diff --git a/ChangeLog.md b/ChangeLog.md
--- a/ChangeLog.md
+++ b/ChangeLog.md
@@ -1,5 +1,9 @@
 # Revision history for long-double
 
+## 0.1.1  -- 2020-06-15
+
+* Add support for aarch64 where long double == _Float128.
+
 ## 0.1  -- 2018-03-09
 
 * First release.
diff --git a/aarch64/Numeric/LongDouble.hs b/aarch64/Numeric/LongDouble.hs
new file mode 100644
--- /dev/null
+++ b/aarch64/Numeric/LongDouble.hs
@@ -0,0 +1,17 @@
+-----------------------------------------------------------------------------
+-- |
+-- Module      :  Numeric.LongDouble
+-- Copyright   :  (C) 2018,2020 Claude Heiland-Allen
+-- License     :  BSD3
+-- Maintainer  :  Claude Heiland-Allen <claude@mathr.co.uk>
+-- Stability   :  experimental
+-- Portability :  non-portable 
+--
+-- This module re-exports the default platform-specfic ABI for C's long double.
+module Numeric.LongDouble ( module Numeric.LongDouble.ARM_128 ) where
+
+import Numeric.LongDouble.ARM_128
+  ( LongDouble()
+  , truncate', round', ceiling', floor'
+  , fromDouble, toDouble, fromInt, toInt
+  )
diff --git a/hs/Numeric/LongDouble/ARM_128.hs b/hs/Numeric/LongDouble/ARM_128.hs
new file mode 100644
--- /dev/null
+++ b/hs/Numeric/LongDouble/ARM_128.hs
@@ -0,0 +1,67 @@
+{-# LANGUAGE GeneralizedNewtypeDeriving #-}
+-----------------------------------------------------------------------------
+-- |
+-- Module      :  Numeric.LongDouble.ARM_128
+-- Copyright   :  (C) 2020 Claude Heiland-Allen
+-- License     :  BSD3
+-- Maintainer  :  Claude Heiland-Allen <claude@mathr.co.uk>
+-- Stability   :  experimental
+-- Portability :  non-portable (assumes long double == _Float128)
+--
+-- This module contains a LongDouble type that is the same as Float128, as on aarch64.
+--
+-- Most code should import Numeric.LongDouble instead, unless a specific ABI
+-- is needed.  This module is for sizeof(long double) == 16, with the type being
+-- a simple alias for _Float128.
+module Numeric.LongDouble.ARM_128
+  ( 
+  -- * long double data type
+    LongDouble(..)
+  -- * RealFrac alternatives
+  , truncate'
+  , round'
+  , ceiling'
+  , floor'
+  -- * Conversions
+  , fromDouble
+  , toDouble
+  , fromInt
+  , toInt
+  ) where
+
+import Data.Coerce (coerce)
+import Foreign (Ptr, with, alloca)
+import qualified Numeric.Float128 as F
+import Foreign.Storable (Storable(..))
+import Numeric (showFloat, readFloat, readSigned)
+import System.IO.Unsafe (unsafePerformIO)
+
+-- | The long double type on aarch64: 128 bits of _Float128 in 128bits of space.
+newtype LongDouble = LD F.Float128
+  deriving (Storable, Eq, Ord, Num, Real, Fractional, RealFrac, Floating, RealFloat)
+
+instance Read LongDouble where
+  readsPrec p = coerce . (readsPrec p :: ReadS F.Float128)
+
+instance Show LongDouble where
+  showsPrec p (LD x) = showsPrec p x
+
+fromInt :: Int -> LongDouble
+fromInt = LD . F.fromInt
+
+toInt :: LongDouble -> Int
+toInt (LD f) = F.toInt f
+
+fromDouble :: Double -> LongDouble
+fromDouble = LD . F.fromDouble
+
+toDouble :: LongDouble -> Double
+toDouble (LD f) = F.toDouble f
+
+-- | Alternate versions of RealFrac methods that
+--   keep the value as a long double.
+truncate', round', ceiling', floor' :: LongDouble -> LongDouble
+truncate' = coerce F.truncate'
+round'    = coerce F.round'
+ceiling'  = coerce F.ceiling'
+floor'    = coerce F.floor'
diff --git a/long-double.cabal b/long-double.cabal
--- a/long-double.cabal
+++ b/long-double.cabal
@@ -1,9 +1,10 @@
 name:                long-double
-version:             0.1
+version:             0.1.1
 synopsis:            FFI bindings for C long double
 description:
   This package provides a LongDouble type, being 80bits of x87 data taking up
   96bits on i386 and 128bits on x86_64.  On arm it is an alias for 64bit double.
+  On aarch64 it is an alias for quadruple precision _Float128.
   It does not provide a CLDouble type usable for FFI without wrapping in Ptr,
   this needs to be done by the compiler.
   See <https://ghc.haskell.org/trac/ghc/ticket/3353>.
@@ -13,15 +14,17 @@
 license-file:        LICENSE
 author:              Claude Heiland-Allen
 maintainer:          claude@mathr.co.uk
-copyright:           (c) 2018 Claude Heiland-Allen
+copyright:           (c) 2018,2020 Claude Heiland-Allen
 category:            Math
 build-type:          Simple
 extra-source-files:
   ChangeLog.md
   i386/Numeric/LongDouble.hs
   x86_64/Numeric/LongDouble.hs
+  aarch64/Numeric/LongDouble.hs
   arm/Numeric/LongDouble.hs
   hs/Numeric/LongDouble/ARM_64.hs
+  hs/Numeric/LongDouble/ARM_128.hs
   hs/Numeric/LongDouble/X87.hs
   hs/Numeric/LongDouble/X87_96.hs
   hs/Numeric/LongDouble/X87_128.hs
@@ -35,13 +38,13 @@
 source-repository this
   type: git
   location: https://code.mathr.co.uk/long-double.git
-  tag: long-double-0.1
+  tag: long-double-0.1.1
 
 library
   exposed-modules:
     Numeric.LongDouble
   build-depends:
-    base >=4.6 && <4.12,
+    base >=4.6 && <4.15,
     integer-gmp >=0.5 && <1.1
   c-sources: c/longdouble.c
   hs-source-dirs: hs
@@ -57,7 +60,12 @@
         hs-source-dirs: arm
         other-modules: Numeric.LongDouble.ARM_64
       else
-        buildable: False
+        if arch(aarch64)
+          hs-source-dirs: aarch64
+          other-modules: Numeric.LongDouble.ARM_128
+          build-depends: float128 >=0.1 && <0.2
+        else
+          buildable: False
   default-language: Haskell2010
   other-extensions:
     CPP,