Accelerated Graphics Port Interface Specification

AGP3.0 Interface Specification
Rev. 1.0
108
4.1.11.1 Isoch Align Transaction
An Isoch Align transaction is similar to a Isoch Read transaction except that it has a special meaning to
the Core-Logic. If the Core-Logic supports time-keeping, it will return an offset value to allow the isoch
device to align its isochronous period to that of the core-logic. The value of this offset is returned in the
lower 8 bits of a 32 bit word. The remaining bits in the word should be ignored. The offset value ranges
from xxxxxx00 through xxxxxx41 (hexadecimal). If the core-logic does not support time-keeping it will
return a value of xxxxxxFF. All other values of the lower 8-bits are reserved.
The core-logic must return the data for an Isoch Align transaction in the first 32-bit word in a common
clock cycle. The remaining 7 words of that common clock cycle are undefined and must be ignored by
the isoch device. A core-logic may drive the remaining 7 words of this cycle with undefined data.
Regardless of the transaction size selected for normal isoch reads, the length field for isoch align will be
000, which selects 32Bytes.
Since Isoch Align is an isochronous transaction, all isoch protocol requirements must be obeyed. Isoch
Align transactions are not included as part of the ISOCH_N value. Isoch_N is defined in the NISTAT AGP
Isochronous Status Register.
4.2 Flow Control Change
Since isochronous and asynchronous transactions may be concurrently processed on the AGP3.0
interface, it is necessary to ensure that a throttled (stalled) asynchronous transaction does not cause
the isochronous latency specification to be violated. Based on the flow-control discussion in Section
2.3.3, the following points emerge:
• Block-level flow control will stall the interface. Buffer-Full flow control only stalls a new asynchronous
data transfer from starting.
• AGP3.0 “Reads and Writes” cannot be flow-controlled at the block-level since the maximum
transaction size is 64 bytes.
• AGP3.0 Fast Writes can be flow-controlled at the block-level since there is no limit to the size of the
transaction.
Two modifications made to the asynchronous transaction flow control are required to ensure proper
isochronous operation. These modifications are described in the following sections.
4.2.1 Isochronous Transactions and Buffer-Full Flow Control
Data transfers for isochronous read or write transaction will ignore buffer-full flow control applied using
RBF and WBF. The core-logic that initiates these transfers may pass a stalled asynchronous data
transfer with an isochronous data transfer. Simply put, the core-logic ignores RBF and WBF when it
initiates an isochronous data transfer.
4.2.2 AGP3.0 Fast Write Flow Control
For AGP “Fast Writes”, the maximum number of wait states introduced during each block-level stall was
not specified. In AGP3.0, the maximum number of Wait states is limited to two common clock cycles.
In other words, the AGP3.0 Master must not suppress TRDY for more than two cycles including the
throttle point cycle. If the AGP3.0 Master needs additional time to receive the Fast Write data, it must
use the Fast Write termination schemes that are defined in the AGP interface specification. The AGP3.0