GT-64260A Design Guide
Doc. No. MV-S300165-00, Rev. A
CONFIDENTIAL
Copyright © 2002 Marvell
Page 30
Document Classification: Proprietary Information
May 21, 2002, Preliminary
Figure 8: Interrupt Pins’ Connectivity
describes the CPU to CPU cache coherency data flow:
Figure 9: CPU to CPU Cache Coherency Data Flow
GT-64260A
CPUInt*
PCI_INT0*
MPP1
MPP0
CPU0
CPU1
CPU_INT0
CPU_INT1
MPP3
MPP2
CPU0
GT-64260A
SDRAM
1. CPU1 initailizes the read cache
block transaction to cache
coherent memory.
CPU1
2. CPU0 signs to CPU1 and to the
GT-64260A that this cache block is
valid in its cache.
3. CPU0 writes the cache
block to SDRAM (cache
block flush).
4. CPU1 reads the
updated block from
SDRAM.