qom: aggressively optimize qom casting

[qemu.git] / cpu-exec.c

diff --git a/cpu-exec.c b/cpu-exec.c
index b87c1c07c65a429992877d06c8c03c2cf8ce22d3..31c089dac0a7db66ac4d6270c293d1294bbc598a 100644 (file)
--- a/cpu-exec.c
+++ b/cpu-exec.c
@@ -23,8 +23,6 @@
 #include "qemu/atomic.h"
 #include "sysemu/qtest.h"
 
-//#define CONFIG_DEBUG_EXEC
-
 bool qemu_cpu_has_work(CPUState *cpu)
 {
     return cpu_has_work(cpu);
@@ -217,6 +215,14 @@ int cpu_exec(CPUArchState *env)
 
     cpu_single_env = env;
 
+    /* As long as cpu_single_env is null, up to the assignment just above,
+     * requests by other threads to exit the execution loop are expected to
+     * be issued using the exit_request global. We must make sure that our
+     * evaluation of the global value is performed past the cpu_single_env
+     * value transition point, which requires a memory barrier as well as
+     * an instruction scheduling constraint on modern architectures.  */
+    smp_mb();
+
     if (unlikely(exit_request)) {
         cpu->exit_request = 1;
     }
@@ -567,14 +573,11 @@ int cpu_exec(CPUArchState *env)
                     env->exception_index = EXCP_INTERRUPT;
                     cpu_loop_exit(env);
                 }
-#if defined(DEBUG_DISAS) || defined(CONFIG_DEBUG_EXEC)
+#if defined(DEBUG_DISAS)
                 if (qemu_loglevel_mask(CPU_LOG_TB_CPU)) {
                     /* restore flags in standard format */
 #if defined(TARGET_I386)
-                    env->eflags = env->eflags | cpu_cc_compute_all(env, CC_OP)
-                        | (DF & DF_MASK);
                     log_cpu_state(env, CPU_DUMP_CCOP);
-                    env->eflags &= ~(DF_MASK | CC_O | CC_S | CC_Z | CC_A | CC_P | CC_C);
 #elif defined(TARGET_M68K)
                     cpu_m68k_flush_flags(env, env->cc_op);
                     env->cc_op = CC_OP_FLAGS;
@@ -585,7 +588,7 @@ int cpu_exec(CPUArchState *env)
                     log_cpu_state(env, 0);
 #endif
                 }
-#endif /* DEBUG_DISAS || CONFIG_DEBUG_EXEC */
+#endif /* DEBUG_DISAS */
                 spin_lock(&tcg_ctx.tb_ctx.tb_lock);
                 tb = tb_find_fast(env);
                 /* Note: we do it here to avoid a gcc bug on Mac OS X when
@@ -597,11 +600,10 @@ int cpu_exec(CPUArchState *env)
                     next_tb = 0;
                     tcg_ctx.tb_ctx.tb_invalidated_flag = 0;
                 }
-#ifdef CONFIG_DEBUG_EXEC
-                qemu_log_mask(CPU_LOG_EXEC, "Trace %p [" TARGET_FMT_lx "] %s\n",
-                             tb->tc_ptr, tb->pc,
-                             lookup_symbol(tb->pc));
-#endif
+                if (qemu_loglevel_mask(CPU_LOG_EXEC)) {
+                    qemu_log("Trace %p [" TARGET_FMT_lx "] %s\n",
+                             tb->tc_ptr, tb->pc, lookup_symbol(tb->pc));
+                }
                 /* see if we can patch the calling TB. When the TB
                    spans two pages, we cannot safely do a direct
                    jump. */