
4–62 Chapter 4: Functional Description
Error Detection and Management
RapidIO MegaCore Function August 2014 Altera Corporation
User Guide
The RapidIO IP core Physical layer transparently manages these errors for you. The
RapidIO specification defines both input and output error detection and recovery
state machines that include handshaking protocols in which the receiving end signals
that an error is detected by sending a
packet-not-accepted
control symbol, the
transmitter then sends an
input-status link-request
control symbol to which the
receiver responds with a
link-response
control symbol to indicate which packet
requires transmission. The input and output error detection and recovery state
machines can be monitored by software that you create to read the status of the
Port
0 Error and Status
CSR (Table 6–10 on page 6–7).
In addition to the registers defined by the specification, the RapidIO IP core provides
several output signals that enable user logic to monitor error detection and the
recovery process. Refer to “Status Packet and Error Monitoring Signals” on page 5–2.
Protocol Violations
Some protocol violations, such as a packet with an unexpected
ackID
or a time-out on
a packet acknowledgment, can use the same error recovery mechanisms as the
transmission errors described in “Physical Layer Error Management” on page 4–61.
Some protocol violations, such as a time-out on a link-request or the RapidIO IP core
receiving a link-response with an
ackID
outside the range of transmitted
ackID
s, can
lead to unrecoverable—or fatal—errors.
Fatal Errors
Fatal errors cause a soft reset of the Physical layer module, which clears all the
transmit buffers and resets the transmission and expected
ackID
to zero. This effect
also can be triggered by software by first writing a one and then a zero to the
PORT_DIS
bit of the
Port 0 Control
CSR (Table 6–11 on page 6–10).
If the link partner is reset when its expected
ackID
is not zero, a fatal error occurs
when the link partner receives the next transmitted packet because the link partner’s
expected
ackID
is reset to zero, which causes a mismatch between the transmitted
ackID
and the expected
ackID
. The fatal error causes a soft reset of the IP core. After
the soft reset completes, transmitted and expected
ackID
s are synchronized and
normal operation resumes. Only the packets that were queued at the time of the fatal
error are lost.
If Send link-request reset-device on fatal errors is turned on in the RapidIO
parameter editor, fatal errors cause the transmitter to send
link-request
control
symbols with
cmd
set to
reset-device
to the link partner.
Logical Layer Error Management
The Logical layer modules only need to process Logical layer errors because errors
detected by the Physical layer are masked from the Logical layer module. Any packet
with an error detected in the Physical layer is dropped in the Physical layer or the
Transport layer before it reaches the Logical layer modules.
The RapidIO specification defines the following common errors and the protocols for
managing them:
■ Malformed request or response packets
■ Unexpected Transaction ID
Commentaires sur ces manuels