Motorola MPC860 PowerQUICC User Manual page 165

Part II. PowerPC Microprocessor Module
Load string and store string instructions may involve operands that are not word-aligned.
As described in ÒAlignment Exception (0x00600)Ó in Chapter 6, ÒExceptions,Ó in The
Programming Environments Manual, a misaligned string operation suffers a performance
penalty compared to a word-aligned operation of the same type.
When a string operation crosses a page boundary, the instruction may be interrupted by a
DSI exception associated with the address translation of the second page. In this case, the
MPC860 performs some or all memory references from the Þrst page and none from the
second before taking the exception. On return from the DSI exception, the load or store
string instruction will re-execute from the beginning. For more information, refer to ÒDSI
Exception (0x00300)Ó in Chapter 6, ÒExceptions,Ó in The Programming Environments
Manual.
6.2.4.3 Branch and Flow Control Instructions
Branch instructions are executed by the branch processing unit (BPU). The BPU receives
branch instructions from the fetch unit and performs condition register (CR) lookahead
operations on conditional branches to resolve them early, achieving the effect of a
zero-cycle branch in many cases.
Some branch instructions can redirect instruction execution conditionally based on the
value of bits in the CR. When the branch processor encounters one of these instructions, it
scans the execution pipelines to determine whether an instruction in progress may affect the
particular CR bit. If no interlock is found, the branch can be resolved immediately by
checking the bit in the CR and taking the action deÞned for the branch instruction.
If an interlock is detected, the branch is considered unresolved and the direction of the
branch is predicted using static branch prediction as described in ÒConditional Branch
ControlÓ in Chapter 4, ÒAddressing Modes and Instruction Set Summary,Ó in the
Programming Environments Manual. The interlock is monitored while instructions are
fetched for the predicted branch. When the interlock is cleared, the branch processor
determines whether the prediction was correct based on the value of the CR bit. If the
prediction is correct, the branch is considered completed and instruction fetching continues.
If the prediction is incorrect, the fetched instructions are purged, and instruction fetching
continues along the alternate path. See Chapter 8, ÒInstruction Timing,Ó for information
about how branches are executed.
6.2.4.3.1 Branch Instruction Address Calculation
Branch instructions can alter the sequence of instruction execution. Instruction addresses
are always assumed to be word aligned; the processor ignores the two low-order bits of the
generated branch target address.
MOTOROLA Chapter 6. MPC860 Instruction Set 6-15