Merge remote-tracking branch 'remotes/rth/tags/pull-tcg-20161101-2' into staging

tcg queued patches

# gpg: Signature made Tue 01 Nov 2016 16:45:42 GMT
# gpg:                using RSA key 0xAD1270CC4DD0279B
# gpg: Good signature from "Richard Henderson <rth7680@gmail.com>"
# gpg:                 aka "Richard Henderson <rth@redhat.com>"
# gpg:                 aka "Richard Henderson <rth@twiddle.net>"
# Primary key fingerprint: 9CB1 8DDA F8E8 49AD 2AFC  16A4 AD12 70CC 4DD0 279B

* remotes/rth/tags/pull-tcg-20161101-2:
  tcg: correct 32-bit tcg_gen_ld8s_i64 sign-extension
  tcg/tcg.h: Improve documentation of TCGv_i32 etc types
  MAINTAINERS: Update PPC status and maintainer
  target-microblaze: Cleanup dec_mul
  tcg: Add tcg_gen_mulsu2_{i32,i64,tl}
  log: Add locking to large logging blocks
  target-openrisc: Do not dump cpu state with -d in_asm
  target-microblaze: Do not dump cpu state with -d in_asm
  target-cris: Do not dump cpu state with -d in_asm

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

29 files changed, 192 insertions, 100 deletions

diff --git a/MAINTAINERS b/MAINTAINERS
index 3fecf458c0..653f52e5b3 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -1504,8 +1504,8 @@ F: tcg/mips/
 F: disas/mips.c
 
 PPC
-M: Vassili Karpov (malc) <av1474@comtv.ru>
-S: Maintained
+M: Richard Henderson <rth@twiddle.net>
+S: Odd Fixes
 F: tcg/ppc/
 F: disas/ppc.c
 
diff --git a/cpu-exec.c b/cpu-exec.c
index 3e408865a8..4188fed3c6 100644
--- a/cpu-exec.c
+++ b/cpu-exec.c
@@ -150,11 +150,13 @@ static inline tcg_target_ulong cpu_tb_exec(CPUState *cpu, TranslationBlock *itb)
 #if defined(DEBUG_DISAS)
     if (qemu_loglevel_mask(CPU_LOG_TB_CPU)
         && qemu_log_in_addr_range(itb->pc)) {
+        qemu_log_lock();
 #if defined(TARGET_I386)
         log_cpu_state(cpu, CPU_DUMP_CCOP);
 #else
         log_cpu_state(cpu, 0);
 #endif
+        qemu_log_unlock();
     }
 #endif /* DEBUG_DISAS */
 
diff --git a/exec.c b/exec.c
index 4d085812ca..b1094c0cd2 100644
--- a/exec.c
+++ b/exec.c
@@ -911,11 +911,13 @@ void cpu_abort(CPUState *cpu, const char *fmt, ...)
     fprintf(stderr, "\n");
     cpu_dump_state(cpu, stderr, fprintf, CPU_DUMP_FPU | CPU_DUMP_CCOP);
     if (qemu_log_separate()) {
+        qemu_log_lock();
         qemu_log("qemu: fatal: ");
         qemu_log_vprintf(fmt, ap2);
         qemu_log("\n");
         log_cpu_state(cpu, CPU_DUMP_FPU | CPU_DUMP_CCOP);
         qemu_log_flush();
+        qemu_log_unlock();
         qemu_log_close();
     }
     va_end(ap2);
diff --git a/include/qemu/log.h b/include/qemu/log.h
index 00bf37fc0f..a50e994c21 100644
--- a/include/qemu/log.h
+++ b/include/qemu/log.h
@@ -51,6 +51,22 @@ static inline bool qemu_loglevel_mask(int mask)
     return (qemu_loglevel & mask) != 0;
 }
 
+/* Lock output for a series of related logs.  Since this is not needed
+ * for a single qemu_log / qemu_log_mask / qemu_log_mask_and_addr, we
+ * assume that qemu_loglevel_mask has already been tested, and that
+ * qemu_loglevel is never set when qemu_logfile is unset.
+ */
+
+static inline void qemu_log_lock(void)
+{
+    qemu_flockfile(qemu_logfile);
+}
+
+static inline void qemu_log_unlock(void)
+{
+    qemu_funlockfile(qemu_logfile);
+}
+
 /* Logging functions: */
 
 /* main logging function
diff --git a/include/sysemu/os-posix.h b/include/sysemu/os-posix.h
index 3cfedbc28b..b0a6c0695b 100644
--- a/include/sysemu/os-posix.h
+++ b/include/sysemu/os-posix.h
@@ -87,4 +87,16 @@ void *qemu_alloc_stack(size_t *sz);
  */
 void qemu_free_stack(void *stack, size_t sz);
 
+/* POSIX and Mingw32 differ in the name of the stdio lock functions.  */
+
+static inline void qemu_flockfile(FILE *f)
+{
+    flockfile(f);
+}
+
+static inline void qemu_funlockfile(FILE *f)
+{
+    funlockfile(f);
+}
+
 #endif
diff --git a/include/sysemu/os-win32.h b/include/sysemu/os-win32.h
index 17aad3b20f..ff18b23db1 100644
--- a/include/sysemu/os-win32.h
+++ b/include/sysemu/os-win32.h
@@ -103,6 +103,21 @@ static inline char *realpath(const char *path, char *resolved_path)
     return resolved_path;
 }
 
+/* ??? Mingw appears to export _lock_file and _unlock_file as the functions
+ * with which to lock a stdio handle.  But something is wrong in the markup,
+ * either in the header or the library, such that we get undefined references
+ * to "_imp___lock_file" etc when linking.  Since we seem to have no other
+ * alternative, and the usage within the logging functions isn't critical,
+ * ignore FILE locking.
+ */
+
+static inline void qemu_flockfile(FILE *f)
+{
+}
+
+static inline void qemu_funlockfile(FILE *f)
+{
+}
 
 /* We wrap all the sockets functions so that we can
  * set errno based on WSAGetLastError()
diff --git a/target-alpha/translate.c b/target-alpha/translate.c
index 03e47765ed..114927b751 100644
--- a/target-alpha/translate.c
+++ b/target-alpha/translate.c
@@ -2994,9 +2994,11 @@ void gen_intermediate_code(CPUAlphaState *env, struct TranslationBlock *tb)
 #ifdef DEBUG_DISAS
     if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)
         && qemu_log_in_addr_range(pc_start)) {
+        qemu_log_lock();
         qemu_log("IN: %s\n", lookup_symbol(pc_start));
         log_target_disas(cs, pc_start, ctx.pc - pc_start, 1);
         qemu_log("\n");
+        qemu_log_unlock();
     }
 #endif
 }
diff --git a/target-arm/translate-a64.c b/target-arm/translate-a64.c
index ded924a0a9..de48747376 100644
--- a/target-arm/translate-a64.c
+++ b/target-arm/translate-a64.c
@@ -11420,11 +11420,13 @@ done_generating:
 #ifdef DEBUG_DISAS
     if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM) &&
         qemu_log_in_addr_range(pc_start)) {
+        qemu_log_lock();
         qemu_log("----------------\n");
         qemu_log("IN: %s\n", lookup_symbol(pc_start));
         log_target_disas(cs, pc_start, dc->pc - pc_start,
                          4 | (bswap_code(dc->sctlr_b) ? 2 : 0));
         qemu_log("\n");
+        qemu_log_unlock();
     }
 #endif
     tb->size = dc->pc - pc_start;
diff --git a/target-arm/translate.c b/target-arm/translate.c
index 718f7d06f3..0ad9070b45 100644
--- a/target-arm/translate.c
+++ b/target-arm/translate.c
@@ -11963,11 +11963,13 @@ done_generating:
 #ifdef DEBUG_DISAS
     if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM) &&
         qemu_log_in_addr_range(pc_start)) {
+        qemu_log_lock();
         qemu_log("----------------\n");
         qemu_log("IN: %s\n", lookup_symbol(pc_start));
         log_target_disas(cs, pc_start, dc->pc - pc_start,
                          dc->thumb | (dc->sctlr_b << 1));
         qemu_log("\n");
+        qemu_log_unlock();
     }
 #endif
     tb->size = dc->pc - pc_start;
diff --git a/target-cris/translate.c b/target-cris/translate.c
index b5ab0a5fb2..b91042743f 100644
--- a/target-cris/translate.c
+++ b/target-cris/translate.c
@@ -3135,29 +3135,6 @@ void gen_intermediate_code(CPUCRISState *env, struct TranslationBlock *tb)
 
     dc->cpustate_changed = 0;
 
-    if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)) {
-        qemu_log(
-                "pc=%x %x flg=%" PRIx64 " bt=%x ds=%u ccs=%x\n"
-                "pid=%x usp=%x\n"
-                "%x.%x.%x.%x\n"
-                "%x.%x.%x.%x\n"
-                "%x.%x.%x.%x\n"
-                "%x.%x.%x.%x\n",
-                dc->pc, dc->ppc,
-                (uint64_t)tb->flags,
-                env->btarget, (unsigned)tb->flags & 7,
-                env->pregs[PR_CCS],
-                env->pregs[PR_PID], env->pregs[PR_USP],
-                env->regs[0], env->regs[1], env->regs[2], env->regs[3],
-                env->regs[4], env->regs[5], env->regs[6], env->regs[7],
-                env->regs[8], env->regs[9],
-                env->regs[10], env->regs[11],
-                env->regs[12], env->regs[13],
-                env->regs[14], env->regs[15]);
-        qemu_log("--------------\n");
-        qemu_log("IN: %s\n", lookup_symbol(pc_start));
-    }
-
     next_page_start = (pc_start & TARGET_PAGE_MASK) + TARGET_PAGE_SIZE;
     num_insns = 0;
     max_insns = tb->cflags & CF_COUNT_MASK;
@@ -3313,10 +3290,14 @@ void gen_intermediate_code(CPUCRISState *env, struct TranslationBlock *tb)
 #if !DISAS_CRIS
     if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)
         && qemu_log_in_addr_range(pc_start)) {
+        qemu_log_lock();
+        qemu_log("--------------\n");
+        qemu_log("IN: %s\n", lookup_symbol(pc_start));
         log_target_disas(cs, pc_start, dc->pc - pc_start,
                          env->pregs[PR_VR]);
         qemu_log("\nisize=%d osize=%d\n",
                  dc->pc - pc_start, tcg_op_buf_count());
+        qemu_log_unlock();
     }
 #endif
 #endif
diff --git a/target-i386/translate.c b/target-i386/translate.c
index 927b366534..324103c885 100644
--- a/target-i386/translate.c
+++ b/target-i386/translate.c
@@ -2432,11 +2432,13 @@ static void gen_unknown_opcode(CPUX86State *env, DisasContext *s)
 
     if (qemu_loglevel_mask(LOG_UNIMP)) {
         target_ulong pc = s->pc_start, end = s->pc;
+        qemu_log_lock();
         qemu_log("ILLOPC: " TARGET_FMT_lx ":", pc);
         for (; pc < end; ++pc) {
             qemu_log(" %02x", cpu_ldub_code(env, pc));
         }
         qemu_log("\n");
+        qemu_log_unlock();
     }
 }
 
@@ -8470,6 +8472,7 @@ done_generating:
     if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)
         && qemu_log_in_addr_range(pc_start)) {
         int disas_flags;
+        qemu_log_lock();
         qemu_log("----------------\n");
         qemu_log("IN: %s\n", lookup_symbol(pc_start));
 #ifdef TARGET_X86_64
@@ -8480,6 +8483,7 @@ done_generating:
             disas_flags = !dc->code32;
         log_target_disas(cs, pc_start, pc_ptr - pc_start, disas_flags);
         qemu_log("\n");
+        qemu_log_unlock();
     }
 #endif
 
diff --git a/target-lm32/translate.c b/target-lm32/translate.c
index 842af63a98..692882f447 100644
--- a/target-lm32/translate.c
+++ b/target-lm32/translate.c
@@ -1148,10 +1148,12 @@ void gen_intermediate_code(CPULM32State *env, struct TranslationBlock *tb)
 #ifdef DEBUG_DISAS
     if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)
         && qemu_log_in_addr_range(pc_start)) {
+        qemu_log_lock();
         qemu_log("\n");
         log_target_disas(cs, pc_start, dc->pc - pc_start, 0);
         qemu_log("\nisize=%d osize=%d\n",
                  dc->pc - pc_start, tcg_op_buf_count());
+        qemu_log_unlock();
     }
 #endif
 }
diff --git a/target-m68k/translate.c b/target-m68k/translate.c
index ee0ffe3e07..9ad974f86a 100644
--- a/target-m68k/translate.c
+++ b/target-m68k/translate.c
@@ -3549,10 +3549,12 @@ void gen_intermediate_code(CPUM68KState *env, TranslationBlock *tb)
 #ifdef DEBUG_DISAS
     if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)
         && qemu_log_in_addr_range(pc_start)) {
+        qemu_log_lock();
         qemu_log("----------------\n");
         qemu_log("IN: %s\n", lookup_symbol(pc_start));
         log_target_disas(cs, pc_start, dc->pc - pc_start, 0);
         qemu_log("\n");
+        qemu_log_unlock();
     }
 #endif
     tb->size = dc->pc - pc_start;
diff --git a/target-microblaze/translate.c b/target-microblaze/translate.c
index 80098ece15..de2090ac71 100644
--- a/target-microblaze/translate.c
+++ b/target-microblaze/translate.c
@@ -581,50 +581,10 @@ static void dec_msr(DisasContext *dc)
     }
 }
 
-/* 64-bit signed mul, lower result in d and upper in d2.  */
-static void t_gen_muls(TCGv d, TCGv d2, TCGv a, TCGv b)
-{
-    TCGv_i64 t0, t1;
-
-    t0 = tcg_temp_new_i64();
-    t1 = tcg_temp_new_i64();
-
-    tcg_gen_ext_i32_i64(t0, a);
-    tcg_gen_ext_i32_i64(t1, b);
-    tcg_gen_mul_i64(t0, t0, t1);
-
-    tcg_gen_extrl_i64_i32(d, t0);
-    tcg_gen_shri_i64(t0, t0, 32);
-    tcg_gen_extrl_i64_i32(d2, t0);
-
-    tcg_temp_free_i64(t0);
-    tcg_temp_free_i64(t1);
-}
-
-/* 64-bit unsigned muls, lower result in d and upper in d2.  */
-static void t_gen_mulu(TCGv d, TCGv d2, TCGv a, TCGv b)
-{
-    TCGv_i64 t0, t1;
-
-    t0 = tcg_temp_new_i64();
-    t1 = tcg_temp_new_i64();
-
-    tcg_gen_extu_i32_i64(t0, a);
-    tcg_gen_extu_i32_i64(t1, b);
-    tcg_gen_mul_i64(t0, t0, t1);
-
-    tcg_gen_extrl_i64_i32(d, t0);
-    tcg_gen_shri_i64(t0, t0, 32);
-    tcg_gen_extrl_i64_i32(d2, t0);
-
-    tcg_temp_free_i64(t0);
-    tcg_temp_free_i64(t1);
-}
-
 /* Multiplier unit.  */
 static void dec_mul(DisasContext *dc)
 {
-    TCGv d[2];
+    TCGv tmp;
     unsigned int subcode;
 
     if ((dc->tb_flags & MSR_EE_FLAG)
@@ -636,13 +596,11 @@ static void dec_mul(DisasContext *dc)
     }
 
     subcode = dc->imm & 3;
-    d[0] = tcg_temp_new();
-    d[1] = tcg_temp_new();
 
     if (dc->type_b) {
         LOG_DIS("muli r%d r%d %x\n", dc->rd, dc->ra, dc->imm);
-        t_gen_mulu(cpu_R[dc->rd], d[1], cpu_R[dc->ra], *(dec_alu_op_b(dc)));
-        goto done;
+        tcg_gen_mul_tl(cpu_R[dc->rd], cpu_R[dc->ra], *(dec_alu_op_b(dc)));
+        return;
     }
 
     /* mulh, mulhsu and mulhu are not available if C_USE_HW_MUL is < 2.  */
@@ -651,30 +609,29 @@ static void dec_mul(DisasContext *dc)
         /* nop??? */
     }
 
+    tmp = tcg_temp_new();
     switch (subcode) {
         case 0:
             LOG_DIS("mul r%d r%d r%d\n", dc->rd, dc->ra, dc->rb);
-            t_gen_mulu(cpu_R[dc->rd], d[1], cpu_R[dc->ra], cpu_R[dc->rb]);
+            tcg_gen_mul_tl(cpu_R[dc->rd], cpu_R[dc->ra], cpu_R[dc->rb]);
             break;
         case 1:
             LOG_DIS("mulh r%d r%d r%d\n", dc->rd, dc->ra, dc->rb);
-            t_gen_muls(d[0], cpu_R[dc->rd], cpu_R[dc->ra], cpu_R[dc->rb]);
+            tcg_gen_muls2_tl(tmp, cpu_R[dc->rd], cpu_R[dc->ra], cpu_R[dc->rb]);
             break;
         case 2:
             LOG_DIS("mulhsu r%d r%d r%d\n", dc->rd, dc->ra, dc->rb);
-            t_gen_muls(d[0], cpu_R[dc->rd], cpu_R[dc->ra], cpu_R[dc->rb]);
+            tcg_gen_mulsu2_tl(tmp, cpu_R[dc->rd], cpu_R[dc->ra], cpu_R[dc->rb]);
             break;
         case 3:
             LOG_DIS("mulhu r%d r%d r%d\n", dc->rd, dc->ra, dc->rb);
-            t_gen_mulu(d[0], cpu_R[dc->rd], cpu_R[dc->ra], cpu_R[dc->rb]);
+            tcg_gen_mulu2_tl(tmp, cpu_R[dc->rd], cpu_R[dc->ra], cpu_R[dc->rb]);
             break;
         default:
             cpu_abort(CPU(dc->cpu), "unknown MUL insn %x\n", subcode);
             break;
     }
-done:
-    tcg_temp_free(d[0]);
-    tcg_temp_free(d[1]);
+    tcg_temp_free(tmp);
 }
 
 /* Div unit.  */
@@ -1670,13 +1627,6 @@ void gen_intermediate_code(CPUMBState *env, struct TranslationBlock *tb)
         cpu_abort(cs, "Microblaze: unaligned PC=%x\n", pc_start);
     }
 
-    if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)) {
-#if !SIM_COMPAT
-        qemu_log("--------------\n");
-        log_cpu_state(CPU(cpu), 0);
-#endif
-    }
-
     next_page_start = (pc_start & TARGET_PAGE_MASK) + TARGET_PAGE_SIZE;
     num_insns = 0;
     max_insns = tb->cflags & CF_COUNT_MASK;
@@ -1820,12 +1770,14 @@ void gen_intermediate_code(CPUMBState *env, struct TranslationBlock *tb)
 #if !SIM_COMPAT
     if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)
         && qemu_log_in_addr_range(pc_start)) {
-        qemu_log("\n");
+        qemu_log_lock();
+        qemu_log("--------------\n");
 #if DISAS_GNU
         log_target_disas(cs, pc_start, dc->pc - pc_start, 0);
 #endif
         qemu_log("\nisize=%d osize=%d\n",
                  dc->pc - pc_start, tcg_op_buf_count());
+        qemu_log_unlock();
     }
 #endif
 #endif
diff --git a/target-mips/translate.c b/target-mips/translate.c
index 55c2ca0c7b..d8dde7a2f5 100644
--- a/target-mips/translate.c
+++ b/target-mips/translate.c
@@ -20043,9 +20043,11 @@ done_generating:
     LOG_DISAS("\n");
     if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)
         && qemu_log_in_addr_range(pc_start)) {
+        qemu_log_lock();
         qemu_log("IN: %s\n", lookup_symbol(pc_start));
         log_target_disas(cs, pc_start, ctx.pc - pc_start, 0);
         qemu_log("\n");
+        qemu_log_unlock();
     }
 #endif
 }
diff --git a/target-openrisc/translate.c b/target-openrisc/translate.c
index 28c944657c..229361aed1 100644
--- a/target-openrisc/translate.c
+++ b/target-openrisc/translate.c
@@ -1651,10 +1651,6 @@ void gen_intermediate_code(CPUOpenRISCState *env, struct TranslationBlock *tb)
     dc->synced_flags = dc->tb_flags = tb->flags;
     dc->delayed_branch = !!(dc->tb_flags & D_FLAG);
     dc->singlestep_enabled = cs->singlestep_enabled;
-    if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)) {
-        qemu_log("-----------------------------------------\n");
-        log_cpu_state(CPU(cpu), 0);
-    }
 
     next_page_start = (pc_start & TARGET_PAGE_MASK) + TARGET_PAGE_SIZE;
     num_insns = 0;
@@ -1754,10 +1750,13 @@ void gen_intermediate_code(CPUOpenRISCState *env, struct TranslationBlock *tb)
 #ifdef DEBUG_DISAS
     if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)
         && qemu_log_in_addr_range(pc_start)) {
-        qemu_log("\n");
+        qemu_log_lock();
+        qemu_log("----------------\n");
+        qemu_log("IN: %s\n", lookup_symbol(pc_start));
         log_target_disas(cs, pc_start, dc->pc - pc_start, 0);
         qemu_log("\nisize=%d osize=%d\n",
                  dc->pc - pc_start, tcg_op_buf_count());
+        qemu_log_unlock();
     }
 #endif
 }
diff --git a/target-ppc/translate.c b/target-ppc/translate.c
index 43505a936c..54f35e9904 100644
--- a/target-ppc/translate.c
+++ b/target-ppc/translate.c
@@ -7211,9 +7211,11 @@ void gen_intermediate_code(CPUPPCState *env, struct TranslationBlock *tb)
         int flags;
         flags = env->bfd_mach;
         flags |= ctx.le_mode << 16;
+        qemu_log_lock();
         qemu_log("IN: %s\n", lookup_symbol(pc_start));
         log_target_disas(cs, pc_start, ctx.nip - pc_start, flags);
         qemu_log("\n");
+        qemu_log_unlock();
     }
 #endif
 }
diff --git a/target-s390x/translate.c b/target-s390x/translate.c
index 1a07d70b21..02bc7058fd 100644
--- a/target-s390x/translate.c
+++ b/target-s390x/translate.c
@@ -5432,9 +5432,11 @@ void gen_intermediate_code(CPUS390XState *env, struct TranslationBlock *tb)
 #if defined(S390X_DEBUG_DISAS)
     if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)
         && qemu_log_in_addr_range(pc_start)) {
+        qemu_log_lock();
         qemu_log("IN: %s\n", lookup_symbol(pc_start));
         log_target_disas(cs, pc_start, dc.pc - pc_start, 1);
         qemu_log("\n");
+        qemu_log_unlock();
     }
 #endif
 }
diff --git a/target-sh4/translate.c b/target-sh4/translate.c
index ca80cf70ca..c89a14733f 100644
--- a/target-sh4/translate.c
+++ b/target-sh4/translate.c
@@ -1927,9 +1927,11 @@ void gen_intermediate_code(CPUSH4State * env, struct TranslationBlock *tb)
 #ifdef DEBUG_DISAS
     if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)
         && qemu_log_in_addr_range(pc_start)) {
+        qemu_log_lock();
 	qemu_log("IN:\n");	/* , lookup_symbol(pc_start)); */
         log_target_disas(cs, pc_start, ctx.pc - pc_start, 0);
 	qemu_log("\n");
+        qemu_log_unlock();
     }
 #endif
 }
diff --git a/target-sparc/translate.c b/target-sparc/translate.c
index a13b76ebd9..2205f89837 100644
--- a/target-sparc/translate.c
+++ b/target-sparc/translate.c
@@ -5796,10 +5796,12 @@ void gen_intermediate_code(CPUSPARCState * env, TranslationBlock * tb)
 #ifdef DEBUG_DISAS
     if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)
         && qemu_log_in_addr_range(pc_start)) {
+        qemu_log_lock();
         qemu_log("--------------\n");
         qemu_log("IN: %s\n", lookup_symbol(pc_start));
         log_target_disas(cs, pc_start, last_pc + 4 - pc_start, 0);
         qemu_log("\n");
+        qemu_log_unlock();
     }
 #endif
 }
diff --git a/target-tilegx/translate.c b/target-tilegx/translate.c
index 11c9732389..9c734eeba3 100644
--- a/target-tilegx/translate.c
+++ b/target-tilegx/translate.c
@@ -2391,6 +2391,7 @@ void gen_intermediate_code(CPUTLGState *env, struct TranslationBlock *tb)
     TCGV_UNUSED_I64(dc->zero);
 
     if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)) {
+        qemu_log_lock();
         qemu_log("IN: %s\n", lookup_symbol(pc_start));
     }
     if (!max_insns) {
@@ -2429,7 +2430,10 @@ void gen_intermediate_code(CPUTLGState *env, struct TranslationBlock *tb)
     tb->size = dc->pc - pc_start;
     tb->icount = num_insns;
 
-    qemu_log_mask(CPU_LOG_TB_IN_ASM, "\n");
+    if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)) {
+        qemu_log("\n");
+        qemu_log_unlock();
+    }
 }
 
 void restore_state_to_opc(CPUTLGState *env, TranslationBlock *tb,
diff --git a/target-tricore/translate.c b/target-tricore/translate.c
index 9a50df9a88..36f734a662 100644
--- a/target-tricore/translate.c
+++ b/target-tricore/translate.c
@@ -8789,9 +8789,11 @@ void gen_intermediate_code(CPUTriCoreState *env, struct TranslationBlock *tb)
 #ifdef DEBUG_DISAS
     if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)
         && qemu_log_in_addr_range(pc_start)) {
+        qemu_log_lock();
         qemu_log("IN: %s\n", lookup_symbol(pc_start));
         log_target_disas(cs, pc_start, ctx.pc - pc_start, 0);
         qemu_log("\n");
+        qemu_log_unlock();
     }
 #endif
 }
diff --git a/target-unicore32/translate.c b/target-unicore32/translate.c
index 09354f92d2..514d460408 100644
--- a/target-unicore32/translate.c
+++ b/target-unicore32/translate.c
@@ -2024,10 +2024,12 @@ done_generating:
 #ifdef DEBUG_DISAS
     if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)
         && qemu_log_in_addr_range(pc_start)) {
+        qemu_log_lock();
         qemu_log("----------------\n");
         qemu_log("IN: %s\n", lookup_symbol(pc_start));
         log_target_disas(cs, pc_start, dc->pc - pc_start, 0);
         qemu_log("\n");
+        qemu_log_unlock();
     }
 #endif
     tb->size = dc->pc - pc_start;
diff --git a/target-xtensa/translate.c b/target-xtensa/translate.c
index fb0fa56f1e..0858c296ea 100644
--- a/target-xtensa/translate.c
+++ b/target-xtensa/translate.c
@@ -3155,10 +3155,12 @@ void gen_intermediate_code(CPUXtensaState *env, TranslationBlock *tb)
 #ifdef DEBUG_DISAS
     if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)
         && qemu_log_in_addr_range(pc_start)) {
+        qemu_log_lock();
         qemu_log("----------------\n");
         qemu_log("IN: %s\n", lookup_symbol(pc_start));
         log_target_disas(cs, pc_start, dc.pc - pc_start, 0);
         qemu_log("\n");
+        qemu_log_unlock();
     }
 #endif
     tb->size = dc.pc - pc_start;
diff --git a/tcg/tcg-op.c b/tcg/tcg-op.c
index bb2bfeef3c..6e2fb3522f 100644
--- a/tcg/tcg-op.c
+++ b/tcg/tcg-op.c
@@ -678,6 +678,33 @@ void tcg_gen_muls2_i32(TCGv_i32 rl, TCGv_i32 rh, TCGv_i32 arg1, TCGv_i32 arg2)
     }
 }
 
+void tcg_gen_mulsu2_i32(TCGv_i32 rl, TCGv_i32 rh, TCGv_i32 arg1, TCGv_i32 arg2)
+{
+    if (TCG_TARGET_REG_BITS == 32) {
+        TCGv_i32 t0 = tcg_temp_new_i32();
+        TCGv_i32 t1 = tcg_temp_new_i32();
+        TCGv_i32 t2 = tcg_temp_new_i32();
+        tcg_gen_mulu2_i32(t0, t1, arg1, arg2);
+        /* Adjust for negative input for the signed arg1.  */
+        tcg_gen_sari_i32(t2, arg1, 31);
+        tcg_gen_and_i32(t2, t2, arg2);
+        tcg_gen_sub_i32(rh, t1, t2);
+        tcg_gen_mov_i32(rl, t0);
+        tcg_temp_free_i32(t0);
+        tcg_temp_free_i32(t1);
+        tcg_temp_free_i32(t2);
+    } else {
+        TCGv_i64 t0 = tcg_temp_new_i64();
+        TCGv_i64 t1 = tcg_temp_new_i64();
+        tcg_gen_ext_i32_i64(t0, arg1);
+        tcg_gen_extu_i32_i64(t1, arg2);
+        tcg_gen_mul_i64(t0, t0, t1);
+        tcg_gen_extr_i64_i32(rl, rh, t0);
+        tcg_temp_free_i64(t0);
+        tcg_temp_free_i64(t1);
+    }
+}
+
 void tcg_gen_ext8s_i32(TCGv_i32 ret, TCGv_i32 arg)
 {
     if (TCG_TARGET_HAS_ext8s_i32) {
@@ -790,7 +817,7 @@ void tcg_gen_ld8u_i64(TCGv_i64 ret, TCGv_ptr arg2, tcg_target_long offset)
 void tcg_gen_ld8s_i64(TCGv_i64 ret, TCGv_ptr arg2, tcg_target_long offset)
 {
     tcg_gen_ld8s_i32(TCGV_LOW(ret), arg2, offset);
-    tcg_gen_sari_i32(TCGV_HIGH(ret), TCGV_HIGH(ret), 31);
+    tcg_gen_sari_i32(TCGV_HIGH(ret), TCGV_LOW(ret), 31);
 }
 
 void tcg_gen_ld16u_i64(TCGv_i64 ret, TCGv_ptr arg2, tcg_target_long offset)
@@ -1748,6 +1775,22 @@ void tcg_gen_muls2_i64(TCGv_i64 rl, TCGv_i64 rh, TCGv_i64 arg1, TCGv_i64 arg2)
     }
 }
 
+void tcg_gen_mulsu2_i64(TCGv_i64 rl, TCGv_i64 rh, TCGv_i64 arg1, TCGv_i64 arg2)
+{
+    TCGv_i64 t0 = tcg_temp_new_i64();
+    TCGv_i64 t1 = tcg_temp_new_i64();
+    TCGv_i64 t2 = tcg_temp_new_i64();
+    tcg_gen_mulu2_i64(t0, t1, arg1, arg2);
+    /* Adjust for negative input for the signed arg1.  */
+    tcg_gen_sari_i64(t2, arg1, 63);
+    tcg_gen_and_i64(t2, t2, arg2);
+    tcg_gen_sub_i64(rh, t1, t2);
+    tcg_gen_mov_i64(rl, t0);
+    tcg_temp_free_i64(t0);
+    tcg_temp_free_i64(t1);
+    tcg_temp_free_i64(t2);
+}
+
 /* Size changing operations.  */
 
 void tcg_gen_extrl_i64_i32(TCGv_i32 ret, TCGv_i64 arg)
diff --git a/tcg/tcg-op.h b/tcg/tcg-op.h
index 89b59e867a..6d044b7c5b 100644
--- a/tcg/tcg-op.h
+++ b/tcg/tcg-op.h
@@ -306,6 +306,7 @@ void tcg_gen_sub2_i32(TCGv_i32 rl, TCGv_i32 rh, TCGv_i32 al,
                       TCGv_i32 ah, TCGv_i32 bl, TCGv_i32 bh);
 void tcg_gen_mulu2_i32(TCGv_i32 rl, TCGv_i32 rh, TCGv_i32 arg1, TCGv_i32 arg2);
 void tcg_gen_muls2_i32(TCGv_i32 rl, TCGv_i32 rh, TCGv_i32 arg1, TCGv_i32 arg2);
+void tcg_gen_mulsu2_i32(TCGv_i32 rl, TCGv_i32 rh, TCGv_i32 arg1, TCGv_i32 arg2);
 void tcg_gen_ext8s_i32(TCGv_i32 ret, TCGv_i32 arg);
 void tcg_gen_ext16s_i32(TCGv_i32 ret, TCGv_i32 arg);
 void tcg_gen_ext8u_i32(TCGv_i32 ret, TCGv_i32 arg);
@@ -482,6 +483,7 @@ void tcg_gen_sub2_i64(TCGv_i64 rl, TCGv_i64 rh, TCGv_i64 al,
                       TCGv_i64 ah, TCGv_i64 bl, TCGv_i64 bh);
 void tcg_gen_mulu2_i64(TCGv_i64 rl, TCGv_i64 rh, TCGv_i64 arg1, TCGv_i64 arg2);
 void tcg_gen_muls2_i64(TCGv_i64 rl, TCGv_i64 rh, TCGv_i64 arg1, TCGv_i64 arg2);
+void tcg_gen_mulsu2_i64(TCGv_i64 rl, TCGv_i64 rh, TCGv_i64 arg1, TCGv_i64 arg2);
 void tcg_gen_not_i64(TCGv_i64 ret, TCGv_i64 arg);
 void tcg_gen_ext8s_i64(TCGv_i64 ret, TCGv_i64 arg);
 void tcg_gen_ext16s_i64(TCGv_i64 ret, TCGv_i64 arg);
@@ -956,6 +958,7 @@ void tcg_gen_atomic_xor_fetch_i64(TCGv_i64, TCGv, TCGv_i64, TCGArg, TCGMemOp);
 #define tcg_gen_sub2_tl tcg_gen_sub2_i64
 #define tcg_gen_mulu2_tl tcg_gen_mulu2_i64
 #define tcg_gen_muls2_tl tcg_gen_muls2_i64
+#define tcg_gen_mulsu2_tl tcg_gen_mulsu2_i64
 #define tcg_gen_atomic_cmpxchg_tl tcg_gen_atomic_cmpxchg_i64
 #define tcg_gen_atomic_xchg_tl tcg_gen_atomic_xchg_i64
 #define tcg_gen_atomic_fetch_add_tl tcg_gen_atomic_fetch_add_i64
@@ -1043,6 +1046,7 @@ void tcg_gen_atomic_xor_fetch_i64(TCGv_i64, TCGv, TCGv_i64, TCGArg, TCGMemOp);
 #define tcg_gen_sub2_tl tcg_gen_sub2_i32
 #define tcg_gen_mulu2_tl tcg_gen_mulu2_i32
 #define tcg_gen_muls2_tl tcg_gen_muls2_i32
+#define tcg_gen_mulsu2_tl tcg_gen_mulsu2_i32
 #define tcg_gen_atomic_cmpxchg_tl tcg_gen_atomic_cmpxchg_i32
 #define tcg_gen_atomic_xchg_tl tcg_gen_atomic_xchg_i32
 #define tcg_gen_atomic_fetch_add_tl tcg_gen_atomic_fetch_add_i32
diff --git a/tcg/tcg.c b/tcg/tcg.c
index 2d3e498bc2..aabf94f365 100644
--- a/tcg/tcg.c
+++ b/tcg/tcg.c
@@ -412,10 +412,12 @@ void tcg_prologue_init(TCGContext *s)
 
 #ifdef DEBUG_DISAS
     if (qemu_loglevel_mask(CPU_LOG_TB_OUT_ASM)) {
+        qemu_log_lock();
         qemu_log("PROLOGUE: [size=%zu]\n", prologue_size);
         log_disas(buf0, prologue_size);
         qemu_log("\n");
         qemu_log_flush();
+        qemu_log_unlock();
     }
 #endif
 }
@@ -2542,9 +2544,11 @@ int tcg_gen_code(TCGContext *s, TranslationBlock *tb)
 #ifdef DEBUG_DISAS
     if (unlikely(qemu_loglevel_mask(CPU_LOG_TB_OP)
                  && qemu_log_in_addr_range(tb->pc))) {
+        qemu_log_lock();
         qemu_log("OP:\n");
         tcg_dump_ops(s);
         qemu_log("\n");
+        qemu_log_unlock();
     }
 #endif
 
@@ -2570,9 +2574,11 @@ int tcg_gen_code(TCGContext *s, TranslationBlock *tb)
 #ifdef DEBUG_DISAS
             if (unlikely(qemu_loglevel_mask(CPU_LOG_TB_OP_IND)
                          && qemu_log_in_addr_range(tb->pc))) {
+                qemu_log_lock();
                 qemu_log("OP before indirect lowering:\n");
                 tcg_dump_ops(s);
                 qemu_log("\n");
+                qemu_log_unlock();
             }
 #endif
             /* Replace indirect temps with direct temps.  */
@@ -2590,9 +2596,11 @@ int tcg_gen_code(TCGContext *s, TranslationBlock *tb)
 #ifdef DEBUG_DISAS
     if (unlikely(qemu_loglevel_mask(CPU_LOG_TB_OP_OPT)
                  && qemu_log_in_addr_range(tb->pc))) {
+        qemu_log_lock();
         qemu_log("OP after optimization and liveness analysis:\n");
         tcg_dump_ops(s);
         qemu_log("\n");
+        qemu_log_unlock();
     }
 #endif
 
diff --git a/tcg/tcg.h b/tcg/tcg.h
index dc1281fb4e..a35e4c4fd4 100644
--- a/tcg/tcg.h
+++ b/tcg/tcg.h
@@ -376,14 +376,36 @@ static inline unsigned get_alignment_bits(TCGMemOp memop)
 
 typedef tcg_target_ulong TCGArg;
 
-/* Define a type and accessor macros for variables.  Using pointer types
-   is nice because it gives some level of type safely.  Converting to and
-   from intptr_t rather than int reduces the number of sign-extension
-   instructions that get implied on 64-bit hosts.  Users of tcg_gen_* don't
-   need to know about any of this, and should treat TCGv as an opaque type.
-   In addition we do typechecking for different types of variables.  TCGv_i32
-   and TCGv_i64 are 32/64-bit variables respectively.  TCGv and TCGv_ptr
-   are aliases for target_ulong and host pointer sized values respectively.  */
+/* Define type and accessor macros for TCG variables.
+
+   TCG variables are the inputs and outputs of TCG ops, as described
+   in tcg/README. Target CPU front-end code uses these types to deal
+   with TCG variables as it emits TCG code via the tcg_gen_* functions.
+   They come in several flavours:
+    * TCGv_i32 : 32 bit integer type
+    * TCGv_i64 : 64 bit integer type
+    * TCGv_ptr : a host pointer type
+    * TCGv : an integer type the same size as target_ulong
+             (an alias for either TCGv_i32 or TCGv_i64)
+   The compiler's type checking will complain if you mix them
+   up and pass the wrong sized TCGv to a function.
+
+   Users of tcg_gen_* don't need to know about any of the internal
+   details of these, and should treat them as opaque types.
+   You won't be able to look inside them in a debugger either.
+
+   Internal implementation details follow:
+
+   Note that there is no definition of the structs TCGv_i32_d etc anywhere.
+   This is deliberate, because the values we store in variables of type
+   TCGv_i32 are not really pointers-to-structures. They're just small
+   integers, but keeping them in pointer types like this means that the
+   compiler will complain if you accidentally pass a TCGv_i32 to a
+   function which takes a TCGv_i64, and so on. Only the internals of
+   TCG need to care about the actual contents of the types, and they always
+   box and unbox via the MAKE_TCGV_* and GET_TCGV_* functions.
+   Converting to and from intptr_t rather than int reduces the number
+   of sign-extension instructions that get implied on 64-bit hosts.  */
 
 typedef struct TCGv_i32_d *TCGv_i32;
 typedef struct TCGv_i64_d *TCGv_i64;
diff --git a/translate-all.c b/translate-all.c
index e6a8b07363..3dd9214904 100644
--- a/translate-all.c
+++ b/translate-all.c
@@ -1355,10 +1355,12 @@ TranslationBlock *tb_gen_code(CPUState *cpu,
 #ifdef DEBUG_DISAS
     if (qemu_loglevel_mask(CPU_LOG_TB_OUT_ASM) &&
         qemu_log_in_addr_range(tb->pc)) {
+        qemu_log_lock();
         qemu_log("OUT: [size=%d]\n", gen_code_size);
         log_disas(tb->tc_ptr, gen_code_size);
         qemu_log("\n");
         qemu_log_flush();
+        qemu_log_unlock();
     }
 #endif