[libunwind][AArch64] Fix libunwind::Registers_arm64::jumpto

[miyuki]

Summary

The AArch64 version of the libunwind function which restores the
CPU state and resumes execution is not interrupt-safe. It restores
the target value of SP before loading the floating-point registers
from the context struct, but that struct is allocated on the stack
which is being deallocated. This means that if an interrupt occurs
during this function, and uses a lot of stack space, it could
overwrite the values about to be loaded into the floating-point
registers.

This patch fixes the issue.

Patch by Oliver Stannard.

[chill]

LGTM. r17 and r18 are be clobbered by a calls, according to AAPCS-64 (5.1.1) and I couldn't find any indication that some Apple ABI would deviate from that.

libunwind/trunk/src/UnwindRegistersRestore.S

DEFINE_LIBUNWIND_FUNCTION(_ZN9libunwind15Registers_arm646jumptoEv)
   // skip restore of x0,x1 for now
   ldp    x2, x3,  [x0, #0x010]
   ldp    x4, x5,  [x0, #0x020]
   ldp    x6, x7,  [x0, #0x030]
   ldp    x8, x9,  [x0, #0x040]
   ldp    x10,x11, [x0, #0x050]
   ldp    x12,x13, [x0, #0x060]
   ldp    x14,x15, [x0, #0x070]
-  ldp    x16,x17, [x0, #0x080]
+  // x16 and x17 were clobbered by the call into the unwinder, so no point in
+  // restoring them.
   ldp    x18,x19, [x0, #0x090]
   ldp    x20,x21, [x0, #0x0A0]
   ldp    x22,x23, [x0, #0x0B0]
   ldp    x24,x25, [x0, #0x0C0]
   ldp    x26,x27, [x0, #0x0D0]
   ldp    x28,x29, [x0, #0x0E0]
   ldr    x30,     [x0, #0x100]  // restore pc into lr
-  ldr    x1,      [x0, #0x0F8]
-  mov    sp,x1                  // restore sp
 
   ldp    d0, d1,  [x0, #0x110]
   ldp    d2, d3,  [x0, #0x120]
   ldp    d4, d5,  [x0, #0x130]
   ldp    d6, d7,  [x0, #0x140]
   ldp    d8, d9,  [x0, #0x150]
   ldp    d10,d11, [x0, #0x160]
   ldp    d12,d13, [x0, #0x170]
   ldp    d14,d15, [x0, #0x180]
   ldp    d16,d17, [x0, #0x190]
   ldp    d18,d19, [x0, #0x1A0]
   ldp    d20,d21, [x0, #0x1B0]
   ldp    d22,d23, [x0, #0x1C0]
   ldp    d24,d25, [x0, #0x1D0]
   ldp    d26,d27, [x0, #0x1E0]
   ldp    d28,d29, [x0, #0x1F0]
   ldr    d30,     [x0, #0x200]
   ldr    d31,     [x0, #0x208]
 
+  // Finally, restore sp. This must be done after the the last read from the
+  // context struct, because it is allocated on the stack, and an exception
+  // could clobber the de-allocated portion of the stack after sp has been
+  // restored.
+  ldr    x16,     [x0, #0x0F8]
   ldp    x0, x1,  [x0, #0x000]  // restore x0,x1
+  mov    sp,x16                 // restore sp
   ret    x30                    // jump to pc
 
 #elif defined(__arm__) && !defined(__APPLE__)
 
 #if !defined(__ARM_ARCH_ISA_ARM)
 #if (__ARM_ARCH_ISA_THUMB == 2)
   .syntax unified
 #endif