diff --git a/include/JSystem/J3DGraphAnimator/J3DModel.h b/include/JSystem/J3DGraphAnimator/J3DModel.h
index a6537a1bf2b..99c75eec7df 100644
--- a/include/JSystem/J3DGraphAnimator/J3DModel.h
+++ b/include/JSystem/J3DGraphAnimator/J3DModel.h
@@ -24,6 +24,19 @@ struct J3DUnkCalc2 {
     virtual void calc(J3DModelData* mpModelData);
 };
 
+struct UserArea {
+    u8 field_0x00[0x60e];
+    s16 field_0x60e;
+    s16 field_0x610;
+    s16 field_0x612;
+    u8 field_0x614[0x14];
+    s16 field_0x628;
+    s16 field_0x62a;
+    u8 field_0x62c[0x12];
+    s16 field_0x63e;
+    s16 field_0x640;
+};
+
 typedef void (*J3DCalcCallBack)(J3DModel*, u32 timing);
 
 class J3DModel {
@@ -77,18 +90,19 @@ public:
     Mtx33* getNrmMtxPtr() const { return mMtxBuffer->getNrmMtxPtr(); }
     Mtx* getDrawMtxPtr() const { return mMtxBuffer->getDrawMtxPtr(); }
     void setBaseScale(const Vec& scale) { mBaseScale = scale; }
-    void setUserArea(u32 area) { mUserArea = area; }
-    u32 getUserArea() const { return mUserArea; }
+    void setUserArea(UserArea* area) { mUserArea = area; }
+    UserArea* getUserArea() const { return mUserArea; }
     Vec* getBaseScale() { return &mBaseScale; }
 
     // is there a better way to handle inlines with same name as non-inlines?
     MtxP i_getAnmMtx(int p1) { return mMtxBuffer->getAnmMtx(p1); }
+    void setAnmMtx(int p1, Mtx mtx) { return mMtxBuffer->setAnmMtx(p1, mtx); }
 
     /* 0x04 */ J3DModelData* mModelData;
     /* 0x08 */ u32 mFlags;
     /* 0x0C */ u32 mDiffFlag;
     /* 0x10 */ J3DCalcCallBack mCalcCallBack;
-    /* 0x14 */ u32 mUserArea;
+    /* 0x14 */ UserArea* mUserArea;
     /* 0x18 */ Vec mBaseScale;
     /* 0x24 */ Mtx mBaseTransformMtx;
     /* 0x54 */ Mtx mInternalView;
diff --git a/include/JSystem/J3DGraphAnimator/J3DMtxBuffer.h b/include/JSystem/J3DGraphAnimator/J3DMtxBuffer.h
index ee497312894..0af381eaec8 100644
--- a/include/JSystem/J3DGraphAnimator/J3DMtxBuffer.h
+++ b/include/JSystem/J3DGraphAnimator/J3DMtxBuffer.h
@@ -25,6 +25,7 @@ public:
     /* 80326EF0 */ void calcBBoardMtx();
 
     MtxP getAnmMtx(int idx) const { return mpAnmMtx[idx]; }
+    void setAnmMtx(int idx, Mtx mtx) { PSMTXCopy(mtx,mpAnmMtx[idx]); }
 
     void setScaleFlag(int idx, u8 flag) { mpScaleFlagArr[idx] = flag; }
     u32* getCurrentViewNoPtr() { return &mCurrentViewNo; }
diff --git a/include/m_Do/m_Do_mtx.h b/include/m_Do/m_Do_mtx.h
index 9d344948f80..0fef06569cb 100644
--- a/include/m_Do/m_Do_mtx.h
+++ b/include/m_Do/m_Do_mtx.h
@@ -15,8 +15,9 @@ void mDoMtx_ZXYrotM(Mtx, s16, s16, s16);
 void mDoMtx_ZrotS(Mtx, s16);
 void mDoMtx_YrotS(Mtx, s16);
 void mDoMtx_XrotS(Mtx, s16);
-void mDoMtx_XrotM(Mtx mtx, s16 x);
+void mDoMtx_XrotM(Mtx, s16);
 void mDoMtx_YrotM(Mtx, s16);
+void mDoMtx_ZrotM(Mtx, s16);
 void mDoMtx_MtxToRot(CMtxP, csXyz*);
 void mDoMtx_lookAt(f32 (*param_0)[4], Vec const* param_1, Vec const* param_2, s16 param_3);
 void mDoMtx_concatProjView(f32 const (*param_0)[4], f32 const (*param_1)[4], f32 (*param_2)[4]);
@@ -37,6 +38,19 @@ inline void mDoMtx_trans(Mtx m, f32 x, f32 y, f32 z) {
     PSMTXTrans(m,x,y,z);
 }
 
+inline void cMtx_XrotM(Mtx mtx, s16 x) {
+    mDoMtx_XrotM(mtx, x);
+}
+
+inline void cMtx_YrotM(Mtx mtx, s16 y) {
+    mDoMtx_YrotM(mtx, y);
+}
+
+inline void cMtx_ZrotM(Mtx mtx, s16 z) {
+    mDoMtx_ZrotM(mtx, z);
+}
+
+
 class mDoMtx_stack_c {
 public:
     /* 8000CCC8 */ static bool push();
diff --git a/rel/d/a/d_a_do/d_a_do.cpp b/rel/d/a/d_a_do/d_a_do.cpp
index d89b62d7a75..efc3bf7725d 100644
--- a/rel/d/a/d_a_do/d_a_do.cpp
+++ b/rel/d/a/d_a_do/d_a_do.cpp
@@ -438,25 +438,47 @@ SECTION_DEAD static char const* const stringBase_8066EFB0 = "Do";
 #pragma pop
 
 /* 80667DA8-80667E68 000148 00C0+00 16/16 0/0 0/0 .text            anm_init__FP8do_classifUcf */
-#ifndef NONMATCHING
 static void anm_init(do_class* i_dogP, int param_1, f32 param_2, u8 param_3, f32 param_4) {
     if (!(i_dogP->field_0x608 > FLOAT_LABEL(lit_3662))) {
         i_dogP->mpMorf->setAnm((J3DAnmTransform*)dComIfG_getObjectRes("Do",param_1),param_3,param_2,param_4,FLOAT_LABEL(lit_3682),FLOAT_LABEL(lit_3683),0);
         i_dogP->field_0x5e4 = param_1;
     }
 }
-#else
-#pragma push
-#pragma optimization_level 0
-#pragma optimizewithasm off
-static asm void anm_init(do_class* i_dogP, int param_1, f32 param_2, u8 param_3, f32 param_4) {
-    nofralloc
-#include "asm/rel/d/a/d_a_do/d_a_do/anm_init__FP8do_classifUcf.s"
-}
-#pragma pop
-#endif
 
 /* 80667E68-80667FE4 000208 017C+00 1/1 0/0 0/0 .text            nodeCallBack__FP8J3DJointi */
+#ifdef NONMATCHING
+// regalloc on user_area
+static int nodeCallBack(J3DJoint* i_jntP, int param_1) {
+    if (param_1 == 0) {
+        int joint_num = i_jntP->getJntNo();
+        J3DModel* model = j3dSys.getModel();
+        UserArea* user_area = model->getUserArea();
+
+        if (user_area) {
+            PSMTXCopy(model->i_getAnmMtx(joint_num),*calc_mtx);
+
+            if (joint_num == 9 || joint_num  == 10) {
+                
+                
+                cMtx_YrotM(*calc_mtx,user_area->field_0x610 + user_area->field_0x628);
+                cMtx_XrotM(*calc_mtx,user_area->field_0x612 + user_area->field_0x62a);
+                cMtx_ZrotM(*calc_mtx,user_area->field_0x60e);
+            } else if (joint_num == 22) {
+                
+                cMtx_YrotM(*calc_mtx,user_area->field_0x640 << 1);
+                cMtx_ZrotM(*calc_mtx,user_area->field_0x63e << 1);
+            } else {
+                cMtx_YrotM(*calc_mtx,user_area->field_0x640);
+                cMtx_ZrotM(*calc_mtx,user_area->field_0x63e);
+            }
+
+            model->setAnmMtx(joint_num,*calc_mtx);
+            PSMTXCopy(*calc_mtx,j3dSys.mCurrentMtx);
+        }
+    }
+    return 1;
+}
+#else
 #pragma push
 #pragma optimization_level 0
 #pragma optimizewithasm off
@@ -465,6 +487,7 @@ static asm void nodeCallBack(J3DJoint* param_0, int param_1) {
 #include "asm/rel/d/a/d_a_do/d_a_do/nodeCallBack__FP8J3DJointi.s"
 }
 #pragma pop
+#endif
 
 /* ############################################################################################## */
 /* 8066EE04-8066EE08 00001C 0004+00 0/8 0/0 0/0 .rodata          @3772 */
diff --git a/rel/d/a/e/d_a_e_wb/d_a_e_wb.cpp b/rel/d/a/e/d_a_e_wb/d_a_e_wb.cpp
index 555b3033c13..dbf7123ec60 100644
--- a/rel/d/a/e/d_a_e_wb/d_a_e_wb.cpp
+++ b/rel/d/a/e/d_a_e_wb/d_a_e_wb.cpp
@@ -4813,7 +4813,8 @@ extern "C" asm void abs__4cXyzCFv() {
 #pragma push
 #pragma optimization_level 0
 #pragma optimizewithasm off
-static asm void cMtx_XrotM(f32 (*param_0)[4], s16 param_1) {
+// static asm void cMtx_XrotM(f32 (*param_0)[4], s16 param_1) {
+extern "C" asm void cMtx_XrotM__FPA4_fs() {
     nofralloc
 #include "asm/rel/d/a/e/d_a_e_wb/d_a_e_wb/cMtx_XrotM__FPA4_fs.s"
 }
diff --git a/rel/d/a/npc/d_a_npc_tr/d_a_npc_tr.cpp b/rel/d/a/npc/d_a_npc_tr/d_a_npc_tr.cpp
index 958b1ff6265..8a5703bca86 100644
--- a/rel/d/a/npc/d_a_npc_tr/d_a_npc_tr.cpp
+++ b/rel/d/a/npc/d_a_npc_tr/d_a_npc_tr.cpp
@@ -560,7 +560,7 @@ static int useHeapInit(fopAc_ac_c* actor) {
         return 0;
     }
 
-    npc_tr->field_0x5b8->setUserArea((u32)actor);
+    npc_tr->field_0x5b8->setUserArea((UserArea*)actor);
 
     for (u16 i = 0; i < npc_tr->field_0x5b8->getModelData()->getJointNum(); i++) {
         npc_tr->field_0x5b8->getModelData()->getJointNodePointer(i)->setCallBack(nodeCallBack);
diff --git a/src/d/a/d_a_npc.cpp b/src/d/a/d_a_npc.cpp
index 723d4d7c9ed..256bd5e961f 100644
--- a/src/d/a/d_a_npc.cpp
+++ b/src/d/a/d_a_npc.cpp
@@ -2036,7 +2036,7 @@ void daNpcT_c::setMtx() {
     mDoMtx_stack_c::scaleM(mScale);
 
     model->i_setBaseTRMtx(mDoMtx_stack_c::get());
-    model->setUserArea((u32)this);
+    model->setUserArea((UserArea*)this);
 
     mMcaMorfAnm[0]->onMorfNone();
     if (cM3d_IsZero(field_0xdfc) != 0) {
@@ -3786,7 +3786,7 @@ void daBaseNpc_c::setMtx(int param_0) {
     model->i_setBaseTRMtx(mDoMtx_stack_c::get());
 
     if (param_0) {
-        model->setUserArea((u32)this);
+        model->setUserArea((UserArea*)this);
     } else {
         model->setUserArea(0);
     }