The ATM Cell Processor (MC92501) is an Asynchronous Transfer Mode (ATM)
layer device composed of dedicated high-performance ingress and egress cell
processors combined with UTOPIA Level 2-compliant physical (PHY) and switch
interface ports (see Block Diagram). The MC92501 is a second generation ATM
cell processor in MotorolaÕs 92500 series. This document provides information
on the new features offered by the second generation ATM cell processor. This
document, combined with MC92500/D, provides the complete speciÞcation for
the ATM cell processor.
GC SUFFIX
GTBGA
CASE 1208
New Features of the MC92501:
¥ Implements ATM Layer Functions for Broadband ISDN According to ATM
Forum UNI 4.0 and TM 4.0 SpeciÞcations, ITU Recommendations, and
Bellcore Recommendations
¥ Provides ABR Relative Rate Marking and EFCI Marking According to TM 4.0
¥ Selective Discard CLP = 1 (or CLP = 0+1) Flow on Selected Connections
¥ UTOPIA Level 2 PHY Interface and UTOPIA ATM Layer Interface
¥ Supports Both Partial Packet Discard (PPD) and Early Packet Discard (EPD)
¥ Change ABR RM Cell Priority
¥ Support for CLP Transparency
Existing MC92500 Features:
¥ Full-Duplex Operation at Data Rates up to 155 Mbit/sec
¥ Performs Internal VPI and VCI Address Compression for up to 64K VCs
¥ CLP-Aware Peak, Average, and Burst-Length Policing with Programmable
Tag/Drop Action Per Policer
¥ Supports up to 16 Physical Links Using Dedicated Ingress/Egress MultiPHY
Control Signals
¥ Each Physical Link Can Be ConÞgured as Either a UNI or NNI Port
¥ Supports Multicast, Multiport Address Translation
¥ Maintains Both Virtual Connection and Physical Link Counters on Both
Ingress and Egress Cell Flows
¥ Provides a Flexible 32-Bit External Memory Port for Context Management
¥ Automated AIS, RDI, CC, and Loopback Functions with Performance
Monitoring Block Test on All 64K Connections
¥ Programmable 32-Bit Microprocessor Interface Supporting Big-Endian or
Little-Endian Bus Formats
¥ Bidirectional UPC or NPC Design with up to Four Leaky Buckets Per
Connection
¥ Supports a Programmable Number of Additional Switch Overhead
Parameters Allowing Adaptation to Any Switch Routing Header Format
¥ Provides Per-Link Cell Counters in Both Directions
ORDERING INFORMATION
MC92501GCGTBGA
This document contains information on a new product. SpeciÞcations and information herein are subject to change without notice.
A typical ATM network consists of user end stations that
transmit and receive 53-byte data cells on virtual connections
(see Figure 1). Physical links and switching systems
interconnect the virtual connections. A virtual connectionÕs
path is established at the beginning of the data transfer,
maintained while the end-stations are communicating, and
torn down after the transfer is complete. This transmission
method increases the transfer speed because the
determination of the path the data will take is done only at the
beginning of the data transfer instead of when each data subblock or packet is transferred.
On a given physical link, each connection is assigned a
unique connection identiÞer. The connection identiÞer is
placed in the header of each cell by the transmitting equipment
and is used by the receiving equipment to route the cell to the
next physical link on the connection path. All cells belonging
to a speciÞc virtual connection follow the identical path from
VCs
Switch
the transmitting end station through the switching systems to
the receiving end station.
An ATM switch contains a high-speed switching fabric that
connects multiple line cards. The switching fabric connects the
input port to the output port based on the switchÕs routing table.
The line card interfaces between the physical medium and the
switching fabric by recovering incoming cells from the arriving
bit stream or converting outgoing cells into a bit stream for
transmission. An ATM swtich partitioned in this fashion can
efÞciently handle multiple physical links by independently
transferring each incoming ATM cell from its source port to its
destination port, based on the switchÕs routing table.
ATM standards divide the tasks to be performed on each
side of the switch fabric into PHY layer and ATM layer tasks.
The PHY layer tasks are dependent on the physical medium
that connects ATM switches. The ATM layer tasks operate at
the cell level and are independent of the physical medium.
Switch
VCs
END STATIONS
SWITCH
CLK REC
Switch
LINE
CARD
PHY
SwitchSwitchSwitch
SWITCHING FABRIC
MC92501
ATM LAYER FUNCTIONS
Figure 1. MC92501 in an ATM Network Application
END STATIONS
LINE
CARD
LINE CARD
MOTOROLA
MC92501
5
1.2.ATM Network Applications
The MC92501 performs the ATM layer functions in an ATM
switch such as cell processing and routing. Since the
MC92501 is an ATM layer device, it is PHY layer independent.
Figure 2 illustrates a typical ATM line card. The MC92501
uses an external memory for storing the cells that it
processes. In addition, the MC92501 offers an option to utilize
an external address compression device accessed via the
same external memory bus.
The microprocessor is used for conÞguration, control, and
status monitoring of the MC92501 and is responsible for
initializing and maintaining the external memory. The
MC92501 is the master of the external memory bus. At regular
intervals, the MC92501 allows the microprocessor to access
the external memory for updating and maintenance.
System RAM can also be located on the line card. The
MC92501 can support a DMA device to allow efÞcient data
transfer to this RAM without processor intervention.
RAM
MICRO-
PROCESSOR
DMA
DMA
The physical interface (PHY-IF) implements the physical
layer functions of the B-ISDN Protocol Reference Model. This
includes the physical medium dependent functions required
to transport ATM cells between the ATM user and the ATM
switch (UNI) or between two ATM switches (NNI). The cells
are transferred between the physical interface and the
MC92501 using the UTOPIA Level 2 standard.
The MC92501 implements B-ISDN UNI/NNI ATM layer
functions required to transfer cells to and from the switch over
virtual connections. These functions include usage
enforcement, address translation, and Operation,
Administration, and Maintenance (OAM) processing. The
MC92501 provides context management for up to 65,536
(64K) Virtual Connections (VCs). The VCs can be either Virtual
Path Connections (VPCs) or Virtual Channel Connections
(VCCs). ATM cells belonging to a particular VCC on a logical
link have the same unique Virtual Path IdentiÞer/Virtual
Channel IdentiÞer (VPI/VCI) value in the cell header. Similarly,
cells belonging to a particular VPC on the same logical link
share a unique VPI.
EXTERNAL
MEMORY
EXTERNAL
ADDRESS
COMPRESSION
MICROPROCESSOR BUS
RECOVERY
LINE CARD
CLOCK
PHY-IF
PHY-IF
PHY IF
MC92501
Figure 2. Typical MC92501 Line Card Application
EXTERNAL
MEMORY
BUS
TO SWITCH
FROM SWITCH
MC92501
6
MOTOROLA
SECTION 2. FUNCTIONAL DESCRIPTION
2.1.System Functional Description
A serial transmission link operating at up to 155.52 Mbit/sec
(PHY) is coupled to the MC92501 via a byte-based interface.
The transmission link timing is adapted to the MC92501 and
switch timing by means of internal cell buffers. A common clock
supplies both the PHY IF and MC92501.
The host microprocessor initializes and provides real-time
control information to the data-ßow chips (PHY IF and
MC92501) using slave accesses.
The MC92501 operates in conjunction with an external
connection memory, which provides one context entry for each
active connection. The entry consists of two types of context
parameters: static and dynamic. The static parameters are
loaded into the context memory when the VC is established,
and are valid for the duration of that connection. The static
parameters include trafÞc descriptors, OAM ßags, and ATM
switch parameters. The dynamic context parameters include
cell counters, UPC/NPC Þelds, and OAM parameters. The
dynamic parameters can be modiÞed while a particular
connection is being processed. The microprocessor can
access the external memory through the MC92501 to collect
trafÞc statistics and to update the OAM parameters. During
normal cell processing, the MC92501 has exclusive access
to the external memory and maintains external memory
coherency.
At user-programmable intervals, the MC92501 provides the
microprocessor with a Òmaintenance slot.Ó During this time,
cell processing is halted and control of the external memory
bus is relinquished. The break in cell processing is made
possible by the difference between the MC92501 cellprocessing rate and the line rate.
The microprocessor can use the maintenance slot for any
of the following tasks:
¥ Connection setup and tear down
¥ Statistics collection
¥ Updating OAM parameters of active connection
The microprocessor is responsible for the external memory
coherency during the maintenance interval.
2.2.MC92501 Functional Description
MC92501 General Features:
¥ Implements ATM layer functions for broadband ISDN
according to CCITT recommendations, ATM Forum
UNI 4.0 and TM 4.0 speciÞcations, and ITU and Bellcore
recommendations.
¥ Provides 155 Mbit/sec throughput capacity and is physical
layer independent.
¥ Optionally supports up to 16 physical links.
¥ Optionally conÞgured as a User Network Interface (UNI) or
Network Node Interface (NNI) on a per-link basis.
¥ Provides Available Bit RateÐRelative Rate (ABRÐRR)
marking and EFCI marking according to TM 4.0.
¥ Supports advanced discard policies such as Selective
Discard, Partial Packet Discard (PPD), Early Packet
Discard (EPD), and Limited Early Packet Discard (Limited
EPD).
¥ Operates in conjunction with an external memory (up to
16 MB) to provide context management for up to 64K virtual
connections.
¥ Provides cell counter coherency on a per-connection basis
by maintaining redundant copies of the counter tables and
dynamically switching between them.
¥ Provides per-link cell counters in both directions.
¥ Provides per-connection Usage Parameter Control (UPC)
or Network Parameter Control (NPC) using a leaky bucket
design with up to four buckets per connection.
¥ Provides support for Operation, Administration, and
Maintenance (OAM) Continuity Check function for all
and OAM performance monitoring on all connections.
¥ Interfaces with either big-endian or little-endian
microprocessors.
¥ Supports cell insertion into the cell streams using direct
access registers which may be written by the
microprocessor or by a DMA device.
¥ Supports copying cells from the cell streams using direct
access registers which may be read by the microprocessor
or by a DMA device.
¥ Supports multicast operation.
2.3.First Generation Features
The MC92501 is a second generation ATM cell processor
that enhances the MC92500 (Þrst generation) functionality.
The MC92501 is backwards-compatible and pincompatible with the MC92500.
This document describes the second generation
enhancements and is meant to supplement the MC92500
speciÞcation. The MC92500 speciÞcation can be ordered from
the Motorola Literature Center by requesting document
MC92500/D.
MOTOROLA
MC92501
7
SECTION 3. PACKET-BASED UPC DISCARD ALGORITHMS
3.1.Introduction
The MC92501 UPC function performs cell-based discard or
packet-based discard according to ATM Forum TM 4.0. It
supports packet discard on VC connections AAL5 packets (not
including OAM cells). The MC92501 also performs Partial
Packet Discard or Early Packet Discard.
The MC92501 offers four modes of UPC operation on a perconnection basis: Cell-Based UPC, Partial Packet Discard
(PPD), Early Packet Discard (EPD), and Limited Early Packet
Discard (Limited EPD). These modes are selected on a perconnection basis using the
Mode
bit in the Common Extension Parameters Table. Packet-
based UPC is enabled globally by the
Enable
bit in the ACR register.
IUOMÑIngress UPC Operation
IPCVÑIngress Features
3.1.1.AAL5 Packet Definition
A packet is deÞned as a stream of user cells belonging to
the same virtual connection that has a series of one or more
cells with the PTI[0] bit set to 0 and the last PTI[0] bit set to 1.
(See Figure 3.)
3.2.Cell-Based UPC
This is the default mode. The MC92501 discards cells on a
per-cell basis as deÞned in MC92500/D.
3.3.Partial Packet Discard (PPD)
According to the PPD algorithm, if a cell is discarded then
all subsequent cells belonging to that packet are discarded up
to but not including the last cell. Following is a detailed
explanation of the UPC function.
¥ The UPC is a two-state machine: discarding and not-
discarding. See Figure 4.
¥ While the UPC is in the not-discarding state, it performs
normal cell-based operation with tagging and policing
counter updates.
¥ The UPC transitions from the not-discarding to the
discarding state on the Þrst discarded cell.
¥ While the UPC is in the discarding state, it does not update
the UPC bucket but it does increment the policing discard
counter.
¥ When in the discarding state and the last cell of a packet
is received, there are two options:
Ñ If
all the cells belonging to that packet were discarded,
then this last cell is discarded.
Ñ If
not all the cells belonging to that packet were
discarded, then this means that the packet was
truncated and this last cell is admitted in order to
delineate the corrupted packet from the next packet.
There is however one exception: if this last cell is
violating cell-based UPC then it is discarded.
Figure 5 illustrates an example for the PPD algorithm. A
UPC policy violation occurs during the transmission of the Þrst
packet. The UPC detects the violation and discards the
remainder of the packet except for the last cell. The last cell
of the Þrst packet is transmitted to avoid the concatenation of
the corrupted packet with the subsequent Packet #2. If the
UPC detects that the Þrst cell of Packet #3 violates its policy
then Packet #3 is truncated. Packet #3Õs last cell is not
transmitted because it cannot be admitted by the cell-based
UPC. Packet #4 is not transmitted either because its Þrst cell
violates the UPC policy.
Figure 3. Delineation of a Packet Within a Cell Stream
LAST CELL OF PACKET ARRIVES
If all cells within a packet are discard
OR
If the last cell is violating UPC policy
then discard the last cell
If not all calls within packet discarded, then admit last cell
Figure 4. UPC Discarding State Machine
PACKET 1
NOT-DISCARDING
DISCARDING
FIRST CELL OF
PACKET 2
CELL VIOLATES
UPC POLICY
MC92501
8
MOTOROLA
Input Stream
UPC Discard Decision
Packet #1Packet #2Packet #3 Packet #4
Output Stream
L
Figure 5. Partial Packet Discard
3.4.Early Packet Discard (EPD)
According to the EPD algorithm, the decision to discard a
packet takes place only at the beginning of a packet. This
means that the complete packet is either fully discarded or
fully passed. The following explains how EPD is implemented.
¥ When the EPD is discarding cells, the buckets are not
updated but the policing discard counter is incremented.
¥ When the EPD decides that a frame should be passed this
means that:
Ñ All tagging buckets continue to work in a cell-based
fashion.
Ñ All discarding buckets perform their calculations as if
the limit parameter is inÞnite, and therefore increment
the bucket content and do not discard any cells. As a
result, their bucket content can be greater than their
bucket limit.
Ñ The MC92501 may increment its police tagging counter.
Figure 6 illustrates an example for the EPD algorithm. A cell
within the Þrst packet violates the UPC, but due to EPD this
packet is fully passed. Since the Þrst cell of Packet #2 violates
the UPC, the second packet is fully discarded. Likewise, cells
within Packet #3 violate the UPC, but this packet is not
discarded. Since the fourth packet comes after a relatively
long time, which allows the UPC buckets to drain, Packet #4Õs
cells do not violate the UPC policy.
3.5.Limited Early Packet Discard (Limited
EPD)
One disadvantage of the EPD algorithm is that once it
decides to admit a packet it cannot change its decision until
the last cell of that packet. In the case of big packets, the switch
can run into congestion. Using the Limited EPD algorithm, a
connection can stop passing cells because of EPD once it
reaches a predeÞned limit. That limit, in the case of the
MC92501, is reached once the
cells. The Þrst bucket should have the same parameters as
one of the other buckets except for the limit, which is bigger.
Figure 7 describes a UPC which contains three buckets.
The Þrst bucket is for limiting EPD, and there are two other
buckets. The Þrst and second buckets share the same
parameters except for the limit. Therefore, their bucket content
is always the same, although the second bucketÕs content is
higher than its limit and cells are admitted by the EPD
algorithm. When the Þrst bucket reaches its limit, then cells
will be discarded.
Figure 8 describes the EPD and Limited EPD functions.
Þrst bucket starts discarding
Input Stream
UPC Discard Decision
Output Stream
MOTOROLA
Packet #1Packet #2
Figure 6. Early Packet Discard
Packet #3
Packet #4
MC92501
9
Limit
Limit
Limit
First BucketSecond BucketThird Bucket
Figure 7. Limited Early Packet Discard
.
Input Stream
First Bucket Limit
Other BucketsÕ Limit
EPDÐUPC Output Stream
Limited EPDÐUPC Output Stream
Figure 8. Difference Between Early Packet Discard and Limited Early Packet Discard
Packet #1Packet #2
Packet #3
Packet #4
MC92501
10
MOTOROLA
SECTION 4. SELECTIVE DISCARD
ATM Forum TM 4.0 deÞnes procedures according to which
cells can be discarded by network elements. A switching
element may discard cells belonging to selected connections
or cells whose CLP = 1 in case of congestion. This function is
called selective discard and it is implemented by the
MC92501. Selective discard is enabled by the
Ingress Congestion NotiÞcation
Control Register (IPLR). Selective discard can be enabled on
bit in the Ingress Processing
ICNGÑGlobal
a per-connection basis by the
Discard Operation Mode
Extension Word. This Þeld determines whether selective
discard is enabled and whether selective discard is performed
on CLP = 1 or on CLP = 0+1 trafÞc. Selective discard can be
enabled globally by the
CFB ConÞguration Register (ACR).
Þeld in the Common Parameters
IPCVÑIngress Enable bit in the ATMC
ISDMÑIngress Selective
MOTOROLA
MC92501
11
SECTION 5. AVAILABLE BIT RATE (ABR) SUPPORT
5.1.Overview and Features
The MC92501 provides a full Available Bit Rate (ABR)
solution for switch behavior relative rate marking and EFCI
marking in accordance with ATM Forum TM 4.0. It also
provides the switch fabric with an interface to increase the RM
cellsÕ trafÞc priority. Following is a list of features:
¥ Performs Relative Rate (RR) marking on Forward
Resource Management (FRM) and/or Backward Resource
Managment (BRM) cells, on selected connections. This
feature is enabled by either setting the ATMC CFB
ConÞguration RegisterÕs (ACR) VP RM Cell PTI (NPRP) bit
or by setting the PTI Þeld in the cellÕs header to Ò110BÓ.
¥ Performs EFCI marking on non-RM cells whose PTI[2] =
0, on selected connections. This feature is enabled by
either control registers or by Þelds that it gets from the
overhead of cells which are received from the switch fabric.
¥ Resets EFCI on non-RM cells whose PTI[2] = 0, on
selected connections.
Header = 5 bytesPayload = 48 bytes
GFC/VPI
VPI
VCI
PTICLPHEC
8
PIDERCCRMCRQLSN
2 x 84 x 84 x 82 x 82 x 8
830 x 8 + 6
¥ Checks CRC on received RM cells and generates CRC for
transmitted RM cells.
¥ Provides different priority to RM cells.
¥ Can copy RM cells to the microprocessor or remove them
from the ßow.
5.2.RM Cell Definition
A cell is an RM cell if and only if at least one of the following
conditions is met:
¥ The cell belongs to a VC connection and its PTI = 6.
¥ The cell belongs to a VP connection, its VCI = 6, and its
PTI = 6.
¥ The cell belongs to a VP connection, its VCI = 6, and the
ATMC CFB ConÞguration Register has the
Cell PTI
bit set.
VPRPÑVP RM
5.3.RM Cell Fields
10
Reserved
CRC-10
DIR
BN
1
1
NOTES:
PID= 1
DIR= Direction
¥ 0 = Forward RM cell
¥ 1 = Backward RM cell
BN= Backward Explicit Congestion
¥ 0 = Generated by source
¥ 1 = Generated not by the source
CI= Congestion Indication
NI= No Increase Bit
ER= Explicit Rate
CCR = Current Cell Rate
MCR = Minimum Cell Rate
CRC - 10
CI 1NI 1RA
1
Figure 9. RM Cell Fields
Reserved
3
MC92501
12
MOTOROLA
5.4.Cell Marking (CI, NI, PTI)
Figure 10 illustrates two MC92501 devices connected to a
switch fabric. In this example, the ABR ßow travels from left to
right. This means that data cells are ßowing from left to right,
FRM cells are ßowing from left to right, and BRM cells are
ßowing from right to left. The switch marks FRM and user cells
Ingress Flow Status
Ingress User Cell Marking (EFCI)
Ingress FRM Cell Marking (CI or NI)
Egress BRM Cell Marking (CI or NI)
EFCI, FRM
BRM
Ingress
Switch Fabric
Egress
MC92501 #1MC92501 #2
ßowing downstream, and BRM cells ßowing upstream. This
switch function can be implemented in the ingress of MC92501
#1 and in the egress of MC92501 #2. MC92501 #1 marks cells
because of the ingress ßow status (for example, ingress ßow
congestion) while MC92501 #2 marks cells because of the
egress ßow status.
Egress Flow Status
Egress User Cell Marking (EFCI)
Egress FRM Cell Marking (CI or NI)
Ingress BRM Cell Marking (CI or NI)
Egress
Ingress
Downstream Direction
Upstream Direction
Figure 10. ABR Flow Cell Marking Example
The MC92501 can take the following actions in response to
the ingress ßow status:
¥ Perform EFCI marking on ingress cells; i.e., set PTI[1] bit
in cells on which PTI[2] = 0.
¥ Set CI or NI in ingress FRM cells.
¥ Set CI or NI in egress BRM cells.
The MC92501 can take the following actions in response to
the egress ßow status:
¥ Perform EFCI marking on egress cells; i.e., set PTI[1] bit
in cells on which PTI[2] = 0.
¥ Set CI or NI in egress FRM cells.
¥ Set CI or NI in ingress BRM cells.
Figure 11 is an overview of the MC92501 marking scheme.
MOTOROLA
MC92501
13
OR
Global Reg.
CellÕs Overhead
Context Bit
Ingress
Status
Collection
Global Registers
Ingress Flow Status
Ingress
Action:
Marking
Global Registers
Context Bits
Cell Type
Set CI
Set NI
Set PTI
Global Reg.
CellÕs Overhead
Context Bit
Egress
Status
Collection
Egress Flow Status
Figure 11. Cell Marking Scheme
There are various ways to inform the MC92501 that it should
mark a cell due to the ingress ßow status or the egress ßow
status. This scheme also shows that the status of the ingress
ßow, the status of the egress ßow, global registers, a context
bit, and the cell type impact the decision of setting CI, NI, and
PTI. Following is a detailed description of each of the function
boxes.
5.4.1.Sources for Ingress Flow Status
The ingress ßow status is gathered from three sources:
global register, cellÕs overhead, or context bit.
5.4.1.1.Ingress Flow Status from Global Register
The switch fabric can notify the MC92501 that it should mark
cells because of the ingress ßow status by setting the
Global Ingress ABR Mark Enable
bit in the Ingress Processing
Control Register (IPLR).
5.4.1.2.Ingress Flow Status from CellÕs Overhead
The switch fabric can notify the MC92501 that it should mark
cells because of the ingress ßow status of connection #n by
setting the
IFSÑOverhead Ingress Flow Status bit in the
overhead of egress cells belonging to that connection. The
location of this bit in the overhead is programmable using the
EIBYÑIFS Byte Location bit and the EIBIÑIFS Bit Location
bit in the Egress Switch Overhead Information Register 1
(ESOIR1). This bit is enabled by the
EIASÑGlobal IFS Enable
bit in the Egress Switch Interface ConÞguration Register
(ESWCR). The MC92501 can be programmed that in such a
case it will mark egress BRM cells.
IAMEÑ
Set CI
Egress
Action:
Marking
Section 5.4.1.2 for details on enabling of
Ingress Flow Status
receives that cell, it copies the bit into the
Ingress Flow Status
Set NI
Set PTI
IFSÑOverhead
bit and its location.) When the MC92501
CIFSÑConnection
bit in the Common Parameters Extension
Word of connection #n. The MC92501 can be programmed
that in such a case it will mark ingress FRM cells or perform
EFCI marking.
5.4.1.4.Logic of Ingress Flow Status
The ingress ßow status equals 1 if:
IAME = 1
IFS = 1 and EIAS = 1 and egress = 1 OR
CIFS = 1 and EIAS = 1 and ingress = 1
Where:
IAME = Global Ingress ABR Mark Enable
IFS = Overhead Ingress Flow Status
EIAS = Global IFS Enable
CIFS = Connection IFS Enable
Egress = Programmed Overhead Egress Bit
Ingress = Programmed Overhead Ingress Bit
5.4.2.Sources for Egress Flow Status
The egress ßow status is gathered from three sources:
global register, cellÕs overhead, and context memory.
5.4.2.1.Egress Flow Status from Global Register
The switch fabric can notify the MC92501 that it should mark
cells because of the egress ßow status by setting the
Global Egress ABR Mark Enable
bit in the Egress Processing
EAMEÑ
Control Register (EPLR).
5.4.1.3.Ingress Flow Status from Context Memory
The switch fabric can notify the MC92501 that it should mark
cells because of the ingress ßow status of connection #n by
setting the
IFSÑOverhead Ingress Flow Status bit in the
overhead of egress cells belonging to that connection. (See
MC92501
14
5.4.2.2.Egress Flow Status from CellÕs Overhead
The switch fabric can notify the MC92501 that it should mark
cells because of the egress ßow status of connection #n by
setting the
EFSÑOverhead Egress Flow Status bit in the
overhead of egress cells belonging to that connection. The
MOTOROLA
location of this bit in the overhead is programmable using the
EEBYÑEFS Byte Location bit and the EEBIÑEFS Bit
Location
Register 1 (ESOIR1). This bit is enabled by the
EFS Enable
Register (ESWCR). The MC92501 can be programmed that
in such a case it will mark egress FRM cells or perform EFCI
marking.
5.4.2.3.Egress Flow Status from Context Memory
cells because of the egress ßow status of connection #n by
setting the
overhead of egress cells belonging to that connection. (See
Section 5.4.2.2 for details on enabling of EFSÑOverheadEgress Flow Status bit and its location.) When the MC92501
receives that cell, it copies the bit into the CEFSÑConnectionEgress Flow Status bit in the Common Parameters Extension
Word of connection #n. The MC92501 can be programmed
that in such a case it will mark ingress BRM cells.
5.4.2.4.Logic of Egress Flow Status
5.4.3.Ingress ABR Marking Bits
ßow status or egress ßow status.
MC92501 can perform one or more of the following:
¥ Set CI bit in an ingress FRM cell Ñ when the ISFCEÑ
¥ Set NI bit in an ingress FRM cell Ñ when the ISFNEÑ
¥ Set PTI[1] bit in an ingress cell whose PTI[2] = 0 Ñ when
MC92501 can perform one or more the following:
¥ Set CI bit in an ingress BRM cell Ñ when the ISBCEÑ
¥ Set NI bit in an ingress BRM cell Ñ when the ISBNEÑ
bit in the Egress Switch Overhead Information
EEASÑGlobal
bit in the Egress Switch Interface ConÞguration
The switch fabric can notify the MC92501 that it should mark
EFSÑOverhead Egress Flow Status bit in the
The egress ßow status equals 1 if:
EAME = 1 OR
EFS = 1 and EEAS = 1 and egress = 1 OR
CEFS = 1 and EEAS = 1 and ingress = 1
Where:
EAME = Global Egress ABR Mark Enable
EFS = Overhead Egress Flow Status
EEAS = Global EFS Enable
CEFS = Connection EFS Enable
Egress = Programmed Overhead Egress Bit
Ingress = Programmed Overhead Ingress Bit
The MC92501 can mark cells as a result of either ingress
In the case where ingress ßow status is asserted, the
Global Ingress Set FRM CI Enable bit in the Ingress
Processing ConÞguration Register (IPCR) is set.
Global Ingress Set FRM NI Enable bit in the IPCR is set.
the ISPEÑGlobal Ingress Set PTI Enable bit in the IPCR
is set.
In the case where egress ßow status is asserted, the
Global Ingress Set BRM CI Enable bit in the IPCR is set.
Global Ingress Set BRM NI Enable bit in the IPCR is set.
All cell marking on the ingress is enabled on a per-
connection basis by the CIMEÑConnection Ingress Marking
Enable bit in the Common Parameters Extension Word.
5.4.3.1.Logic of Ingress ABR Marking Bits
The CI bit is set if:
FRM cell and CIME = 1 and ingress ßow status = 1
and ISFCE = 1 OR
BRM cell and CIME = 1 and egress ßow status = 1
and ISBCE = 1
The NI bit is set if:
FRM cell and CIME = 1 and ingress ßow status = 1 and
ISFNE = 1 OR
BRM cell and CIME = = 1 and egress ßow status = 1 and
ISBNE = 1
The PTI[1] bit is set if:
PTI[2] = 0 and CIME = 1 and ingress ßow status = 1 and
ISPE = 1
Where:
CIME = Connections Ingress Marking Enable
FRM Cell = Cell marked as FRM cell
BRM Cell = Cell marked as BRM cell
Ingress Flow Status = Set as deÞned in Section 5.4.1.4
Egress Flow Status = Set as deÞned in Section 5.4.2.4
ISFCE = Global Ingress Set FRM CI Enable
ISFNE = Global Ingress Set FRM NI Enable
ISPE = Global Ingress Set PTI Enable
ISBCE = Global Ingress Set BRM CI Enable
ISBNE = Global Ingress Set BRM NI Enable
5.4.4.Egress ABR Marking Bits
The MC92501 can mark cells as a result of either ingress
ßow status or egress ßow status.
In the case where egress ßow status is asserted, the
MC92501 can perform one or more of the following:
¥ Set CI bit in an egress FRM cell Ñ when the ESFCEÑ
Global Egress Set FRM CI Enable bit in the Egress
Processing ConÞguration Register (EPCR) is set.
¥ Set NI bit in an egress FRM cell Ñ when the ESFNEÑ
Global Egress Set FRM NI Enable bit in the EPCR is set.
¥ Set PTI[1] bit in an egress cell whose PTI[2] = 0 Ñ when
the ESPEÑGlobal Egress Set PTI Enable bit in the EPCR
is set.
In the case where ingress ßow status is asserted, the
MC92501 can perform one or more the following:
¥ Set CI bit in an egress BRM cell Ñ when the ESBCEÑ
Global Egress Set BRM CI Enable bit in the EPCR is set.
¥ Set NI bit in an egress BRM cell Ñ when the ESBNEÑ
Global Egress Set BRM NI Enable bit in the EPCR is set.
All cell marking on the egress is enabled on a per-
connection basis by the CEMEÑConnection Egress MarkingEnable bit in the Common Parameters Extension Word.
MOTOROLA
MC92501
15
5.4.4.1.Logic of Egress ABR Marking Bits
The CI bit is set if:
FRM cell and CEME = 1 and egress ßow status = 1
and ESFCE = 1 OR
BRM cell and CEME = 1 and ingress ßow status = 1
and ESBCE = 1
The NI bit is set if:
FRM cell and CEME = 1 and egress ßow status = 1 and
ESFNE = 1 OR
BRM cell and CEME = 1 and ingress ßow status = 1 and
ESBNE = 1
The PTI[1] bit is set if:
PTI[2] = 0 and CEME = 1 and egress ßow status = 1 and
ESPE = 1
Where:
CEME = Connections Egress Marking Enable
FRM Cell = Cell marked as FRM cell
BRM Cell = Cell marked as BRM cell
Ingress Flow Status = Set as deÞned in Section 5.4.1.4
Egress Flow Status = Set as deÞned in Section 5.4.2.4
ESFCE = Global Egress Set FRM CI Enable
ESFNE = Global Egress Set FRM NI Enable
ESPE = Global Egress Set PTI Enable
ESBCE = Global Egress Set BRM CI Enable
ESBNE = Global Egress Set BRM NI Enable
5.4.5.Cell Marking Examples
Figure 12, Figure 13, and Figure 14 provide examples for
CI and NI marking.
NOTE 1
NOTE 3
NOTE 4
MC92501 #1MC92501 #2
NOTES:
1. Initially the microprocessor conÞgures the MC92501 #1 as follows:
¥ Sets the ISFCEÑGlobal Ingress Set FRM CI Enable bit.
¥ Sets the CIMEÑConnection Ingress Marking Enable bit for selected ABR connections.
2. The switch fabric informs the microprocessor that ingress ABR queues have reached
some limit.
3. The microprocessor sets the IAMEÑGlobal Ingress ABR Mark Enable bit.
4. The MC92501 sets the CI bit for FRM cells belonging to the selected ABR connections.
Microprocessor
NOTE 2
Switch Fabric
Downstream Direction
Upstream Direction
Figure 12. Enable Marking CI Bits of Ingress FRM Cells
MC92501
16
MOTOROLA
NOTE 1
Microprocessor
NOTE 3
MC92501 #1MC92501 #2
NOTES:
1. Initially the microprocessor conÞgures the MC92501 #1 as follows:
¥ Sets the EIASÑGlobal IFS Enable bit.
¥ Sets the ESBNEÑGlobal Egress Set BRM NI Enable bit.
¥ Programs the location of the IFSÑOverhead Ingress Flow Status bit by writing to the EIBYÑIFS Byte Location and the EIBIÑIFS
Bit Location Þelds.
On connection setup the microprocessor conÞgures the MC92501 #1 as follows:
¥ Sets the CEMEÑConnection Egress Marking Enable bit for selected ABR connections.
2. The switch detects that the ingress queue of connection #n has reached a limit. It sets ingress ßow status bit on the overhead of
cells belonging to that connection.
3. The MC92501 sets the NI bit of BRM cell belonging to connection #n.
NOTE 2
Switch Fabric
Downstream Direction
Upstream Direction
Figure 13. Egress Flow Contains Ingress Flow Status and Causes the MC92501 to Mark BRM Cell NI Field
Microprocessor
NOTE 1
NOTE 2
NOTE 3
Switch Fabric
MC92501 #1MC92501 #2
Downstream Direction
Upstream Direction
NOTES:
1. Initially the microprocessor conÞgures the MC92501 #2 as follows:
¥ Sets the EEASÑGlobal EFS Enable bit.
¥ Sets the ISBCEÑGlobal Ingress Set BRM CI Enable bit.
¥ Programs the location of the EFSÑOverhead Egress Flow Status bit by writing to the EEBYÑEFS Byte Location and the EEBIÑEFS Bit Location Þelds.
On connection setup the microprocessor conÞgures the MC92501 #2 as follows:
¥ Sets the CEMEÑConnection Egress Marking Enable bit for selected ABR connections.
2. The switch detects that the egress queue of connection #n has reached a limit. It sets egress ßow status bit on the overhead of cells
belonging to that connection.
3. The MC92501 copies the overhead egress ßow status bit into the CEFSÑConnection Egress Flow Status bit and effectively sets it.
4. The MC92501 sets the CI bit of BRM cells belonging to connection #n.
NOTE 4
Figure 14. Egress Flow Contains Egress Flow Status for Connection #n Causes the MC92501 to Mark CI Bit
of All Ingress BRM Cells Belonging to That Connection
MOTOROLA
MC92501
17
5.5.Ingress Switch ABR Priority Interface
The MC92501 deÞnes an 8-bit Þeld which can be overlayed
on bits of the ingress switch parameters belonging to RM cells.
In applications where the overlayed switch parameter Þeld is
a priority Þeld which is used by the switch fabric, RM cells can
gain higher priority in passing the switch and thus enable
shortening the feedback loop for ABR.
The MC92501 performs this Þeld overlay if one of the
following occurs:
¥ An ingress BRM cell is received and both the IBOEÑ
Ingress BRM Overlay Enable bit and the IROEÑIngress
RM Overlay Enable bit are set.
WRITE HERE IF CELL IS A BRM
Byte #8Byte #9Byte #10Byte #11
1
Switch Params #2
Byte #4Byte #5Byte #6Byte #7
Switch Params #1
Byte #0Byte #1Byte #2Byte #3
Switch Params #0
001
¥ An ingress FRM cell is received and both the IFOEÑ
Ingress FRM Overlay Enable bit and the the IROEÑ
Ingress RM Overlay Enable bit are set.
Once the MC92501 is enabled, it uses the ROLÑRM
Overlay Location Þeld in order to locate one byte out of 12
bytes on the ingress switch parameters. This byte is overlaid
by the ROFÑRM Overlay Þeld only on bits which are enabled
by the ROMÑRM Overlay Mask Þeld.
See Section 11.4.12 for Þeld descriptions.
5.5.1.An Example
Figure 15 demonstrates the ingress switch APB priority
interface, supposing that the ROF Þeld = 11001101, the ROM
Þeld = 01111000, the ROL Þeld = 9, the IBOE bit is set, and
that the current cell is a backward RM cell.
Figure 15. Example of Ingress Switch ABR Priority Interface
5.6.Egress Reset EFCI
The MC92501 resets PTI[1] on a cell which meet the
following conditions:
¥ Its PTI[2] = 0.
¥ It is a non-RM cell.
¥ The EREFÑEgress Reset EFCI bit in the Context
Parameters Extension Table for the connection to which
the cell belongs is set.
¥ This feature can be used as part of Òdestination behavior.Ó
Egress Reset EFCI is enabled globally by the EPCVÑEgress Features Enable bit in the ACR.
MC92501
18
MOTOROLA
SECTION 6. CLP TRANSPARENCY
6.1.Overview
The trafÞc management speciÞcation deÞnes two network
operation models with relation to CLP = 1 ßow: CLP
transparent and CLP signiÞcant. A connection which is CLP
transparent does not have different Cell Loss Ratio (CLR) for
CLP = 0 or CLP = 1 trafÞc, and therefore does not prefer
discarding CLP = 1 over CLP = 0 cells on congestion. Current
switch fabrics do distinguish globally between CLP = 0 trafÞc
and CLP = 1 trafÞc with the CLP = 1 trafÞc being more
susceptible to discarding in case of congestion.
The MC92501 solves the problem in the following manner:
¥ If a cell belongs to a connection which supports CLP
transparency (the ICTEÑIngress CLP TransparencyEnable bit in the Common Parameters Extension Word is
set), then the MC92501 moves the CLP from the cellÕs
ICTVÑIngress CLP Transparency Value
NOTE 1
PayloadOverheadHeader
header to IOCLPÑIngress Overhead CLP bit in the cellÕs
overhead and assigns 0 to the header CLP.
¥ The cell is forwarded to the switch fabric. The switch fabric
considers the cell as if it has CLP = 0.
¥ On the egress side, the MC92501 reconstructs the header
CLP from the EOCLPÑEgress Overhead CLP bit in the
cellÕs overhead.
In order to support other applications, the MC92501 function
is extended as follows. It assigns the ICTVÑIngress CLPTransparency Value bit (deÞned in the Common Parameters
Extension Word) to the cellÕs header instead of 0. If the ICTVÑIngress CLP Transparency Value bit = 0 then we are back at
the CLP transparency application. CLP Transparency is
enabled globally by the IPCVÑIngress Features Enable bit in
the ACR register.
NOTE 2
Overhead
Header
Payload
Switch Fabric
MC92501 #1MC92501 #2
NOTES:
1. If a cell belongs to a connection which supports CLP transparency (the ICTEÑIngress CLP Transparency Enable bit
is set), then MC92501 moves the CLP from the cellÕs header to the IOCLPÑIngress Overhead CLP bit in the cellÕs
overhead and assigns the ICTVÑIngress CLP Transparency Value bit to the header CLP.
2. If the CIFSÑConnection Ingress Flow Status bit is set then the MC92501 reconstructs the header CLP from the
EOCLPÑEgress Overhead CLP bit in the cellÕs overhead.
Figure 16. CLP Transparency with a CLP Significant Switch Fabric
The IOCLPÑIngress Overhead CLP bit location is
programmable using the OCBLÑIOCLP Bit Location Þeld in
the Ingress Switch Interface ConÞguration Register (ISWCR).
The EOCLPÑEgress Overhead bit location is
programmable using the EOBYÑEOCLP Byte Location Þeld
and the EOBIÑEOCLP Bit Location Þeld in the Egress Switch
Overhead Information Register 1 (ESOIR1).
MOTOROLA
MC92501
19
SECTION 7. INDIRECT EXTERNAL MEMORY ACCESS
7.1.Overview
The MC92500 allows the processor to access its external
memory for the duration of one cell processing time out of N,
where N is programmable. The MC92501 will additionally
allow the processor to access external memory while
operating on the cell stream. The indirect access takes place
at least once in every cell processing slot. The indirect access
is not performed during maintenance.
7.2.User Interface
Indirect external memory access is performed using two
registers: the Indirect External Memory Access Address
Register (IAAR) and the Indirect External Memory Access
Data Register (IADR).
7.2.1.Write Access
In order to write to the external memory, the processor
should poll the IABÑIndirect External Memory Access Busy
bit in the IAAR register to verify the status of the IAAR and
IADR registers. If IAB is clear, then the processor can write
the address data and status into the appropriate registers. The
IADÑIndirect External Memory Address DIR bit is is set to 0
for a write operation. Writing to the IAAR register triggers the
MC92501 to wait for a dedicated clock to write the data into
the external memory using the given address and data. Once
the MC92501 Þnishes writing, it clears the IABÑIndirectExternal Memory Access Busy bit in the IAAR register.
7.2.2.Read Access
In order to read from the external memory, the processor
should poll the IABÑIndirect External Memory Access Busy
bit in the IAAR register to verify that it may write the IAAR
register. If IAB is clear, then the processor can write the
address, size, and direction into the appropriate registers. For
a read operation, the IADÑIndirect External Memory AddressDIR bit is set to 1. Writing to the IAAR register triggers the
MC92501 to wait for a dedicated clock, and read the data from
external memory using the given address and write the data
into the IADR register. Once the data was written into the IADR
register, the MC92501 clears the IABÑIndirect ExternalMemory Access Busy bit in the IAAR register. The processor
then may read the data from the IADR register.
The address space which is covered by this interface
includes all the non-destructive external memory access and
an external address compression device.
NOTE
Indirect write access to an external memory
space, which can be written by the MC92501, is
not recommended. For example, an indirect write
access to a ßag-table record of an active
connection is not recommended. It is advisable to
use the maintenance cell slot for this purpose.
Table 1 summarizes indirect access Þelds:
Table 1. Indirect Access Fields
IADÑ
Indirect
External
Memory
Access DIR
00xxWrite IADR[31:00] to external memory word bits [31:00]
0100Write IADR[31:16] to external memory word [31:16]
0101Write IADR[15:00] to external memory word [15:00]
0110Write IADR[15:00] to external memory word [15:00]
0111Write IADR[31:16] to external memory word [31:16]
1xxxRead external memory word bits [31:00] to IADR[31:00]
IAWÑ
Indirect
External
Memory
Access Size
Least
SigniÞcant
Bit of IAAÑ
Indirect
External
Memory
Access
Address
DOÑData
Order
Function
MC92501
20
MOTOROLA
SECTION 8. IMPROVED HOST INTERFACE
8.1.Overview
In order to improve the MC92501 interface to the
microprocessor, the following features were added:
8.1.1.An Additional MDTACK
The MDT
on which there are two MDT
ACK signals enable a glueless interface to systems
ACK signals and their combination
conveys the bus width of the slave. The MDT
driven only when the MDT
ACK0 signal is driven and when the
MDCÑMDTACK Drive Control is set.
8.1.2.Programmable MREQ
The MC92500 generates three DREQ
for external memory request, MCIREQ for cell insertion
request, and MCOREQ
for cell extraction request. The
MC92501 changes the functionality (and the name) of the
three DREQ
lines. Each MREQ[n] line may be connected to
each of the internal requests: external memory, cell insertion,
or cell extraction. If a system contains three DMAs, then each
DMA may be connected to a different DMA DREQ
WSSMÑWord Select Signals
Mode = 0
Signal
ACK1 signal is
Signals
signals: EMMREQ
signal. In
Table 2. Host Interface Fields
WSSMÑWord Select Signals
Mode = 1
and
DOÑData Order = 0
systems with less than three DMAs, each such DMA should
switch between the MREQ
lines. This switching can be done
when using the MC92500 through the use of glue logic. In
these cases, this feature enables a glueless interface. The
MREQ
signals default to the MC92500 conÞguration so that
backwards compatibility is maintained.
8.1.3.Update the Definition of MWSH
and MWSL
Signals
MWSH
and MWSL are used in MC92500 as word select
during external memory accesses. Some systems assert
these signals only during write cycles. The MC92501 uses
these signals only on external memory write accesses. Read
accesses are always performed regardless of these signals.
Moreover, another use of these signals is for systems to use
them for Address[1] bit and SIZE. The MC92501 supports
another deÞnition for the same pins: MWSH
Address[1] while MWSL
can serve as SIZE. The WSSMÑ
pin can serve as
Word Select Signals Mode bit selects which mode is
supported. Table 2 describes the pinsÕ functionality:
WSSMÑWord Select Signals
Mode = 1
and
DOÑData Order = 1
Function
MWSHMWSLA1SizeA1Size
00x0x0Write D(31:0)
010111Write D(31:16)
101101Write D(15:00)
MOTOROLA
MC92501
21
SECTION 9. EGRESS OVERHEAD MANIPULATION
The MC92501 supports the following features:
¥ The size of the ECI Þeld used by the egress cell processing
block can be programmed by writing to the ECESÑEgressCell Processing Block ECI Size Þeld in the Egress
Overhead Manipulation Register (EGOMR).
¥ The size of the MTTS Þeld used by the egress cell
processing block can be programmed by writing to the
ECTSÑEgress Cell Processing Block MTTS Size Þeld in
the EGOMR.
¥ The M bit used by the egress cell processing block can be
either the M bit which was extracted from the cellÕs
overhead, the logical not of the M bit which was extracted
from the cellÕs overhead, or 1 or 0 by programming the
ECMSÑEgress Cell Processing Block M Bit Source Þeld
in the EGOMR.
¥ In ECI on Header mode, the ECI is extracted from the ATM
cell header. The header VPI Þeld size can be programmed
to either 12 bits or 8 bits using the VPSÑVPI Size in ECIon Header Mode bit of the Egress Switch Interface
ConÞguration Register (ESWCR).
MC92501
22
MOTOROLA
SECTION 10. UTOPIA LEVEL 2 PHY INTERFACE
The MC92501 PHY interface can be programmed to
support UTOPIA Level 2. It allows for operation of one TxClav
and one RxClav signal.
On the ingress direction, the MC92501 supports address
polling on up to 16 physical links. It scans all the links in a
round robin fashion and decides from which PHY to read the
next cell.
UTOPIA Level 2 is enabled on ingress by programming the
IUMÑIngress UTOPIA Mode bit of the Ingress PHY
ConÞguration Register (IPHCR). In this mode, Receive PHY
ID 0-3/Receive Address 0-3 (RXPHYID0-RXPHYID3/
RXADD0-RXADDR3) signals function as RXADDR0-
RXADDR4.
NOTE
RXADDR4 is a new functional signal in MC92501
that was a NC (No Connect) on the MC92500.
PHY Device
PHY Device
TxAddr
UTOPIA Level 2 is enabled on egress by programming the
EUMÑEgress UTOPIA Mode bit of the Egress PHY
ConÞguration Register (EPHCR). In this mode, the Transmit
PHY ID 0-3 /Transmit Address 0-3 (TXPHYID0-TXPHYID3/
TXADDR0-TXADDR3) signals and the Transmit Next PHY ID
Valid/Transmit Address 4 (TXPHYIDV/TXADDR4) signal are
used as TXADDR0-TXADDR4. The MC92501 polls the link of
the cell in its egress PHY IF FIFO, and when enabled it outputs
the cell to the link PHY. The MC92501 performs address
polling on all other links as well in order to enable external
logic to monitor the PHYÕs status.
Figure 17 illustrates an application on which external logic
monitors the RxClav signal while the MC92501 is polling the
PHY devices. Based on this information, the external logic can
input cells to the MC92501 egress ßow.
MC92501
PHY Device
PHY Device
Feedback
TxAddr
TxClav
External Logic
Figure 17. Feedback Using TxClav
MOTOROLA
MC92501
23
SECTION 11. REGISTER DESCRIPTIONS
This section describes the registers which were added or modiÞed for the MC92501.
11.1.General Register List
Table 3 contains all the registers which were added and their addresses.
FQFÑFMC Queue Full
This bit reports that the internal FMC queue is full. The reason for this is that the FMC generation rate is higher then the allocated
insertion bandwidth. Insertion rate is controlled by the insertion leaky bucket.
11.2.2.Interrupt Mask Register (IMR)
The FQFE bit has been added.
31302928272625242322212019181716
OMECMEMSEE0000SPDE00FQFECIQEECEQRECEQIECEQLECEQFE
1514131211109876543210
00UDIE50 UDIE40 UDIE30UDIE20UDIE1E
SPEE
UDIE0E
SHEE
000IPPEEIPHEEFQOEFQEOE MNAEE
FQFEÑFMC Queue Full Interrupt Enable
When FQF and FQFE are set, an interrupt is generated.
MC92501
24
MOTOROLA
11.2.3.ATMC CFB Revision Register (ARR)
The AMRV Þeld has been updated.
Table 4. Values of ATMC CFB Revision Fields
AMRVASRVATMC CFB Revision
000001000000MC92501 (MC92500 Revision B)
11.2.4.MC92501 Revision Register (RR)
The MRV Þeld has been updated.
Table 5. Values of the MC92501 Revision Fields
IDMRVSRVMC92501 Revision
10000000000000000000000001000000MC92501 (MC92500 Revision B)
11.3.Control Registers
The following registers have been added.
11.3.1.Ingress Processing Control Register (IPLR)
This register has the following structure:
31302928272625242322212019181716
0000000000000000
1514131211109876543210
00000000000000ICNGIAME
ICNGÑGlobal Ingress Congestion NotiÞcation
This bit notiÞes the MC92501 whether there is congestion in the ingress ßow. See Section 4.
0 = No ingress congestion.
1 = Ingress congestion. The MC92501 performs selective discard according to per-connection CIMEÑConnection Ingress
Marking Enable bit.
IAMEÑGlobal Ingress ABR Mark Enable
This bit, when set, indicates that current ingress ßow status implies that the MC92501 should perform RR marking and/or EFCI
marking if enabled. See Section 5.4.1.
11.3.2.Egress Processing Control Register (EPLR)
This register has the following structure:
31302928272625242322212019181716
0000000000000000
1514131211109876543210
000000000000000EAME
EAMEÑGlobal Egress ABR Mark Enable
This bit, when set, indicates that current egress ßow status implies that the MC92501 should perform RR marking and/or EFCI
This register contains the address, width, and busy bit for accessing the MC92501 external memory or the external memory
device. Refer to Section 7 for details. The register has the following structure:
31302928272625242322212019181716
IABIADIAW000IAASIAA
1514131211109876543210
IAA0
IABÑIndirect External Memory Access Busy
This bit indicates that indirect external memory access mechanism is busy.
0 = Indirect access mechanism is free and therefore indirect external memory access data register can be accessed.
1 = Indirect access mechanism is busy and therefore indirect access data register should not be accessed.
IADÑIndirect External Memory Access DIR
This bit indicates indirect access direction.
IAAÑIndirect External Memory Access Address
This Þeld indicates bits 23:1 of the address within the address space speciÞed in the IAASÑIndirect External Memory Access
Address Space Þeld.
11.3.4.Indirect External Memory Access Data Register (IADR)
This register contains the data which should be written to the external memory in case of an indirect write access or the data
that was last read from external memory in case of an indirect read access. Refer to Section 7 for details.
MC92501
26
MOTOROLA
11.4.Configuration Register
The following registers have been updated. The Þelds that have been added are in bold.
IGCTEÑGlobal Ingress CLP Transparency Enable
This bit enables CLP transparency function on the ingress. See Section 6 for details.
ICCRÑIngress Check CRC on RM Cells
This bit determines whether the CRC of RM cells that are received in the ingress is checked.
0 = The CRC of RM cells that are recevied in the ingress is not checked.
1 = The CRC of RM cells that are received in the ingress is checked and if it is not okay, then the cell is removed and can be
copied to the microprocessor.
IRCRÑIngress Recalculate CRC on RM Cells
This bit determines whether the CRC of ingress RM cells is recalculated.
0 = The CRC of ingress RM cells is not recalculated.
1 = The CRC of ingress RM cells is recalculated.
ISFCEÑGlobal Ingress Set FRM CI Enable
This bit enables setting CI bit in forward RM cells received in ingress. See Section 5.4.3 for details.
0 = Setting CI bit in forward RM cells received in ingress is disabled.
1 = Setting CI bit in forward RM cells received in ingress is enabled.
ISFNEÑGlobal Ingress Set FRM NI Enable
This bit enables setting NI bit in forward RM cells received in ingress. See Section 5.4.3 for details.
0 = Setting NI bit in forward RM cells received in ingress is disabled.
1 = Setting NI bit in forward RM cells received in ingress is enabled.
ISPEÑGlobal Ingress Set PTI Enable
This bit enables setting PTI[1] bit in cells with PTI[2] = 0 which are received in ingress. See Section 5.4.3 for details.
0 = Setting PTI[1] bit in cells with PTI[2] = 0 which are received in ingress is disabled.
1 = Setting PTI[1] bit in cells with PTI[2] = 0 which are received in ingress is enabled.
ISBCEÑGlobal Ingress Set BRM CI Enable
This bit enables setting CI bit in backward RM cells received in ingress. See Section 5.4.3 for details.
0 = Setting CI bit in backward RM cells received in ingress is disabled.
1 = Setting CI bit in backward RM cells received in ingress is enabled.
ISBNEÑGlobal Ingress Set BRM NI Enable
This bit enables setting NI bit in backward RM cells received in ingress. See Section 5.4.3 for details.
0 = Setting NI bit in backward RM cells received in ingress is disabled.
1 = Setting NI bit in backward RM cells received in ingress is enabled.
IROEÑIngress RM Overlay Enable
This bit enables updating switch parameter words in the case of RM cells. See Section 5.5 for details.
IPCVÑIngress Features Enable
This bit should be set when the following features are used: packet-based UPC, selective discard, and CLP transparency.
EGCTEÑGlobal Egress CLP Transparency Enable
This bit enables CLP transparency function on the egress. See Section 6 for details.
ECCRÑEgress Check CRC on RM Cells
This bit determines whether the CRC of RM cells that are received in the egress is checked.
0 = The CRC of RM cells that are recevied in the egress is not checked.
1 = The CRC of RM cells that are received in the egress is checked and if it is not okay, then the cell is removed and can be
copied to the microprocessor.
ERCRÑEgress Recalculate CRC on RM Cells
This bit determines whether the CRC of egress RM cells is recalculated.
0 = The CRC of egress RM cells is not recalculated.
1 = The CRC of egress RM cells is recalculated.
ESFCEÑGlobal Egress Set FRM CI Enable
This bit enables setting CI bit in forward RM cells received in egress. See Section 5.4.2.
0 = Setting CI bit in forward RM cells received in egress is disabled.
1 = Setting CI bit in forward RM cells received in egress is enabled.
ESFNEÑGlobal Egress Set FRM NI Enable
This bit enables setting NI bit in forward RM cells received in egress. See Section 5.4.2.
0 = Setting NI bit in forward RM cells received in egress is disabled.
1 = Setting NI bit in forward RM cells received in egress is enabled.
ESPEÑGlobal Egress Set PTI Enable
This bit enables setting PTI[1] bit in cells with PTI[2] = 0 which are received in egress. See Section 5.4.2.
0 = Setting PTI[1] bit in cells with PTI[2] = 0 which are received in egress is disabled.
1 = Setting PTI[1] bit in cells with PTI[2] = 0 which are received in egress is enabled.
ESBCEÑGlobal Egress Set BRM CI Enable
This bit enables setting CI bit in backward RM cells received in egress. See Section 5.4.2.
0 = Setting CI bit in backward RM cells received in egress is disabled.
1 = Setting CI bit in backward RM cells received in egress is enabled.
ESBNEÑGlobal Egress Set BRM NI Enable
This bit enables setting NI bit in backward RM cells received in egress. See Section 5.4.2.
0 = Setting NI bit in backward RM cells received in egress is disabled.
1 = Setting NI bit in backward RM cells received in egress is enabled.
EPCCÑEgress Policing Counters Control
This Þeld determines which counters appear in the Policing Counters Table if egress UPC is enabled. (The UPCFÑUPC Flow
bit in the ACR is set.) It also determines the size of each record in the table.
000 = The policing table does not exist.
001 = The policing table contains three counters and one reserved long word: DSCD0, DSCD1, TAG, Reserved.
010 = The policing table contains three counters: DSCD0, DSCD1, TAG.
011 = The policing table contains two counters: DSCD, TAG.
100 = The policing table contains one counter: TAG.
101 = The policing table contains one counter: DSCD.
110 = Reserved
111 = Reserved
MC92501
28
MOTOROLA
EIPÑEgress Insertion Priority
This bit determines the priority between inserted/generated cells and egress received cells.
Note that insertion is always limited by the leaky bucket mechanism.
0 = Inserted/generated cellsÕ priority is higher than egress received cells.
1 = Egress received cellsÕ priority is higher than inserted/generated cells.
EPCVÑEgress Features Enable
This bit should be set when the reset EFCI feature is activated. See Section 5.6.
11.4.3.ATMC CFB Configuration Register (ACR)
31302928272625242322212019181716
ATCSPCCOMCINPCEGPCDVTCFLGCOAMCVPRPFTMCRRPPMAC
1514131211109876543210
UPCF000000000000000
VPRPÑVP RM Cell PTI
This bit determines whether a cell is a VP RM cell only if its PTI = 6.
0 = A cell is a VP RM cell if and only if it belongs to a VP connection, its VCI = 6, and its PTI = 6.
1 = A cell is a VP RM cell if and only if it belongs to a VP connection and its VCI = 6.
CRRPÑVC RM Cell Removal Point
This bit determines whether a VC cell whose PTI = 6 or 7 is removed at the OAM termination point, or whether its removal is
subjected to the per-connection enable bits for PTI = 6 or PTI = 7.
0 = A VC cell whose PTI = 6 or 7 is removed at the OAM termination point as deÞned by the EEOTÑEgress End-to-End
OAM Termination bit in the egress and by the IEOTÑIngress End-to-End OAM Termination bit in the ingress.
1 = A VC cell is removed at the egress if the EP6RÑEgress PTI 6 Remove bit is set and its PTI = 6 or if the EP7RÑEgress
PTI 7 Remove bit is set and its PTI = 7. A VC cell is removed at the ingress if the IP6RÑIngress PTI 6 Remove bit is
set and its PTI = 6 or if the IP7RÑIngress PTI 7 Remove bit is set and its PTI = 7.
PMACÑPM on All Connections
This bit determines whether the OAM performance monitoring test can be done on all connections or on 64 connections.
0 = Performance monitoring can be done only on 64 selected connections.
1 = Performance monitoring can be done on all connections.
UPCFÑUPC Flow
This bit determines whether the UPC is active in the ingress ßow or in the egress ßow.
EIASÑGlobal IFS Enable
This bit enables the MC92501 to use theIFSÑOverhead Ingress Flow Status bit in the egress switch overhead. See Section 5.4.1
for details.
0 = The IFSÑOverhead Ingress Flow Status bit is not deÞned in the egress overhead Þelds so it cannot trigger ABR cell
marking.
1 = The IFSÑOverhead Ingress Flow Status bit is deÞned in the egress overhead Þelds and is used by the MC92501 for
marking cells.
MOTOROLA
MC92501
29
EEASÑGlobal EFS Enable
This bit enables the MC92501 to use the EFSÑOverhead Egress Flow Status bit in the egress switch overhead. See
Section 5.4.2 for details.
0 = The EFSÑOverhead Egress Flow Status bit is not deÞned in the egress overhead Þelds so it cannot trigger ABR cell
marking.
1 = The EFSÑOverhead Egress Flow Status bit is deÞned in the egress overhead Þelds and is used by the MC92501 for
marking cells.
VPSÑVPI Size in ECI on Header Mode
This bit determines the size of the VPI Þeld for ECI on Header mode (IHAF = 1). See Section 9 for details.
0 = VPI size is 12 bits
1 = VPI size is 8 bits
11.4.5.Egress Switch Overhead Information Register 0 (ESOIR0)
This register name was ESOIR on MC92500.
The following deÞnition is changed:
MTBI-MTTS Bit Location
This Þeld indicates the location of the MTTS Þeld within the byte speciÞed by the MTBY-MTTS Byte Location Þeld.
0 = MTTS equals the value that resides in bits 7:5 of the byte pointed to by the MTBY-MTTS Byte Location Þeld.
1 = MTTS equals the value that resides in bits 7:6 of the byte pointed to by the MTBY-MTTS Byte Location Þeld.
2 = MTTS equals the value that resides in bit 7 of the byte pointed to by the MTBY-MTTS Byte Location Þeld.
3 = MTTS equals the value that resides in bits 3:0 of the byte pointed to by the MTBY-MTTS Byte Location Þeld.
4 = MTTS equals the value that resides in bits 4:1 of the byte pointed to by the MTBY-MTTS Byte Location Þeld.
5 = MTTS equals the value that resides in bits 5:2 of the byte pointed to by the MTBY-MTTS Byte Location Þeld.
6 = MTTS equals the value that resides in bits 6:3 of the byte pointed to by the MTBY-MTTS Byte Location Þeld.
7 = MTTS equals the value that resides in bits 7:4 of the byte pointed to by the MTBY-MTTS Byte Location Þeld.
Note that this deÞnition is backwards-compatible to the deÞnition in MC92500.
WSSMÑWord Select Signals Mode
This bit deÞnes the functionality of the MP Word Write Enable High / Address 1 (MWSH/A1) and the MP Word Write Enable
Low / SIZE (MWSL/SIZE) signals. See Section 8.1.3 for details.
0 = MWSH/A1 functions as MWSH-word write enable high and MWSL/SIZE functions as MWSL-word write enable low.
1 = MWSH/A1 functions as A1 and MWSL/SIZE functions as SIZE.
RQ0ÑMREQ0
This Þeld deÞnes the functionality of the MP Request 0 (MREQ0
Signal Functionality
) signal. See Section 8.1.2 for details.
00 = Cell in request
01 = Cell in request
10 = Cell out request
11 = External memory request
RQ1ÑMREQ1
This Þeld deÞnes the functionality of the MP Request 1 (MREQ1
Signal Functionality
) signal. See Section 8.1.2 for details.
00 = Cell out request
01 = Cell in request
10 = Cell out request
11 = External memory request
MC92501
30
MOTOROLA
RQ2ÑMREQ2 Signal Functionality
This Þeld deÞnes the functionality of the MP Request 2 (MREQ2
) signal. See Section 8.1.2 for details.
00 = External memory request
01 = Cell in request
10 = Cell out request
11 = External memory request
MDCÑMDTACK Drive Control
This bit determines which MDTACK signals are driven.
The maximum value for the MSDR is therefore 511 instead of 63. This means that the maintenance request signals can be
asserted as much as 511 clocks (or 8 cell processing slots) before the CM bit.
11.4.8.Ingress PHY Configuration Register (IPHCR)
31302928272625242322212019181716
0000000000000000
1514131211109876543210
00000000000IUMINVPDIPOMIPPRIPLP
IUMÑIngress UTOPIA Mode
This bit deÞnes the UTOPIA level mode of the ingress PHY. See Section 10.
0 = UTOPIA Level 1
1 = UTOPIA Level 2
11.4.9.Egress PHY Configuration Register (EPHCR)
31302928272625242322212019181716
0000000000000000
1514131211109876543210
00000000000EUMEPFCEPOMECGEEGIC
EUMÑEgress UTOPIA Mode
This bit deÞnes the UTOPIA level mode of the egress PHY. See Section 10.
0 = UTOPIA Level 1
1 = UTOPIA Level 2
MOTOROLA
MC92501
31
11.4.10. MC92501 General Configuration Register (GCR)
31302928272625242322212019181716
000000000000000PHIDC
1514131211109876543210
000000000CMPC00ILCC00ELCC
CMPCÑContext Parameters Extension Table Control
This bit determines the existence of the Context Parameters Extension Table in external memory. See Section 12.1 for details.
0 = The Common Parameters Table does not exist.
1 = The Common Parameters Table exists.
11.4.11. Egress Switch Overhead Information Register 1 (ESOIR1)
This register determines the location of the overhead information in the data structure received from the switch. The register
has the following structure:
31302928272625242322212019181716
00000000EOBYEOBI
1514131211109876543210
EIBYEIBIEEBYEEBI
EOBYÑEOCLP Byte Location
This Þeld contains the byte number of the switch data structure in which the EOCLPÑEgress Overhead CLP bit in the cellÕs
overhead. The byte on which STXSOC is asserted is byte number 0. See Section 6 for details.
EOBIÑEOCLP Bit Location
This Þeld contains the number of the EOCLPÑEgress Overhead CLP bit in the cellÕs overhead. The most signiÞcant bit is
number 7, and the least signiÞcant bit is number 0. See Section 6 for details.
EIBYÑIFS Byte Location
This Þeld contains the byte number of the switch data structure in which the IFSÑOverhead Ingress Flow Status bit can be
found (overhead, header, and HEC bytes). The byte on which STXSOC is asserted is byte number 0. See Section 5.4.1.2 for details.
EIBIÑIFS Bit Location
This Þeld contains the number of the IFSÑOverhead Ingress Flow Status bit within the byte speciÞed by the EEBYÑEFS Byte
Location Þeld. The most signiÞcant bit is number 7, and the least signiÞcant bit is number 0. See Section 5.4.1.2 for details.
EEBYÑEFS Byte Location
This Þeld contains the byte number of the switch data structure in which the EFSÑOverhead Egress Flow Status bit can be
found (overhead, header, and HEC bytes). The byte on which STXSOC is asserted is byte number 0. See Section 5.4.2.2 for details.
EEBIÑEFS Bit Location
This Þeld contains the number of the EFSÑOverhead Egress Flow Status bit within the byte speciÞed by the EIBYÑIFS Byte
Location Þeld. The most signiÞcant bit is number 7, and the least signiÞcant bit is number 0. See Section 5.4.2.2 for details.
This register contains all the parameters which are related to RM cell overlay. Refer to Section 5.5 for details. The register has
the following structure:
31302928272625242322212019181716
IBOEIFOE00ROLROM
1514131211109876543210
00000000ROF
IBOEÑIngress BRM Overlay Enable
This bit determines whether the MC92501 overlays the ROFÑRM Overlay Þeld on the switch parameters for ingress backward
RM cells.
0 = Switch parameters are not overlayed when a backward RM cell is received in the ingress.
1 = Switch parameters are overlayed when a backward RM cell is received in the ingress.
IFOEÑIngress FRM Overlay Enable
This bit determines whether the MC92501 overlays the ROFÑRM Overlay Þeld on the switch parameters for ingress forward
RM cells.
0 = Switch parameters are not overlayed when a forward RM cell is received in the ingress.
1 = Switch parameters are overlayed when a forward RM cell is received in the ingress.
ROLÑRM Overlay Location
This Þeld contains the number of the switch parameters byte which should be overlayed.
ROMÑRM Overlay Mask
This Þeld contains the byte mask which serves for overlaying the ROFÑRM Overlay Þeld over the ingress switch parameters byte.
ROFÑRM Overlay
This Þeld contains the byte which is overlayed on the ingress switch parameters byte. Each bit in this Þeld is overlayed on the
corresponding bit in the ingress switch parameters only if it is enabled by the corresponding bit in the ROMÑRM Overlay Mask Þeld.
11.4.13. CLP Transparency Overlay Register (CTOR)
This register contains the location of the IOCLPÑIngress Overhead CLP bit in the ingress switch parameters. See Section 6
for details. The register has the following structure:
31302928272625242322212019181716
0000000000000000
1514131211109876543210
000000000OCBIOCBL
OCBLÑIOCLP Byte Location
This Þeld contains the byte number within the switch parameter word on which the IOCLPÑIngress Overhead CLP bit is located.
The most signiÞcant byte is number 0, and the least signiÞcant byte is number 3.
OCBIÑIOCLP Bit Location
This Þeld contains the number of the IOCLPÑIngress Overhead CLP bit within the byte speciÞed by the OCBLÑIOCLP Byte
Location Þeld. The most signiÞcant bit is number 7, and the least signiÞcant bit is number 0.
This register contains Þelds for manipulating egress overhead Þelds. See Section 9 for details.
31302928272625242322212019181716
0000000000000000
1514131211109876543210
00000000 ECMSECTSECES
ECMSÑ Egress Cell Processing Block M Bit Source
This Þeld contains the source for the M bit which is used by the egress cell processing block.
00 = The M bit used by the egress cell processing block is taken from the M bit which is extracted from the switch cell data
structure.
01 = The M bit used by the egress cell processing block is taken from the logical NOT of the M bit which is extracted from
the switch cell data structure.
10 = The M bit used by the egress cell processing block is 0.
11 = The M bit used by the egress cell processing block is 1.
ECTSÑ Egress Cell Processing Block MTTS Size
This Þeld contains the size of the MTTS Þeld which is used by the egress cell processing block.
0 = The MTTS Þeld which is used by the egress cell processing block is the MTTS Þeld, which is extracted from the switch
cell data structure.
1 = The MTTS Þeld which is used by the egress cell processing block is the least signiÞcant bit of the MTTS Þeld, which is
extracted from the switch cell data structure.
2 = The MTTS Þeld which is used by the egress cell processing block is the two least signiÞcant bits of the MTTS Þeld, which
are extracted from the switch cell data structure.
3 = The MTTS Þeld which is used by the egress cell processing block is the three least signiÞcant bits of the MTTS Þeld,
which are extracted from the switch cell data structure.
ECESÑ Egress Cell Processing Block ECI Size
This Þeld contains the size of the ECI Þeld which is used by the egress cell processing block.
0 = The ECI Þeld which is used by the egress cell processing block is the ECI Þeld, which is extracted from the switch cell
data structure.
1 = Reserved
2 = Reserved
3 = Reserved
4 = Reserved
5 = Reserved
6 = The ECI Þeld which is used by the egress cell processing block is the six least signiÞcant bits of the ECI Þeld, which are
extracted from the switch cell data structure.
7 = The ECI Þeld which is used by the egress cell processing block is the seven least signiÞcant bits of the ECI Þeld, which
are extracted from the switch cell data structure.
8 = The ECI Þeld which is used by the egress cell processing block is the eight least signiÞcant bits of the ECI Þeld, which
are extracted from the switch cell data structure.
9 = The ECI Þeld which is used by the egress cell processing block is the nine least signiÞcant bits of the ECI Þeld, which
are extracted from the switch cell data structure.
10 = The ECI Þeld which is used by the egress cell processing block is the 10 least signiÞcant bits of the ECI Þeld, which are
extracted from the switch cell data structure.
11 = The ECI Þeld which is used by the egress cell processing block is the 11 least signiÞcant bits of the ECI Þeld, which are
extracted from the switch cell data structure.
12 = The ECI Þeld which is used by the egress cell processing block is the 12 least signiÞcant bits of the ECI Þeld, which are
extracted from the switch cell data structure.
13 = The ECI Þeld which is used by the egress cell processing block is the 13 least signiÞcant bits of the ECI Þeld, which are
extracted from the switch cell data structure.
14 = The ECI Þeld which is used by the egress cell processing block is the 14 least signiÞcant bits of the ECI Þeld, which are
extracted from the switch cell data structure.
15 = The ECI Þeld which is used by the egress cell processing block is the 15 least signiÞcant bits of the ECI Þeld, which are
extracted from the switch cell data structure.
MC92501
34
MOTOROLA
SECTION 12. EXTERNAL MEMORY DESCRIPTION
The following table has been added.
12.1.Context Parameters Extension Table
Each context Parameters Extension Table record contains one word, the Common Parameters Extension Word.
12.1.1.Common Parameters Extension Word
31302928272625242322212019181716
BKPT[21:12]00CIFSCEFSECTECEME
1514131211109876543210
0PDV00EREF0ISDMUOM0ICTVICTECIME
BKPT[21:12]ÑBucket Pointer[21:12]
When VPAP-VP on all connections, this Þeld contains bits 21 to 12 of the bucket point. When VPAP-VP on all connections is
reset, this Þeld is reserved and should be 0.
CIFSÑConnection Ingress Flow Status
The MC92501 copies the IFSÑOverhead Ingress Flow Status bit to this bit. This bit is used by the ingress processing block
for ABR cell marking. This bit is therefore intended for the MC92501Õs internal use. See Section 5.4.1 for details.
CEFSÑConnection Egress Flow Status
The MC92501 copies the EFSÑOverhead Egress Flow Status bit to this bit. This bit is used by the ingress processing block
for ABR cell marking. This bit is therefore intended for the MC92501Õs internal use. See Section 5.4.2 for details.
ECTEÑEgress CLP Transparency Enable
This bit determines whether CLP should be copied from the EOCLPÑEgress Overhead CLP bit in the cellÕs overhead bit to
the cell header. See Section 6 for details.
0 = CLP should not be copied from the switch overhead to the cell header.
1 = CLP should be copied from the switch overhead to the cell header.
CEMEÑConnection Egress Marking Enable
This bit enables marking of cells which are received in the egress. See Section 5.4.2.
0 = Marking of cells which are received in the egress is disabled.
1 = Marking of cells which are received in the egress is enabled.
IPDVÑIngress Packet Discard Variables
This Þeld is accessed only by the MC92501.
EREFÑEgress Reset EFCI
This bit determines if PTI[1] of an egress cell is to be reset.
0 = PTI[1] of an egress cell is not reset.
1 = PTI[1] of an egress cell is to be reset.
ISDMÑIngress Selective Discard Operation Mode
This Þeld determines the selective discard operation mode. See Section 3.
00 = No selective discard.
01 = Reserved.
10 = Selective discard on CLP = 1 ßow.
11 = Selective discard on CLP = 0+1 ßow.
UOMÑUPC Operation Mode
This Þeld determines the UPC operation mode.
00 = Cell-based UPC
01 = Partial Packet Discard (PPD). See Section 3.3 for details.
10 = Early Packet Discard (EPD). See Section 3.4 for details.
11 = Limited EPD. SeeSection 3.5 for details.
MOTOROLA
MC92501
35
ICTVÑIngress CLP Transparency Value
This bit determines the value that should be written to a cellÕs header if the ICTEÑIngress CLP Transparency Enable bit is set.
See Section 6 for details.
ICTEÑIngress CLP Transparency Enable
This bit determines whether CLP should be copied to the IOCLPÑIngress Overhead CLP bit and whether the ICTVÑIngress
CLP Transparency Value bit should be written to the cell header CLP. See Section 6 for details.
0 = The ingress header CLP bit is not touched.
1 = CLP should be copied from the cell header to the ingress switch parameters. The ICTVÑIngress CLP Transparency
Value bit should be written to the cell header CLP.
CIMEÑConnection Ingress Marking Enable
This bit enables marking of cells which are received in the ingress. See Section 5.4.3 for details.
0 = Marking of cells which are received in the ingress is disabled.
1 = Marking of cells which are received in the ingress is enabled.
12.2.CONTEXT PARAMETERS TABLE
Some bits have been added, and some bit deÞnitons have been updated in the Egress Parameters Word and the Ingress
EP6RÑEgress PTI 6 Remove
When this bit is set and the CRRP-VC RM cell removal point is set, then an egress cell whose PTI = 6 is removed, provided
that the connection is a VC connection.
EP7RÑEgress PTI 7 Remove
When this bit is set and the RRP-RM cell removal point is set, then an egress cell whose PTI = 7 is removed, provided that the
connection is a VC connection.
EEOTÑEgress End-to-End OAM Termination
When this bit is set, the egress ßow is treated as the terminating point of the OAM end-to-end cell ßow for the connection.
Additionally, if the CRRPÑVC RM Cell Removal Point bit is reset, then cells with PTI = 6 or 7 are removed at this point.
IP6RÑIngress PTI 6 Remove
When this bit is set and the RRP-RM Cell Removal Point is set, then an ingress cell whose PTI = 6 is removed, provided that
the connection is a VC connection.
IP7RÑIngress PTI 7 Remove
When this bit is set and the RRP-RM Cell Removal Point is set, then an ingress cell whose PTI = 7 is removed, provided that
the connection is a VC connection.
IEOTÑIngress End-to-End OAM Termination
When this bit is set, the ingress ßow is treated as the terminating point of the OAM end-to-end cell ßow for the connection.
Additionally, if the CRRPÑVC RM Cell Removal Point bit is reset, then cells which belong to a VC connection and whose PTI =
6 or 7 are removed at this point.
MC92501
36
MOTOROLA
12.2.3.Common Parameters
The size and the location of some of the Þelds is changed according to PMAC-PM on all connections.
When VPAP-VP on all connections is reset, the structure of the common parameters is the structure of the MC92500.
31302928272625242322212019181716
IOPVEOPVOAM_ptr[5:0]NBKBKT_PTR(21:16)
1514131211109876543210
BKT_ptr(15:0)
When PMAC-PM on all connections is set, the structure is as follows:
31302928272625242322212019181716
IOPVEOPVNBKBKT_PTR(11:00)
1514131211109876543210
OAM_Ptr(15:0)
MOTOROLA
MC92501
37
SECTION 13. DATA STRUCTURES
This section presents the data structures which where added or updated.
13.1.General Fields
13.1.1.Reason
The deÞnition of Reason 01001 is changed: ÒA CRC error was detectedÓ instead of ÒA CRC error was detected (OAM cells only)Ó.
MC92501
38
MOTOROLA
SECTION 14. SIGNAL DESCRIPTION
The following are the pins which have been added or whose
deÞnition has been changed.
14.1.Microproccessor Signals (MP)
The following signal deÞnitions have been updated.
The MC92500 MDT
Acknowledge0 (MDT
Acknowledge0 (MDT
MP Data Acknowledge0 (MDT
MP Data Acknowledge1 (MDT
MDT
ACK0 and MDTACK1 are three-state output signals
used to indicate when the data on MDATA is valid during a
read access from the MC92501. At the end of each access,
these signals are actively pulled up and then released. The
user may program the MC92501 not to drive these signals
during certain types of accesses. See Section 11.4.6 for
details. These signals are active low and the outputs are
asynchronous to the MCLK.
MP Cell Request Options
MREQ0
MCOREQ
signals are programmable to one of the following options:
1. MP Cell In Request
MREQ[n]
DMA device as a control line indicating when to start a new
cell insertion cycle into the MC92501. It is asserted whenever
the cell insertion register array is available to be written. This
signal is active low, and the output is on the falling edge of
MCLK.
2. MP Cell Out Request
MREQ[n]
DMA device as a control line indicating when to start a new
cell extraction cycle from the MC92501. It is asserted
whenever the cell extraction register array is available to be
read. The microprocessor control register (MPCTLR) contains
the number of maintenance accesses performed in a single
maintenance slot. It is active low, and the output is on the falling
edge of MCLK.
3. External Memory Maintenance Request
MREQ[n]
programmable number of clock cycles before the start of an
external memory maintenance cycle (see Section 11.4.7). It
is negated after a programmable number of maintenance
accesses have been performed. It is active low, and the output
is on the falling edge of MCLK.
MP Request 0 (MREQ0
This output signal can be programmed to one of the above
three options. Its default value is the Þrst option: MP Cell In
Request (MCIREQ
See Section 8.1.2 and Section 11.4.6 for details.
, MREQ1, and MREQ2 signals replace MCIREQ,
, and EMMREQ, respectively. Each of the MREQ[n]
is an output signal that can be used by an external
is an output signal that may be used by an external
is an output signal that can be asserted a
ACK signal is renamed to MP Data
ACK0) and the MP Data
ACK1) signal has been added.
ACK0),
ACK1)
)
).
MP Request 1 (MREQ1
This output signal can be programmed to one of the above
three options. Its default value is the second option: MP Cell
Out Request (MCOREQ
See Section 8.1.2 and Section 11.4.6 for details.
MP Request 2 (MREQ2
This output signal can be programmed to one of the above
three options. Its default value is the third option: External
Memory Maintenance Request (EMMREQ
See Section 8.1.2 and Section 11.4.6 for details.
The default values of MREQ0, MREQ1, and
MREQ2
EMMREQ
make the MC92501 backwards-compatible with
the MC92500.
MP Word Write Enable High / Address 1 (MWSH
This input signal can be programmed by the WSSMÑWord
Select Signals Mode bit to one of the following modes:
1. Write-Enable Mode: This signal indicates that the high
word is being written. During a maintenance write access, the
value detected on MWSH
EMBSH
always asserted. This signal is active low.
2. Add1-Size Mode: This signal serves as address 1 during
a maintenance write access. During a read access, this signal
is ignored.
This signal is sampled by the MC92501 on the falling edge
of MCLK.
See Section 8.1.3 and Section 11.4.6 for details.
MP Word Write Enable Low / SIZE (MWSL
This input signal can be programmed by the WSSMÑWord
Select Signals Mode bit to one of the following modes:
1. Write-Enable Mode: This signal indicates that the low
word is being written. During a maintenance write access, the
value detected on MWSL
EMBSL
always asserted. This signal is active low.
2. Add1-Size Mode: This signal indicates the size of the
maintenance write access which is either 32-bit or 16-bit
access. During a read access, this signal is ignored and the
access width is 32 bits.
This signal is sampled by the MC92501 on the falling edge
of MCLK.
See Section 8.1.3 and Section 11.4.6 for details.
All cell extraction register, cell insertion register,
and general register accesses are long-word (32bit) accesses, so both MWSH
SIZE should be asserted low for these write
accesses when write-enable mode is selected.
signals are MCIREQ, MCOREQ, and
, respectively. These default values
signal. During the read access, the EMBSH signal is
signal. During the read access, the EMBSL signal is
)
).
)
).
NOTE
/A1)
/A1 is driven on the appropriate
/SIZE)
/SIZE is driven on the appropriate
NOTE
/A1 and MWSL/
MOTOROLA
MC92501
39
14.2.Ingress PHY Signals
The deÞnition of RXPHYID0-3 has been updated. The
RXADDR4 signal has been added.
Receive PHY ID 0-3/Receive Address 0-3
(RXPHYID0ÐRXPHYID3/RXADDR0ÐRXADDR3)
This bus has two modes depending on the IUMÑIngress
UTOPIA Mode bit of the Ingress PHY ConÞguration Register
(IPHCR):
¥ In UTOPIA Level 1 Ñ The RXPHYID0ÐRXPHYID3 input
bus indicates the ID number of the PHY device currently
transferring data to the MC92501. If only a single PHY device is supported, this bus should be tied low. This bus is
sampled along with the Þrst octet of each cell.
¥ In UTOPIA Level 2 Ñ The RXADD0ÐRXADDR3 output bus
that indicates the four least signiÞcant bits of the ID number
of the PHY device which is being polled or selected by the
MC92501. See Section 10 for details.
Receive Address 4 (RXADDR4)
This signal is an output signal that indicates the most
signiÞcant bit of the ID number of the PHY device which is
being polled or selected by the MC92501. See Section 10 for
details.
14.3.Egress PHY Signals
The TXPHYID0-3 deÞnition has been updated and renamed
to TXPH. TXPHYID
signal of the MC92500.
V/TXADDR4 signal replaces TXPHYIDV
Transmit PHY ID 0-3 / Transmit Address 0-3
(TXPHYID0ÐTXPHYID3/TXADDR0ÐTXADDR3)
This bus has two modes depending on the EUMÑEgress
UTOPIA Mode bit:
¥ In UTOPIA Level 1 Ñ The TXPHYID0ÐTXPHYID3 output
bus indicates the ID number of the PHY device to which
either the current cell or the next cell is directed. The functionality is controlled by the MC92500 General ConÞguration Register (GCR).
¥ In UTOPIA Level 2 Ñ The TXADDR0ÐTXADDR3 output
bus indicates the four less signiÞcant bits of the ID number
of the PHY device which is being polled or selected by the
MC92501. See Section 10 for details.
Transmit Next PHY ID Valid/Transmit Address 4
(TXPHYID
This bit has two modes depending on the EUMÑEgress
UTOPIA Mode bit:
¥ In UTOPIA Level 1 Ñ The TXPHYID
low, indicates that TXPHYID
cellÕs ID) is valid. If TXPHYID
current cell, TXPHYID
¥ In UTOPIA Level 2 Ñ The TXADDR4 output signal indi-
cates the most signiÞcant bit of the ID number of the PHY
device which is being polled or selected by the MC92501.
See Section 10 for details.
V/TXADDR4)
V output signal, when
(when conÞgured as the next
is conÞgured to refer to the
V is not used.
MC92501
40
MOTOROLA
SECTION 15. TEST OPERATION
15.1.Device Identification Register
The code for the MC92501 is changed to: 0100_0001_1100_0011_1010_0000_0001_1101.
15.2.Boundary Scan Register
Table 6. Boundary Scan Bit DeÞnition
Signal NameI/O Cell TypeSystem ModeScan Bit #Output
STXCLKinin360
STXCLAVbidirout359358
STXSOCbidirin357356
STXPRTYbidirin355354
STXDATA7bidirin353352
STXDATA6bidirin351350
STXDATA5bidirin349348
STXDATA4bidirin347346
STXDATA3bidirin345344
Enable
STXDATA2bidirin343342
STXDATA1bidirin341340
STXDATA0bidirin339338
STXENBbidirin337336
TXENBbidirout335334
TXFULLinin333
TXCCLRbidirin332331
TXPHYIDVthree-stateout330
TXPRTYbidirout329328
TXSOCbidirout327326
TXDATA7bidirout325324
TXDATA6bidirout323322
TXDATA5bidirout321320
TXDATA4bidirout319318
TXDATA3bidirout317316
TXDATA2bidirout315314
TXDATA1bidirout313312
MOTOROLA
TXDATA0bidirout311310
TXPHYID3three-stateout309
TXPHYID2three-stateout308
TXPHYID1three-stateout307
MC92501
41
Table 6. Boundary Scan Bit DeÞnition
Signal NameI/O Cell TypeSystem ModeScan Bit #Output
Enable
TXPHYID0three-stateout306
MDATA31bidirbidir305304enscan1
MDATA30bidirbidir303302enscan1
MDATA29bidirbidir301300enscan1
MDATA28bidirbidir299298enscan1
MDATA27bidirbidir297296enscan1
MDATA26bidirbidir295294enscan1
MDATA25bidirbidir293292enscan1
MDATA24bidirbidir291290enscan1
MDATA23bidirbidir289288enscan1
MDATA22bidirbidir287286enscan1
MDATA21bidirbidir285284enscan1
MDATA20bidirbidir283282enscan1
MDATA19bidirbidir281280enscan1
MDATA18bidirbidir279278enscan1
MDATA17bidirbidir277276enscan1
MDATA16bidirbidir275274enscan1
MDATA15bidirbidir273272enscan1
MDATA14bidirbidir271270enscan1
MDATA13bidirbidir269268enscan1
MDATA12bidirbidir267266enscan1
MDATA11bidirbidir265264enscan1
MDATA10bidirbidir263262enscan1
MDATA9bidirbidir261260enscan1
MDATA8bidirbidir259258enscan1
MDATA7bidirbidir257256enscan1
MDATA6bidirbidir255254enscan1
MDATA5bidirbidir253252enscan1
MDATA4bidirbidir251250enscan1
MDATA3bidirbidir249248enscan1
MDATA2bidirbidir247246enscan1
MC92501
42
MDATA1bidirbidir245244enscan1
MDATA0bidirbidir243242enscan1
MADD25bidirin241240
MOTOROLA
Table 6. Boundary Scan Bit DeÞnition
Signal NameI/O Cell TypeSystem ModeScan Bit #Output
Enable
MADD24bidirin239238
MADD23bidirin237236
MADD22bidirin235234
MADD21bidirin233232
MADD20bidirin231230
MADD19bidirin229228
MADD18bidirin227226
MADD17bidirin225224
MADD16bidirin223222
MADD15bidirin221220
MADD14bidirin219218
MADD13bidirin217216
MADD12bidirin215214
MADD11bidirin213212
MADD10bidirin211210
MADD9bidirin209208
MADD8bidirin207206
MADD7bidirin205204
MADD6bidirin203202
MADD5bidirin201200
MADD4bidirin199198
MADD3bidirin197196
MADD2bidirin195194
MSELinin193
MREQ0three-stateout192
MREQ1three-stateout191
MDTACK0three-statethree-state190 enscan2
MINTthree-stateout189
MREQ2three-stateout188
MCLKinin187
MWRinin186
MOTOROLA
MWSHinin185
MWSLinin184
MDSinin183
MC92501
43
Table 6. Boundary Scan Bit DeÞnition
Signal NameI/O Cell TypeSystem ModeScan Bit #Output
Enable
SRXENBinin182
SRXDATA7bidirthree-state181180enscan4
SRXDATA6bidirthree-state179178enscan4
SRXDATA5bidirthree-state177176enscan4
SRXDATA4bidirthree-state175174enscan4
SRXDATA3bidirthree-state173172enscan4
SRXDATA2bidirthree-state171170enscan4
SRXDATA1bidirthree-state169168enscan4
SRXDATA0bidirthree-state167166enscan4
SRXCLKinin165 enscan4
SRXCLAVbidirout164163
SRXSOCbidirthree-state162161enscan4
SRXPRTYbidirthree-state160159enscan4
MDTACK1three-statethree-state158 enscan6
RXADDR4three-statethree-state157 enscan3
RXSOCbidirin156155
RXENBbidirout154153
RXEMPTYbidirin152151
RXPHYID3bidirbidir150 149enscan3
RXPHYID2bidirbidir148 147enscan3
RXPHYID1bidirbidir146 145enscan3
RXPHYID0bidirbidir144 143enscan3
RXPRTYbidirin142141
RXDATA7bidirin140 139
RXDATA6bidirin138 137
RXDATA5bidirin136 135
RXDATA4bidirin134 133
RXDATA3bidirin132 131
RXDATA2bidirin130 129
RXDATA1bidirin128 127
RXDATA0bidirin126 125
MC92501
44
EMDATA31bidirbidir124123enscan5
EMDATA30bidirbidir122121enscan5
EMDATA29bidirbidir120119enscan5
MOTOROLA
Table 6. Boundary Scan Bit DeÞnition
Signal NameI/O Cell TypeSystem ModeScan Bit #Output
Enable
EMDATA28bidirbidir118117enscan5
EMDATA27bidirbidir116115enscan5
EMDATA26bidirbidir114113enscan5
EMDATA25bidirbidir112111enscan5
EMDATA24bidirbidir110109enscan5
EMDATA23bidirbidir108107enscan5
EMDATA22bidirbidir106105enscan5
EMDATA21bidirbidir104103enscan5
EMDATA20bidirbidir102101enscan5
EMDATA19bidirbidir10099enscan5
EACENthree-stateout98
EMWRthree-stateout97
EMDATA18bidirbidir9695enscan5
EMDATA17bidirbidir9493enscan5
EMDATA16bidirbidir9291enscan5
EMDATA15bidirbidir9089enscan5
EMDATA14bidirbidir8887enscan5
EMDATA13bidirbidir8685enscan5
EMDATA12bidirbidir8483enscan5
EMDATA11bidirbidir8281enscan5
EMDATA10bidirbidir8079enscan5
EMDATA9bidirbidir7877enscan5
EMDATA8bidirbidir7675enscan5
EMDATA7bidirbidir7473enscan5
EMDATA6bidirbidir7271enscan5
EMDATA5bidirbidir7069enscan5
EMDATA4bidirbidir6867enscan5
EMDATA3bidirbidir6665enscan5
EMDATA2bidirbidir6463enscan5
EMDATA1bidirbidir6261enscan5
EMDATA0bidirbidir6059enscan5
MOTOROLA
EMADD23bidirout5857
EMADD22bidirout5655
EMADD21bidirout5453
MC92501
45
Table 6. Boundary Scan Bit DeÞnition
Signal NameI/O Cell TypeSystem ModeScan Bit #Output
Enable
EMADD20bidirout5251
EMADD19bidirout5049
EMADD18bidirout4847
EMADD17bidirout4645
EMADD16bidirout4443
EMADD15bidirout4241
EMADD14bidirout4039
EMADD13bidirout3837
EMADD12bidirout3635
EMADD11bidirout3433
EMADD10bidirout3231
EMADD9bidirout3029
EMADD8bidirout2827
EMADD7bidirout2625
EMADD6bidirout2423
EMADD5bidirout2221
EMADD4bidirout2019
EMADD3bidirout1817
EMADD2bidirout1615
EMBSH0three-stateout14
EMBSH1three-stateout13
EMBSH2three-stateout12
EMBSH3three-stateout11
EMBSL0three-stateout10
EMBSL1three-stateout9
EMBSL2three-stateout8
EMBSL3three-stateout7
ARSTinin6
enscan1(core macro)5
enscan2(core macro)4
enscan3(core macro)3
MC92501
46
enscan4(core macro)2
enscan5(core macro)1
enscan6(core macro)0
MOTOROLA
SECTION 16. ELECTRICAL CHARACTERISTICS
This device contains protection circuitry to
guard against damage due to high static voltages
or electric Þelds. However, precautions must be
taken to avoid applications of any voltage higher
than maximum rated voltages to this high-impedance circuit. For proper operation it is recommended that Vin and V
out
be constrained to 0 £
(Vin, V
out
) £ 5.5 V.
Unused outputs must always be tied to an appropriate logic voltage level (e.g., either VSS or
VDD). Unused outputs must be left open.
16.1.Electrical Specification for Clocks and Interfaces
Electrical speciÞcations for the clocks, microprocessor interface timing, PHY interface timing, switch interface timing, and
external memory interface timing are identical to the MC92500. Please refer to document MC92500/D for speciÞc values.
16.2.DC Electrical Characteristics
ABSOLUTE MAXIMUM RATINGS (See Note)
SymbolParameterValue/Value RangeUnit
V
V
V
out
T
NOTE: Maximum ratings are those values beyond which damage to the device may occur.
Motorola reserves the right to make changes without further notice to any products herein. Motorola makes no warranty, representation or guarantee regarding
the suitability of its products for any particular purpose, nor does Motorola assume any liability arising out of the application or use of any product or circuit, and
speciÞcally disclaims any and all liability, including without limitation consequential or incidental damages. ÒTypicalÓ parameters can and do vary in different
applications and actual performance may vary over time. All operating parameters, including ÒTypicalsÓ, must be validated for each customer application by
customerÕs technical experts. Motorola does not convey any license under its patent rights nor the rights of others. Motorola products are not designed, intended,
or authorized for use as components in systems intended for surgical implant into the body, or other applications intended to suppor t or sustain life, or for any
other application in which the failure of the Motorola product could create a situation where personal injury or death may occur. Should Buyer purchase or use
Motorola products for any such unintended or unauthorized application, Buyer shall indemnify and hold Motorola and its ofÞcers, employees, subsidiaries,
afÞliates, and distributors harmless against all claims, costs, damages, and expenses, and reasonable attorney fees arising out of, directly or indirectly, any claim
of personal injury or death associated with such unintended or unauthorized use, even if such claim alleges that Motorola was negligent regarding the design or
manufacture of the part. Motorola and are registered trademarks of Motorola, Inc. Motorola, Inc. is an Equal Opportunity/AfÞrmative Action Employer.
How to reach us:
USA/EUROPE/Locations Not Listed: Motorola Literature Distribution;
P.O. Box 5405; Denver, Colorado 80217. 303-675-2140 or 1-800-441-2447