AM437x RTOS QSPI flash 4-byte addressing

David Carter

Configuration:

H/W: AM437x IDK (TMDXIDK437X) using Macronix MX66L51235FMI-10G 512Mbit QSPI flash

S/W: CCS 6.1.2.00015, am437x PDK 1.0.4, SYS/BIOS 6.45.1.29

Hello,

I am using the AM437x IDK to develop a SYS/BIOS application requiring access to the full 64MB address space offered by the on-board QSPI flash. I am experiencing problems with using the supplied flash driver to access memory addresses higher than 16MB (i.e. requiring 4-byte addressing).

For example, taking the supplied example project 'QSPI_BasicExample_idkAM437x_armExampleProject' and modifying the project to perform its test read/write operations at 16MB offset (instead of 0MB offset) results in invalid data being returned.

Changes made to the project to produce this error state are as follows:

(main_qspi_flash_read_write.c)

//unsigned int addrValue = 0x000000U;
unsigned int addrValue = 0x1000000U;

//unsigned int blockNumber = 0U; /* Block number */
unsigned int blockNumber = 256U; /* Block number */

(.cfg)

/* Define the base address of the 1 Meg page the peripheral resides in. */
var peripheralBaseAddr = 0x31000000;

/* Configure the corresponding MMU page descriptor accordingly */
Mmu.setFirstLevelDescMeta(peripheralBaseAddr,
peripheralBaseAddr,
peripheralAttrs);

Are there any issues which would prevent me using the SYS/BIOS driver in this manner? I am able to successfully access the entire address space of the device using the equivalent Starterware driver.

Thank you in advance for your input.

David

over 8 years ago

0 Biser Gatchev-XID over 8 years ago

TI__Guru**** 393215 points

The RTOS team have been notified. They will respond here.

0 lding over 8 years ago in reply to Biser Gatchev-XID

TI__Guru* 95265 points

Hi,

Thnaks for finding this out! We tried to access the 16MB address offset and didn't work. A ticket was opened and we will update you.

Regards, Eric

0 Rahul Prabhu over 8 years ago

TI__Guru** 116170 points

David,

Can you please check the spec of the flash device. You may need to write a command to put the flash into 4 byte address mode by writing to the configuration register of the flash as the flash device typically uses 3 byte (24 bit addressing).

Regards,
Rahul

0 David Carter over 8 years ago in reply to Rahul Prabhu

Prodigy 160 points

Rahul,

Many thanks for your response. The flash device does require the driver to put it into (and take it out of) 4-byte address mode - this looks like it is being handled by the driver associated with the 'QSPI_BasicExample_idkAM437x_armExampleProject' example. In 'S25FL.c' the function S25FLFlash_Enable4ByteAddrMode(S25FL_Handle flashHandle, bool enable4ByteMode) performs this task, and the commands sent to the flash are correct for the device:

QSPI_LIB_CMD_ENTER_4_BYTE_ADDR = 0xB7U,
/**< Command to enter 4 byte addressing mode */
QSPI_LIB_CMD_EXIT_4_BYTE_ADDR = 0xE9U,
/**< Command to exit 4 byte addressing mode */

Regards,

David

0 Rahul Prabhu over 8 years ago in reply to David Carter

TI__Guru** 116170 points

David,

Even if the QSPI flash is being set to 4-byte address mode correctly, the driver is still using read and write command set that correspond to 3 byte addressing in the S25FL.c file. Note from the datasheet the read and write commands for 4 byte addressing are different so when the Flash device switches to 4 byte address mode, the code should use 4 byte addressing commands to access higher memory region on the flash device. the 3 byte addressing commands will also work but only for memory regions less than 16MB as the 3 byte addressing command will ignore the 4th address byte.

Check command on page 18 and page 19 here:
www.macronix.com/.../MX66L51235F, 3V, 512Mb, v1.1.pdf

and check this against the commands listed in qspiLibCmd in the file S25FL.h. you will notice that only 3 byte addressing commands are listed.

Regards,
Rahul

0 Rahul Prabhu over 8 years ago in reply to Rahul Prabhu

TI__Guru** 116170 points

David,

We have discussed this analysis with the development team and they agree with the above assessment that the flash commands in the flash driver needs to be updated to incorporate the 4 byte addressing commands. This issue is being tracked and will be resolved in 1Q17 release of the software. Please let us know if you want a fix for this issue before that release and we can provide you a patch.

Regards,
Rahul

0 David Carter over 8 years ago in reply to Rahul Prabhu

Prodigy 160 points

Rahul,

Many thanks for your response and for investigating this issue. Yes, if you are able to provide a patch ahead of the forthcoming release that would be much appreciated.

Kind regards,
David

0 David Carter over 8 years ago in reply to Rahul Prabhu

Prodigy 160 points

Rahul,
Are you able to advise when you will be able to issue me a patch for this driver?
Kind regards,
David

0 Rahul Prabhu over 8 years ago in reply to David Carter

TI__Guru** 116170 points

David,

I am checking with the developer regarding this issue and will get back to you soon. Before the holidays, we had this fix tentatively scheduled to be available by mid Feb but if this timeline doesn`t work for you then I can see if I can get some priority on this fix.

Regards,

Rahul

0 David Carter over 8 years ago in reply to Rahul Prabhu

Prodigy 160 points

Rahul,

Thanks for your response. If there is any way you could get this fix for me by the end of Jan it would greatly assist with our integration testing.

Best regards,

David

0 Rogerio Almeida over 8 years ago in reply to David Carter

TI__Mastermind 26215 points

Rahul,

You have mentioned mid February and customer asked for late January...

Customer is in their final stages, very close to go to production...any news on the fix?

0 Rahul Prabhu over 8 years ago in reply to Rogerio Almeida

TI__Guru** 116170 points

Rogerio,

The development team is unable to pull this update in by end of January. They are currently targeting early Feb as the earliest when they can provide the fix.

Regards,

Rahul

0 Rogerio Almeida over 8 years ago in reply to Rahul Prabhu

TI__Mastermind 26215 points

Thanks Rahul.

0 David Carter over 8 years ago in reply to Rogerio Almeida

Prodigy 160 points

Hi Rahul,

The availability of this QSPI driver fix is now approaching critical-path on our product development - as mid-Feb has now passed are you able to provide a further update and anticipated availability?

Many thanks,
David

0 Rahul Prabhu over 8 years ago in reply to David Carter

TI__Guru** 116170 points

David,

The issue was escalated to the development team so it is being worked on at the highest priority so that they can resolve this ASAP. The latest I heard from the development team is that they hope to have the fix hosted on the external GIT as soon as the fix and the internal testing is completed. I will try to get an update for you on this by the end of this week.

Regards,
Rahul