[Qemu-devel] [PATCH 01/16] target-i386: Split fxsave/fxrstor implementation

[Richard Henderson]

We will be able to reuse these pieces for XSAVE/XRSTOR.

Signed-off-by: Richard Henderson <address@hidden>
---
 target-i386/fpu_helper.c | 175 ++++++++++++++++++++++++++---------------------
 target-i386/helper.h     |   4 +-
 target-i386/translate.c  |   4 +-
 3 files changed, 101 insertions(+), 82 deletions(-)

diff --git a/target-i386/fpu_helper.c b/target-i386/fpu_helper.c
index 2d54b47..a7da370 100644
--- a/target-i386/fpu_helper.c
+++ b/target-i386/fpu_helper.c
@@ -1115,89 +1115,89 @@ void cpu_x86_frstor(CPUX86State *env, target_ulong ptr, 
int data32)
 }
 #endif
 
-static void do_fxsave(CPUX86State *env, target_ulong ptr, int data64,
-                      uintptr_t retaddr)
+static void do_xsave_fpu(CPUX86State *env, target_ulong ptr, uintptr_t ra)
 {
-    int fpus, fptag, i, nb_xmm_regs;
-    floatx80 tmp;
+    int fpus, fptag, i;
     target_ulong addr;
 
-    /* The operand must be 16 byte aligned */
-    if (ptr & 0xf) {
-        raise_exception_ra(env, EXCP0D_GPF, retaddr);
-    }
-
     fpus = (env->fpus & ~0x3800) | (env->fpstt & 0x7) << 11;
     fptag = 0;
     for (i = 0; i < 8; i++) {
         fptag |= (env->fptags[i] << i);
     }
-    cpu_stw_data_ra(env, ptr, env->fpuc, retaddr);
-    cpu_stw_data_ra(env, ptr + 2, fpus, retaddr);
-    cpu_stw_data_ra(env, ptr + 4, fptag ^ 0xff, retaddr);
-#ifdef TARGET_X86_64
-    if (data64) {
-        cpu_stq_data_ra(env, ptr + 0x08, 0, retaddr); /* rip */
-        cpu_stq_data_ra(env, ptr + 0x10, 0, retaddr); /* rdp */
-    } else
-#endif
-    {
-        cpu_stl_data_ra(env, ptr + 0x08, 0, retaddr); /* eip */
-        cpu_stl_data_ra(env, ptr + 0x0c, 0, retaddr); /* sel  */
-        cpu_stl_data_ra(env, ptr + 0x10, 0, retaddr); /* dp */
-        cpu_stl_data_ra(env, ptr + 0x14, 0, retaddr); /* sel  */
-    }
+    cpu_stw_data_ra(env, ptr, env->fpuc, ra);
+    cpu_stw_data_ra(env, ptr + 2, fpus, ra);
+    cpu_stw_data_ra(env, ptr + 4, fptag ^ 0xff, ra);
+
+    /* In 32-bit mode this is eip, sel, dp, sel.
+       In 64-bit mode this is rip, rdp.
+       But in either case we don't write actual data, just zeros.  */
+    cpu_stq_data_ra(env, ptr + 0x08, 0, ra); /* eip+sel; rip */
+    cpu_stq_data_ra(env, ptr + 0x10, 0, ra); /* edp+sel; rdp */
 
     addr = ptr + 0x20;
     for (i = 0; i < 8; i++) {
-        tmp = ST(i);
-        helper_fstt(env, tmp, addr, retaddr);
+        floatx80 tmp = ST(i);
+        helper_fstt(env, tmp, addr, ra);
         addr += 16;
     }
+}
+
+static void do_xsave_mxcsr(CPUX86State *env, target_ulong ptr, uintptr_t ra)
+{
+    cpu_stl_data_ra(env, ptr + 0x18, env->mxcsr, ra); /* mxcsr */
+    cpu_stl_data_ra(env, ptr + 0x1c, 0x0000ffff, ra); /* mxcsr_mask */
+}
+
+static void do_xsave_sse(CPUX86State *env, target_ulong ptr, uintptr_t ra)
+{
+    int i, nb_xmm_regs;
+    target_ulong addr;
+
+    if (env->hflags & HF_CS64_MASK) {
+        nb_xmm_regs = 16;
+    } else {
+        nb_xmm_regs = 8;
+    }
+
+    addr = ptr + 0xa0;
+    for (i = 0; i < nb_xmm_regs; i++) {
+        cpu_stq_data_ra(env, addr, env->xmm_regs[i].ZMM_Q(0), ra);
+        cpu_stq_data_ra(env, addr + 8, env->xmm_regs[i].ZMM_Q(1), ra);
+        addr += 16;
+    }
+}
+
+void helper_fxsave(CPUX86State *env, target_ulong ptr)
+{
+    uintptr_t ra = GETPC();
+
+    /* The operand must be 16 byte aligned */
+    if (ptr & 0xf) {
+        raise_exception_ra(env, EXCP0D_GPF, ra);
+    }
+
+    do_xsave_fpu(env, ptr, ra);
 
     if (env->cr[4] & CR4_OSFXSR_MASK) {
-        /* XXX: finish it */
-        cpu_stl_data_ra(env, ptr + 0x18, env->mxcsr, retaddr); /* mxcsr */
-        cpu_stl_data_ra(env, ptr + 0x1c, 0x0000ffff, retaddr); /* mxcsr_mask */
-        if (env->hflags & HF_CS64_MASK) {
-            nb_xmm_regs = 16;
-        } else {
-            nb_xmm_regs = 8;
-        }
-        addr = ptr + 0xa0;
+        do_xsave_mxcsr(env, ptr, ra);
         /* Fast FXSAVE leaves out the XMM registers */
         if (!(env->efer & MSR_EFER_FFXSR)
             || (env->hflags & HF_CPL_MASK)
             || !(env->hflags & HF_LMA_MASK)) {
-            for (i = 0; i < nb_xmm_regs; i++) {
-                cpu_stq_data_ra(env, addr, env->xmm_regs[i].ZMM_Q(0), retaddr);
-                cpu_stq_data_ra(env, addr + 8, env->xmm_regs[i].ZMM_Q(1), 
retaddr);
-                addr += 16;
-            }
+            do_xsave_sse(env, ptr, ra);
         }
     }
 }
 
-void helper_fxsave(CPUX86State *env, target_ulong ptr, int data64)
-{
-    do_fxsave(env, ptr, data64, GETPC());
-}
-
-static void do_fxrstor(CPUX86State *env, target_ulong ptr, int data64,
-                       uintptr_t retaddr)
+static void do_xrstor_fpu(CPUX86State *env, target_ulong ptr, uintptr_t ra)
 {
-    int i, fpus, fptag, nb_xmm_regs;
-    floatx80 tmp;
+    int i, fpus, fptag;
     target_ulong addr;
 
-    /* The operand must be 16 byte aligned */
-    if (ptr & 0xf) {
-        raise_exception_ra(env, EXCP0D_GPF, retaddr);
-    }
-
-    cpu_set_fpuc(env, cpu_lduw_data_ra(env, ptr, retaddr));
-    fpus = cpu_lduw_data_ra(env, ptr + 2, retaddr);
-    fptag = cpu_lduw_data_ra(env, ptr + 4, retaddr);
+    cpu_set_fpuc(env, cpu_lduw_data_ra(env, ptr, ra));
+    fpus = cpu_lduw_data_ra(env, ptr + 2, ra);
+    fptag = cpu_lduw_data_ra(env, ptr + 4, ra);
     env->fpstt = (fpus >> 11) & 7;
     env->fpus = fpus & ~0x3800;
     fptag ^= 0xff;
@@ -1207,39 +1207,58 @@ static void do_fxrstor(CPUX86State *env, target_ulong 
ptr, int data64,
 
     addr = ptr + 0x20;
     for (i = 0; i < 8; i++) {
-        tmp = helper_fldt(env, addr, retaddr);
+        floatx80 tmp = helper_fldt(env, addr, ra);
         ST(i) = tmp;
         addr += 16;
     }
+}
+
+static void do_xrstor_mxcsr(CPUX86State *env, target_ulong ptr, uintptr_t ra)
+{
+    cpu_set_mxcsr(env, cpu_ldl_data_ra(env, ptr + 0x18, ra));
+}
+
+static void do_xrstor_sse(CPUX86State *env, target_ulong ptr, uintptr_t ra)
+{
+    int i, nb_xmm_regs;
+    target_ulong addr;
+
+    if (env->hflags & HF_CS64_MASK) {
+        nb_xmm_regs = 16;
+    } else {
+        nb_xmm_regs = 8;
+    }
+
+    addr = ptr + 0xa0;
+    for (i = 0; i < nb_xmm_regs; i++) {
+        env->xmm_regs[i].ZMM_Q(0) = cpu_ldq_data_ra(env, addr, ra);
+        env->xmm_regs[i].ZMM_Q(1) = cpu_ldq_data_ra(env, addr + 8, ra);
+        addr += 16;
+    }
+}
+
+void helper_fxrstor(CPUX86State *env, target_ulong ptr)
+{
+    uintptr_t ra = GETPC();
+
+    /* The operand must be 16 byte aligned */
+    if (ptr & 0xf) {
+        raise_exception_ra(env, EXCP0D_GPF, ra);
+    }
+
+    do_xrstor_fpu(env, ptr, ra);
 
     if (env->cr[4] & CR4_OSFXSR_MASK) {
-        /* XXX: finish it */
-        cpu_set_mxcsr(env, cpu_ldl_data_ra(env, ptr + 0x18, retaddr));
-        /* cpu_ldl_data_ra(env, ptr + 0x1c, retaddr); */
-        if (env->hflags & HF_CS64_MASK) {
-            nb_xmm_regs = 16;
-        } else {
-            nb_xmm_regs = 8;
-        }
-        addr = ptr + 0xa0;
-        /* Fast FXRESTORE leaves out the XMM registers */
+        do_xrstor_mxcsr(env, ptr, ra);
+        /* Fast FXRSTOR leaves out the XMM registers */
         if (!(env->efer & MSR_EFER_FFXSR)
             || (env->hflags & HF_CPL_MASK)
             || !(env->hflags & HF_LMA_MASK)) {
-            for (i = 0; i < nb_xmm_regs; i++) {
-                env->xmm_regs[i].ZMM_Q(0) = cpu_ldq_data_ra(env, addr, 
retaddr);
-                env->xmm_regs[i].ZMM_Q(1) = cpu_ldq_data_ra(env, addr + 8, 
retaddr);
-                addr += 16;
-            }
+            do_xrstor_sse(env, ptr, ra);
         }
     }
 }
 
-void helper_fxrstor(CPUX86State *env, target_ulong ptr, int data64)
-{
-    do_fxrstor(env, ptr, data64, GETPC());
-}
-
 void cpu_get_fp80(uint64_t *pmant, uint16_t *pexp, floatx80 f)
 {
     CPU_LDoubleU temp;
diff --git a/target-i386/helper.h b/target-i386/helper.h
index 3a25c3b..6109e46 100644
--- a/target-i386/helper.h
+++ b/target-i386/helper.h
@@ -185,8 +185,8 @@ DEF_HELPER_3(fstenv, void, env, tl, int)
 DEF_HELPER_3(fldenv, void, env, tl, int)
 DEF_HELPER_3(fsave, void, env, tl, int)
 DEF_HELPER_3(frstor, void, env, tl, int)
-DEF_HELPER_3(fxsave, void, env, tl, int)
-DEF_HELPER_3(fxrstor, void, env, tl, int)
+DEF_HELPER_FLAGS_2(fxsave, TCG_CALL_NO_WG, void, env, tl)
+DEF_HELPER_FLAGS_2(fxrstor, TCG_CALL_NO_WG, void, env, tl)
 
 DEF_HELPER_FLAGS_1(clz, TCG_CALL_NO_RWG_SE, tl, tl)
 DEF_HELPER_FLAGS_1(ctz, TCG_CALL_NO_RWG_SE, tl, tl)
diff --git a/target-i386/translate.c b/target-i386/translate.c
index 986954b..9781543 100644
--- a/target-i386/translate.c
+++ b/target-i386/translate.c
@@ -7505,7 +7505,7 @@ static target_ulong disas_insn(CPUX86State *env, 
DisasContext *s,
                 break;
             }
             gen_lea_modrm(env, s, modrm);
-            gen_helper_fxsave(cpu_env, cpu_A0, tcg_const_i32(dflag == MO_64));
+            gen_helper_fxsave(cpu_env, cpu_A0);
             break;
         case 1: /* fxrstor */
             if (mod == 3 || !(s->cpuid_features & CPUID_FXSR) ||
@@ -7516,7 +7516,7 @@ static target_ulong disas_insn(CPUX86State *env, 
DisasContext *s,
                 break;
             }
             gen_lea_modrm(env, s, modrm);
-            gen_helper_fxrstor(cpu_env, cpu_A0, tcg_const_i32(dflag == MO_64));
+            gen_helper_fxrstor(cpu_env, cpu_A0);
             break;
         case 2: /* ldmxcsr */
         case 3: /* stmxcsr */
-- 
2.5.0