Fix erroneous input mutation in linalg routines

[stuartarchibald]

This fixes an issue where inputs were not correctly being copied on the way into linear algebra routines that destroy data during computation. The issue stemmed from np.asfortranarray being called on 'A' ordered arrays, and, in the case that they were already 'F' ordered at run time, no copy would be made. The following implements a new helper function to ensure that 'A' ordered arrays have an appropriate copy made at runtime, fixes up the implementations that were incorrect to now use this, and adds tests for functions that would hit these code paths.

Fixes #5870
Fixes #3368
Fixes #5933

[sklam]

Can you add testing to _copy_to_fortran_order?

for example:

diff --git a/numba/tests/test_linalg.py b/numba/tests/test_linalg.py
index 235ff8ad9..61cc768e3 100644
--- a/numba/tests/test_linalg.py
+++ b/numba/tests/test_linalg.py
@@ -8,7 +8,7 @@ import platform
 
 import numpy as np
 
-from numba import jit
+from numba import jit, njit, typeof
 from numba.core import errors
 from numba.tests.support import (TestCase, tag, needs_lapack, needs_blas,
                                  _is_armv7l, skip_ppc64le_issue4026)
@@ -1577,8 +1577,8 @@ class TestLinalgLstsq(TestLinalgSystems):
         empties = [
         [(0, 1), (1,)], # empty A, valid b
         [(1, 0), (1,)], # empty A, valid b
-        [(1, 1), (0,)], # valid A, empty 1D b 
-        [(1, 1), (1, 0)],  # valid A, empty 2D b 
+        [(1, 1), (0,)], # valid A, empty 1D b
+        [(1, 1), (1, 0)],  # valid A, empty 2D b
         ]
 
         for A, b in empties:
@@ -2646,5 +2646,46 @@ class TestBasics(TestLinalgSystems):  # TestLinalgSystems for 1d test
         self.assert_error(cfunc, args, msg, err=errors.TypingError)
 
 
+class TestHelpers(TestCase):
+    def test_copy_to_fortran_order(self):
+        from numba.np.linalg import _copy_to_fortran_order
+
+        def check(udt, expectfn, shapes, dtypes, orders):
+            for shape, dtype, order in product(shapes, dtypes, orders):
+                a = np.arange(np.prod(shape)).reshape(shape, order=order)
+
+                r = udt(a)
+                # check correct operation
+                self.assertPreciseEqual(expectfn(a), r)
+                # check new copy has made
+                self.assertNotEqual(a.ctypes.data, r.ctypes.data)
+
+        @njit
+        def direct_call(a):
+            return _copy_to_fortran_order(a)
+
+        shapes = [(3, 4), (3, 2, 5)]
+        dtypes = [np.intp]
+        orders = ['C', 'F']
+        check(direct_call, np.asfortranarray, shapes, dtypes, orders)
+
+
+        @njit
+        def slice_to_any(a):
+            # make a 'any' layout slice
+            sliced = a[::2][0]
+            return _copy_to_fortran_order(sliced)
+
+        shapes = [(3, 3, 4), (3, 3, 2, 5)]
+        dtypes = [np.intp]
+        orders = ['C', 'F']
+
+        def expected_slice_to_any(a):
+            # make a 'any' layout slice
+            sliced = a[::2][0]
+            return np.asfortranarray(sliced)
+
+        check(slice_to_any, expected_slice_to_any, shapes, dtypes, orders)
+
 if __name__ == '__main__':
     unittest.main()

[sklam]

Everything else looks good.

[stuartarchibald]

Thanks for the review, patch suggested here #5879 (comment) is applied as afa20e0.

[sklam]

Thanks for the patch!