qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] Bug AioContext polling is not implemented on Windows wi


From: Peter Xu
Subject: Re: [Qemu-devel] Bug AioContext polling is not implemented on Windows with qemu-system-ppc.exe
Date: Thu, 22 Mar 2018 16:52:50 +0800
User-agent: Mutt/1.9.1 (2017-09-22)

On Thu, Mar 22, 2018 at 09:28:34AM +0100, Howard Spoelstra wrote:
> > Hi, Howard,
> >
> > Thanks for reporting the problem.  Could you try applying below patch
> > to your rc0 tree and rebuild QEMU, to see whether it fixes your problem?
> >
> > -----------------------------
> > diff --git a/iothread.c b/iothread.c
> > index 1b3463cb00..9f52270ae1 100644
> > --- a/iothread.c
> > +++ b/iothread.c
> > @@ -31,11 +31,15 @@ typedef ObjectClass IOThreadClass;
> >  #define IOTHREAD_CLASS(klass) \
> >     OBJECT_CLASS_CHECK(IOThreadClass, klass, TYPE_IOTHREAD)
> >
> > +#ifdef CONFIG_LINUX
> >  /* Benchmark results from 2016 on NVMe SSD drives show max polling times 
> > around
> >   * 16-32 microseconds yield IOPS improvements for both iodepth=1 and 
> > iodepth=32
> >   * workloads.
> >   */
> >  #define IOTHREAD_POLL_MAX_NS_DEFAULT 32768ULL
> > +#else
> > +#define IOTHREAD_POLL_MAX_NS_DEFAULT 0ULL
> > +#endif
> >
> >  static __thread IOThread *my_iothread;
> >
> > diff --git a/util/aio-win32.c b/util/aio-win32.c
> > index d6d5e02f00..a67b00c6ad 100644
> > --- a/util/aio-win32.c
> > +++ b/util/aio-win32.c
> > @@ -410,5 +410,7 @@ void aio_context_setup(AioContext *ctx)
> >  void aio_context_set_poll_params(AioContext *ctx, int64_t max_ns,
> >                                   int64_t grow, int64_t shrink, Error 
> > **errp)
> >  {
> > -    error_setg(errp, "AioContext polling is not implemented on Windows");
> > +    if (max_ns) {
> > +        error_setg(errp, "AioContext polling is not implemented on 
> > Windows");
> > +    }
> >  }
> > -----------------------------
> >
> > Thanks,
> >
> > --
> > Peter Xu
> 
> Hi Peter,
> 
> Thanks, that patch does indeed fix the problem.

Thank you for testing that.  I'll post a formal patch soon.  Hopefully
it can reach rc1.

> 
> Best,
> Howard

-- 
Peter Xu



reply via email to

[Prev in Thread] Current Thread [Next in Thread]