accel/tcg: Precompute curr_cflags into cpu->tcg_cflags
The primary motivation is to remove a dozen insns along
the fast-path in tb_lookup. As a byproduct, this allows
us to completely remove parallel_cpus.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index 389ec09..9522f60 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -6481,6 +6481,16 @@
/* Grab a mutex so that thread setup appears atomic. */
pthread_mutex_lock(&clone_lock);
+ /*
+ * If this is our first additional thread, we need to ensure we
+ * generate code for parallel execution and flush old translations.
+ * Do this now so that the copy gets CF_PARALLEL too.
+ */
+ if (!(cpu->tcg_cflags & CF_PARALLEL)) {
+ cpu->tcg_cflags |= CF_PARALLEL;
+ tb_flush(cpu);
+ }
+
/* we create a new CPU instance. */
new_env = cpu_copy(env);
/* Init regs that differ from the parent. */
@@ -6521,14 +6531,6 @@
sigprocmask(SIG_BLOCK, &sigmask, &info.sigmask);
cpu->random_seed = qemu_guest_random_seed_thread_part1();
- /* If this is our first additional thread, we need to ensure we
- * generate code for parallel execution and flush old translations.
- */
- if (!parallel_cpus) {
- parallel_cpus = true;
- tb_flush(cpu);
- }
-
ret = pthread_create(&info.thread, &attr, clone_func, &info);
/* TODO: Free new CPU state if thread creation failed. */
