BUG: Workaround segfault in Apple Accelerate framework SGEMV

SGEMV in Accelerate framework will segfault on MacOS X version 10.9
(aka Mavericks) if arrays are not aligned to 32 byte boundaries and the
CPU supports AVX instructions. This can produce segfaults in np.dot.
This patch overshadows the symbols cblas_sgemv, sgemv_ and sgemv
exported by Accelerate to produce the correct behavior. The MacOS X
version and CPU specs are checked on module import. If Mavericks and AVX
are detected the call to SGEMV is emulated with a call to SGEMM if the
arrays are not 32 byte aligned. If the exported symbols cannot be
overshadowed on module import, a fatal error is produced and the process
aborts. All the fixes are in a self-contained C file and do not alter
the multiarray C code. The patch is not applied unless NumPy is
configured to link with Apple's Accelerate framework.

1 files changed, 3 insertions, 0 deletions

diff --git a/numpy/core/setup.py b/numpy/core/setup.py
index 11b443cf8..7a82f1e35 100644
--- a/numpy/core/setup.py
+++ b/numpy/core/setup.py
@@ -12,6 +12,7 @@ from os.path import join
 from numpy.distutils import log
 from distutils.dep_util import newer
 from distutils.sysconfig import get_config_var
+from numpy.build_utils.apple_accelerate import uses_accelerate_framework, get_sgemv_fix
 
 from setup_common import *
 
@@ -838,6 +839,8 @@ def configuration(parent_package='',top_path=None):
         multiarray_src.extend([join('src', 'multiarray', 'cblasfuncs.c'),
                                join('src', 'multiarray', 'python_xerbla.c'),
                               ])
+        if uses_accelerate_framework(blas_info):
+            multiarray_src.extend(get_sgemv_fix())
     else:
         extra_info = {}