zepto-forks/meta-retro
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

mame: update to 0178

Browse Source 
Signed-off-by: Andreas Müller <schnitzeltony@googlemail.com>
This commit is contained in:
Andreas Müller
2016-10-19 01:16:16 +02:00
parent fe0ea15c2f
commit f91f9c08bf
3 changed files with 2 additions and 370 deletions
Download Patch File Download Diff File Expand all files Collapse all files

327
recipes-emulators/mame/files/0002-float4_neon.h-refactor-buildins-for-later-gcc-s.patch
View File

@@ -1,327 +0,0 @@
From 917f9e35b110c84b7748e47d7cf27097c04f8e49 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andreas=20M=C3=BCller?= <schnitzeltony@googlemail.com>
Date: Fri, 1 Jan 2016 20:48:10 +0100
Subject: [PATCH] float4_neon.h: refactor buildins for later gcc's
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
taken from [1-2]
[1] https://patchwork.ozlabs.org/patch/410050/
[2] https://gcc.gnu.org/ml/gcc-patches/2014-04/txtfStfgLNxV5.txt
Upstream-Status: Pending
Signed-off-by: Andreas Müller <schnitzeltony@googlemail.com>
---
3rdparty/bx/include/bx/float4_neon.h | 74 ++++++++++++++++++++++++------------------------
1 file changed, 37 insertions(+), 37 deletions(-)
diff --git a/3rdparty/bx/include/bx/float4_neon.h b/3rdparty/bx/include/bx/float4_neon.h
index 3b6fa18..df42814 100644
--- a/3rdparty/bx/include/bx/float4_neon.h
+++ b/3rdparty/bx/include/bx/float4_neon.h
@@ -96,22 +96,22 @@ IMPLEMENT_TEST(xyzw , xyzw);
BX_FLOAT4_FORCE_INLINE float float4_x(float4_t _a)
{
- return __builtin_neon_vget_lanev4sf(_a, 0, 3);
+ return __builtin_neon_vget_lanev4sf(_a, 0);
}
BX_FLOAT4_FORCE_INLINE float float4_y(float4_t _a)
{
- return __builtin_neon_vget_lanev4sf(_a, 1, 3);
+ return __builtin_neon_vget_lanev4sf(_a, 1);
}
BX_FLOAT4_FORCE_INLINE float float4_z(float4_t _a)
{
- return __builtin_neon_vget_lanev4sf(_a, 2, 3);
+ return __builtin_neon_vget_lanev4sf(_a, 2);
}
BX_FLOAT4_FORCE_INLINE float float4_w(float4_t _a)
{
- return __builtin_neon_vget_lanev4sf(_a, 3, 3);
+ return __builtin_neon_vget_lanev4sf(_a, 3);
}
BX_FLOAT4_FORCE_INLINE float4_t float4_ld(const void* _ptr)
@@ -179,14 +179,14 @@ IMPLEMENT_TEST(xyzw , xyzw);
BX_FLOAT4_FORCE_INLINE float4_t float4_itof(float4_t _a)
{
const _i32x4_t itof = __builtin_neon_vreinterpretv4siv4sf(_a);
- const float4_t result = __builtin_neon_vcvtv4si(itof, 1);
+ const float4_t result = __builtin_neon_vcvtsv4si(itof);
return result;
}
BX_FLOAT4_FORCE_INLINE float4_t float4_ftoi(float4_t _a)
{
- const _i32x4_t ftoi = __builtin_neon_vcvtv4sf(_a, 1);
+ const _i32x4_t ftoi = __builtin_neon_vcvtsv4sf(_a);
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(ftoi);
return result;
@@ -194,32 +194,32 @@ IMPLEMENT_TEST(xyzw , xyzw);
BX_FLOAT4_FORCE_INLINE float4_t float4_add(float4_t _a, float4_t _b)
{
- return __builtin_neon_vaddv4sf(_a, _b, 3);
+ return __builtin_neon_vaddv4sf(_a, _b);
}
BX_FLOAT4_FORCE_INLINE float4_t float4_sub(float4_t _a, float4_t _b)
{
- return __builtin_neon_vsubv4sf(_a, _b, 3);
+ return __builtin_neon_vsubv4sf(_a, _b);
}
BX_FLOAT4_FORCE_INLINE float4_t float4_mul(float4_t _a, float4_t _b)
{
- return __builtin_neon_vmulv4sf(_a, _b, 3);
+ return __builtin_neon_vmulfv4sf(_a, _b);
}
BX_FLOAT4_FORCE_INLINE float4_t float4_rcp_est(float4_t _a)
{
- return __builtin_neon_vrecpev4sf(_a, 3);
+ return __builtin_neon_vrecpev4sf(_a);
}
BX_FLOAT4_FORCE_INLINE float4_t float4_rsqrt_est(float4_t _a)
{
- return __builtin_neon_vrsqrtev4sf(_a, 3);
+ return __builtin_neon_vrsqrtev4sf(_a);
}
BX_FLOAT4_FORCE_INLINE float4_t float4_cmpeq(float4_t _a, float4_t _b)
{
- const _i32x4_t tmp = __builtin_neon_vceqv4sf(_a, _b, 3);
+ const _i32x4_t tmp = __builtin_neon_vceqv4sf(_a, _b);
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp);
return result;
@@ -227,7 +227,7 @@ IMPLEMENT_TEST(xyzw , xyzw);
BX_FLOAT4_FORCE_INLINE float4_t float4_cmplt(float4_t _a, float4_t _b)
{
- const _i32x4_t tmp = __builtin_neon_vcgtv4sf(_b, _a, 3);
+ const _i32x4_t tmp = __builtin_neon_vcgtv4sf(_b, _a);
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp);
return result;
@@ -235,7 +235,7 @@ IMPLEMENT_TEST(xyzw , xyzw);
BX_FLOAT4_FORCE_INLINE float4_t float4_cmple(float4_t _a, float4_t _b)
{
- const _i32x4_t tmp = __builtin_neon_vcgev4sf(_b, _a, 3);
+ const _i32x4_t tmp = __builtin_neon_vcgev4sf(_b, _a);
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp);
return result;
@@ -243,7 +243,7 @@ IMPLEMENT_TEST(xyzw , xyzw);
BX_FLOAT4_FORCE_INLINE float4_t float4_cmpgt(float4_t _a, float4_t _b)
{
- const _i32x4_t tmp = __builtin_neon_vcgtv4sf(_a, _b, 3);
+ const _i32x4_t tmp = __builtin_neon_vcgtv4sf(_a, _b);
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp);
return result;
@@ -251,7 +251,7 @@ IMPLEMENT_TEST(xyzw , xyzw);
BX_FLOAT4_FORCE_INLINE float4_t float4_cmpge(float4_t _a, float4_t _b)
{
- const _i32x4_t tmp = __builtin_neon_vcgev4sf(_a, _b, 3);
+ const _i32x4_t tmp = __builtin_neon_vcgev4sf(_a, _b);
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp);
return result;
@@ -259,19 +259,19 @@ IMPLEMENT_TEST(xyzw , xyzw);
BX_FLOAT4_FORCE_INLINE float4_t float4_min(float4_t _a, float4_t _b)
{
- return __builtin_neon_vminv4sf(_a, _b, 3);
+ return __builtin_neon_vminfv4sf(_a, _b);
}
BX_FLOAT4_FORCE_INLINE float4_t float4_max(float4_t _a, float4_t _b)
{
- return __builtin_neon_vmaxv4sf(_a, _b, 3);
+ return __builtin_neon_vmaxfv4sf(_a, _b);
}
BX_FLOAT4_FORCE_INLINE float4_t float4_and(float4_t _a, float4_t _b)
{
const _i32x4_t tmp0 = __builtin_neon_vreinterpretv4siv4sf(_a);
const _i32x4_t tmp1 = __builtin_neon_vreinterpretv4siv4sf(_b);
- const _i32x4_t tmp2 = __builtin_neon_vandv4si(tmp0, tmp1, 0);
+ const _i32x4_t tmp2 = tmp0 & tmp1;
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp2);
return result;
@@ -281,7 +281,7 @@ IMPLEMENT_TEST(xyzw , xyzw);
{
const _i32x4_t tmp0 = __builtin_neon_vreinterpretv4siv4sf(_a);
const _i32x4_t tmp1 = __builtin_neon_vreinterpretv4siv4sf(_b);
- const _i32x4_t tmp2 = __builtin_neon_vbicv4si(tmp0, tmp1, 0);
+ const _i32x4_t tmp2 = tmp0 & ~tmp1;
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp2);
return result;
@@ -291,7 +291,7 @@ IMPLEMENT_TEST(xyzw , xyzw);
{
const _i32x4_t tmp0 = __builtin_neon_vreinterpretv4siv4sf(_a);
const _i32x4_t tmp1 = __builtin_neon_vreinterpretv4siv4sf(_b);
- const _i32x4_t tmp2 = __builtin_neon_vorrv4si(tmp0, tmp1, 0);
+ const _i32x4_t tmp2 = tmp0 | tmp1;
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp2);
return result;
@@ -301,7 +301,7 @@ IMPLEMENT_TEST(xyzw , xyzw);
{
const _i32x4_t tmp0 = __builtin_neon_vreinterpretv4siv4sf(_a);
const _i32x4_t tmp1 = __builtin_neon_vreinterpretv4siv4sf(_b);
- const _i32x4_t tmp2 = __builtin_neon_veorv4si(tmp0, tmp1, 0);
+ const _i32x4_t tmp2 = tmp0 ^ tmp1;
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp2);
return result;
@@ -312,7 +312,7 @@ IMPLEMENT_TEST(xyzw , xyzw);
if (__builtin_constant_p(_count) )
{
const _i32x4_t tmp0 = __builtin_neon_vreinterpretv4siv4sf(_a);
- const _i32x4_t tmp1 = __builtin_neon_vshl_nv4si(tmp0, _count, 0);
+ const _i32x4_t tmp1 = __builtin_neon_vshl_nv4si(tmp0, _count);
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp1);
return result;
@@ -320,7 +320,7 @@ IMPLEMENT_TEST(xyzw , xyzw);
const _i32x4_t tmp0 = __builtin_neon_vreinterpretv4siv4sf(_a);
const _i32x4_t shift = __builtin_neon_vdup_nv4si( (__builtin_neon_si)_count);
- const _i32x4_t tmp1 = __builtin_neon_vshlv4si(tmp0, shift, 1);
+ const _i32x4_t tmp1 = __builtin_neon_vshlsv4si(tmp0, shift);
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp1);
return result;
@@ -331,7 +331,7 @@ IMPLEMENT_TEST(xyzw , xyzw);
if (__builtin_constant_p(_count) )
{
const _i32x4_t tmp0 = __builtin_neon_vreinterpretv4siv4sf(_a);
- const _i32x4_t tmp1 = __builtin_neon_vshr_nv4si(tmp0, _count, 0);
+ const _i32x4_t tmp1 = __builtin_neon_vshru_nv4si(tmp0, _count);
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp1);
return result;
@@ -339,7 +339,7 @@ IMPLEMENT_TEST(xyzw , xyzw);
const _i32x4_t tmp0 = __builtin_neon_vreinterpretv4siv4sf(_a);
const _i32x4_t shift = __builtin_neon_vdup_nv4si( (__builtin_neon_si)-_count);
- const _i32x4_t tmp1 = __builtin_neon_vshlv4si(tmp0, shift, 1);
+ const _i32x4_t tmp1 = __builtin_neon_vshlsv4si(tmp0, shift);
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp1);
return result;
@@ -350,7 +350,7 @@ IMPLEMENT_TEST(xyzw , xyzw);
if (__builtin_constant_p(_count) )
{
const _i32x4_t tmp0 = __builtin_neon_vreinterpretv4siv4sf(_a);
- const _i32x4_t tmp1 = __builtin_neon_vshr_nv4si(tmp0, _count, 1);
+ const _i32x4_t tmp1 = __builtin_neon_vshrs_nv4si(tmp0, _count);
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp1);
return result;
@@ -358,7 +358,7 @@ IMPLEMENT_TEST(xyzw , xyzw);
const _i32x4_t tmp0 = __builtin_neon_vreinterpretv4siv4sf(_a);
const _i32x4_t shift = __builtin_neon_vdup_nv4si( (__builtin_neon_si)-_count);
- const _i32x4_t tmp1 = __builtin_neon_vshlv4si(tmp0, shift, 1);
+ const _i32x4_t tmp1 = __builtin_neon_vshlsv4si(tmp0, shift);
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp1);
return result;
@@ -366,19 +366,19 @@ IMPLEMENT_TEST(xyzw , xyzw);
BX_FLOAT4_FORCE_INLINE float4_t float4_madd(float4_t _a, float4_t _b, float4_t _c)
{
- return __builtin_neon_vmlav4sf(_c, _a, _b, 3);
+ return __builtin_neon_vmlav4sf(_c, _a, _b);
}
BX_FLOAT4_FORCE_INLINE float4_t float4_nmsub(float4_t _a, float4_t _b, float4_t _c)
{
- return __builtin_neon_vmlsv4sf(_c, _a, _b, 3);
+ return __builtin_neon_vmlsv4sf(_c, _a, _b);
}
BX_FLOAT4_FORCE_INLINE float4_t float4_icmpeq(float4_t _a, float4_t _b)
{
const _i32x4_t tmp0 = __builtin_neon_vreinterpretv4siv4sf(_a);
const _i32x4_t tmp1 = __builtin_neon_vreinterpretv4siv4sf(_b);
- const _i32x4_t tmp2 = __builtin_neon_vceqv4si(tmp0, tmp1, 1);
+ const _i32x4_t tmp2 = __builtin_neon_vceqv4si(tmp0, tmp1);
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp2);
return result;
@@ -388,7 +388,7 @@ IMPLEMENT_TEST(xyzw , xyzw);
{
const _i32x4_t tmp0 = __builtin_neon_vreinterpretv4siv4sf(_a);
const _i32x4_t tmp1 = __builtin_neon_vreinterpretv4siv4sf(_b);
- const _i32x4_t tmp2 = __builtin_neon_vcgtv4si(tmp1, tmp0, 1);
+ const _i32x4_t tmp2 = __builtin_neon_vcgtv4si(tmp1, tmp0);
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp2);
return result;
@@ -398,7 +398,7 @@ IMPLEMENT_TEST(xyzw , xyzw);
{
const _i32x4_t tmp0 = __builtin_neon_vreinterpretv4siv4sf(_a);
const _i32x4_t tmp1 = __builtin_neon_vreinterpretv4siv4sf(_b);
- const _i32x4_t tmp2 = __builtin_neon_vcgtv4si(tmp0, tmp1, 1);
+ const _i32x4_t tmp2 = __builtin_neon_vcgtv4si(tmp0, tmp1);
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp2);
return result;
@@ -408,7 +408,7 @@ IMPLEMENT_TEST(xyzw , xyzw);
{
const _i32x4_t tmp0 = __builtin_neon_vreinterpretv4siv4sf(_a);
const _i32x4_t tmp1 = __builtin_neon_vreinterpretv4siv4sf(_b);
- const _i32x4_t tmp2 = __builtin_neon_vminv4si(tmp0, tmp1, 1);
+ const _i32x4_t tmp2 = __builtin_neon_vminsv4si(tmp0, tmp1);
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp2);
return result;
@@ -418,7 +418,7 @@ IMPLEMENT_TEST(xyzw , xyzw);
{
const _i32x4_t tmp0 = __builtin_neon_vreinterpretv4siv4sf(_a);
const _i32x4_t tmp1 = __builtin_neon_vreinterpretv4siv4sf(_b);
- const _i32x4_t tmp2 = __builtin_neon_vmaxv4si(tmp0, tmp1, 1);
+ const _i32x4_t tmp2 = __builtin_neon_vmaxsv4si(tmp0, tmp1);
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp2);
return result;
@@ -428,7 +428,7 @@ IMPLEMENT_TEST(xyzw , xyzw);
{
const _i32x4_t tmp0 = __builtin_neon_vreinterpretv4siv4sf(_a);
const _i32x4_t tmp1 = __builtin_neon_vreinterpretv4siv4sf(_b);
- const _i32x4_t tmp2 = __builtin_neon_vaddv4si(tmp0, tmp1, 1);
+ const _i32x4_t tmp2 = tmp0 + tmp1;
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp2);
return result;
@@ -438,7 +438,7 @@ IMPLEMENT_TEST(xyzw , xyzw);
{
const _i32x4_t tmp0 = __builtin_neon_vreinterpretv4siv4sf(_a);
const _i32x4_t tmp1 = __builtin_neon_vreinterpretv4siv4sf(_b);
- const _i32x4_t tmp2 = __builtin_neon_vsubv4si(tmp0, tmp1, 1);
+ const _i32x4_t tmp2 = tmp0 - tmp1;
const float4_t result = __builtin_neon_vreinterpretv4sfv4si(tmp2);
return result;
--
2.5.0

39
recipes-emulators/mame/files/0003-bx-float4_neon.h-Remove-neon_reinterpret-use-casts.patch
View File

@@ -1,39 +0,0 @@
From b3db2aa6b762711015f2d504006b44e698a4de0f Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andreas=20M=C3=BCller?= <schnitzeltony@googlemail.com>
Date: Sun, 31 Jul 2016 01:36:52 +0200
Subject: [PATCH] bx/float4_neon.h: Remove neon_reinterpret, use casts
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
See gcc6;
commit da08cb5504f0b791ada8889b16620055672999b8
Author: alalaw01 <alalaw01@138bc75d-0d04-0410-961f-82ee72b054a4>
Date: Thu Feb 4 14:49:02 2016 +0000
[ARM] Remove neon_reinterpret, use casts
Signed-off-by: Andreas Müller <schnitzeltony@googlemail.com>
---
3rdparty/bx/include/bx/float4_neon.h | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/3rdparty/bx/include/bx/float4_neon.h b/3rdparty/bx/include/bx/float4_neon.h
index c3545d0..811c4ff 100644
--- a/3rdparty/bx/include/bx/float4_neon.h
+++ b/3rdparty/bx/include/bx/float4_neon.h
@@ -6,6 +6,10 @@
#ifndef BX_FLOAT4_NEON_H_HEADER_GUARD
#define BX_FLOAT4_NEON_H_HEADER_GUARD
+/* gcc6 hacks */
+#define __builtin_neon_vreinterpretv4sfv4si(a) ((float4_t)a)
+#define __builtin_neon_vreinterpretv4siv4sf(a) ((_i32x4_t)a)
+
namespace bx
{
typedef __builtin_neon_sf float4_t __attribute__( (__vector_size__(16) ) );
--
2.5.5

6
recipes-emulators/mame/mame_0176.bb → recipes-emulators/mame/mame_0178.bb
View File

@@ -7,11 +7,9 @@ LIC_FILES_CHKSUM = "file://LICENSE.md;md5=89436197a991695e278e47475b4ff7ae"
SRC_URI = " \
https://github.com/mamedev/mame/archive/${BPN}${PV}.tar.gz \
file://0001-use-pkg-config-for-finding-sdl-library-settings.patch \
file://0002-float4_neon.h-refactor-buildins-for-later-gcc-s.patch \
file://0003-bx-float4_neon.h-Remove-neon_reinterpret-use-casts.patch \
"
SRC_URI[md5sum] = "b2aed655e1ea2bccf26bf96f189257ba"
SRC_URI[sha256sum] = "e8837ae9c21ac6ca289c0214747a6d7ff7cc4683b9426377f42cda318fddcb25"
SRC_URI[md5sum] = "89ee2eeb14a990f4d8aeddf6b835a4d5"
SRC_URI[sha256sum] = "e48df1fbdd8e2a5c3ad87bf3cfa793c619ebbabc2462f5a5569c0f1d29ef0d37"
S = "${WORKDIR}/${BPN}-${BPN}${PV}"