STM32G474: Descriptions on access paths to SRAM2 is inconsistent within a data sheet.

[Yoichi Shinoda]

DS12589 Rev 1 (for G431) and DS12288 Rev 1 (for G474) both states in Section 3.5 that SRAM2 can be accessed by CM4 core through System Bus or through the I-Code/D-Code bus. However, Figure 2 "Multi-AHB bus matrix" suggests that SRAM is only accessible by CM4 core through System Bus.

[waclawek.jan]

That's most probably a copy/paste error, there's no indication that SRAM2 could be remapped to the I/D address space.

JW

@Imen DAHMEN​ 

[Imen.D]

Hello,

We will correct the figure in the coming release of datasheets :

The SRAM2 can be accessed by the CPU I-bus and D-bus after remap (like SRAM1). SRAM2 is supplied independently and can be retained in Standby mode.

Thanks for bringing this typo to our attention.

Best Regards,

Imen

[Imen.D]

Hello Jan,

Sorry, after deep check, the original picture is correct and we will modify the description in the documents.

The SRAM2 is not accessible through I/D bus. The remap is possible only for SRAM1 / FMC / QUADSPI and the Flash.

Kind Regards,

Imen

[waclawek.jan]

Hi Imen ,

Thanks for the clarification.

> after remap (like SRAM1).

How exactly? Does it remap together with SRAM1, as its continuation; or does it remap separately being mapped onto 0x00000000?

I grab on the oportunity and abuse this thread to ask a related question (I keep asking here for years) for the 'F4 family: RM0090 rev.17 ([edit] and also rev.18[/edit]) in chapter 2.3.1 Embedded SRAM says:

The CPU can access the SRAM1, SRAM2, and SRAM3 through the System Bus or through

the I-Code/D-Code buses when boot from SRAM is selected or when physical remap is

selected

However, exactly as here in the case of 'G4, there is no indication in the bus matrix diagrams, the Memory mapping vs. Boot mode/physical remap tables, nor in SYSCFG_MEMRMP.MEM_MODE description, that SRAM2 (and SRAM3 for the 'F42x/'F43x) could be remapped.

Can ST please clarify?

Thanks,

Jan

@Imen DAHMEN​ 

[waclawek.jan]

Hi @Imen DAHMEN​  ,

thanks for the clarification.

Can please wording of chapter 2.3.1 Embedded SRAM in RM0090 be fixed accordingly, too?

Thanks,

Jan

[Imen.D]

Hi @Community member​ ,

Yes, I also raised this typo internally for correction in the coming release of RM0090.

Many thanks for bringing this to our attention.

Best Regards,

Imen

[waclawek.jan]

Thanks, Imen.

Jan