System Control
ARM DDI 0500D
Copyright © 2013-2014 ARM. All rights reserved.
4-25
ID021414
Non-Confidential
Table 4-29
shows the ID_MMFR2_EL1 bit assignments.
To access the ID_MMFR2_EL1:
MRS <Xt>, ID_MMFR2_EL1 ; Read ID_MMFR2_EL1 into Xt
Register access is encoded as follows:
Table 4-29 ID_MMFR2_EL1 bit assignments
Bits
Name
Function
[31:28]
HWAccFlg
Hardware access flag. Indicates support for a hardware access flag, as part of the VMSAv7 implementation:
0x0
Not supported.
[27:24]
WFIStall
Wait For Interrupt Stall. Indicates the support for
Wait For Interrupt
(WFI) stalling:
0x1
Support for WFI stalling.
[23:20]
MemBarr
Memory Barrier. Indicates the supported CP15 memory barrier operations.
0x2
Supported CP15 memory barrier operations are:
•
Data Synchronization Barrier
(DSB).
•
Instruction Synchronization Barrier
(ISB).
•
Data Memory Barrier
(DMB).
[19:16]
UniTLB
Unified TLB. Indicates the supported TLB maintenance operations, for a unified TLB implementation.
0x6
Supported unified TLB maintenance operations are:
•
Invalidate all entries in the TLB.
•
Invalidate TLB entry by MVA.
•
Invalidate TLB entries by ASID match.
•
Invalidate instruction TLB and data TLB entries by MVA All ASID. This is a shared
unified TLB operation.
•
Invalidate Hyp mode unified TLB entry by MVA.
•
Invalidate entire Non-secure EL1 and EL0 unified TLB.
•
Invalidate entire Hyp mode unified TLB.
•
TLBIMVALIS
,
TLBIMVAALIS
,
TLBIMVALHIS
,
TLBIMVAL
,
TLBIMVAAL
, and
TLBIMVALH
.
•
TLBIIPAS2IS
,
TLBIIPAS2LIS
,
TLBIIPAS2
, and
TLBIIPAS2L
.
[15:12]
HvdTLB
Harvard TLB. Indicates the supported TLB maintenance operations, for a Harvard TLB implementation:
0x0
Not supported.
[11:8]
LL1HvdRng
L1 Harvard cache Range. Indicates the supported L1 cache maintenance range operations, for a Harvard
cache implementation:
0x0
Not supported.
[7:4]
L1HvdBG
L1 Harvard cache Background fetch. Indicates the supported L1 cache background prefetch operations, for
a Harvard cache implementation:
0x0
Not supported.
[3:0]
L1HvdFG
L1 Harvard cache Foreground fetch. Indicates the supported L1 cache foreground prefetch operations, for
a Harvard cache implementation:
0x0
Not supported.
Table 4-30 ID_MMFR2_EL1 access encoding
op0
op1
CRn
CRm
op2
11
000
0000
0001
110