TM4C1294NCZAD: Serial Flash for a file system using a Tiva CPU

Interesting. Mouser shows over 2,000 available and the datasheet does not indicate the part is no longer recommended for new designs.

My goal is to prototype a file system with a flash part and not use a microSD card. 

Thank you for providing the latest flash parts that can be used with the Tiva.

I also have a DK-TM4C129x eval board with both microSD and a SPI flash part (MX66L5).  There are 2 examples for using the file system but from the readme files talks about using microSD only and not the MX66L5. 

The examples are fatsd_DK_TM4C129X_TI_TivaTM4C129XNCZAD and fatsdraw_DK_TM4C129X_TI_TivaTM4C129XNCZAD.

I verified both examples using the microSD card.  Should these two examples work if the microSD card is removed and the JP7 jumpers are changed so the MX66L5 can be used?

In looking at the examples, I see that the function Board_initSDSPI() initializes the CPU pin PH4 that the microSD uses as a chip select.  If I were to remove the microSD card and use the MX66L5 should that work if I change the chip select to use PQ1 since PQ1 is the chip select to the MX66L5? 

If the MX66L5 can be used for testing these examples where in the example app is the chip select PH4 being set and cleared?

Thanks,

Doug

Would you expect I could download and build the examples the wiki references in CCS 8.2.0?

I downloaded the example projects and unzipped them and then imported them into CCS8.2.0 and I get the following errors when I try to build the projects linked in the wiki..

I assume the fix is to download and install an older version of TI-RTOS to make this work.  I have 2.16.0.08 and 2.16.1.14 installed:

What should I install to test these examples?

Thanks,

Doug

From the problems tab in CCS 8.2.0.

Description Resource Path Location Type
can't locate the package 'ti.sysbios.fatfs' along the path: 'C:/TI/tirtos_tivac_2_16_00_08/packages;C:/TI/tirtos_tivac_2_16_00_08/products/tidrivers_tivac_2_16_00_08/packages;C:/TI/tirtos_tivac_2_16_00_08/products/bios_6_45_01_29/packages;C:/TI/tirtos_tivac_2_16_00_08/products/ndk_2_25_00_09/packages;C:/TI/tirtos_tivac_2_16_00_08/products/uia_2_00_05_50/packages;C:/TI/tirtos_tivac_2_16_00_08/products/ns_1_11_00_10/packages;C:/ccs8p2/ccsv8/ccs_base;C:/TI/xdctools_3_32_00_06_core/packages;..;'. Ensure that the package path is set correctly. .xdchelp /serialFlash_TM4C129_LP line 53 C/C++ Problem
gmake: *** [build-201117329] Error 2 serialFlash_TM4C129_LP    C/C++ Problem
gmake: Target 'all' not remade because of errors. serialFlash_TM4C129_LP    C/C++ Problem
gmake[1]: *** [build-201117329-inproc] Error 1 serialFlash_TM4C129_LP    C/C++ Problem
Invalid project path: Include path not found (C:\TI\tirtos_tivac_2_16_00_08\products\TivaWare_C_Series-2.1.0.12573c). serialFlash_TM4C129_LP  pathentry Path Entry Problem
Product 'TI-RTOS for TivaC' v2.12.1.33 is not currently installed. A compatible version 2.16.0.08 will be used. serialFlash_TM4C129_LP  serialFlash_TM4C129_LP Problem
This project was created using a version of compiler that is not currently installed - 5.2.2 [ARM]. Another version of the compiler will be used during build - 18.1.3.LTS. See 'Help > Install New Software' and select 'Code Generation Tools Updates' to check if this compiler is available through a CCS update. Visit <a href="liveaction:OpenAppCenter">CCS App Center</a> to get the latest compiler support. Or <a href="software-dl.ti.com/.../a> and install the compiler, then register it with CCS through 'Preferences > CCS > Build > Compilers'. serialFlash_TM4C129_LP  properties Problem

From The Console window:

Building file: "../serialflash.cfg"

Invoking: XDCtools

"C:/TI/xdctools_3_32_00_06_core/xs" --xdcpath="C:/TI/tirtos_tivac_2_16_00_08/packages;C:/TI/tirtos_tivac_2_16_00_08/products/tidrivers_tivac_2_16_00_08/packages;C:/TI/tirtos_tivac_2_16_00_08/products/bios_6_45_01_29/packages;C:/TI/tirtos_tivac_2_16_00_08/products/ndk_2_25_00_09/packages;C:/TI/tirtos_tivac_2_16_00_08/products/uia_2_00_05_50/packages;C:/TI/tirtos_tivac_2_16_00_08/products/ns_1_11_00_10/packages;C:/ccs8p2/ccsv8/ccs_base;" xdc.tools.configuro -o configPkg -t ti.targets.arm.elf.M4F -p ti.platforms.tiva:TM4C1294NCPDT -r release -c "C:/ccs8p2/ccsv8/tools/compiler/ti-cgt-arm_18.1.3.LTS" --compileOptions "-mv7M4 --code_state=16 --float_support=FPv4SPD16 --abi=eabi -me --include_path=\"C:/ccs8p2/ccsv8/tools/compiler/ti-cgt-arm_18.1.3.LTS/include\" --include_path=\"C:/TI/tirtos_tivac_2_16_00_08/products/TivaWare_C_Series-2.1.0.12573c\" -g --gcc --define=ccs=\"ccs\" --define=PART_TM4C1294NCPDT --define=ccs --define=TIVAWARE --diag_warning=225 --diag_wrap=off --display_error_number --gen_func_subsections=on " "../serialflash.cfg"

configuring serialflash.xem4f from package/cfg/serialflash_pem4f.cfg ...

subdir_rules.mk:19: recipe for target 'build-201117329-inproc' failed

js: "C:/Tiva File system testing/Tiva FS example/serialFlash_TM4C129_LP/serialflash.cfg", line 53: xdc.services.global.XDCException: xdc.PACKAGE_NOT_FOUND: can't locate the package 'ti.sysbios.fatfs' along the path: 'C:/TI/tirtos_tivac_2_16_00_08/packages;C:/TI/tirtos_tivac_2_16_00_08/products/tidrivers_tivac_2_16_00_08/packages;C:/TI/tirtos_tivac_2_16_00_08/products/bios_6_45_01_29/packages;C:/TI/tirtos_tivac_2_16_00_08/products/ndk_2_25_00_09/packages;C:/TI/tirtos_tivac_2_16_00_08/products/uia_2_00_05_50/packages;C:/TI/tirtos_tivac_2_16_00_08/products/ns_1_11_00_10/packages;C:/ccs8p2/ccsv8/ccs_base;C:/TI/xdctools_3_32_00_06_core/packages;..;'. Ensure that the package path is set correctly.

"./package/cfg/serialflash_pem4f.cfg", line 184

xdctools_3_32_00_06_core\gmake.exe: *** [package/cfg/serialflash_pem4f.xdl] Error 1

js: "C:/TI/xdctools_3_32_00_06_core/packages/xdc/tools/Cmdr.xs", line 51: Error: xdc.tools.configuro: configuration failed due to earlier errors (status = 2); 'linker.cmd' deleted.

gmake[1]: *** [build-201117329-inproc] Error 1

subdir_rules.mk:16: recipe for target 'build-201117329' failed

gmake: *** [build-201117329] Error 2

gmake: Target 'all' not remade because of errors.

**** Build Finished ****

This project was created using a version of XDCtools that is not currently installed: 3.31.0.24_core. Please install the XDCtools of this version, or migrate the project to one of the supported versions. serialFlash_TM4C129_LP  serialFlash_TM4C129_LP Problem

This is a project I have been working on.export2.zip

I have not been able to get this example to work correctly.   I can confirm that the project serialFlash_TM4C129_LP builds and executes correctly and I think that project uses the same hardware as this example .

The project posted in the link above serialFlash_FatFS_TM4C129_LP builds with a warning on line 114

fresult = f_mkfs(at45dbHandle->driveNumber, 0, AT45DBFatFs_FLASH_SECTOR_SIZE );

With warning "Description Resource Path Location Type  #169-D argument of type "UInt" is incompatible with parameter of type "const TCHAR *" serialflash.c /serialFlash_FatFS_TM4C129_LP line 114 C/C++ Problem"

When I step through the code the it gets to the function f_mkfs() below and it aborts because  f_mkfs() returns FR_INVALID_DRIVE.

AT45DBFatFs_init();

AT45DBFatFs_Params_init(&params);

at45dbHandle = AT45DBFatFs_open(spiHandle, Board_CS, &params, DRIVE_NUM);

if (spiHandle == NULL) {

System_abort("AT45DBFatFs_open failed\n");

}

fresult = f_open(&src, inputfile, FA_WRITE|FA_READ);

if (fresult != FR_OK) {

if (fresult == FR_NO_FILESYSTEM) {

System_printf("Formatting Flash disk\n");

fresult = f_mkfs(at45dbHandle->driveNumber, 0, AT45DBFatFs_FLASH_SECTOR_SIZE );

I looked at the documentation for TI RTOS and it points to this online documentation but in the link the f_mkfs() has 5 arguments in the function call so that's different that the example code.

Can someone help me get this example working?

Thanks,

Doug:

Hi Doug,

The version of FatFS in TI-RTOS for TivaC 2.16 is R0.11a (the 2.12 version had R0.08a). The one on the elm site is newer (R0.13c). The .11a version only has three parameters to f_mkfs...but they are different from 08a!! The first parameter is a string instead of a value...doh! Sorry for missing this the first time. I must have ran the 2.12 FatFS one and it called f_mkfs properly so when I ran the 2.16 version the f_mkfs was never called. When I just tried calling the non-FatFs one (and killed the filesystem) and then tried the 2.16 FatFs one, I got the same result as you.

I changed the call to f_mkfs to (again I hard-coded it for a quick test).

fresult = f_mkfs("0", 0, AT45DBFatFs_FLASH_SECTOR_SIZE)

and then it worked. I checked out the other two compiler warnings and they we ok. I went ahead and changed them also 

before:

DRESULT             AT45DBFatFs_diskRead(UChar drv, UChar *buf, ULong sector,
                                         UChar count);
DRESULT             AT45DBFatFs_diskWrite(UChar drv, const UChar *buf,
                                          ULong sector, UChar count);

after:

DRESULT             AT45DBFatFs_diskRead(BYTE drv, BYTE *buf, DWORD sector,
                                         UINT count);
DRESULT             AT45DBFatFs_diskWrite(BYTE drv, const BYTE *buf,
                                          DWORD sector, UINT count);

and then I changed the actual function calls to match.

Can you give that a try and see if it works?

Todd

Hi Doug,

The first three are sample code and we don't typically document them much.

The last three are from FatFS. You can download the desired FatFs version (check the FatFs header files to see the version) from http://elm-chan.org/fsw/ff/archives.html. The zip has the documentation for that version.  In hindsight, we should have included this documentation in the TI-RTOS product.

Todd

Todd,

Where is the number of sectors specified or isn't it specified?   I see the call:

f_mkfs("0", 0, AT45DBFatFs_FLASH_SECTOR_SIZE);

specifies the sector size but where is the number of sectors specified?

The sector size is defined as:

#define AT45DBFatFs_FLASH_SECTOR_SIZE 512

I assume if I get a larger flash device I would need to specify the size of the new part some place.

Thanks,

Doug

Doug,

Number of sectors (pages) is in AT45DB.h: AT45DB161_NUMBER_OF_PAGES. This is returned to FatFs in the AT45DBFatFs_diskIOctrl() call when the ctrl message is GET_SECTOR_COUNT.

Yes, if you use a new part, these would have to be adjusted accordingly.

Todd

Thanks for the help on this!!

One more question, as per a recommendation above from Bob on  Jun 17, 2019 3:08 PM, I assume I can use any of the parts listed in the link below with only minor changes to the software setup (sector size/number of sectors. possible SPI speed).

www.digikey.com/.../774

If I wanted more memory that 16 Mbit, is there a part you would recommend?  The largest part I see is 16 Meg and that may be OK for our application but if I needed more that 16Mbit are there any parts that would be recommended I could use?

Thanks,

Doug

Hi Doug,

Sorry, I cannot be much help on this one since I'm not familiar with SPI Flash devices (other than the AT45DB one).

Todd

I took the working example "serialFlash_FatFS_TM4C129_LP" and replaced the flash part with the flash part in the data sheet in the link below.

https://www.winbond.com/resource-files/w25q128fv%20rev.l%2008242015.pdf

In the application "serialFlash_FatFS_TM4C129_LP" I made the following changes:

1) I changed AT45DBFatFs_FLASH_SECTOR_SIZE from 512 to 4192

2) I changed AT45DB161_NUMBER_OF_PAGES from 4096 to 65535.

3) I changed AT45DB161_PAGE_SIZE from 528 to 256.

With the three changes listed above the example serialFlash_FatFS_TM4C129_LP no longer works correctly and it fails on the f_open() call.

Is one of my parameters wrong or is there another parameter I need to change to use this memory chip?

Thanks,

Doug

I don't think that is a valid sector size based on the following comment in ff.h

WORD	ssize;			/* Bytes per sector (512, 1024, 2048 or 4096) */

That's well noted - appears poster has 'illegally merged/contracted.'    (4096 & 8192 - yielding the errant 4192)

Firm/I have 'no interest' in '4C129' - yet many such Flash memory devices provide their 'Vendor/Model Number' - and the existing code (if it interrogates the Flash memory) may not 'recognize' the upgraded device...

Todd, thanks for pointing out that the limit of AT45DB161_NUMBER_OF_PAGES is (512, 1024, 2048 or 4096).

From the document on the flash chip it states:

"The W25Q128FV array is organized into 65,536 programmable pages of 256-bytes each. Up to 256 bytes can be programmed at a time. Pages can be erased in groups of 16 (4KB sector erase), groups of 128 (32KB block erase), groups of 256 (64KB block erase) or the entire chip (chip erase). The W25Q128FV has 4,096 erasable sectors and 256 erasable blocks respectively. The small 4KB sectors allow for greater flexibility in applications that require data and parameter storage. (See Figure 2.)" 

Is the flash chip I am trying to make work not going to be easily configured to work with the FFS just by changing a few #defines or am I using the work value of AT45DB161_NUMBER_OF_PAGES for this part?

cb1_mobile, Good catch, that was a typo.

Thanks,

Doug

Hi Doug,

Since the device page size is 256, you'll need to set the sector size to 512. Then when a read/write/etc. comes down for 512, you'll need to act on two adjacent 256 byte pages. Does that make sense?

Todd

I don’t really understand this so I’ll go through where I got the numbers and maybe you can point out what I am missing.

If I understand this correctly the chip is 16Mbyte in size and the page size is 256 Bytes so:

Each Page is = 256 Bytes

65,535 pages x 256 bytes per page = 16 M Bytes Chip

Each Sector is 4096K Bytes

Each Block = 65,535 (4KB Sectors * 16 Sectors)

4096 sectors * 4 K Byte/sector = 16 M Bytes Chip

16 pages X 256 Bytes / Page = 4K Byte erase block

128 pages X 256 Bytes / Page = 32K Byte erase block

256 pages X 256 Bytes / Page = 64K Byte erase block

I used:

#define AT45DBFatFs_FLASH_SECTOR_SIZE   4096

#define AT45DB161_NUMBER_OF_PAGES       65535 <<<As you pointed out this is invalid

#define AT45DB161_PAGE_SIZE           256

Is the AT45DB161_NUMBER_OF_PAGES the total number of pages on the chip because if it is then 65,535 would be what the flash chip really has? If the AT45DB161_NUMBER_OF_PAGES is the number of pages per sector then it would be 16.  Looking at the code though it looks like this value AT45DB161_NUMBER_OF_PAGES is put in for the sector count.

One more thing we found in the data sheet is the flash chip has a write enable instruction 0x06 that it says:

“The Write Enable instruction (Figure 5) sets the WriteEnable Latch (WEL) bit in the Status Register to a 1. The WEL bit must be set prior to every Page Program, Quad Page Program, Sector Erase, Block Erase, Chip Erase,Write Status Register and Erase/Program Security Registers instruction. The Write Enable instruction is entered by driving /CSlow, shifting the instruction code “06h” into the Data Input (DI) pin on the rising edge of CLK, and then driving /CShigh.”

Would this write enable command require changes to the low level to make this work with this part?

Is this chips something that could be “easily used” with the Tiva and TI ROTOS by just modifying some #defines or should I find another part that would be a better fit?

Thanks,

Doug

Todd,

Doug and I met up today to talk more about this.  The Atmel device substituted for the original memory TI selected for this, which has since been labeled 
'not intended for new designs' is not a drop-in memory device from a software perspective.  Doug has confirmed that.  We suspect the memory partitioning and sectoring in general are different, therefore, not a drop-in.

Has your team done work with any other substitute memory devices that you can suggest to Doug, or frankly other customers in the same predicament?  We would welcome any input you can extend.  We realize this is memory and not something TI manufactures themselves, but we are hopeful from a TI-RTOS and overall project standpoint, that your team can provide some additional guidance.  Ultimately, we want to find a larger memory density device to the original, that provides as close to drop-in as possible, or at least with identifiable changes to the RTOS project and corresponding software.

Thank you,
Chris