matter what. So i went and looked at all the calls for ReadStatus in my project and this popped up:

From here it seems there's one callback that will ensure the SD_Read passes and thats BSP_SD_ReadCpltCallback. Although this is never called anywhere in the project and therefore SD_Read always fails. Therefore it always remain 0 and the LL read function fails. This is cube generated.

Is this a mistake from my side or CubeMX??

I'm using the latest cube(V 4.0 and the latest Hal(1.8.0) for the stm32F7

f_mount-sdcard sdcard cubemx-project stm32f7-hal stm32f7

[Imen.D · Accepted solution]

Posted on November 29, 2017 at 16:21

Hello

Reishauer.Ian

,

Iwill raise this issue internally forcheck and investigation.

Best Regards

Imen

[con3]

Posted on November 29, 2017 at 16:34

Thank you, the write function seems fine as it internally sets the write flag to 1, it looks like this was just forgotten in the read function

[greg239955_stm1]

Posted on November 30, 2017 at 23:52

Same issue here.

Furthermore, the Cube generated sd_diskio.c doesn't call BSP_SD_Init() during SD_initialize().

[con3]

Posted on December 02, 2017 at 00:47

Same issue, after I implemented the fix to SD_initialize, as well as SD_read and SD_write I can successfully mount and open files. F_write still fails which seems very odd(cause I can create files), but trying to debug it. Once I figure out why its bugging out I should be able to completely use the FATFS system.

[franckngatcha]

Posted on December 02, 2017 at 13:00

same issue with 

V 4.22.1 and the latest Hal(1.8.0) for the stm32F7

[Amel NASRI]

Posted on December 06, 2017 at 17:16

Hi

Reishauer.Ian

‌,

It seems you identified root cause of your issue as stated in

https://community.st.com/0D50X00009XkXg5SAF

.

For others having similar issues, please do the following:

1) Re-defineSD_initialize as following (in case code is generated with CubeMX 4.23):

Depending on you application requirement:

FreeRTOS not used:

In the generated file 'your_project/src/sd_diskio.c', re-define 'SD_initialize' as following:

DSTATUS SD_initialize(BYTE lun)

{ if(BSP_SD_Init() == MSD_OK) { Stat = SD_CheckStatus(lun); } return Stat; }

FreeRTOS used:

In the generated file 'your_project/src/sd_diskio.c', re-define 'SD_initialize' as following:

DSTATUS SD_initialize(BYTE lun)

{ Stat = STA_NOINIT; /* * check that the kernel has been started before continuing * as the osMessage API will fail otherwise */ if(osKernelRunning()) { if(BSP_SD_Init() == MSD_OK) { Stat = SD_CheckStatus(lun); } /* * if the SD is correctly initialized, create the operation queue */ if (Stat != STA_NOINIT) { osMessageQDef(SD_Queue, QUEUE_SIZE, uint16_t); SDQueueID = osMessageCreate (osMessageQ(SD_Queue), NULL); } } return Stat; }

2) Apply the fix suggested by

Reishauer.Ian

‌ in the other discussions.

If you still face issues, please let us know about them.

-Amel

[con3]

Posted on December 06, 2017 at 17:52

Hi Amel, Both posts I Made are issues.

The one being the flag never being set (ReadStatus =1) and the state of the sdcard that is never returned to Ready.

I hope these help anyone that has a similar issue.

-Piet

[Amel NASRI]

Posted on December 08, 2017 at 11:42

Ok

Reishauer.Ian

‌,

As already said by

DAHMEN.IMEN

‌, the case is already raised internally.

Similar issue is already discussed in

https://community.st.com/0D50X00009XkWyLSAV

and a solution is suggeste there.

You can test it and let us know if it is working for you.

-Amel

[con3]

Posted on December 08, 2017 at 11:58

Hi Amel,

My issues are completely resolved as mentioned after implementing the changes I have mentioned above. I understand the issues have been raised internally, I was merely replying to you.

Kind Regards,

Piet