DERP/DORP Implementation for pgpe dd2

Change-Id: Iccfab131b8704666b2a45c20d6326f2007acced9
RTC:173229
Reviewed-on: http://ralgit01.raleigh.ibm.com/gerrit1/41634
Tested-by: Jenkins Server <pfd-jenkins+hostboot@us.ibm.com>
Tested-by: PPE CI <ppe-ci+hostboot@us.ibm.com>
Reviewed-by: Michael S. Floyd <mfloyd@us.ibm.com>
Reviewed-by: Brian T. Vanderpool <vanderp@us.ibm.com>
Reviewed-by: Gregory S. Still <stillgs@us.ibm.com>
Reviewed-on: http://ralgit01.raleigh.ibm.com/gerrit1/41636
Tested-by: FSP CI Jenkins <fsp-CI-jenkins+hostboot@us.ibm.com>
Reviewed-by: Sachin Gupta <sgupta2m@in.ibm.com>

src/import/chips/p9/procedures/ppe/pk/ppe42/div32.S

@@ -5,7 +5,7 @@
 #
 # OpenPOWER sbe Project
 #
-# Contributors Listed Below - COPYRIGHT 2016
+# Contributors Listed Below - COPYRIGHT 2016,2017
 # [+] International Business Machines Corp.
 #
 #
@@ -22,6 +22,7 @@
 # permissions and limitations under the License.
 #
 # IBM_PROLOG_END_TAG
+#if (!defined(PSTATE_GPE) || NIMBUS_DD_LEVEL==1)
     .nolist
 #include "ppe42_asm.h"
     .list
@@ -78,5 +79,5 @@ __umodsi3:
     mtctr       %r10
     li          %r9,1
     b           udivmodsi4_loop
-
+#endif
 

src/import/chips/p9/procedures/ppe/pk/ppe42/math.c

@@ -5,7 +5,7 @@
 /*                                                                        */
 /* OpenPOWER sbe Project                                                  */
 /*                                                                        */
-/* Contributors Listed Below - COPYRIGHT 2016                             */
+/* Contributors Listed Below - COPYRIGHT 2016,2017                        */
 /* [+] International Business Machines Corp.                              */
 /*                                                                        */
 /*                                                                        */
@@ -37,6 +37,75 @@ extern "C"
 //}
 
 
+#ifdef PSTATE_GPE
+#if (NIMBUS_DD_LEVEL != 1)
+
+#include "ocb_register_addresses.h"
+#define out64(addr, data) \
+    {\
+        unsigned long long __d = (data); \
+        unsigned long* __a = (unsigned long*)(addr); \
+        asm volatile \
+        (\
+         "stvd %1, %0 \n" \
+         : "=o"(*__a) \
+         : "r"(__d) \
+        ); \
+    }
+
+#define in64(addr) \
+    ({\
+        unsigned long long __d; \
+        unsigned long* __a = (unsigned long*)(addr); \
+        asm volatile \
+        (\
+         "lvd %0, %1 \n" \
+         :"=r"(__d) \
+         :"o"(*__a) \
+        ); \
+        __d; \
+    })
+
+
+unsigned long udivmodsi4(unsigned long long _a,
+                         unsigned long _mod)
+{
+
+    out64(OCB_DERP, _a);
+
+    do
+    {
+        _a = in64(OCB_DORP);
+    }
+    while((~_a) == 0);
+
+    if(_mod)
+    {
+        return (unsigned long)_a;
+    }
+
+    return (unsigned long)(_a >> 32);
+}
+
+unsigned long __udivsi3(unsigned long _a, unsigned long _b)
+{
+    unsigned long long v =
+        ((unsigned long long)_a) << 32 |
+        ((unsigned long long)_b);
+
+    return udivmodsi4(v, 0);
+}
+
+unsigned long __umodsi3(unsigned long _a, unsigned long _b)
+{
+    unsigned long long v =
+        ((unsigned long long)_a) << 32 |
+        ((unsigned long long)_b);
+    return udivmodsi4(v, 1);
+}
+#endif
+#endif
+
 // 32 bit signed divide
 int __divsi3(int _a, int _b)
 {
@@ -54,7 +123,7 @@ int __divsi3(int _a, int _b)
         neg = !neg;
     }
 
-    int c = __udivsi3((unsigned long)_a, (unsigned long)_b);
+    int c = (int)__udivsi3((unsigned long)_a, (unsigned long)_b);
 
     if(neg)
     {
@@ -64,6 +133,7 @@ int __divsi3(int _a, int _b)
     return c;
 }
 
+
 // 32 bit unsigned mutiply
 unsigned long __umulsi3(unsigned long _a, unsigned long _b)
 {
@@ -139,6 +209,7 @@ unsigned long long __muldi3(unsigned long long _a, unsigned long long _b)
     return sum;
 }
 
+
 //float __mulsf3(float _a , float _b)
 //{
 //    // floating point math