MCU-PLUS-SDK-AM263X: Sysconfig raises an exception when I try to "load form JSON"

Hi Suven,

I have assigned your thread to the Sysconfig expert, but due to the holiday season, the response may be delayed. Sorry for any inconvenience!

Best regards,

Ming 

Suven,

   Have you checked these Flash devices do support SFDP??

The porting SOP request json file extract from qspi_flash_diag execution result. On current EVMs, test result shows FLASHs on both EVMs  don't support SFDP.

The json files from SDK which is used for EVMs may contain some unexpected items (missing or extra or both) that may make syscfg crash.

You need to check FLASH datasheet first to confirm them support SDFP.     

 S25FL128 on control card

S25HL512 on LP-AM263x

Regards

Andre

Dear Andre,

OK, I'll try SOP again after got our Micron NOR flash that support SFDP.

Thanks

Suven

Hello Suven,

Please let us know your results once alternate flash part is available for testing.

Best Regards,

Zackary Fleenor

Dear Fleenor,

  I still got same failure as below captured after I used "load from JSON" button in CCS.

My environment  as below:

My Json file that got from qspi_flash_diag as below attachment.

MT25QL512ABB.zip

 I'd like to change the customer flash part in sbl_uart_uniflash and sbl_qspi example.

 Could you help to give it a try?

Thanks

Suven 

Hello Suven,

I have looped in our Flash experts to review and provide feedback.

In the meantime, could you please download the latest MCU+SDK for AM263x (v08.05.00.24) and retest?

download mcu_plus_sdk_am263x_08_05_00_24-windows-x64-installer.exe  — 57438 K (ti.com)

Best Regards,

Zackary Fleenor

Dear Fleenor,

 It got same failure with 08_05_00_24.

Thanks

Suven

Aakash,

     Any suggestion, quick workaround to overcome this problem before new SDK release?

Regards

Andre

Dear Askash,

  I got same failure after followed Kowshik's comment to modify environment variable and  flash_v0.syscfg.js in sdk 08_04_00_17/08_05_00_24.

Thanks

Suven

Hi Suven Su,

This fix might take more time, can you fill these details manually by reading off the table.

That will accelerate the fix faster.

Best Regards,
Aakash

Dear Askash,

    I don't quite understand what you mean of " can you fill these details manually by reading off the table."? 

Could your more detail what you want me to do to help it?

Thanks

Suven

Hi Suven Su,

The table already provided a table for its configuration. You can update the values by fetching it from the table in the sysconfig directly (instead of import from json).

Even that would give you the same generated code.

Best Regards,
Aakash

Dear Askash,

 I got your point now, but I got below error when I tried to fill the content with my json file in sysconfig.

Thanks

Suven

Hi Suven Su,

The flash size supported in AM263x is only 32MB (256 Mbits). You are trying to configure it 64MB (512 Mbits) which is causing an issue here.

Best Regards,
Aakash

Dear Aakash,

  Our customer consigns MT25QL512ABB parts in our project.(It is 512Mb/64MB)

Can you provide us document that specify the flash size limitation with AM263x?

Thanks

Suven

Askash,

     It doesn't make sense to have such limitation. If FLASH provides 4 byte address mode, it can support up to 4GB addressing. I don't see any limitation on AM263x QSPI controller. Can you please check where is this limitation coming from? syscfg? if it is syscfg tool limitation, please advice how to cancel this limitation.

Regards

Andre

Askash,

     It doesn't make sense to have such limitation. If FLASH provides 4 byte address mode, it can support up to 4GB addressing. I don't see any limitation on AM263x QSPI controller. Can you please check where is this limitation coming from? syscfg? if it is syscfg tool limitation, please advice how to cancel this limitation.

Regards

Andre

Askash,

    I did some quick test by qspi_flash_demo, indeed, it will fail if we try to access flash address offset > 8MB. Then next questions, why do we put EVM 16MB QSPI FLASH on LP-AM263x and CC-AM263x? why do we limit 32MB on syscfg not 8MB?

regards

Andre

Hi AndreTseng,

AndreTseng said:

Then next questions, why do we put EVM 16MB QSPI FLASH on LP-AM263x and CC-AM263x?

This is an issue identified very recently. This was never intended design and hence is part of our errata.

AndreTseng said:

why do we limit 32MB on syscfg not 8MB?

This is definitely a software bug in sysconfig. If the flash is configured in memory map mode, only then the issue is suppose to be valid. The limit should only be imposed for Memory Map mode and not Config Mode. IMO, in config mode, the flash can be accessed for entire flash size.

Best Regards,
Aakash