Running QSPI Samples on STM32F7 EVAL board

[LRodr]

I'm trying to run the QSPI samples from the STMCube_FW_F7_V1.15.0 package on my STM32F7779I-EVAL board. They run with no error but there's point where the data read back is compared to the data written and they never match. I tried both the interrupt driven and dma driven samples. Thoughts?

[Tesla DeLorean]

Check there aren't multiple memories used/supported by the board, and that you are building for the correct one.

Check jumpers and solder-bridge/option settings for proper QSPI connectivity.

[LRodr]

There are certainly multiple memories supported by the board but I haven't edited the sample so I expect that it's using what it needs. The firmware package is the one provided by STM for this board. The sample is loaded from the directory "Projects\STM32F769I_EVAL\Examples\QSPI" under the firmware package folder. I realize that this directory is for an STM32F769I while I have an STM32F779I, but I gather that they are compatible. They even share the same reference document.

[Tesla DeLorean]

>>I expect that it's using what it needs.

Yet it doesn't work. I think you misunderstand my point. You need to check the IC on the board, and confirm it is the same part# the code expects. ST sometime uses multiple different parts depending on availability, obsolescence or BOM options.

BSP code also frequently has bugs/typos, you might want to walk it. Expect to have to sink several man-hours into it.

Caching and MPU settings can impact things, as can where in memory the linker places things. You're building with what tool-chain?

[Andreas Bolsch]

"never match" is as useful as saying simply "it doesn't work". What's the problem *precisely*? Do your read only 0xFF, or 0x00, correct data but "shifted" by some bits or bytes or what else? The last two issues could be e.g . due to different number of dummy cycles (in QSPI interface and flash chip). Without a precise description nobody will be able to help.