qemu-devel
[Top][All Lists]
Advanced

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

Re: [Qemu-devel] [PATCH 2/7] virtio-blk: handle virtio_blk_handle_reques


From: Greg Kurz
Subject: Re: [Qemu-devel] [PATCH 2/7] virtio-blk: handle virtio_blk_handle_request() errors
Date: Wed, 21 Sep 2016 17:01:32 +0200

On Wed, 21 Sep 2016 16:28:34 +0200
Cornelia Huck <address@hidden> wrote:

> On Wed, 21 Sep 2016 15:14:09 +0200
> Greg Kurz <address@hidden> wrote:
> 
> > All these errors are caused by a buggy guest: let's switch the device to
> > the broken state instead of terminating QEMU.
> > 
> > Signed-off-by: Greg Kurz <address@hidden>
> > ---
> >  hw/block/virtio-blk.c |   27 +++++++++++++++++----------
> >  1 file changed, 17 insertions(+), 10 deletions(-)
> > 
> > diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c
> > index 3a6112fbf4c4..1285d196a40f 100644
> > --- a/hw/block/virtio-blk.c
> > +++ b/hw/block/virtio-blk.c
> > @@ -468,30 +468,32 @@ static bool virtio_blk_sect_range_ok(VirtIOBlock *dev,
> >      return true;
> >  }
> > 
> > -void virtio_blk_handle_request(VirtIOBlockReq *req, MultiReqBuffer *mrb)
> > +int virtio_blk_handle_request(VirtIOBlockReq *req, MultiReqBuffer *mrb)  
> 
> Unrelated to your patch: It seems there are no callers (left) outside
> of this file; should the function be made static?
> 

Yeah you're right. The dataplane does not call this function anymore since:

"03de2f527499 virtio-blk: do not use vring in dataplane"

I'll fix this in a preparatory patch.

> Related to your patch: You didn't change the prototype in the header :)
> 

Oops... caught in the act of not running a final build before sending :)

> (...)
> 
> > @@ -586,7 +589,9 @@ void virtio_blk_handle_vq(VirtIOBlock *s, VirtQueue *vq)
> >      blk_io_plug(s->blk);
> > 
> >      while ((req = virtio_blk_get_request(s, vq))) {
> > -        virtio_blk_handle_request(req, &mrb);
> > +        if (virtio_blk_handle_request(req, &mrb)) {
> > +            return;  
> 
> Does the missing blk_io_unplug() have any side-effects outside of this
> device, which is broken anyway?
> 

Ugh, this is not intentional... and since it may go down many paths, I
won't bet there are no side effects.

I'll fix this with an explicit err_out: label.

> > +        }
> >      }
> > 
> >      if (mrb.num_reqs) {  
> 
> 




reply via email to

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