We had started relying on GMP 5.x (for usage of mpz_powm_sec), but this
is pretty painful on RHEL-esque targets, which still use GMP 4.x.

In the mean time while we're still supporting this, it's easier to just
fallback to mpz_powm when _sec is unavailable, and emit a WARNING for
using the primitive.

This also installs a header, HsIntegerGmp.h, which clients could use for
a fallback.

As a side note, this will probably also help Debian oldstable users who
might have outdated GMP versions (which I believe is the cause for #8666.)

Reviewed-by: Herbert Valerio Riedel <hvr@gnu.org>
Signed-off-by: Austin Seipp <austin@well-typed.com>

High-level pseudo code of what the code was supposed to implement:

    quotRem' :: Integer -> Integer -> (Integer,Integer)
    quotRem' a b@(S# _)
      | b < 0     = negFst . uncurry quotRem' . negSnd $ (a,b)
      | otherwise = quotRemUI a (fromIntegral (abs b))

    divMod' :: Integer -> Integer -> (Integer,Integer)
    divMod' a b@(S# _)
      | b < 0      = negSnd . uncurry divMod' . negBoth $ (a,b)
      | otherwise  = divModUI a (fromIntegral b)

    negFst  (q,r) = (-q,r)
    negSnd  (q,r) = ( q,-r)
    negBoth (q,r) = (-q,-r)

    -- quotRemUI and divModUI represent GMP's `mpz_{f,t}div_qr_ui()`
    quotRemUI, divModUI :: Integer -> Word -> (Integer,Integer)

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

Herbert, this is a merge candidate to the 7.8 branch

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

Among other things, this unhides `GHC.Integer` and re-groups the export list.
Moreover, the internal representation of `Integer` is explained a bit more,
and `/Since: 0.5.1.0/` annotations have been added.

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

This is due to `mpz_*()` functions having @long@ arguments which are
32bit on IL32P64, whereas `Int#` and `Word#` are 64bit wide, causing all
sorts of malfunction due to truncation.

This affects mostly the new big/small-int primops introduced in the
course of #8647, so when `SIZEOF_W != SIZEOF_LONG` we simply fall back
to using the big/big-int primops. big/small primops implemented via the
low-level `mpn_*()` GMP operations are not affected, as those use
`mp_limb_t` arguments.

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

This allows to simplify code in `float.c` and to encode the
"sizeof(mp_limb_t) != sizeof(W_)" assumption as a compile-time
assertion.

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

I'm a bit surprised though this didn't cause linkage errors

(the incorrect import statement was introduced in 7bdcadda)

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

Signed-off-by: Austin Seipp <austin@well-typed.com>

This is similiar to what has been done in [af2ba9c8/integer-gmp] for
`gmpz_tdiv_{q,r,qr}_ui` (re #8647); However, the gain is more modest
here, as performance-conscious code tends to use `quot`/`rem` rather
than `div`/`mod`:

     Program    Size    Allocs   Runtime   Elapsed  TotalMem
 -------------------------------------------------------------
   primetest   +0.3%     -2.4%      0.06      0.06     +0.0%
         rsa   +0.2%     -3.3%      0.02      0.02     +0.0%

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

We now allocate a 1-limb mpz_t on the stack instead of doing a more
expensive heap-allocation (especially if the heap-allocated copy becomes
garbage right away); this addresses #8647.

In order to delay heap allocations of 1-limb `ByteArray#`s instead of
the previous `(# Int#, ByteArray# #)` pair, a 3-tuple
`(# Int#, ByteArray#, Word# #)` is returned now. This tuple is given the
type-synonym `MPZ#`.

This 3-tuple representation uses either the 1st and the 2nd element, or
the 1st and the 3rd element to represent the limb(s) (NB: undefined
`ByteArray#` elements must not be accessed as they don't point to a
proper `ByteArray#`, see also `DUMMY_BYTE_ARR`); more specifically, the
following encoding is used (where `⊥` means undefined/unused):

 -  (#  0#, ⊥, 0## #) -> value = 0
 -  (#  1#, ⊥, w   #) -> value = w
 -  (# -1#, ⊥, w   #) -> value = -w
 -  (#  s#, d, 0## #) -> value = J# s d

The `mpzToInteger` helper takes care of converting `MPZ#` into an
`Integer`, and allocating a 1-limb `ByteArray#` in case the
value (`w`/`-w`) doesn't fit the `S# Int#` representation).

The following nofib benchmarks benefit from this optimization:

        Program      Size    Allocs   Runtime   Elapsed  TotalMem
 ------------------------------------------------------------------
     bernouilli     +0.2%     -5.2%      0.12      0.12     +0.0%
         gamteb     +0.2%     -1.7%      0.03      0.03     +0.0%
          kahan     +0.3%    -13.2%      0.17      0.17     +0.0%
         mandel     +0.2%    -24.6%      0.04      0.04     +0.0%
          power     +0.2%     -2.6%     -2.0%     -2.0%     -8.3%
      primetest     +0.1%    -17.3%      0.06      0.06     +0.0%
            rsa     +0.2%    -18.5%      0.02      0.02     +0.0%
            scs     +0.1%     -2.9%     -0.1%     -0.1%     +0.0%
         sphere     +0.3%     -0.8%      0.03      0.03     +0.0%
         symalg     +0.2%     -3.1%      0.01      0.01     +0.0%
 ------------------------------------------------------------------
            Min     +0.1%    -24.6%     -4.6%     -4.6%     -8.3%
            Max     +0.3%     +0.0%     +5.9%     +5.9%     +4.5%
 Geometric Mean     +0.2%     -1.0%     +0.2%     +0.2%     -0.0%

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

This fixes the actual cause for #8661, i.e. a mismatch between the
actual arity of the Cmm implementation and the arity declared in the
foreign import statement.

This also reverts [a3878d17/integer-gmp] as the workaround isn't needed
anymore.

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

Forgot to add this chunk to the commit [a3878d17/integer-gmp]

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

I still need to investigated, but for some reason not yet obvious to me,
commit [af2ba9c8/integer-gmp] (re #8647) seems to have triggered #8661
on linux/32

This commit disables the use of the `quotRemIntegerWord#` primop on
32bit (which seems to trigger the issue).

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

This avoids allocating this special value over and over again every
time it's needed, and therefore this addresses #8647.

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

This is useful as `quot`/`rem` are often used with small-int divisors,
like when computing the digits of an `Integer`. This optimization
reduces allocations in the following `nofib` benchmarks:

      Program        Size    Allocs   Runtime   Elapsed  TotalMem
   -----------------------------------------------------------------
        power       +0.3%     -0.8%     -1.2%     -1.2%     +0.0%
    primetest       +0.3%     -3.9%      0.07      0.07     +0.0%
          rsa       +0.3%     -4.0%      0.02      0.02     +0.0%
       symalg       +0.2%     -1.4%      0.01      0.01     +0.0%

This addresses #8647

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

See the comments and #8102. The basic gist of it seems to be that the
build system follows an implied rule from somewhere to directly build a
C file, which doesn't have a dependency on the in-tree gmp.h that we
build. As a result, the C file compilation races against the GMP build,
causing an error.

This is a pretty unsatisfactory hack, but for Windows and OS X machines
where we more often build in-tree GMPs, it's quite important.

Authored-by: Kazu Yamamoto <kazu@iij.ad.jp>
Signed-off-by: Austin Seipp <austin@well-typed.com>

This adds `{plus,minus}IntegerInt#` which help to reduce temporary
allocations in `plusInteger` and `minusInteger`.

This and the previous commit introducing `timesIntegerInt#` (i.e. baeeef7a)
result in reduced allocations for the following nofib benchmarks on Linux/amd64:

         Program      Size    Allocs   Runtime   Elapsed  TotalMem
  ------------------------------------------------------------------
      bernouilli     +0.0%     -4.2%      0.12      0.12     +0.0%
           kahan     +0.1%    -12.6%      0.17      0.17     +0.0%
        pidigits     +0.0%     -0.5%     -4.7%     -4.5%     +0.0%
           power     +0.0%     -2.7%     +3.1%     +3.1%     +9.1%
       primetest     +0.0%     -4.2%      0.07      0.07     +0.0%
             rsa     +0.0%     -4.1%      0.02      0.02     +0.0%
             scs     +0.0%     -2.6%     -0.8%     -0.7%     +0.0%
  ------------------------------------------------------------------
             Min     +0.0%    -12.6%     -4.7%     -4.5%     -5.0%
             Max     +0.1%     +0.2%     +3.1%     +3.1%     +9.1%
  Geometric Mean     +0.1%     -0.3%     -0.0%     +0.0%     +0.1%
  ------------------------------------------------------------------

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

This primop helps reducing allocation by being able to pass one `S#`
argument directly to the GMP multiplication primitive without needing to
promote (and thus allocate a `ByteArray#` as well) the `J#` first.

This benefits a few nofib benchmarks wrt to allocations (having most
impact on `kahan` resulting in about 10% less allocations)

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

This factors out the recurring task of converting mpz_t structures
to/from Int#/ByteArrays# pairs and makes the code more readable.

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

GCC is able to generate better code when using `memcpy` instead of
manually copying bytes in a loop. Otoh, `stgAllocForGMP` is typically
called for enlarging initial single-limb structures (see also #8647 for
more information) and so this minor optimization won't be very visible
in measurements.

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

Otoh, `divModInt#` is not a proper primop (it's implemented as wrapper
around `quotRemInt#` in `GHC.Base`), so we can't do the same for
`divModInteger` yet.

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

This should improve the check for a suitable installed GMP version, to
implicitly check if GMP provides __gmpz_powm_sec() which has been added
in GMP version 5.0.0 and switch to the in-tree GMP 5.0.3 source otherwise.

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

These are supposed to be called with `J#`-kind `Integer`s, so check that
constructor first.

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

These follow closely the existing implementations for
`importIntegerFromByteArray` and `exportIntegerToMutableByteArray`.

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

This renames to more verbose names which include the type these
operations import/export from/to:

 - `importIntegerFromByteArray`, and
 - `exportIntegerToMutableByteArray`.

This follows the naming convention used for other primitive operations,
such as the recently added `copyMutableByteArrayToAddr` operation.

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

This is a follow-up to e94799c9 fixing the Cmm implementation of the
primops based on suggestions by Duncan Coutts.

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

This adds the following new (internal) primitives

{{{#!hs
sizeInBaseInteger :: Integer -> Int# -> Word#

exportInteger :: Integer -> MutableByteArray# s -> Word# -> Int#
                 -> State# s -> (# State# s, Word# #)

importInteger :: ByteArray# -> Word# -> Word# -> Int# -> Integer
}}}

The import/export primitives support selecting most/least significant
byte first order as well as using an offset into the byte-arrays.

See Haddock comments for more details.

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

This exposes `mpz_probab_prime_p()` and `mpz_nextprime()` as
`testPrimeInteger` and `nextPrimeInteger` respectively and is especially
useful for cryptographic algorithms such as RSA.

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

This is a follow-up to 97c101b7 which
introduced the "ordinary" `powModInteger` operation.

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

This sets a sensible cabal category (i.e. `Numerical`), extends
`extra-tmp-{files,files}` to make this package self-contained, updates
the bug-report URL, and cleans up the `{-# LANGUAGE #-}` pragma usage in
the source code.

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

These files are taken from GNU automake 1.13.3

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

The extended GCD computation is useful to have for implementing
algorithms such as the chinese reminder theorem.

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>

This exposes the GMP functions `mpz_pow_ui()`, `mpz_powm()`, and
`mpz_invert()` as `powInteger`, `powModInteger`, and `recipModInteger`
respectively in the module `GHC.Integer.GMP.Internals`.

Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>