On 09/24/2016 04:52 AM, Paolo Bonzini wrote:
----- Original Message -----
From: "Richard Henderson" <[email protected]>
To: "Paolo Bonzini" <[email protected]>, [email protected]
Cc: "serge fdrv" <[email protected]>, [email protected], "alex bennee"
<[email protected]>, "sergey fedorov"
<[email protected]>
Sent: Friday, September 23, 2016 8:23:46 PM
Subject: Re: [Qemu-devel] [PATCH 16/16] cpus-common: lock-free fast path for
cpu_exec_start/end
On 09/23/2016 12:31 AM, Paolo Bonzini wrote:
+ if (atomic_read(&other_cpu->running)) {
...
+ atomic_set(&cpu->running, true);
...
+ cpu->running = false;
...
+ cpu->running = true;
Inconsistent use of atomics. I don't see that the cpu_list_lock protects the
last two lines in any way.
It does:
qemu_mutex_lock(&qemu_cpu_list_lock);
What I meant is that I don't see that the mutex avoids the need for atomic_set.
but I can change it anyway to atomic_set.
Thanks,
r~
