[ARM64/Unix] (#9500)

* [Arm64/Unix] Update arm64 *.S files to match *.asm

* [Arm64/Unix] Fix CONTEXTToNativeContext()

* [Arm64/Unix] ThrowExceptionFromContextInternal

* [Arm64/Unix] Preserve x8 argument register

* [ARM64/Unix] Add CFI directives

Add native unwind info

* [Arm64/Unix] Fix RtlRestoreContext

* [Arm64/Unix] Restore FP from CurrentContextPointers

* [Arm64/Unix] fix pointer math

* [Arm64/Unix] Fix CallDescrWorkerInternal personality

* [Arm64/Unix] More Fp fixups

* [Arm64/Unix] CallEHFunclet machine state

Restore non-volatile machine state in CallEHFunclet

* [Arm64/Unix] CallDescrWorkerInternal

Use empty stack slot to save argument

* [Arm64/Unix] RtlVirtualUnwind update pointers

* [Arm64] LazyMachState fixes

* [Arm64/Unix] disable USE_REDIRECT_FOR_GCSTRESS

When FEATURE_PAL is enableds USE_REDIRECT_FOR_GCSTRESS
is not supported

* [Arm64] ClearRegDisplayArgumentAndScratchRegisters()

* [Arm64] Remove unnecesary copy in TransitionFrame

* [Arm64/Unix] Fix comment per review

* [Arm64/Unix] move constants per review

* [Arm64/Unix] Use ldp per review

Also fix indentation

* [Arm64/Unix] Fix indentation per review

* [Arm64/Unix] Remove m_Unwound per review comments

* [Arm64/Unix] Use PREPARE_EXTERNAL_VAR to access globals

* [Arm64/Unix] Fix more whitespace per earlier review comments

7 files changed, 235 insertions, 156 deletions

diff --git a/src/pal/inc/unixasmmacrosarm64.inc b/src/pal/inc/unixasmmacrosarm64.inc
index 359f27f878..ed73748e41 100644
--- a/src/pal/inc/unixasmmacrosarm64.inc
+++ b/src/pal/inc/unixasmmacrosarm64.inc
@@ -31,8 +31,8 @@ C_FUNC(\Name):
 .endm
 
 .macro LEAF_END_MARKED Name, Section
-        .global C_FUNC(\Name\()_End)
 C_FUNC(\Name\()_End):
+        .global C_FUNC(\Name\()_End)
         LEAF_END \Name, \Section
 .endm
 
@@ -48,18 +48,23 @@ C_FUNC(\Name\()_End):
 
 .macro EPILOG_STACK_FREE Size
         add sp, sp, \Size
+        .cfi_adjust_cfa_offset -\Size
 .endm
 
 .macro EPILOG_STACK_RESTORE
         mov sp, fp
+        .cfi_restore sp
 .endm
 
 .macro PROLOG_SAVE_REG reg, ofs
         str \reg, [sp, \ofs]
+        .cfi_rel_offset \reg, \ofs
 .endm
 
 .macro PROLOG_SAVE_REG_PAIR reg1, reg2, ofs
         stp \reg1, \reg2, [sp, \ofs]
+        .cfi_rel_offset \reg1, \ofs
+        .cfi_rel_offset \reg2, \ofs + 8
         .ifc \reg1, fp
         mov fp, sp
         .endif
@@ -67,6 +72,9 @@ C_FUNC(\Name\()_End):
 
 .macro PROLOG_SAVE_REG_PAIR_INDEXED reg1, reg2, ofs
         stp \reg1, \reg2, [sp, \ofs]!
+        .cfi_adjust_cfa_offset -\ofs
+        .cfi_rel_offset \reg1, 0
+        .cfi_rel_offset \reg2, 8
         .ifc \reg1, fp
         mov fp, sp
         .endif
@@ -74,14 +82,20 @@ C_FUNC(\Name\()_End):
 
 .macro EPILOG_RESTORE_REG reg, ofs
         ldr \reg, [sp, \ofs]
+        .cfi_restore \reg1
 .endm
 
 .macro EPILOG_RESTORE_REG_PAIR reg1, reg2, ofs
         ldp \reg1, \reg2, [sp, \ofs]
+        .cfi_restore \reg1
+        .cfi_restore \reg2
 .endm
 
 .macro EPILOG_RESTORE_REG_PAIR_INDEXED reg1, reg2, ofs
         ldp \reg1, \reg2, [sp], \ofs
+        .cfi_restore \reg1
+        .cfi_restore \reg2
+        .cfi_adjust_cfa_offset -\ofs
 .endm
 
 .macro EPILOG_RETURN
@@ -94,14 +108,14 @@ C_FUNC(\Name\()_End):
 
 //-----------------------------------------------------------------------------
 // Define the prolog for a TransitionFrame-based method. This macro should be called first in the method and
-// comprises the entire prolog (i.e. don't modify SP after calling this).The locals must be 8 byte aligned 
+// comprises the entire prolog (i.e. don't modify SP after calling this).The locals must be 8 byte aligned
 //
 // Stack layout:
 //
 // (stack parameters)
 // ...
 // fp
-// lr 
+// lr
 // CalleeSavedRegisters::x28
 // CalleeSavedRegisters::x27
 // CalleeSavedRegisters::x26
@@ -133,6 +147,7 @@ C_FUNC(\Name\()_End):
 .macro PROLOG_WITH_TRANSITION_BLOCK extraLocals = 0, SaveFPArgs = 1
 
         __PWTB_FloatArgumentRegisters = \extraLocals
+        __PWTB_SaveFPArgs = \SaveFPArgs
 
         .if ((__PWTB_FloatArgumentRegisters % 16) != 0)
                 __PWTB_FloatArgumentRegisters = __PWTB_FloatArgumentRegisters + 8
@@ -140,74 +155,114 @@ C_FUNC(\Name\()_End):
 
         __PWTB_TransitionBlock = __PWTB_FloatArgumentRegisters
 
-        .if \SaveFPArgs > 0
+        .if (__PWTB_SaveFPArgs == 1)
             __PWTB_TransitionBlock = __PWTB_TransitionBlock + SIZEOF__FloatArgumentRegisters
         .endif
 
         __PWTB_StackAlloc = __PWTB_TransitionBlock
-        __PWTB_ArgumentRegisters = __PWTB_StackAlloc + 96 
-
-        PROLOG_SAVE_REG_PAIR_INDEXED   fp, lr, #-160
-        // Spill callee saved registers 
-        PROLOG_SAVE_REG_PAIR   x19, x20, #16
-        PROLOG_SAVE_REG_PAIR   x21, x22, #32
-        PROLOG_SAVE_REG_PAIR   x23, x24, #48
-        PROLOG_SAVE_REG_PAIR   x25, x26, #64
-        PROLOG_SAVE_REG_PAIR   x27, x28, #80
-       
+        __PWTB_ArgumentRegisters = __PWTB_StackAlloc + 96
+
+        PROLOG_SAVE_REG_PAIR_INDEXED   fp, lr, -176
+        // Spill callee saved registers
+        PROLOG_SAVE_REG_PAIR   x19, x20, 16
+        PROLOG_SAVE_REG_PAIR   x21, x22, 32
+        PROLOG_SAVE_REG_PAIR   x23, x24, 48
+        PROLOG_SAVE_REG_PAIR   x25, x26, 64
+        PROLOG_SAVE_REG_PAIR   x27, x28, 80
+
         // Allocate space for the rest of the frame
         PROLOG_STACK_ALLOC  __PWTB_StackAlloc
-        
+
         // Spill argument registers.
         SAVE_ARGUMENT_REGISTERS        sp, __PWTB_ArgumentRegisters
 
-        .if \SaveFPArgs > 0
+        .if (__PWTB_SaveFPArgs == 1)
                 SAVE_FLOAT_ARGUMENT_REGISTERS sp, \extraLocals
         .endif
 
 .endm
 
 //-----------------------------------------------------------------------------
-// The Following sets of SAVE_*_REGISTERS expect the memory to be reserved and 
+// The Following sets of SAVE_*_REGISTERS expect the memory to be reserved and
 // base address to be passed in $reg
 //
 
 // Reserve 64 bytes of memory before calling  SAVE_ARGUMENT_REGISTERS
-.macro SAVE_ARGUMENT_REGISTERS reg, ofs 
+.macro SAVE_ARGUMENT_REGISTERS reg, ofs
 
         stp                    x0, x1, [\reg, #(\ofs)]
+       .cfi_rel_offset x0, \ofs
+       .cfi_rel_offset x1, \ofs + 8
         stp                    x2, x3, [\reg, #(\ofs + 16)]
+       .cfi_rel_offset x2, \ofs + 16
+       .cfi_rel_offset x3, \ofs + 24
         stp                    x4, x5, [\reg, #(\ofs + 32)]
+       .cfi_rel_offset x4, \ofs + 32
+       .cfi_rel_offset x5, \ofs + 40
         stp                    x6, x7, [\reg, #(\ofs + 48)]
+       .cfi_rel_offset x6, \ofs + 48
+       .cfi_rel_offset x7, \ofs + 56
+        str                    x8, [\reg, #(\ofs + 64)]
+       .cfi_rel_offset x8, \ofs + 64
 
 .endm
 
 // Reserve 64 bytes of memory before calling  SAVE_FLOAT_ARGUMENT_REGISTERS
-.macro SAVE_FLOAT_ARGUMENT_REGISTERS reg, ofs 
+.macro SAVE_FLOAT_ARGUMENT_REGISTERS reg, ofs
 
         stp                    d0, d1, [\reg, #(\ofs)]
+       .cfi_rel_offset d0, \ofs + 0
+       .cfi_rel_offset d1, \ofs + 8
         stp                    d2, d3, [\reg, #(\ofs + 16)]
+       .cfi_rel_offset d2, \ofs + 16
+       .cfi_rel_offset d3, \ofs + 24
         stp                    d4, d5, [\reg, #(\ofs + 32)]
+       .cfi_rel_offset d4, \ofs + 32
+       .cfi_rel_offset d5, \ofs + 40
         stp                    d6, d7, [\reg, #(\ofs + 48)]
+       .cfi_rel_offset d6, \ofs + 48
+       .cfi_rel_offset d7, \ofs + 56
 
 .endm
 
-.macro RESTORE_ARGUMENT_REGISTERS reg, ofs 
+.macro RESTORE_ARGUMENT_REGISTERS reg, ofs
 
         ldp                    x0, x1, [\reg, #(\ofs)]
+        .cfi_restore x0
+        .cfi_restore x1
         ldp                    x2, x3, [\reg, #(\ofs + 16)]
+        .cfi_restore x2
+        .cfi_restore x3
         ldp                    x4, x5, [\reg, #(\ofs + 32)]
+        .cfi_restore x4
+        .cfi_restore x5
         ldp                    x6, x7, [\reg, #(\ofs + 48)]
+        .cfi_restore x6
+        .cfi_restore x7
+        ldr                    x8, [\reg, #(\ofs + 64)]
+        .cfi_restore x8
 
 .endm
 
-.macro RESTORE_FLOAT_ARGUMENT_REGISTERS reg, ofs 
+.macro RESTORE_FLOAT_ARGUMENT_REGISTERS reg, ofs
 
         ldp                    d0, d1, [\reg, #(\ofs)]
+        .cfi_restore d0
+        .cfi_restore d1
         ldp                    d2, d3, [\reg, #(\ofs + 16)]
+        .cfi_restore d2
+        .cfi_restore d3
         ldp                    d4, d5, [\reg, #(\ofs + 32)]
+        .cfi_restore d4
+        .cfi_restore d5
         ldp                    d6, d7, [\reg, #(\ofs + 48)]
+        .cfi_restore d6
+        .cfi_restore d7
+
+.endm
 
+.macro EPILOG_BRANCH Target
+        b \Target
 .endm
 
 .macro EPILOG_BRANCH_REG reg
@@ -216,40 +271,42 @@ C_FUNC(\Name\()_End):
 
 .endm
 
-//-----------------------------------------------------------------------------
-// Provides a matching epilog to PROLOG_WITH_TRANSITION_BLOCK and ends by preparing for tail-calling.
-// Since this is a tail call argument registers are restored.
-//
-.macro EPILOG_WITH_TRANSITION_BLOCK_TAILCALL extraLocals = 0, SaveFPArgs =1
 
-        __PWTB_FloatArgumentRegisters = \extraLocals
+.macro EPILOG_WITH_TRANSITION_BLOCK_RETURN
 
-        .if ((__PWTB_FloatArgumentRegisters % 16) != 0)
-                __PWTB_FloatArgumentRegisters = __PWTB_FloatArgumentRegisters + 8
-        .endif
+        EPILOG_STACK_FREE                 __PWTB_StackAlloc
 
-        __PWTB_TransitionBlock = __PWTB_FloatArgumentRegisters
+        EPILOG_RESTORE_REG_PAIR   x19, x20, 16
+        EPILOG_RESTORE_REG_PAIR   x21, x22, 32
+        EPILOG_RESTORE_REG_PAIR   x23, x24, 48
+        EPILOG_RESTORE_REG_PAIR   x25, x26, 64
+        EPILOG_RESTORE_REG_PAIR   x27, x28, 80
+        EPILOG_RESTORE_REG_PAIR_INDEXED fp, lr, 176
+        ret
 
-        .if \SaveFPArgs > 0
-            __PWTB_TransitionBlock = __PWTB_TransitionBlock + SIZEOF__FloatArgumentRegisters
-        .endif
+.endm
 
-        __PWTB_StackAlloc = __PWTB_TransitionBlock
-        __PWTB_ArgumentRegisters = __PWTB_StackAlloc + 96 
 
-        .if \SaveFPArgs > 0
-                RESTORE_FLOAT_ARGUMENT_REGISTERS  sp, __PWTB_FloatArgumentRegisters
+//-----------------------------------------------------------------------------
+// Provides a matching epilog to PROLOG_WITH_TRANSITION_BLOCK and ends by preparing for tail-calling.
+// Since this is a tail call argument registers are restored.
+//
+.macro EPILOG_WITH_TRANSITION_BLOCK_TAILCALL
+
+        .if (__PWTB_SaveFPArgs == 1)
+            RESTORE_FLOAT_ARGUMENT_REGISTERS  sp, __PWTB_FloatArgumentRegisters
         .endif
 
         RESTORE_ARGUMENT_REGISTERS        sp, __PWTB_ArgumentRegisters
+
         EPILOG_STACK_FREE                 __PWTB_StackAlloc
-        
-        EPILOG_RESTORE_REG_PAIR   x19, x20, #16
-        EPILOG_RESTORE_REG_PAIR   x21, x22, #32
-        EPILOG_RESTORE_REG_PAIR   x23, x24, #48
-        EPILOG_RESTORE_REG_PAIR   x25, x26, #64
-        EPILOG_RESTORE_REG_PAIR   x27, x28, #80
-        EPILOG_RESTORE_REG_PAIR_INDEXED fp, lr, #160
+
+        EPILOG_RESTORE_REG_PAIR   x19, x20, 16
+        EPILOG_RESTORE_REG_PAIR   x21, x22, 32
+        EPILOG_RESTORE_REG_PAIR   x23, x24, 48
+        EPILOG_RESTORE_REG_PAIR   x25, x26, 64
+        EPILOG_RESTORE_REG_PAIR   x27, x28, 80
+        EPILOG_RESTORE_REG_PAIR_INDEXED fp, lr, 176
 
 .endm
 
@@ -273,8 +330,8 @@ __RedirectionFuncName SETS "|?RedirectedHandledJITCaseFor":CC:"$reason":CC:"@Thr
         IMPORT $__RedirectionFuncName
 
         NESTED_ENTRY $__RedirectionStubFuncName
-        PROLOG_SAVE_REG_PAIR    fp, lr, #-16
-        sub sp, sp, #16                          // stack slot for CONTEXT * and padding 
+        PROLOG_SAVE_REG_PAIR    fp, lr, -16
+        sub sp, sp, #16                          // stack slot for CONTEXT * and padding
 
         //REDIRECTSTUB_SP_OFFSET_CONTEXT is defined in asmconstants.h and is used in GetCONTEXTFromRedirectedStubStackFrame
         //If CONTEXT is not saved at 0 offset from SP it must be changed as well.
diff --git a/src/pal/src/arch/arm64/asmconstants.h b/src/pal/src/arch/arm64/asmconstants.h
new file mode 100644
index 0000000000..b2bf74461f
--- /dev/null
+++ b/src/pal/src/arch/arm64/asmconstants.h
@@ -0,0 +1,95 @@
+// Licensed to the .NET Foundation under one or more agreements.
+// The .NET Foundation licenses this file to you under the MIT license.
+// See the LICENSE file in the project root for more information.
+
+#ifndef __PAL_ARM64_ASMCONSTANTS_H__
+#define __PAL_ARM64_ASMCONSTANTS_H__
+
+#define CONTEXT_ARM64   0x00400000L
+
+#define CONTEXT_CONTROL_BIT (0)
+#define CONTEXT_INTEGER_BIT (1)
+#define CONTEXT_FLOATING_POINT_BIT (2)
+#define CONTEXT_DEBUG_REGISTERS_BIT (3)
+
+#define CONTEXT_CONTROL (CONTEXT_ARM64 | (1L << CONTEXT_CONTROL_BIT))
+#define CONTEXT_INTEGER (CONTEXT_ARM64 | (1 << CONTEXT_INTEGER_BIT))
+#define CONTEXT_FLOATING_POINT  (CONTEXT_ARM64 | (1 << CONTEXT_FLOATING_POINT_BIT))
+#define CONTEXT_DEBUG_REGISTERS (CONTEXT_ARM64 | (1 << CONTEXT_DEBUG_REGISTERS_BIT))
+
+#define CONTEXT_FULL (CONTEXT_CONTROL | CONTEXT_INTEGER | CONTEXT_FLOATING_POINT)
+
+
+#define CONTEXT_ContextFlags 0
+#define CONTEXT_Cpsr         CONTEXT_ContextFlags+4
+#define CONTEXT_X0           CONTEXT_Cpsr+4
+#define CONTEXT_X1           CONTEXT_X0+8
+#define CONTEXT_X2           CONTEXT_X1+8
+#define CONTEXT_X3           CONTEXT_X2+8
+#define CONTEXT_X4           CONTEXT_X3+8
+#define CONTEXT_X5           CONTEXT_X4+8
+#define CONTEXT_X6           CONTEXT_X5+8
+#define CONTEXT_X7           CONTEXT_X6+8
+#define CONTEXT_X8           CONTEXT_X7+8
+#define CONTEXT_X9           CONTEXT_X8+8
+#define CONTEXT_X10          CONTEXT_X9+8
+#define CONTEXT_X11          CONTEXT_X10+8
+#define CONTEXT_X12          CONTEXT_X11+8
+#define CONTEXT_X13          CONTEXT_X12+8
+#define CONTEXT_X14          CONTEXT_X13+8
+#define CONTEXT_X15          CONTEXT_X14+8
+#define CONTEXT_X16          CONTEXT_X15+8
+#define CONTEXT_X17          CONTEXT_X16+8
+#define CONTEXT_X18          CONTEXT_X17+8
+#define CONTEXT_X19          CONTEXT_X18+8
+#define CONTEXT_X20          CONTEXT_X19+8
+#define CONTEXT_X21          CONTEXT_X20+8
+#define CONTEXT_X22          CONTEXT_X21+8
+#define CONTEXT_X23          CONTEXT_X22+8
+#define CONTEXT_X24          CONTEXT_X23+8
+#define CONTEXT_X25          CONTEXT_X24+8
+#define CONTEXT_X26          CONTEXT_X25+8
+#define CONTEXT_X27          CONTEXT_X26+8
+#define CONTEXT_X28          CONTEXT_X27+8
+#define CONTEXT_Fp           CONTEXT_X28+8
+#define CONTEXT_Lr           CONTEXT_Fp+8
+#define CONTEXT_Sp           CONTEXT_Lr+8
+#define CONTEXT_Pc           CONTEXT_Sp+8
+#define CONTEXT_NEON_OFFSET  CONTEXT_Pc+8
+#define CONTEXT_V0           0
+#define CONTEXT_V1           CONTEXT_V0+16
+#define CONTEXT_V2           CONTEXT_V1+16
+#define CONTEXT_V3           CONTEXT_V2+16
+#define CONTEXT_V4           CONTEXT_V3+16
+#define CONTEXT_V5           CONTEXT_V4+16
+#define CONTEXT_V6           CONTEXT_V5+16
+#define CONTEXT_V7           CONTEXT_V6+16
+#define CONTEXT_V8           CONTEXT_V7+16
+#define CONTEXT_V9           CONTEXT_V8+16
+#define CONTEXT_V10          CONTEXT_V9+16
+#define CONTEXT_V11          CONTEXT_V10+16
+#define CONTEXT_V12          CONTEXT_V11+16
+#define CONTEXT_V13          CONTEXT_V12+16
+#define CONTEXT_V14          CONTEXT_V13+16
+#define CONTEXT_V15          CONTEXT_V14+16
+#define CONTEXT_V16          CONTEXT_V15+16
+#define CONTEXT_V17          CONTEXT_V16+16
+#define CONTEXT_V18          CONTEXT_V17+16
+#define CONTEXT_V19          CONTEXT_V18+16
+#define CONTEXT_V20          CONTEXT_V19+16
+#define CONTEXT_V21          CONTEXT_V20+16
+#define CONTEXT_V22          CONTEXT_V21+16
+#define CONTEXT_V23          CONTEXT_V22+16
+#define CONTEXT_V24          CONTEXT_V23+16
+#define CONTEXT_V25          CONTEXT_V24+16
+#define CONTEXT_V26          CONTEXT_V25+16
+#define CONTEXT_V27          CONTEXT_V26+16
+#define CONTEXT_V28          CONTEXT_V27+16
+#define CONTEXT_V29          CONTEXT_V28+16
+#define CONTEXT_V30          CONTEXT_V29+16
+#define CONTEXT_V31          CONTEXT_V30+16
+#define CONTEXT_FLOAT_CONTROL_OFFSET  CONTEXT_V31
+#define CONTEXT_Fpcr         0
+#define CONTEXT_Fpsr         CONTEXT_Fpcr+4
+
+#endif
diff --git a/src/pal/src/arch/arm64/context2.S b/src/pal/src/arch/arm64/context2.S
index a64e62c94d..e62a9ac4d9 100644
--- a/src/pal/src/arch/arm64/context2.S
+++ b/src/pal/src/arch/arm64/context2.S
@@ -8,87 +8,7 @@
 //
 
 #include "unixasmmacros.inc"
-
-#define CONTEXT_ARM64   0x00400000L
-
-#define CONTEXT_CONTROL (CONTEXT_ARM64 | 0x1L)
-#define CONTEXT_INTEGER (CONTEXT_ARM64 | 0x2L)
-#define CONTEXT_FLOATING_POINT  (CONTEXT_ARM64 | 0x4L)
-#define CONTEXT_DEBUG_REGISTERS (CONTEXT_ARM64 | 0x8L)
-
-#define CONTEXT_FULL (CONTEXT_CONTROL | CONTEXT_INTEGER | CONTEXT_FLOATING_POINT)
-
-#define CONTEXT_ContextFlags 0
-#define CONTEXT_Cpsr         CONTEXT_ContextFlags+4
-#define CONTEXT_X0           CONTEXT_Cpsr+4
-#define CONTEXT_X1           CONTEXT_X0+8
-#define CONTEXT_X2           CONTEXT_X1+8
-#define CONTEXT_X3           CONTEXT_X2+8
-#define CONTEXT_X4           CONTEXT_X3+8
-#define CONTEXT_X5           CONTEXT_X4+8
-#define CONTEXT_X6           CONTEXT_X5+8
-#define CONTEXT_X7           CONTEXT_X6+8
-#define CONTEXT_X8           CONTEXT_X7+8
-#define CONTEXT_X9           CONTEXT_X8+8
-#define CONTEXT_X10          CONTEXT_X9+8
-#define CONTEXT_X11          CONTEXT_X10+8
-#define CONTEXT_X12          CONTEXT_X11+8
-#define CONTEXT_X13          CONTEXT_X12+8
-#define CONTEXT_X14          CONTEXT_X13+8
-#define CONTEXT_X15          CONTEXT_X14+8
-#define CONTEXT_X16          CONTEXT_X15+8
-#define CONTEXT_X17          CONTEXT_X16+8
-#define CONTEXT_X18          CONTEXT_X17+8
-#define CONTEXT_X19          CONTEXT_X18+8
-#define CONTEXT_X20          CONTEXT_X19+8
-#define CONTEXT_X21          CONTEXT_X20+8
-#define CONTEXT_X22          CONTEXT_X21+8
-#define CONTEXT_X23          CONTEXT_X22+8
-#define CONTEXT_X24          CONTEXT_X23+8
-#define CONTEXT_X25          CONTEXT_X24+8
-#define CONTEXT_X26          CONTEXT_X25+8
-#define CONTEXT_X27          CONTEXT_X26+8
-#define CONTEXT_X28          CONTEXT_X27+8
-#define CONTEXT_Fp           CONTEXT_X28+8
-#define CONTEXT_Lr           CONTEXT_Fp+8
-#define CONTEXT_Sp           CONTEXT_Lr+8
-#define CONTEXT_Pc           CONTEXT_Sp+8
-#define CONTEXT_NEON_OFFSET  CONTEXT_Pc+8
-#define CONTEXT_V0           0
-#define CONTEXT_V1           CONTEXT_V0+16
-#define CONTEXT_V2           CONTEXT_V1+16
-#define CONTEXT_V3           CONTEXT_V2+16
-#define CONTEXT_V4           CONTEXT_V3+16
-#define CONTEXT_V5           CONTEXT_V4+16
-#define CONTEXT_V6           CONTEXT_V5+16
-#define CONTEXT_V7           CONTEXT_V6+16
-#define CONTEXT_V8           CONTEXT_V7+16
-#define CONTEXT_V9           CONTEXT_V8+16
-#define CONTEXT_V10          CONTEXT_V9+16
-#define CONTEXT_V11          CONTEXT_V10+16
-#define CONTEXT_V12          CONTEXT_V11+16
-#define CONTEXT_V13          CONTEXT_V12+16
-#define CONTEXT_V14          CONTEXT_V13+16
-#define CONTEXT_V15          CONTEXT_V14+16
-#define CONTEXT_V16          CONTEXT_V15+16
-#define CONTEXT_V17          CONTEXT_V16+16
-#define CONTEXT_V18          CONTEXT_V17+16
-#define CONTEXT_V19          CONTEXT_V18+16
-#define CONTEXT_V20          CONTEXT_V19+16
-#define CONTEXT_V21          CONTEXT_V20+16
-#define CONTEXT_V22          CONTEXT_V21+16
-#define CONTEXT_V23          CONTEXT_V22+16
-#define CONTEXT_V24          CONTEXT_V23+16
-#define CONTEXT_V25          CONTEXT_V24+16
-#define CONTEXT_V26          CONTEXT_V25+16
-#define CONTEXT_V27          CONTEXT_V26+16
-#define CONTEXT_V28          CONTEXT_V27+16
-#define CONTEXT_V29          CONTEXT_V28+16
-#define CONTEXT_V30          CONTEXT_V29+16
-#define CONTEXT_V31          CONTEXT_V30+16
-#define CONTEXT_FLOAT_CONTROL_OFFSET  CONTEXT_V31
-#define CONTEXT_Fpcr         0
-#define CONTEXT_Fpsr         CONTEXT_Fpcr+4
+#include "asmconstants.h"
 
 // Incoming:
 //  x0: Context*
@@ -115,10 +35,8 @@ LEAF_ENTRY CONTEXT_CaptureContext, _TEXT
     ldr x2, [sp, 24]
     str w2, [x0, CONTEXT_Cpsr]
     stp fp, lr, [x0, CONTEXT_Fp]
-    add sp, sp, #32
-    mov x2, sp
+    add x2, sp, #32
     stp x2, lr, [x0, CONTEXT_Sp]
-    sub sp, sp, #32
 
 LOCAL_LABEL(Done_CONTEXT_CONTROL): 
     // we dont clobber x1 in the CONTEXT_CONTROL case
@@ -224,14 +142,8 @@ LEAF_ENTRY RtlRestoreContext, _TEXT
     // since we potentially clobber x0 below, we'll bank it in x16
     mov x16, x0
 
-    ldr w2, [x16, CONTEXT_ContextFlags]
-    // clangs assembler doesn't seem to support the mov Wx, imm32 yet
-    movz w3, #0x40, lsl #16
-    movk w3, #0x4
-    mov w4, w3
-    and w3, w2, w3
-    cmp w3, w4
-    b.ne LOCAL_LABEL(No_Restore_CONTEXT_FLOATING_POINT)
+    ldr w17, [x16, CONTEXT_ContextFlags]
+    tbz w17, #CONTEXT_FLOATING_POINT_BIT, LOCAL_LABEL(No_Restore_CONTEXT_FLOATING_POINT)
 
     add x16, x16,   CONTEXT_NEON_OFFSET
     ldp q0, q1,   [x16, CONTEXT_V0]
@@ -256,12 +168,7 @@ LEAF_ENTRY RtlRestoreContext, _TEXT
     sub x16, x16,   CONTEXT_NEON_OFFSET
 
 LOCAL_LABEL(No_Restore_CONTEXT_FLOATING_POINT):
-    movz w2, #0x40, lsl #16
-    movk w2, #0x2
-    mov w3, w2
-    and w2, w1, w2
-    cmp w2, w3
-    b.ne LOCAL_LABEL(No_Restore_CONTEXT_INTEGER)
+    tbz w17, #CONTEXT_INTEGER_BIT, LOCAL_LABEL(No_Restore_CONTEXT_INTEGER)
 
     ldp x0, x1,   [x16, CONTEXT_X0]
     ldp x2, x3,   [x16, CONTEXT_X2]
@@ -279,12 +186,7 @@ LOCAL_LABEL(No_Restore_CONTEXT_FLOATING_POINT):
     ldr x28,      [x16, CONTEXT_X28]
 
 LOCAL_LABEL(No_Restore_CONTEXT_INTEGER):
-    movz w2, #0x40, lsl #16
-    movk w2, #0x2
-    mov w3, w2
-    and w2, w1, w2
-    cmp w2, w3
-    b.ne LOCAL_LABEL(No_Restore_CONTEXT_CONTROL)
+    tbz w17, #CONTEXT_CONTROL_BIT, LOCAL_LABEL(No_Restore_CONTEXT_CONTROL)
 
     ldr w17, [x16, CONTEXT_Cpsr]
     msr nzcv, x17
@@ -293,8 +195,8 @@ LOCAL_LABEL(No_Restore_CONTEXT_INTEGER):
     mov sp, x17
     ldr x17, [x16, CONTEXT_Pc]
     br x17
-
+ 
 LOCAL_LABEL(No_Restore_CONTEXT_CONTROL):
-    ret
+   ret
 
 LEAF_END RtlRestoreContext, _TEXT
diff --git a/src/pal/src/arch/arm64/exceptionhelper.S b/src/pal/src/arch/arm64/exceptionhelper.S
index 4fdcfc5eb1..480846eb61 100644
--- a/src/pal/src/arch/arm64/exceptionhelper.S
+++ b/src/pal/src/arch/arm64/exceptionhelper.S
@@ -3,7 +3,30 @@
 // See the LICENSE file in the project root for more information.
 
 #include "unixasmmacros.inc"
+#include "asmconstants.h"
 
+//////////////////////////////////////////////////////////////////////////
+//
+// This function creates a stack frame right below the target frame, restores all callee
+// saved registers, SP, and LR from the passed in context.
+// Then it uses the ThrowExceptionHelper to throw the passed in exception from that context.
+// EXTERN_C void ThrowExceptionFromContextInternal(CONTEXT* context, PAL_SEHException* ex);
 LEAF_ENTRY ThrowExceptionFromContextInternal, _TEXT
-    EMIT_BREAKPOINT
+    // Save the FP & LR to the stack so that the unwind can work at the instruction after
+    // loading the FP from the context, but before loading the SP from the context.
+    stp fp, lr,  [sp, -16]!
+
+    ldp x19,x20, [x0, #(CONTEXT_X19)]
+    ldp x21,x22, [x0, #(CONTEXT_X21)]
+    ldp x23,x24, [x0, #(CONTEXT_X23)]
+    ldp x24,x25, [x0, #(CONTEXT_X24)]
+    ldp x26,x27, [x0, #(CONTEXT_X26)]
+    ldp x28,fp,  [x0, #(CONTEXT_X28)]
+    ldr lr,      [x0, #(CONTEXT_Pc)]
+    ldr x2,      [x0, #(CONTEXT_Sp)]
+    mov sp, x2
+
+    // The PAL_SEHException pointer
+    mov x0, x1
+    b   EXTERNAL_C_FUNC(ThrowExceptionHelper)
 LEAF_END ThrowExceptionFromContextInternal, _TEXT
diff --git a/src/pal/src/exception/seh-unwind.cpp b/src/pal/src/exception/seh-unwind.cpp
index aeb6fa4b6a..1f20ee0cad 100644
--- a/src/pal/src/exception/seh-unwind.cpp
+++ b/src/pal/src/exception/seh-unwind.cpp
@@ -244,6 +244,7 @@ static void GetContextPointers(unw_cursor_t *cursor, unw_context_t *unwContext,
     GetContextPointer(cursor, unwContext, UNW_AARCH64_X26, &contextPointers->X26);
     GetContextPointer(cursor, unwContext, UNW_AARCH64_X27, &contextPointers->X27);
     GetContextPointer(cursor, unwContext, UNW_AARCH64_X28, &contextPointers->X28);
+    GetContextPointer(cursor, unwContext, UNW_AARCH64_X29, &contextPointers->Fp);
 #else
 #error unsupported architecture
 #endif
diff --git a/src/pal/src/include/pal/context.h b/src/pal/src/include/pal/context.h
index 6857c130ee..08fa05da7d 100644
--- a/src/pal/src/include/pal/context.h
+++ b/src/pal/src/include/pal/context.h
@@ -248,8 +248,7 @@ inline void *FPREG_Xstate_Ymmh(const ucontext_t *uc)
 
 #define MCREG_Sp(mc)      ((mc).sp)
 #define MCREG_Pc(mc)      ((mc).pc)
-#define MCREG_PState(mc)  ((mc).pstate)
-#define MCREG_Cpsr(mc)    ((mc).cpsr)
+#define MCREG_Cpsr(mc)    ((mc).pstate)
 #else
     // For FreeBSD, as found in x86/ucontext.h
 #define MCREG_Rbp(mc)	    ((mc).mc_rbp)
diff --git a/src/pal/src/thread/context.cpp b/src/pal/src/thread/context.cpp
index bee6ddd449..98867c9554 100644
--- a/src/pal/src/thread/context.cpp
+++ b/src/pal/src/thread/context.cpp
@@ -127,6 +127,8 @@ typedef int __ptrace_request;
         ASSIGN_REG(R12)
 #elif defined(_ARM64_)
 #define ASSIGN_CONTROL_REGS \
+        ASSIGN_REG(Cpsr)    \
+        ASSIGN_REG(Fp)      \
         ASSIGN_REG(Sp)      \
         ASSIGN_REG(Lr)      \
         ASSIGN_REG(Pc)