Sami Mujawar 2f16993c25 ArmPkg: ARM v8.2 updates for detecting FP
The ARMv8.2-FP16 extension introduces support for half precision
floating point and the processor ID registers have been updated to
enable detection of the implementation.

The possible values for the FP bits in ID_AA64PFR0_EL1[19:16] are:
  - 0000 : Floating-point is implemented.
  - 0001 : Floating-point including Half-precision support is
           implemented.
  - 1111 : Floating-point is not implemented.
  - All other values are reserved.

Previously ArmEnableVFP() compared the FP bits with 0000b to see if
the FP was implemented, before enabling FP. Modified this check to
enable the FP if the FP bits 19:16 are not 1111b.

Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Sami Mujawar <sami.mujawar@arm.com>
Signed-off-by: Evan Lloyd <evan.lloyd@arm.com>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
2017-09-15 16:56:08 +01:00

496 lines
15 KiB
ArmAsm

#------------------------------------------------------------------------------
#
# Copyright (c) 2008 - 2010, Apple Inc. All rights reserved.<BR>
# Copyright (c) 2011 - 2017, ARM Limited. All rights reserved.
# Copyright (c) 2016, Linaro Limited. All rights reserved.
#
# This program and the accompanying materials
# are licensed and made available under the terms and conditions of the BSD License
# which accompanies this distribution. The full text of the license may be found at
# http://opensource.org/licenses/bsd-license.php
#
# THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS,
# WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR IMPLIED.
#
#------------------------------------------------------------------------------
#include <Chipset/AArch64.h>
#include <AsmMacroIoLibV8.h>
.set CTRL_M_BIT, (1 << 0)
.set CTRL_A_BIT, (1 << 1)
.set CTRL_C_BIT, (1 << 2)
.set CTRL_SA_BIT, (1 << 3)
.set CTRL_I_BIT, (1 << 12)
.set CTRL_V_BIT, (1 << 12)
.set CPACR_VFP_BITS, (3 << 20)
ASM_FUNC(ArmInvalidateDataCacheEntryByMVA)
dc ivac, x0 // Invalidate single data cache line
ret
ASM_FUNC(ArmCleanDataCacheEntryByMVA)
dc cvac, x0 // Clean single data cache line
ret
ASM_FUNC(ArmCleanDataCacheEntryToPoUByMVA)
dc cvau, x0 // Clean single data cache line to PoU
ret
ASM_FUNC(ArmInvalidateInstructionCacheEntryToPoUByMVA)
ic ivau, x0 // Invalidate single instruction cache line to PoU
ret
ASM_FUNC(ArmCleanInvalidateDataCacheEntryByMVA)
dc civac, x0 // Clean and invalidate single data cache line
ret
ASM_FUNC(ArmInvalidateDataCacheEntryBySetWay)
dc isw, x0 // Invalidate this line
ret
ASM_FUNC(ArmCleanInvalidateDataCacheEntryBySetWay)
dc cisw, x0 // Clean and Invalidate this line
ret
ASM_FUNC(ArmCleanDataCacheEntryBySetWay)
dc csw, x0 // Clean this line
ret
ASM_FUNC(ArmInvalidateInstructionCache)
ic iallu // Invalidate entire instruction cache
dsb sy
isb
ret
ASM_FUNC(ArmEnableMmu)
EL1_OR_EL2_OR_EL3(x1)
1: mrs x0, sctlr_el1 // Read System control register EL1
b 4f
2: mrs x0, sctlr_el2 // Read System control register EL2
b 4f
3: mrs x0, sctlr_el3 // Read System control register EL3
4: orr x0, x0, #CTRL_M_BIT // Set MMU enable bit
EL1_OR_EL2_OR_EL3(x1)
1: tlbi vmalle1
dsb nsh
isb
msr sctlr_el1, x0 // Write back
b 4f
2: tlbi alle2
dsb nsh
isb
msr sctlr_el2, x0 // Write back
b 4f
3: tlbi alle3
dsb nsh
isb
msr sctlr_el3, x0 // Write back
4: isb
ret
ASM_FUNC(ArmDisableMmu)
EL1_OR_EL2_OR_EL3(x1)
1: mrs x0, sctlr_el1 // Read System Control Register EL1
b 4f
2: mrs x0, sctlr_el2 // Read System Control Register EL2
b 4f
3: mrs x0, sctlr_el3 // Read System Control Register EL3
4: and x0, x0, #~CTRL_M_BIT // Clear MMU enable bit
EL1_OR_EL2_OR_EL3(x1)
1: msr sctlr_el1, x0 // Write back
tlbi vmalle1
b 4f
2: msr sctlr_el2, x0 // Write back
tlbi alle2
b 4f
3: msr sctlr_el3, x0 // Write back
tlbi alle3
4: dsb sy
isb
ret
ASM_FUNC(ArmDisableCachesAndMmu)
EL1_OR_EL2_OR_EL3(x1)
1: mrs x0, sctlr_el1 // Get control register EL1
b 4f
2: mrs x0, sctlr_el2 // Get control register EL2
b 4f
3: mrs x0, sctlr_el3 // Get control register EL3
4: mov x1, #~(CTRL_M_BIT | CTRL_C_BIT | CTRL_I_BIT) // Disable MMU, D & I caches
and x0, x0, x1
EL1_OR_EL2_OR_EL3(x1)
1: msr sctlr_el1, x0 // Write back control register
b 4f
2: msr sctlr_el2, x0 // Write back control register
b 4f
3: msr sctlr_el3, x0 // Write back control register
4: dsb sy
isb
ret
ASM_FUNC(ArmMmuEnabled)
EL1_OR_EL2_OR_EL3(x1)
1: mrs x0, sctlr_el1 // Get control register EL1
b 4f
2: mrs x0, sctlr_el2 // Get control register EL2
b 4f
3: mrs x0, sctlr_el3 // Get control register EL3
4: and x0, x0, #CTRL_M_BIT
ret
ASM_FUNC(ArmEnableDataCache)
EL1_OR_EL2_OR_EL3(x1)
1: mrs x0, sctlr_el1 // Get control register EL1
b 4f
2: mrs x0, sctlr_el2 // Get control register EL2
b 4f
3: mrs x0, sctlr_el3 // Get control register EL3
4: orr x0, x0, #CTRL_C_BIT // Set C bit
EL1_OR_EL2_OR_EL3(x1)
1: msr sctlr_el1, x0 // Write back control register
b 4f
2: msr sctlr_el2, x0 // Write back control register
b 4f
3: msr sctlr_el3, x0 // Write back control register
4: dsb sy
isb
ret
ASM_FUNC(ArmDisableDataCache)
EL1_OR_EL2_OR_EL3(x1)
1: mrs x0, sctlr_el1 // Get control register EL1
b 4f
2: mrs x0, sctlr_el2 // Get control register EL2
b 4f
3: mrs x0, sctlr_el3 // Get control register EL3
4: and x0, x0, #~CTRL_C_BIT // Clear C bit
EL1_OR_EL2_OR_EL3(x1)
1: msr sctlr_el1, x0 // Write back control register
b 4f
2: msr sctlr_el2, x0 // Write back control register
b 4f
3: msr sctlr_el3, x0 // Write back control register
4: dsb sy
isb
ret
ASM_FUNC(ArmEnableInstructionCache)
EL1_OR_EL2_OR_EL3(x1)
1: mrs x0, sctlr_el1 // Get control register EL1
b 4f
2: mrs x0, sctlr_el2 // Get control register EL2
b 4f
3: mrs x0, sctlr_el3 // Get control register EL3
4: orr x0, x0, #CTRL_I_BIT // Set I bit
EL1_OR_EL2_OR_EL3(x1)
1: msr sctlr_el1, x0 // Write back control register
b 4f
2: msr sctlr_el2, x0 // Write back control register
b 4f
3: msr sctlr_el3, x0 // Write back control register
4: dsb sy
isb
ret
ASM_FUNC(ArmDisableInstructionCache)
EL1_OR_EL2_OR_EL3(x1)
1: mrs x0, sctlr_el1 // Get control register EL1
b 4f
2: mrs x0, sctlr_el2 // Get control register EL2
b 4f
3: mrs x0, sctlr_el3 // Get control register EL3
4: and x0, x0, #~CTRL_I_BIT // Clear I bit
EL1_OR_EL2_OR_EL3(x1)
1: msr sctlr_el1, x0 // Write back control register
b 4f
2: msr sctlr_el2, x0 // Write back control register
b 4f
3: msr sctlr_el3, x0 // Write back control register
4: dsb sy
isb
ret
ASM_FUNC(ArmEnableAlignmentCheck)
EL1_OR_EL2(x1)
1: mrs x0, sctlr_el1 // Get control register EL1
b 3f
2: mrs x0, sctlr_el2 // Get control register EL2
3: orr x0, x0, #CTRL_A_BIT // Set A (alignment check) bit
EL1_OR_EL2(x1)
1: msr sctlr_el1, x0 // Write back control register
b 3f
2: msr sctlr_el2, x0 // Write back control register
3: dsb sy
isb
ret
ASM_FUNC(ArmDisableAlignmentCheck)
EL1_OR_EL2_OR_EL3(x1)
1: mrs x0, sctlr_el1 // Get control register EL1
b 4f
2: mrs x0, sctlr_el2 // Get control register EL2
b 4f
3: mrs x0, sctlr_el3 // Get control register EL3
4: and x0, x0, #~CTRL_A_BIT // Clear A (alignment check) bit
EL1_OR_EL2_OR_EL3(x1)
1: msr sctlr_el1, x0 // Write back control register
b 4f
2: msr sctlr_el2, x0 // Write back control register
b 4f
3: msr sctlr_el3, x0 // Write back control register
4: dsb sy
isb
ret
ASM_FUNC(ArmEnableStackAlignmentCheck)
EL1_OR_EL2(x1)
1: mrs x0, sctlr_el1 // Get control register EL1
b 3f
2: mrs x0, sctlr_el2 // Get control register EL2
3: orr x0, x0, #CTRL_SA_BIT // Set SA (stack alignment check) bit
EL1_OR_EL2(x1)
1: msr sctlr_el1, x0 // Write back control register
b 3f
2: msr sctlr_el2, x0 // Write back control register
3: dsb sy
isb
ret
ASM_FUNC(ArmDisableStackAlignmentCheck)
EL1_OR_EL2_OR_EL3(x1)
1: mrs x0, sctlr_el1 // Get control register EL1
b 4f
2: mrs x0, sctlr_el2 // Get control register EL2
b 4f
3: mrs x0, sctlr_el3 // Get control register EL3
4: bic x0, x0, #CTRL_SA_BIT // Clear SA (stack alignment check) bit
EL1_OR_EL2_OR_EL3(x1)
1: msr sctlr_el1, x0 // Write back control register
b 4f
2: msr sctlr_el2, x0 // Write back control register
b 4f
3: msr sctlr_el3, x0 // Write back control register
4: dsb sy
isb
ret
// Always turned on in AArch64. Else implementation specific. Leave in for C compatibility for now
ASM_FUNC(ArmEnableBranchPrediction)
ret
// Always turned on in AArch64. Else implementation specific. Leave in for C compatibility for now.
ASM_FUNC(ArmDisableBranchPrediction)
ret
ASM_FUNC(AArch64AllDataCachesOperation)
// We can use regs 0-7 and 9-15 without having to save/restore.
// Save our link register on the stack. - The stack must always be quad-word aligned
stp x29, x30, [sp, #-16]!
mov x29, sp
mov x1, x0 // Save Function call in x1
mrs x6, clidr_el1 // Read EL1 CLIDR
and x3, x6, #0x7000000 // Mask out all but Level of Coherency (LoC)
lsr x3, x3, #23 // Left align cache level value - the level is shifted by 1 to the
// right to ease the access to CSSELR and the Set/Way operation.
cbz x3, L_Finished // No need to clean if LoC is 0
mov x10, #0 // Start clean at cache level 0
Loop1:
add x2, x10, x10, lsr #1 // Work out 3x cachelevel for cache info
lsr x12, x6, x2 // bottom 3 bits are the Cache type for this level
and x12, x12, #7 // get those 3 bits alone
cmp x12, #2 // what cache at this level?
b.lt L_Skip // no cache or only instruction cache at this level
msr csselr_el1, x10 // write the Cache Size selection register with current level (CSSELR)
isb // isb to sync the change to the CacheSizeID reg
mrs x12, ccsidr_el1 // reads current Cache Size ID register (CCSIDR)
and x2, x12, #0x7 // extract the line length field
add x2, x2, #4 // add 4 for the line length offset (log2 16 bytes)
mov x4, #0x400
sub x4, x4, #1
and x4, x4, x12, lsr #3 // x4 is the max number on the way size (right aligned)
clz w5, w4 // w5 is the bit position of the way size increment
mov x7, #0x00008000
sub x7, x7, #1
and x7, x7, x12, lsr #13 // x7 is the max number of the index size (right aligned)
Loop2:
mov x9, x4 // x9 working copy of the max way size (right aligned)
Loop3:
lsl x11, x9, x5
orr x0, x10, x11 // factor in the way number and cache number
lsl x11, x7, x2
orr x0, x0, x11 // factor in the index number
blr x1 // Goto requested cache operation
subs x9, x9, #1 // decrement the way number
b.ge Loop3
subs x7, x7, #1 // decrement the index
b.ge Loop2
L_Skip:
add x10, x10, #2 // increment the cache number
cmp x3, x10
b.gt Loop1
L_Finished:
dsb sy
isb
ldp x29, x30, [sp], #0x10
ret
ASM_FUNC(ArmDataMemoryBarrier)
dmb sy
ret
ASM_FUNC(ArmDataSynchronizationBarrier)
dsb sy
ret
ASM_FUNC(ArmInstructionSynchronizationBarrier)
isb
ret
ASM_FUNC(ArmWriteVBar)
EL1_OR_EL2_OR_EL3(x1)
1: msr vbar_el1, x0 // Set the Address of the EL1 Vector Table in the VBAR register
b 4f
2: msr vbar_el2, x0 // Set the Address of the EL2 Vector Table in the VBAR register
b 4f
3: msr vbar_el3, x0 // Set the Address of the EL3 Vector Table in the VBAR register
4: isb
ret
ASM_FUNC(ArmReadVBar)
EL1_OR_EL2_OR_EL3(x1)
1: mrs x0, vbar_el1 // Set the Address of the EL1 Vector Table in the VBAR register
ret
2: mrs x0, vbar_el2 // Set the Address of the EL2 Vector Table in the VBAR register
ret
3: mrs x0, vbar_el3 // Set the Address of the EL3 Vector Table in the VBAR register
ret
ASM_FUNC(ArmEnableVFP)
// Check whether floating-point is implemented in the processor.
mov x1, x30 // Save LR
bl ArmReadIdPfr0 // Read EL1 Processor Feature Register (PFR0)
mov x30, x1 // Restore LR
ubfx x0, x0, #16, #4 // Extract the FP bits 16:19
cmp x0, #0xF // Check if FP bits are '1111b',
// i.e. Floating Point not implemented
b.eq 4f // Exit when VFP is not implemented.
// FVP is implemented.
// Make sure VFP exceptions are not trapped (to any exception level).
mrs x0, cpacr_el1 // Read EL1 Coprocessor Access Control Register (CPACR)
orr x0, x0, #CPACR_VFP_BITS // Disable FVP traps to EL1
msr cpacr_el1, x0 // Write back EL1 Coprocessor Access Control Register (CPACR)
mov x1, #AARCH64_CPTR_TFP // TFP Bit for trapping VFP Exceptions
EL1_OR_EL2_OR_EL3(x2)
1:ret // Not configurable in EL1
2:mrs x0, cptr_el2 // Disable VFP traps to EL2
bic x0, x0, x1
msr cptr_el2, x0
ret
3:mrs x0, cptr_el3 // Disable VFP traps to EL3
bic x0, x0, x1
msr cptr_el3, x0
4:ret
ASM_FUNC(ArmCallWFI)
wfi
ret
ASM_FUNC(ArmReadMpidr)
mrs x0, mpidr_el1 // read EL1 MPIDR
ret
// Keep old function names for C compatibilty for now. Change later?
ASM_FUNC(ArmReadTpidrurw)
mrs x0, tpidr_el0 // read tpidr_el0 (v7 TPIDRURW) -> (v8 TPIDR_EL0)
ret
// Keep old function names for C compatibilty for now. Change later?
ASM_FUNC(ArmWriteTpidrurw)
msr tpidr_el0, x0 // write tpidr_el0 (v7 TPIDRURW) -> (v8 TPIDR_EL0)
ret
// Arch timers are mandatory on AArch64
ASM_FUNC(ArmIsArchTimerImplemented)
mov x0, #1
ret
ASM_FUNC(ArmReadIdPfr0)
mrs x0, id_aa64pfr0_el1 // Read ID_AA64PFR0 Register
ret
// Q: id_aa64pfr1_el1 not defined yet. What does this funtion want to access?
// A: used to setup arch timer. Check if we have security extensions, permissions to set stuff.
// See: ArmPkg/Library/ArmArchTimerLib/AArch64/ArmArchTimerLib.c
// Not defined yet, but stick in here for now, should read all zeros.
ASM_FUNC(ArmReadIdPfr1)
mrs x0, id_aa64pfr1_el1 // Read ID_PFR1 Register
ret
// VOID ArmWriteHcr(UINTN Hcr)
ASM_FUNC(ArmWriteHcr)
msr hcr_el2, x0 // Write the passed HCR value
ret
// UINTN ArmReadHcr(VOID)
ASM_FUNC(ArmReadHcr)
mrs x0, hcr_el2
ret
// UINTN ArmReadCurrentEL(VOID)
ASM_FUNC(ArmReadCurrentEL)
mrs x0, CurrentEL
ret
// UINT32 ArmReadCntHctl(VOID)
ASM_FUNC(ArmReadCntHctl)
mrs x0, cnthctl_el2
ret
// VOID ArmWriteCntHctl(UINT32 CntHctl)
ASM_FUNC(ArmWriteCntHctl)
msr cnthctl_el2, x0
ret
ASM_FUNCTION_REMOVE_IF_UNREFERENCED