Re: [Qemu-devel] [RFC] use little granularity lock to substitue qemu_mutex_lock_iothread

[Jan Kiszka]

On 2012-06-21 16:49, Liu Ping Fan wrote:
> Nowadays, we use qemu_mutex_lock_iothread()/qemu_mutex_unlock_iothread() to
> protect the race to access the emulated dev launched by vcpu threads & 
> iothread.
> 
> But this lock is too big. We can break it down.
> These patches separate the CPUArchState's protection from the other devices, 
> so we
> can have a per-cpu lock for each CPUArchState, not the big lock any longer.

Anything that reduces lock dependencies is generally welcome. But can
you specify in more details what you gain, and under which conditions?

I'm skeptical if this is the right area to start. With the in-kernel
irqchip enabled, no CPUArchState field is touched during normal
operations (unless I missed something subtle in the past). At the same
time, this locking is unfortunately fairly complex and invasive, so not
"cheap" to integrate.

IMO more interesting is breaking out some I/O path, e.g. from a NIC to a
network backend, and get this processed in a separate thread without
touching the BQL (Big QEMU Lock). I've experimental patches for this
here, but they need rebasing and polishing.

Thanks,
Jan

-- 
Siemens AG, Corporate Technology, CT T DE IT 1
Corporate Competence Center Embedded Linux