![Marvell PXA300 Скачать руководство пользователя страница 163](http://html.mh-extra.com/html/marvell/pxa300/pxa300_developers-manual_1734615163.webp)
69rlq62d-f714peg4 * Memec (Headquar
ter
s) - Unique
Tec
h,
Insight,
Impact
MAR
VELL CONFIDENTIAL,
UNDER ND
A# 12101050
69rlq62d-f714peg4 * Memec (Headquar
ter
s) - Unique
Tec
h,
Insight,
Impact
MAR
VELL CONFIDENTIAL,
A# 12101050
69r
lq62d-f714peg4 * Memec (Headquar
ters) - Unique
T
ech, Insight, Impact * UNDER ND
A# 12101050
MAR
VELL CONFIDENTIAL - UNA
UTHORIZED DISTRIB
UTION OR USE STRICTL
Y PR
OHIBITED
Slave Clock Control Unit
Copyright © 12/13/06 Marvell
CONFIDENTIAL
Doc. No. MV-TBD-00 Rev. A
December 13, 2006
Document Classification: Proprietary Information
Page 163
Not approved by Document Control. For review only.
If a status bit does not exist for that operation, then software must read back the control bit that was written. This
time interval is sufficient for the operation to have completed.
Finally, this read-back protocol should be followed for data bits. All data bits that are written should be read back
and verified before a control bit that uses that data bit is programmed.
7.3.1
Application Subsystem Clock Configuration Register
(ACCR)
The core clock is the base clock from which the core frequency is derived. The
Configuration Register (ACCR)”
, defined in
, controls this frequency.
Writing to various bits in the ACCR triggers actions in the clock control unit. For example, setting the SFLFS
field triggers a change of frequency for the internal SRAM controller clocks. These are not instantaneous
operations.
•
As a general rule, therefore, software must refrain from writing any other bits until the operations initiated
by the first write have completed. This completion can be detected by polling the corresponding status bits
in the ACSR status register. When the two match, the operations are deemed to have completed and another
operation can be initiated.
•
This rule does not prevent software from writing multiple fields simultaneously. It also does not apply to
writing values to purely data fields, such as the L and N fields. Purely data fields require additional operation
before they can be used (for example, a write to the coprocessor CLKCFG register).
•
Finally, software must refrain from initiating actions over any coprocessor interface while ACCR-initiated
operations are in progress. Thus, a core turbo change can not be initiated while a DMC frequency change is
in progress. Similarly, a low-power mode entry command can not be issued while the system PLL is being
disabled.
The ACCR includes the following bits:
•
Core PLL run-mode-to-oscillator ratio (XL)—Creates the nominal core PLL run-mode frequency by
multiplying the processor (13 MHz) oscillator by XL (core PLL run-mode frequency is 13 MHz * XL).
•
Core PLL turbo-mode-to-run-mode ratio (XN)—Creates the nominal core PLL turbo-mode frequency by
multiplying the run-mode frequency by XN (core PLL turbo-mode frequency is 13 MHz * XL * XN).
Note:
The value of XN must always be less than or equal to the value of the MTS field in the
“Application Subsystem Power Status/Configuration Register (ASCR)”
.
•
Power-mode change-clock enable (PCCE)—Enables the processor to use the ring oscillator output clock as
a clock source when transitioning from any low-power mode to D0 mode.
When PCCE = 1, the processor begins operation using the ring oscillator output while the PLLs are enabled
and locking. Once the core and system PLL outputs are locked and stable, the processor generates an
interrupt to the core if the PCIE bit in
“Application Subsystem Interrupt Control/Status Register (AICSR)”
is
set to 0b1. At this point, software has the option of switching the clocks over to the PLLs by performing a
D0CS exit operation. If the PCIE bit in
“Application Subsystem Interrupt Control/Status Register (AICSR)”
is set to a 0b0, no interrupt is generated and software must poll the status of the PLLs and their lock bits to
determine when to switch over to PLL-based operation.
When PCCE = 0, the processor operation does not begin until the core and system PLLs are enabled and
locked. Once the PLL outputs are locked and stable, the processor clocks are started with the PLL outputs.
In this case, no software intervention is necessary (or possible, given the absence of clocks).