Re: [Qemu-devel] virtqueue corruption in emulation mode?

[Stefan Hajnoczi <stefanha@gmail.com>]

On Mon, Sep 26, 2011 at 07:16:56PM -0500, Sinha, Ani wrote:
> I am using the virtqueue (virtqueue_pop, virtqueue_push etc) in the emulated mode (non-kvm mode) from an IO thread (a separate thread different from main QEMU thread). What I am observing is that the virtqueue memory seems to get corrupt. Either qemu crashes while performing virtqueue_push() (virtqueue_push() -> virtqueue_fill() ->bring_used_idx()->lduw_phys()->qemu_get_ram_ptr()->"bad ram offset") or crashes when the guest accesses a bad memory while using virtqueue. Now this never ever happens when I run QEMU in KVM mode (/dev/kvm present) OR when I use my functions from within the main qemu thread. I am unable to figure out why this is happening. I have looked into my code over and over again and I can't seem to explain this behavior. Can any of you guys give me any inkling?

QEMU is not thread-safe in general.  It uses a big lock to protect most
of its internal state.

When you say "an IO thread" it sounds like you spawn a new thread
outside the big lock (qemu_global_mutex).  You cannot call the existing
virtqueue functions outside the big lock because they traverse (and
modify!) the memory management data structures.

Please call new threads "helper threads" or something other than "IO
thread" because I/O thread has a specific meaning in QEMU.  It's the
event loop thread that execute main_loop_wait() and dispatches fd
handlers when select(2) returns.  This will prevent confusion.

If you follow the way that existing virtio devices are implemented there
should be no problem.

Stefan