TDA4VM-Q1: [OSPI FLASH] How to disable the function of mapping 0x50000000 to the 0 address of OSPI Flash?

Hi Tongxin,

I am not really sure if I am getting the question right.
You can use INDAC mode if you don't want the OSPI memory to be mapped at 0x50000000

Regards,
Parth

Hi ,we updater the config like

    .dacEnable = FALSE,

    .xipEnable = FALSE,

    .ospiClkSpeed = 133333333U,

    .dtrEnable = FALSE,

    .phyEnable = FALSE,

and in this config ,TDA4 will reset when func OSPI_waitReadSramLvl call CSL_ospiGetSramLvl.

7/5> Meeting planned for July 6th. 

John 

Per call July-7

1)

Tongxin Wang said:

Currently, we found that any kernel can read and write data from OSPI flash through 0x50000000, which we believe poses significant risks.

TI - Check alternative strategies for memory protection and share the details.

Tongxin Wang said:

dacEnable = FALSE,

    .xipEnable = FALSE,

    .ospiClkSpeed = 133333333U,

    .dtrEnable = FALSE,

    .phyEnable = FALSE,

and in this config ,TDA4 will reset when func OSPI_waitReadSramLvl call CSL_ospiGetSramLvl.

ZF to refer to OSPI driver example and check the difference between these configurations.

Regards,
Parth

HI Tongxin,

Tongxin Wang said:

Currently, we found that any kernel can read and write data from OSPI flash through 0x50000000, which we believe poses significant risks.

How are you confirming that OSPI memory is being corrupted?

Regards,
Parth

Per call July 7

1) Booting works fine

2) Failure is seen in application runtime - (Reading/Writing data from OSPI - MCU1_0)

• Is data is written properly? - How?

            What else is running in the system?

• MAIN R5 -> Camera application
• A72 -> Qnx

Tested the R5 application in standalone environment?

ZF to confirm if the data is written properly in OSPI.
ZF to try running MCU R5 application in a standalone environment.

Parth Nagpal said:

ZF to refer to OSPI driver example and check the difference between these configurations.

Ongoing action.

Regards,
Parth

Hi，Parth：

We have configured MPU to protect the address space of 128MB starting from 0x50000000.

Currently we are performing the stress test and reproducing the issue, I think we can suspend this thread for now.