| Electronic Components Datasheet Search |
|
AVR32AP Datasheet(PDF) 27 Page - ATMEL Corporation |
|
|
|||||||||||||||||||||||||||||
AVR32AP Datasheet(HTML) 27 Page - ATMEL Corporation |
|
27 / 181 page ![]() 27 32001A–AVR32–06/06 AVR32 Figure 3-4. Forwarding destinations 3.9 Hazards not handled by the hardware All hazards occurring between normal arithmetical, logical, load-store and change-of-flow instructions are handled automatically by hardware. There are, however, a few instruction sequences which must be sequenced by the user. These sequences are described in this chap- ter. The programmer can assume that any instruction sequence other than the sequences explicitly mentioned in this chapter will work without any special consideration. 3.9.1 Accessing system registers with mtsr and mfsr The mtsr instruction writes the contents of a register into a system register. The system registers control the behaviour of the CPU. The programmer must make sure that any mtsr instruction has committed and has altered the state of the system in the desired way before issuing any new instructions that depend on this new state. This can be done by inserting nop instructions, or other instructions that do not depend on the new state generated by the mtsr instruction. Table 2-2, “System Registers implemented in AVR32 AP,” on page 10 details the timing for writes into the different system registers. The system registers are written as the mtsr instruction leaves the pipeline stage described in the table. The system registers are read as the mfsr instruction leaves the pipeline stage described in the table. As soon as a system register is read by mfsr, it can be forwarded as any regular register file register. Some of the system registers are located inside modules on the TCB bus. These are written when the mtsr instruction leaves the D pipeline stage. Instructions depending on a mtsr to these system registers being committed must therefore wait in the IS stage until the effects of the mtsr is guaranteed to be visible to the instruction. The following code demonstrates a write to the ASID field of TLBEHI, followed by a rete to an address which requires the new ASID to be visi- ble. A nop is inserted to guarantee that the mtsr leaves the D stage at the same time as rete leaves the A1 stage. In the following cycle, the icache will start fetching at the specified address A1 M1 M2 D WB A2 Integer unit Load-store unit IF Prefetch unit DA Reg File Issue Pointer To cache |
|
Link URL |
| Does ALLDATASHEET help your business so far? [ DONATE ] |
About Alldatasheet | Advertisement | Contact us | Privacy Policy | Link to Datasheet | Link Exchange | Manufacturer List All Rights Reserved©Alldatasheet.com |
| Russian : Alldatasheetru.com | Korean : Alldatasheet.co.kr | Spanish : Alldatasheet.es | French : Alldatasheet.fr | Italian : Alldatasheetit.com Portuguese : Alldatasheetpt.com | Polish : Alldatasheet.pl | Vietnamese : Alldatasheet.vn Indian : Alldatasheet.in | Mexican : Alldatasheet.com.mx | British : Alldatasheet.co.uk | New Zealand : Alldatasheet.co.nz |
|
Family Site : ic2ic.com |
icmetro.com |