3 files changed, 75 insertions, 1 deletions

diff --git a/host/lib/convert/CMakeLists.txt b/host/lib/convert/CMakeLists.txt
index 363555f45..bec88b520 100644
--- a/host/lib/convert/CMakeLists.txt
+++ b/host/lib/convert/CMakeLists.txt
@@ -94,8 +94,11 @@ IF(CMAKE_COMPILER_IS_GNUCXX)
 ENDIF(CMAKE_COMPILER_IS_GNUCXX)
 
 IF(HAVE_ARM_NEON_H)
+    ENABLE_LANGUAGE(ASM)
+
     LIBUHD_APPEND_SOURCES(
         ${CMAKE_CURRENT_SOURCE_DIR}/convert_with_neon.cpp
+        ${CMAKE_CURRENT_SOURCE_DIR}/convert_neon.S
     )
 ENDIF()
 
diff --git a/host/lib/convert/convert_neon.S b/host/lib/convert/convert_neon.S
new file mode 100644
index 000000000..8cbe82bde
--- /dev/null
+++ b/host/lib/convert/convert_neon.S
@@ -0,0 +1,37 @@
+//
+// Copyright 2014 Ettus Research LLC
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU General Public License as published by
+// the Free Software Foundation, either version 3 of the License, or
+// (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU General Public License for more details.
+//
+// You should have received a copy of the GNU General Public License
+// along with this program.  If not, see <http://www.gnu.org/licenses/>.
+//
+
+	.arch armv7-a
+	.fpu neon
+	.syntax unified
+	.text
+	.align 2
+	.global neon_item32_sc16_swap_16n
+	.type neon_item32_sc16_swap_16n, %function
+neon_item32_sc16_swap_16n:
+.loop_swap:
+	vld2.16      {q0, q1}, [r0]!
+	vld2.16      {q2, q3}, [r0]!
+	vswp               q0, q1
+	vswp               q2, q3
+	vst2.16      {q0, q1}, [r1]!
+	vst2.16      {q2, q3}, [r1]!
+	subs               r2, #1
+	bne          .loop_swap
+	bx                 lr
+	.size neon_item32_sc16_swap_16n, .-neon_item32_sc16_swap_16n
+	.section .note.GNU-stack,"",%progbits
diff --git a/host/lib/convert/convert_with_neon.cpp b/host/lib/convert/convert_with_neon.cpp
index e994d97a6..f1c7773ec 100644
--- a/host/lib/convert/convert_with_neon.cpp
+++ b/host/lib/convert/convert_with_neon.cpp
@@ -1,5 +1,5 @@
 //
-// Copyright 2011-2012 Ettus Research LLC
+// Copyright 2011-2014 Ettus Research LLC
 //
 // This program is free software: you can redistribute it and/or modify
 // it under the terms of the GNU General Public License as published by
@@ -19,6 +19,12 @@
 #include <uhd/utils/byteswap.hpp>
 #include <arm_neon.h>
 
+extern "C" {
+void neon_item32_sc16_swap_16n(void *, void *, int iter);
+}
+
+static const int SIMD_WIDTH = 16;
+
 using namespace uhd::convert;
 
 DECLARE_CONVERTER(fc32, 1, sc16_item32_le, 1, PRIORITY_SIMD){
@@ -58,3 +64,31 @@ DECLARE_CONVERTER(sc16_item32_le, 1, fc32, 1, PRIORITY_SIMD){
 
     item32_sc16_to_xx<uhd::htowx>(input+i, output+i, nsamps-i, scale_factor);
 }
+
+DECLARE_CONVERTER(sc16, 1, sc16_item32_le, 1, PRIORITY_SIMD){
+    const sc16_t *input = reinterpret_cast<const sc16_t *>(inputs[0]);
+    item32_t *output = reinterpret_cast<item32_t *>(outputs[0]);
+
+    size_t i = nsamps / SIMD_WIDTH;
+
+    if (i)
+        neon_item32_sc16_swap_16n((void *) input, (void *) output, i);
+
+    i *= SIMD_WIDTH;
+
+    xx_to_item32_sc16<uhd::htowx>(input+i, output+i, nsamps-i, scale_factor);
+}
+
+DECLARE_CONVERTER(sc16_item32_le, 1, sc16, 1, PRIORITY_SIMD){
+    const item32_t *input = reinterpret_cast<const item32_t *>(inputs[0]);
+    sc16_t *output = reinterpret_cast<sc16_t *>(outputs[0]);
+
+    size_t i = nsamps / SIMD_WIDTH;
+
+    if (i)
+        neon_item32_sc16_swap_16n((void *) input, (void *) output, i);
+
+    i *= SIMD_WIDTH;
+
+    item32_sc16_to_xx<uhd::wtohx>(input+i, output+i, nsamps-i, scale_factor);
+}