
ST40RA
A Interconnect architecture
A.1.5
LMI2 arbiter: (CPU, GPDMA, PCI, EMPI)
The default configuration (after reset) as to be to work fixed priority mode in the following priority
order:
● PCI,
● EMPI,
● GPDMA,
● CPU buffer (although the CPU requests are not supposed to go in that node to be send in the
LMI, it has to be managed in order to avoid deadlock).
The priority order have to be programmable and the latency checking algorithm can be enabled for
GPDMA, PCI, EMPI.
A.1.6
Return arbitration
The possibilities of the return arbitration are simpler than for the request arbitration. The arbiter is
not programmable but a specific arbitration can be chosen when implementing it.
The arbitration mode chosen is the fixed priority. For each arbiter (one per initiator), the order is the
following: LMI then other targets for the arbiters in node 1 and LMI, EMI, PCI, peripheral subsystem
for the arbiters of node 2.
A.2
Interconnect registers
A summary of registers is given in
Table 1. Addresses in the table are offset from the interconnect
base address at 0x1B05 0000.
Table 1: Interconnect register summary
Address
offset
Name
Function
0x010
LATENCY_LMI1_ENABLE
0x018
LMI1_CPU_PRI
0x020
LATENCY_LMI1_VALUE
Defines priority and latency value for the node 2 in the LMI1 arbiter, see
0x110
LATENCY_LMI2_ENABLE
0x118
LMI2_CPU_PRI
0x120
LMI2_LATENCY_PCI
Defines priority and latency value for PCI initiator in the PCI arbiter, see
0x128
LMI2_LATENCY_EMPI
Defines priority and latency value for EMPI initiator in the PCI arbiter, see
0x130
LMI2_LATENCY_GPDMA
Defines priority and latency value for GPDMA initiator in the PCI arbiter, see
0x210
LATENCY_EMI_ENABLE
0x218
EMI_CPU_PRI
0x220
EMI_LATENCY_PCI
Defines priority and latency value for PCI initiator in the EMI arbiter, see
0x228
EMI_LATENCY_EMPI
Defines priority and latency value for EMPI initiator in the EMI arbiter, see