Re: [Qemu-devel] [PATCH] xen-platform: Replace assert() with appropriate error reporting

[Eduardo Habkost]

On Tue, Oct 20, 2015 at 11:01:16AM +0100, Stefano Stabellini wrote:
> On Mon, 19 Oct 2015, Paolo Bonzini wrote:
> > On 19/10/2015 20:39, Eduardo Habkost wrote:
> > > Commit dbb7405d8caad0814ceddd568cb49f163a847561 made it possible to
> > > trigger an assert using "-device xen-platform". Replace it with
> > > appropriate error reporting.
> > > 
> > > Before:
> > > 
> > >   $ qemu-system-x86_64 -device xen-platform
> > >   qemu-system-x86_64: hw/i386/xen/xen_platform.c:391: 
> > > xen_platform_initfn: Assertion `xen_enabled()' failed.
> > >   Aborted (core dumped)
> > >   $
> > > 
> > > After:
> > > 
> > >   $ qemu-system-x86_64 -device xen-platform
> > >   qemu-system-x86_64: -device xen-platform: xen-platform device requires 
> > > the Xen accelerator
> > >   qemu-system-x86_64: -device xen-platform: Device initialization failed
> > >   $
> > > 
> > > Signed-off-by: Eduardo Habkost <address@hidden>
> > > ---
> > >  hw/i386/xen/xen_platform.c | 6 +++++-
> > >  1 file changed, 5 insertions(+), 1 deletion(-)
> > > 
> > > diff --git a/hw/i386/xen/xen_platform.c b/hw/i386/xen/xen_platform.c
> > > index 8682c42..5667f29 100644
> > > --- a/hw/i386/xen/xen_platform.c
> > > +++ b/hw/i386/xen/xen_platform.c
> > > @@ -33,6 +33,7 @@
> > >  #include "trace.h"
> > >  #include "exec/address-spaces.h"
> > >  #include "sysemu/block-backend.h"
> > > +#include "qemu/error-report.h"
> > >  
> > >  #include <xenguest.h>
> > >  
> > > @@ -388,7 +389,10 @@ static int xen_platform_initfn(PCIDevice *dev)
> > >      uint8_t *pci_conf;
> > >  
> > >      /* Device will crash on reset if xen is not initialized */
> > > -    assert(xen_enabled());
> > > +    if (!xen_enabled()) {
> > > +        error_report("xen-platform device requires the Xen accelerator");
> > 
> > You can use PCIDeviceClass's realize (see Stefano's
> > http://thread.gmane.org/gmane.comp.emulators.qemu/369998) to achieve
> > better error propagation.
> 

Oops, I misunderstood the previous message and I thought Paolo meant the
xen_enabled() check was being done too early, and not at realize time. I
wasn't aware that PCIDeviceClass::realize existed.

> Eduardo, if you are happy with that, I'll wait for your respin :-)

Yes, I will send a new version based on your conversion to k->realize.
Thanks!

-- 
Eduardo