qemu-devel
[Top][All Lists]
Advanced

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

Re: [Qemu-devel] [PATCH 06/11] pseries: Improve error handling in find_u


From: David Gibson
Subject: Re: [Qemu-devel] [PATCH 06/11] pseries: Improve error handling in find_unknown_sysbus_device()
Date: Mon, 14 Dec 2015 12:21:45 +1100
User-agent: Mutt/1.5.24 (2015-08-30)

On Fri, Dec 11, 2015 at 08:15:39AM -0700, Eric Blake wrote:
> On 12/10/2015 05:11 PM, David Gibson wrote:
> > Use error_setg() to return an error instead of using an explicit exit().
> > 
> > Signed-off-by: David Gibson <address@hidden>
> > ---
> >  hw/ppc/spapr.c | 9 +++++----
> >  1 file changed, 5 insertions(+), 4 deletions(-)
> > 
> 
> > +++ b/hw/ppc/spapr.c
> > @@ -1107,6 +1107,7 @@ static void spapr_reset_htab(sPAPRMachineState 
> > *spapr, Error **errp)
> >  
> >  static int find_unknown_sysbus_device(SysBusDevice *sbdev, void *opaque)
> >  {
> 
> >      if (!matched) {
> > -        error_report("Device %s is not supported by this machine yet.",
> > -                     qdev_fw_name(DEVICE(sbdev)));
> > -        exit(1);
> > +        error_setg(errp,
> > +                   "Device %s is not supported by this machine yet",
> > +                   qdev_fw_name(DEVICE(sbdev)));
> >      }
> >  
> >      return 0;
> 
> It looks like find_unknown_sysbus_device is designed to be called in a
> loop, and that returning 0 lets the loop continue.
> 
> > @@ -1151,7 +1152,7 @@ static void ppc_spapr_reset(void)
> >      uint32_t rtas_limit;
> >  
> >      /* Check for unknown sysbus devices */
> > -    foreach_dynamic_sysbus_device(find_unknown_sysbus_device, NULL);
> > +    foreach_dynamic_sysbus_device(find_unknown_sysbus_device, 
> > &error_abort);
> 
> If a caller passes something other than &error_abort as the opaque, AND
> the error condition occurs more than once in the loop iteration, then
> you really need to change find_unknown_sysbus_device() to return
> non-zero after raising error on the first failure, so that the loop
> doesn't continue on to a second pass and attempt an error_setg() onto an
> already-set error.

Ah, good catch.  I'll correct in the next spin.

> Of course, since _this_ patch uses &error_abort as the only client, the
> loop will never continue after the first failure, but it's more robust
> to be clean without having to audit the callers.
> 



-- 
David Gibson                    | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au  | minimalist, thank you.  NOT _the_ _other_
                                | _way_ _around_!
http://www.ozlabs.org/~dgibson

Attachment: signature.asc
Description: PGP signature


reply via email to

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