Re: [Patch 2/2] tools/libxc: Prevent erroneous success from xc_domain_restore

[Andrew Cooper <andrew.cooper3@citrix.com>]

On 04/02/14 17:22, Ian Campbell wrote:
> On Tue, 2014-02-04 at 17:16 +0000, Andrew Cooper wrote:
>>>                  goto out;
>>>              }
>>>          } else {
>>> -            rc = -1;
> Mostly looks good but I'm not sure about this change 
>
> We get here on input error (toolstack data available but no callback
> provided) which is neither migration success nor failure, it's a bug in
> the caller. So arguably returning a separate failure from
> success/unsuccess makes sense.
>
> I'd have thought it ought to set errno (too EINVAL perhaps) too, but
> lets not mess with that now.
>
>
> Ian.
>

Hilariously, it turns out that xc_domain_restore() is specified to
return 0 on success and -1 on failure.  From what I can tell, this is
the sole action which would cause xc_domain_restore() to return anything
other than 0 or 1.

I think fixing this should fall into the bucket of "sanitisation of
libxc error paths".

~Andrew