cpus: ignore ESRCH in qemu_cpu_kick_thread()
authorLaurent Vivier <lvivier@redhat.com>
Wed, 2 Jan 2019 14:16:03 +0000 (15:16 +0100)
committerPaolo Bonzini <pbonzini@redhat.com>
Tue, 5 Feb 2019 15:50:16 +0000 (16:50 +0100)
We can have a race condition between qemu_cpu_kick_thread() and
qemu_kvm_cpu_thread_fn() when we hotunplug a CPU. In this case,
qemu_cpu_kick_thread() can try to kick a thread that is exiting.
pthread_kill() returns an error and qemu is stopped by an exit(1).

   qemu:qemu_cpu_kick_thread: No such process

We can ignore safely this error.

Signed-off-by: Laurent Vivier <lvivier@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
cpus.c

diff --git a/cpus.c b/cpus.c
index b09b7027126183338463268f868b5dddf6a3e679..154daf57dcb02ab546ddc3ddafb833b240418158 100644 (file)
--- a/cpus.c
+++ b/cpus.c
@@ -1778,7 +1778,7 @@ static void qemu_cpu_kick_thread(CPUState *cpu)
     }
     cpu->thread_kicked = true;
     err = pthread_kill(cpu->thread->thread, SIG_IPI);
-    if (err) {
+    if (err && err != ESRCH) {
         fprintf(stderr, "qemu:%s: %s", __func__, strerror(err));
         exit(1);
     }