AMD AMD-K6-2/500AFX Data Sheet - Page 255
L1 Cache Inhibit, Purpose
View all AMD AMD-K6-2/500AFX manuals
Add to My Manuals
Save this manual to your list of manuals |
Page 255 highlights
21850J/0-February 2000 Preliminary Information AMD-K6®-2 Processor Data Sheet 11.4 Purpose The following states have no effect on the normal or test operation of the processor other than as shown in Figure 86 on page 233: s Run-Test/Idle-This state is an idle state between scan operations. s Select-DR-Scan-This is the initial state of the test data register state transitions. s Select-IR-Scan-This is the initial state of the Instruction Register state transitions. s Exit1-DR-This state is entered to terminate the shifting process and enter the Update-DR state. s Exit1-IR-This state is entered to terminate the shifting process and enter the Update-IR state. s Pause-DR-This state is entered to temporarily stop the shifting process of a Test Data Register. s Pause-IR-This state is entered to temporarily stop the shifting process of the Instruction Register. s Exit2-DR-This state is entered in order to either terminate the shifting process and enter the Update-DR state or to resume shifting following the exit from the Pause-DR state. s Exit2-IR-This state is entered in order to either terminate the shifting process and enter the Update-IR state or to resume shifting following the exit from the Pause-IR state. L1 Cache Inhibit The AMD-K6-2 processor provides a means for inhibiting the normal operation of its L1 instruction and data caches while still supporting an external cache. This capability allows system designers to disable the L1 cache during the testing and debug of an external cache. If the Cache Inhibit bit (bit 3) of Test Register 12 (TR12) is set to 0, the processor's L1 cache is enabled and operates as described in "Cache Organization" on page 179. If the Cache Inhibit bit is set to 1, the L1 cache is disabled and no new cache lines are allocated. Even though new allocations do not occur, valid L1 cache lines remain valid and are read by the processor when a requested address hits a cache line. In addition, the processor continues to support inquire cycles initiated by the Chapter 11 Test and Debug 235