Other Parts Discussed in Thread: SYSCONFIG
Hello,
we tried updateing the SDK to 08.05 and postponed the new flash-driver in 08.04 already. So we wanted to finally update it but experiences problems as described in https://e2e.ti.com/support/microcontrollers/arm-based-microcontrollers-group/arm-based-microcontrollers/f/arm-based-microcontrollers-forum/1189521/mcu-plus-sdk-am243x-loadcpus-sometimes-asserts-in-dma-access/4521992#4521992
with the IS25WX256T it does not seem to work, at least the read data is not correct.
I will copy the observation from the other thread:
so I compared the init of both sdks. First I let the old one run, which works. I copied all the issued flash-cmds for config inside Flash_norOspiOpen and copied the register values to compare them. The only difference I see is the order of the address-byte setting. But the registers seem to be the same in the end. I let both run until
since this is the point where we get different results when reading the flash.
Every indendation means this is the value/call inside the function above.
for the old working drivers:
for the new updated - not working - driver:
I also noticed that, when trying to do a CPU reset via ccs and load the image again after the call of the mentioned OSPI_phyReadAttackVector, the first read in SDR config won't work anymore and thus we can't read out the flash-id. We do - independently of the driver - have the same sequence you use in your driver-example:
This always worked with the old driver. With the new driver we need a power reset of the device to get this working. Else the read stuff is not correct. I guess the driver configures something different here but I don't know what.
Also some screenshots of how the read out memory looks after the call of OSPI_phyReadAttackVector at 0x60000000 via memory browser:
old driver:
it seems only some values match and then somehow all memory after that has the same value. Maybe that helps to identify the problem better.
Additionally here is our new configuration that I migrated:
is there maybe another point inside the sdk that was changed that could have an impact here? But all other drivers are working fine. mcspi, pruicss, icssemac, clock, hwtimer, adc, gpios and so on are working as expected.
Also pls mention: I removed the subsequent Flash_norOspiReadId of the new driver, which was used to calculate the readDataCapDelay, since this never worked (it counted down to 0). I just set it hard to 3 which was the value of the old driver. But also this did not work.
The setting of the old driver were:
We need the flash-driver to be working with the new sdk.
Best regards
Felix