target-i386: do not memcpy in and out of xmm_regs

After the next patch, we will move the high parts of AVX and AVX512 registers
in the same array as the SSE registers.  This will make it impossible to
memcpy an array of 128-bit values in and out of xmm_regs in one swoop.
Use a for loop instead.

Similarly, always use XMM_Q in translate.c.  This avoids introducing bugs
such as the one fixed in the previous patch.

Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target-i386/kvm.c

diff --git a/target-i386/kvm.c b/target-i386/kvm.c
index f92edfe..cf9f331 100644
--- a/target-i386/kvm.c
+++ b/target-i386/kvm.c
@@ -1019,7 +1019,10 @@
         fpu.ftwx |= (!env->fptags[i]) << i;
     }
     memcpy(fpu.fpr, env->fpregs, sizeof env->fpregs);
-    memcpy(fpu.xmm, env->xmm_regs, sizeof env->xmm_regs);
+    for (i = 0; i < CPU_NB_REGS; i++) {
+        stq_p(&fpu.xmm[i][0], env->xmm_regs[i].XMM_Q(0));
+        stq_p(&fpu.xmm[i][8], env->xmm_regs[i].XMM_Q(1));
+    }
     fpu.mxcsr = env->mxcsr;
 
     return kvm_vcpu_ioctl(CPU(cpu), KVM_SET_FPU, &fpu);
@@ -1045,6 +1048,7 @@
     CPUX86State *env = &cpu->env;
     struct kvm_xsave* xsave = env->kvm_xsave_buf;
     uint16_t cwd, swd, twd;
+    uint8_t *xmm;
     int i, r;
 
     if (!kvm_has_xsave()) {
@@ -1065,8 +1069,6 @@
     memcpy(&xsave->region[XSAVE_CWD_RDP], &env->fpdp, sizeof(env->fpdp));
     memcpy(&xsave->region[XSAVE_ST_SPACE], env->fpregs,
             sizeof env->fpregs);
-    memcpy(&xsave->region[XSAVE_XMM_SPACE], env->xmm_regs,
-            sizeof env->xmm_regs);
     xsave->region[XSAVE_MXCSR] = env->mxcsr;
     *(uint64_t *)&xsave->region[XSAVE_XSTATE_BV] = env->xstate_bv;
     memcpy(&xsave->region[XSAVE_YMMH_SPACE], env->ymmh_regs,
@@ -1079,6 +1081,13 @@
             sizeof env->opmask_regs);
     memcpy(&xsave->region[XSAVE_ZMM_Hi256], env->zmmh_regs,
             sizeof env->zmmh_regs);
+
+    xmm = (uint8_t *)&xsave->region[XSAVE_XMM_SPACE];
+    for (i = 0; i < CPU_NB_REGS; i++, xmm += 16) {
+        stq_p(xmm,     env->xmm_regs[i].XMM_Q(0));
+        stq_p(xmm+8,   env->xmm_regs[i].XMM_Q(1));
+    }
+
 #ifdef TARGET_X86_64
     memcpy(&xsave->region[XSAVE_Hi16_ZMM], env->hi16_zmm_regs,
             sizeof env->hi16_zmm_regs);
@@ -1384,7 +1393,10 @@
         env->fptags[i] = !((fpu.ftwx >> i) & 1);
     }
     memcpy(env->fpregs, fpu.fpr, sizeof env->fpregs);
-    memcpy(env->xmm_regs, fpu.xmm, sizeof env->xmm_regs);
+    for (i = 0; i < CPU_NB_REGS; i++) {
+        env->xmm_regs[i].XMM_Q(0) = ldq_p(&fpu.xmm[i][0]);
+        env->xmm_regs[i].XMM_Q(1) = ldq_p(&fpu.xmm[i][8]);
+    }
     env->mxcsr = fpu.mxcsr;
 
     return 0;
@@ -1395,6 +1407,7 @@
     CPUX86State *env = &cpu->env;
     struct kvm_xsave* xsave = env->kvm_xsave_buf;
     int ret, i;
+    const uint8_t *xmm;
     uint16_t cwd, swd, twd;
 
     if (!kvm_has_xsave()) {
@@ -1421,8 +1434,6 @@
     env->mxcsr = xsave->region[XSAVE_MXCSR];
     memcpy(env->fpregs, &xsave->region[XSAVE_ST_SPACE],
             sizeof env->fpregs);
-    memcpy(env->xmm_regs, &xsave->region[XSAVE_XMM_SPACE],
-            sizeof env->xmm_regs);
     env->xstate_bv = *(uint64_t *)&xsave->region[XSAVE_XSTATE_BV];
     memcpy(env->ymmh_regs, &xsave->region[XSAVE_YMMH_SPACE],
             sizeof env->ymmh_regs);
@@ -1434,6 +1445,13 @@
             sizeof env->opmask_regs);
     memcpy(env->zmmh_regs, &xsave->region[XSAVE_ZMM_Hi256],
             sizeof env->zmmh_regs);
+
+    xmm = (const uint8_t *)&xsave->region[XSAVE_XMM_SPACE];
+    for (i = 0; i < CPU_NB_REGS; i++, xmm += 16) {
+        env->xmm_regs[i].XMM_Q(0) = ldq_p(xmm);
+        env->xmm_regs[i].XMM_Q(1) = ldq_p(xmm+8);
+    }
+
 #ifdef TARGET_X86_64
     memcpy(env->hi16_zmm_regs, &xsave->region[XSAVE_Hi16_ZMM],
             sizeof env->hi16_zmm_regs);

target-i386/translate.c

diff --git a/target-i386/translate.c b/target-i386/translate.c
index 5af4300..9ebdf4b 100644
--- a/target-i386/translate.c
+++ b/target-i386/translate.c
@@ -2621,10 +2621,10 @@
 
 static inline void gen_op_movo(int d_offset, int s_offset)
 {
-    tcg_gen_ld_i64(cpu_tmp1_i64, cpu_env, s_offset);
-    tcg_gen_st_i64(cpu_tmp1_i64, cpu_env, d_offset);
-    tcg_gen_ld_i64(cpu_tmp1_i64, cpu_env, s_offset + 8);
-    tcg_gen_st_i64(cpu_tmp1_i64, cpu_env, d_offset + 8);
+    tcg_gen_ld_i64(cpu_tmp1_i64, cpu_env, s_offset + offsetof(XMMReg, XMM_Q(0)));
+    tcg_gen_st_i64(cpu_tmp1_i64, cpu_env, d_offset + offsetof(XMMReg, XMM_Q(0)));
+    tcg_gen_ld_i64(cpu_tmp1_i64, cpu_env, s_offset + offsetof(XMMReg, XMM_Q(1)));
+    tcg_gen_st_i64(cpu_tmp1_i64, cpu_env, d_offset + offsetof(XMMReg, XMM_Q(1)));
 }
 
 static inline void gen_op_movq(int d_offset, int s_offset)