Because of the Thanksgiving holiday in the U.S., TI E2E™ design support forum responses may be delayed the week of November 23.

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TMS570LC4357: TMS570LC4357 and internal memory select conflict in documentation

Part Number: TMS570LC4357

Hello,

could you clarify PS[31:30] memory select signal in TMS570LC4357?
It could be problem on documentation only.

See to SPNS195C Table 6-25. Module Registers / Memories Memory Map
I try to see SPNU563A Table 2-2. Module Registers / Memories Memory-Map and it contain same information.

It define PS[31:30] memory select signal as: 

  • CPGMAC Slave (Ethernet Slave) PS[30]-PS[31] 0xFCF7_8000~0xFCF7_87FF
  • PCR3 registers PS[31:30] 0xFFF7_8000~0xFFF7_87FF

This mean that this signal is active for two absolutely different memory ranges. Could you check it? Which signal is needed for CPGMAC Slave and which for PCR3 registers?

Jiri

  • Same problem is on PPSE[4]–PPSE[5] memory selects
    It is:
    * PCR2 registers PPSE[4]–PPSE[5] 0xFCFF_1000 ~ 0xFCFF_17FF
    * PCR1 registers PPSE[4]–PPSE[5] 0xFFFF_1000 ~ 0xFFFF_17FF
  • And again PPSE[6]:
    * NMPU (EMAC) PPSE[6] 0xFCFF_1800 ~ 0xFCFF_18FF
    * NMPU (PS_SCR_S) PPSE[6] 0xFFFF_1800 ~ 0xFFFF_19FF
    * NMPU (DMA Port A) PPSE[6] 0xFFFF_1A00 ~ 0xFFFF_1BFF
  • One more bug in documentation:
    see SPNS195C Table 6-25. Module Registers / Memories Memory Map
    in this row
    Pin Mux Control (IOMM) PPSE[7] 0xFFFF_1C00 0xFFFF_1FFF 2KB 1KB
    Frame size is 1KB not 2KB
  • Jiri,

    The base addresses are different for the individual slaves to be accessed. The address decoder evaluates all 32 bits of the CPU address to generate the correct select. Only one slave is addressed for a given 32-bit address.

    I agree with your comment on the PPSE[7] frame being 1KB instead of 2KB mentioned in the datasheet.

    Regards,
    Sunil
  • I undestant, that it isn't problem for device addressing.
    But it is problem for PCR unit setting. We need to disable access to all unused devices.
    It this PS_, PPS_, PPSE_, PCS_, PPCS_ numbers valid from PCR setting point of view? (PSxxMSTID_L/H and PPSxxMSTID_L/H registers)
    It looks strange that one bit enable/disable two absolutely independent devices together. Strange because most of option in PCR are unused, there isn't pressure to share same bits. Strange because its named "memory select" in datasheet.

    Jiri

  • It is not problem for addressing.
    We need it to fulfill requirements from Safety Manual SPNU540A chapter
    7.92 PCR Access Management: Protection Mode and MasterID Filtering
    And from this point of view conflicting memory select is problem. Theoretically we don't have problem that one signal enable two peripherals, but there is no documentation about this unexpected behavior and it is problem.
  • Now I found that for ex. PS[31:30] have only same ID but there are in different PCR module. Therefore there is no conflict.

    PCR3 registers is PS[31:30] under PCR3 (Peripheral Segment 3)
    CPGMAC Slave (Ethernet Slave) if PS[30]-PS[31] under another PCR2 (Peripheral Segment 2)

    Correct?
  • Hi Jiri,

    Yes, that is correct. The memory map table in the datasheet is organized per the base address of the memory selects.

    Regards,
    Sunil