Mass Storage Device reads only first sector of memory

On the stack, MSC_MemoryRead() is called repeatedly, even for single sector read.
The sector size is 512 bytes (MSC_BlockSize), but the max packet size of the bulk endpoint (EP) is 64 bytes. 512 bytes sector is split into 8 transactions (packets) over the bulk EP. ie. MSC_MemoryRead() is called 8 times for single sector read.

When MSC_MemoryRead() is called once, 64 bytes packet is passed to the endpoint by USB_WriteEP(). MSC_MemoryRead() moves "Offset" variables for the next 64 bytes packet.

In your code, NANDFLASHPageRead() reads out 512 bytes at a time.
Is there any problem when NANDFLASHPageRead() is repeatedly called?

Tsuneo

On the stack, MSC_MemoryRead() is called repeatedly, even for single sector read.
The sector size is 512 bytes (MSC_BlockSize), but the max packet size of the bulk endpoint (EP) is 64 bytes. 512 bytes sector is split into 8 transactions (packets) over the bulk EP. ie. MSC_MemoryRead() is called 8 times for single sector read.

When MSC_MemoryRead() is called once, 64 bytes packet is passed to the endpoint by USB_WriteEP(). MSC_MemoryRead() moves "Offset" variables for the next 64 bytes packet.

In your code, NANDFLASHPageRead() reads out 512 bytes at a time.
Is there any problem when NANDFLASHPageRead() is repeatedly called?

Tsuneo

Hello Tsuneo,

Reading the sector multiple times is no problem.
I tried various modifications today. For now I am able to read the first 128 sectors (MSC-blocks) without any problem.
What I did was to number the sectors in the flash. The first two bytes contain the block- and page-address.
The output of disk probe and WinHex shows correct the sector-numbers.
Therefor I had to modify the LBA, which seems pretty weird. In the example the LBA is extracted off the CBW and then multiplied by the sectorsize to get the final byte-address.
In my case the LBA had to be divided by 512 to get the byte-adress. I can't really explain the reason, I've just found out (by numbering the sectors), that the MSC read every 512th sector.

By now the problem has moved. I can read the first 128 sectors, but the next 64 sectors being read start at sector 0 again with an offset of 128 Byte. Same thing happens every 128 sectors.
Maybe I'm just blind for the problem, so here's the code:

void MSC_MemoryRead (void)
{
        u32 n;

        for(n=0;n<16384;n++) Buffer[n]=0x00;
        if (Length > MSC_MAX_PACKET)    n = MSC_MAX_PACKET;
        else    n = Length;

        if ((Offset + n) > MSC_MemorySize)
        {
                n = MSC_MemorySize - Offset;
                BulkStage = MSC_BS_DATA_IN_LAST_STALL;
        }

        NANDFLASHPageRead(Offset/512+FAT_offset_pages<<9 , &Buffer[0], 512);           // entsprechende Page einlesen

        USB_WriteEP(MSC_EP_IN, &Buffer[Offset%512], n);
        Offset += n;
        Length -= n;

        CSW.dDataResidue-=n;

    if(Length==0)
        BulkStage = MSC_BS_DATA_IN_LAST;

        if (BulkStage != MSC_BS_DATA_IN) {

    CSW.bStatus = CSW_CMD_PASSED;
  }
}

As there seems to be no way to debug the MSC-functions inside the IDE, i'm a bit stuck right here. Maybe I'm blind and just don't see the/my mistake.

Greetings
Frank

Hello Tsuneo

finally I made it working. I forgot brackets in the makro FAT_offset_pages. This totally screwed up the addressing.

For now I can read the device using Disk Probe and WinHex, what doesn't work is opening a testfile via windows explorer. I guess, that windows needs correct working write-functions for opening a file.

Thanks a lot!
Frank