Implement new integer-gmp2 from scratch (re #9281)

This is done as a separate `integer-gmp2` backend library because it
turned out to become a complete rewrite from scratch.

Due to the different (over)allocation scheme and potentially different
accounting (via the new `{shrink,resize}MutableByteArray#` primitives),
some of the nofib benchmarks actually results in increased allocation
numbers (but not necessarily an increase in runtime!).  I believe the
allocation numbers could improve if `{resize,shrink}MutableByteArray#`
could be optimised to reallocate in-place more efficiently.

Here are the more apparent changes in the latest nofib comparision
between `integer-gmp` and `integer-gmp2`:

  ------------------------------------------------------------------
          Program     Size    Allocs   Runtime   Elapsed  TotalMem
  ------------------------------------------------------------------
              ...
       bernouilli    +1.6%    +15.3%     0.132     0.132      0.0%
              ...
     cryptarithm1    -2.2%      0.0%     -9.7%     -9.7%      0.0%
              ...
            fasta    -0.7%     -0.0%    +10.9%    +10.9%      0.0%
              ...
            kahan    +0.6%    +38.9%     0.169     0.169      0.0%
              ...
             lcss    -0.7%     -0.0%     -6.4%     -6.4%      0.0%
              ...
           mandel    +1.6%    +33.6%     0.049     0.049      0.0%
              ...
         pidigits    +0.8%     +8.5%     +3.9%     +3.9%      0.0%
            power    +1.4%    -23.8%    -18.6%    -18.6%    -16.7%
              ...
        primetest    +1.3%    +50.1%     0.085     0.085      0.0%
              ...
              rsa    +1.6%    +53.4%     0.026     0.026      0.0%
              ...
              scs    +1.2%     +6.6%     +6.5%     +6.6%    +14.3%
              ...
           symalg    +1.0%     +9.5%     0.010     0.010      0.0%
              ...
        transform    -0.6%     -0.0%     -5.9%     -5.9%      0.0%
              ...
  ------------------------------------------------------------------
              Min    -2.3%    -23.8%    -18.6%    -18.6%    -16.7%
              Max    +1.6%    +53.4%    +10.9%    +10.9%    +14.3%
   Geometric Mean    -0.3%     +1.9%     -0.8%     -0.8%     +0.0%

(see P35 / https://phabricator.haskell.org/P35 for full report)

By default, `INTEGER_LIBRARY=integer-gmp2` is active now, which results
in the package `integer-gmp-1.0.0.0` being registered in the package db.
The previous `integer-gmp-0.5.1.0` can be restored by setting
`INTEGER_LIBRARY=integer-gmp` (but will probably be removed altogether
for GHC 7.12). In-tree GMP support has been stolen from the old
`integer-gmp` (while unpatching the custom memory-allocators, as well as
forcing `-fPIC`)

A minor hack to `ghc-cabal` was necessary in order to support two different
`integer-gmp` packages (in different folders) with the same package key.

There will be a couple of follow-up commits re-implementing some features
that were dropped to keep D82 minimal, as well as further
clean-ups/improvements.

More information can be found via #9281 and
https://ghc.haskell.org/trac/ghc/wiki/Design/IntegerGmp2

Reviewed By: austin, rwbarton, simonmar

Differential Revision: https://phabricator.haskell.org/D82

9 files changed, 19 insertions, 11 deletions

diff --git a/testsuite/tests/ghci/scripts/ghci025.stdout b/testsuite/tests/ghci/scripts/ghci025.stdout
index e5654b3734..532a3347cc 100644
--- a/testsuite/tests/ghci/scripts/ghci025.stdout
+++ b/testsuite/tests/ghci/scripts/ghci025.stdout
@@ -56,9 +56,9 @@ Prelude.length ::
   Data.Foldable.Foldable t => forall a. t a -> GHC.Types.Int
 -- imported via T
 data T.Integer
-  = integer-gmp-0.5.1.0:GHC.Integer.Type.S# GHC.Prim.Int#
-  | integer-gmp-0.5.1.0:GHC.Integer.Type.J# GHC.Prim.Int#
-                                            GHC.Prim.ByteArray#
+  = integer-gmp-1.0.0.0:GHC.Integer.Type.S# !GHC.Prim.Int#
+  | integer-gmp-1.0.0.0:GHC.Integer.Type.Jp# {-# UNPACK #-}integer-gmp-1.0.0.0:GHC.Integer.Type.BigNat
+  | integer-gmp-1.0.0.0:GHC.Integer.Type.Jn# {-# UNPACK #-}integer-gmp-1.0.0.0:GHC.Integer.Type.BigNat
 T.length :: Data.ByteString.Internal.ByteString -> GHC.Types.Int
 :browse! T
 -- defined locally
diff --git a/testsuite/tests/lib/integer/all.T b/testsuite/tests/lib/integer/all.T
index 7b5e5f2dbe..55154265fc 100644
--- a/testsuite/tests/lib/integer/all.T
+++ b/testsuite/tests/lib/integer/all.T
@@ -1,7 +1,8 @@
 test('integerBits', normal, compile_and_run, [''])
 test('integerConversions', normal, compile_and_run, [''])
+## 'integerGmpInternals' disabled till the extra primitives are re-implemented
 # skip ghci as it doesn't support unboxed tuples
-test('integerGmpInternals', [reqlib('integer-gmp'), omit_ways('ghci')], compile_and_run, [''])
+# test('integerGmpInternals', [reqlib('integer-gmp'), omit_ways('ghci')], compile_and_run, [''])
 test('integerConstantFolding',
      [ extra_clean(['integerConstantFolding.simpl'])
      , when(compiler_debugged(), expect_broken(8525))],
diff --git a/testsuite/tests/llvm/should_compile/all.T b/testsuite/tests/llvm/should_compile/all.T
index e9154196f0..b630645f1e 100644
--- a/testsuite/tests/llvm/should_compile/all.T
+++ b/testsuite/tests/llvm/should_compile/all.T
@@ -7,7 +7,7 @@ setTestOpts(f)
 
 test('T5054', reqlib('hmatrix'), compile, ['-package hmatrix'])
 test('T5054_2', reqlib('hmatrix'), compile, ['-package hmatrix'])
-test('T5486', reqlib('integer-gmp'), compile, [''])
+# test('T5486', reqlib('integer-gmp'), compile, [''])
 test('T5681', normal, compile, [''])
 test('T6158', [reqlib('vector'), reqlib('primitive')], compile, ['-package vector -package primitive'])
 test('T7571', cmm_src, compile, [''])
diff --git a/testsuite/tests/perf/should_run/all.T b/testsuite/tests/perf/should_run/all.T
index 749e8aefe6..d8af52bbef 100644
--- a/testsuite/tests/perf/should_run/all.T
+++ b/testsuite/tests/perf/should_run/all.T
@@ -169,8 +169,9 @@ test('T5549',
      [stats_num_field('bytes allocated',
                       [(wordsize(32), 3362958676, 5),
                     # expected value: 3362958676 (Windows)
-                       (wordsize(64), 6725846120, 5)]),
+                       (wordsize(64), 8193140752, 5)]),
                     # expected value: 6725846120 (amd64/Linux)
+                    #                 8193140752 (amd64/Linux) integer-gmp2
       only_ways(['normal'])
       ],
      compile_and_run,
diff --git a/testsuite/tests/perf/space_leaks/all.T b/testsuite/tests/perf/space_leaks/all.T
index ac60c8fa58..af7eefccc5 100644
--- a/testsuite/tests/perf/space_leaks/all.T
+++ b/testsuite/tests/perf/space_leaks/all.T
@@ -4,7 +4,8 @@ test('space_leak_001',
      # Now it's: 3 (amd64/Linux)
      #           4 (x86/OS X)
      #           5 (x86/Linux)
-     [stats_num_field('peak_megabytes_allocated', (4, 1)),
+     [stats_num_field('peak_megabytes_allocated', (3, 1)),
+                        # 3 (amd64/Linux, integer-gmp2)
       stats_num_field('max_bytes_used',
           [(wordsize(64), 440000, 15),
                         # 440224 (amd64/Linux)
@@ -14,11 +15,12 @@ test('space_leak_001',
            (wordsize(32), 405650, 10)]),
              # 2013-02-10 372072 (x86/OSX)
              # 2013-02-10 439228 (x86/OSX)
-      stats_num_field('bytes allocated', (9079316016, 1)),
+      stats_num_field('bytes allocated', (11315747416, 1)),
                         # expected value: 9079316016 (amd64/Linux)
                         #                 9331570416 (x86/Linux)
                         #                 9329073952 (x86/OS X)
                         #                 9327959840 (x86/Windows)
+                        #                 11315747416 (amd64/Lnx, integer-gmp2)
       omit_ways(['profasm','profthreaded','threaded1','threaded2'])
       ],
      compile_and_run,
diff --git a/testsuite/tests/safeHaskell/check/pkg01/all.T b/testsuite/tests/safeHaskell/check/pkg01/all.T
index 604a5cc777..f121b99add 100644
--- a/testsuite/tests/safeHaskell/check/pkg01/all.T
+++ b/testsuite/tests/safeHaskell/check/pkg01/all.T
@@ -40,6 +40,7 @@ test('safePkg01',
       normalise_errmsg_fun(ignoreLdOutput),
       normalise_fun(
         normaliseArrayPackage,
+        normaliseIntegerPackage,
         normaliseBytestringPackage)],
      run_command,
      ['$MAKE -s --no-print-directory safePkg01 ' + make_args])
diff --git a/testsuite/tests/simplCore/should_run/T5603.hs b/testsuite/tests/simplCore/should_run/T5603.hs
index 635de33d6a..c1545d2c39 100644
--- a/testsuite/tests/simplCore/should_run/T5603.hs
+++ b/testsuite/tests/simplCore/should_run/T5603.hs
@@ -12,4 +12,5 @@ main = (encodeDouble 0 :: Double) `seq` return ()
 {-# INLINE encodeDouble #-}
 encodeDouble :: Integer -> Double
 encodeDouble (S# _)   = D# 3.0##
-encodeDouble (J# _ _) = D# 4.0##
+encodeDouble (Jp# _)  = D# 4.0##
+encodeDouble (Jn# _)  = D# 5.0##
diff --git a/testsuite/tests/typecheck/should_fail/T5095.stderr b/testsuite/tests/typecheck/should_fail/T5095.stderr
index 82613e64d9..701bd761d3 100644
--- a/testsuite/tests/typecheck/should_fail/T5095.stderr
+++ b/testsuite/tests/typecheck/should_fail/T5095.stderr
@@ -56,8 +56,10 @@ T5095.hs:9:11:
       instance Eq Ordering -- Defined in ‘GHC.Classes’
       instance Eq Word -- Defined in ‘GHC.Classes’
       instance Eq a => Eq [a] -- Defined in ‘GHC.Classes’
+      instance Eq integer-gmp-1.0.0.0:GHC.Integer.Type.BigNat
+        -- Defined in ‘integer-gmp-1.0.0.0:GHC.Integer.Type’
       instance Eq Integer
-        -- Defined in ‘integer-gmp-0.5.1.0:GHC.Integer.Type’
+        -- Defined in ‘integer-gmp-1.0.0.0:GHC.Integer.Type’
     (The choice depends on the instantiation of ‘a’
      To pick the first instance above, use IncoherentInstances
      when compiling the other instance declarations)
diff --git a/testsuite/tests/typecheck/should_fail/tcfail072.stderr b/testsuite/tests/typecheck/should_fail/tcfail072.stderr
index d5eb4aa87f..c9b1d10b2b 100644
--- a/testsuite/tests/typecheck/should_fail/tcfail072.stderr
+++ b/testsuite/tests/typecheck/should_fail/tcfail072.stderr
@@ -10,6 +10,6 @@ tcfail072.hs:23:13:
       instance Ord a => Ord (Maybe a) -- Defined in ‘GHC.Base’
       instance Ord () -- Defined in ‘GHC.Classes’
       instance (Ord a, Ord b) => Ord (a, b) -- Defined in ‘GHC.Classes’
-      ...plus 22 others
+      ...plus 23 others
     In the expression: g A
     In an equation for ‘g’: g (B _ _) = g A