Xen devel - Apr 2013 - [PATCH] libxl: adjust point of backend name resolution

Resolution of a backend name to a domid needs to happen a little earlier
in some cases.

For example, if a domU is specified as a backend for a
disk and, as previously written, libxl__device_disk_setdefault() calls
libxl__resolve_domid() last, then disk->backend_domid still equals
LIBXL_TOOLSTACK_DOMID when libxl__device_disk_set_backend() is called.
This results in libxl__device_disk_set_backend() making an incorrect
attempt to validate the target by calling stat() on a file on dom0,
resulting in ERROR_INVAL (see libxl_device.c lines 239-248), which
prevents creation of the frontend domain.

Likewise, libxl__device_nic_setdefault() previously made use of
nic->backend_domid before it was set.

Signed-off-by: Eric Shelton <eshelton@pobox.com>
Cc: Daniel De Graaf <dgdegra@tycho.nsa.gov>

---

 tools/libxl/libxl.c |   11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/tools/libxl/libxl.c b/tools/libxl/libxl.c
index 30302c7..fcab76a 100644
--- a/tools/libxl/libxl.c
+++ b/tools/libxl/libxl.c
@@ -1978,10 +1978,10 @@ int libxl__device_disk_setdefault(libxl__gc
*gc, libxl_device_disk *disk)
 {
     int rc;

-    rc = libxl__device_disk_set_backend(gc, disk);
-    if (rc) return rc;
-
     rc = libxl__resolve_domid(gc, disk->backend_domname,
&disk->backend_domid);
+    if (rc < 0) return rc;
+
+    rc = libxl__device_disk_set_backend(gc, disk);
     return rc;
 }

@@ -2812,6 +2812,10 @@ int libxl__device_nic_setdefault(libxl__gc *gc,
libxl_device_nic *nic,
         LOG(ERROR, "unable to get current hotplug scripts execution
setting");
         return run_hotplug_scripts;
     }
+
+    rc = libxl__resolve_domid(gc, nic->backend_domname,
&nic->backend_domid);
+    if (rc < 0) return rc;
+
     if (nic->backend_domid != LIBXL_TOOLSTACK_DOMID &&
run_hotplug_scripts) {
         LOG(ERROR, "cannot use a backend domain different than %d if"
                    "hotplug scripts are executed from libxl",
@@ -2837,7 +2841,6 @@ int libxl__device_nic_setdefault(libxl__gc *gc,
libxl_device_nic *nic,
         abort();
     }

-    rc = libxl__resolve_domid(gc, nic->backend_domname,
&nic->backend_domid);
     return rc;
 }

--
1.7.10.2 (Apple Git-33)


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

On Tue, 2013-04-30 at 16:03 +0100, Eric Shelton wrote:
> Resolution of a backend name to a domid needs to happen a little earlier
> in some cases.
> 
> For example, if a domU is specified as a backend for a
> disk and, as previously written, libxl__device_disk_setdefault() calls
> libxl__resolve_domid() last, then disk->backend_domid still equals
> LIBXL_TOOLSTACK_DOMID when libxl__device_disk_set_backend() is called.
> This results in libxl__device_disk_set_backend() making an incorrect
> attempt to validate the target by calling stat() on a file on dom0,
> resulting in ERROR_INVAL (see libxl_device.c lines 239-248), which
> prevents creation of the frontend domain.
> 
> Likewise, libxl__device_nic_setdefault() previously made use of
> nic->backend_domid before it was set.
> 
> Signed-off-by: Eric Shelton <eshelton@pobox.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
> Cc: Daniel De Graaf <dgdegra@tycho.nsa.gov>
Daniel, any comments?

On 04/30/2013 11:12 AM, Ian Campbell wrote:
> On Tue, 2013-04-30 at 16:03 +0100, Eric Shelton wrote:
>> Resolution of a backend name to a domid needs to happen a little
earlier
>> in some cases.
>>
>> For example, if a domU is specified as a backend for a
>> disk and, as previously written, libxl__device_disk_setdefault() calls
>> libxl__resolve_domid() last, then disk->backend_domid still equals
>> LIBXL_TOOLSTACK_DOMID when libxl__device_disk_set_backend() is called.
>> This results in libxl__device_disk_set_backend() making an incorrect
>> attempt to validate the target by calling stat() on a file on dom0,
>> resulting in ERROR_INVAL (see libxl_device.c lines 239-248), which
>> prevents creation of the frontend domain.
>>
>> Likewise, libxl__device_nic_setdefault() previously made use of
>> nic->backend_domid before it was set.
>>
>> Signed-off-by: Eric Shelton <eshelton@pobox.com>
>
> Acked-by: Ian Campbell <ian.campbell@citrix.com>
>
>> Cc: Daniel De Graaf <dgdegra@tycho.nsa.gov>
>
> Daniel, any comments?
>
Yes, the backend_domname conversion needs to happen earlier or the checks
against LIBXL_TOOLSTACK_DOMID won''t work right - I actually had run
into
this and was bypassing it by specifying a script, so this fix is needed.

Reviewed-by: Daniel De Graaf <dgdegra@tycho.nsa.gov>

On Tue, 2013-04-30 at 18:55 +0100, Daniel De Graaf
wrote:
> On 04/30/2013 11:12 AM, Ian Campbell wrote:
> > On Tue, 2013-04-30 at 16:03 +0100, Eric Shelton wrote:
> >> Resolution of a backend name to a domid needs to happen a little
earlier
> >> in some cases.
> >>
> >> For example, if a domU is specified as a backend for a
> >> disk and, as previously written, libxl__device_disk_setdefault()
calls
> >> libxl__resolve_domid() last, then disk->backend_domid still
equals
> >> LIBXL_TOOLSTACK_DOMID when libxl__device_disk_set_backend() is
called.
> >> This results in libxl__device_disk_set_backend() making an
incorrect
> >> attempt to validate the target by calling stat() on a file on
dom0,
> >> resulting in ERROR_INVAL (see libxl_device.c lines 239-248), which
> >> prevents creation of the frontend domain.
> >>
> >> Likewise, libxl__device_nic_setdefault() previously made use of
> >> nic->backend_domid before it was set.
> >>
> >> Signed-off-by: Eric Shelton <eshelton@pobox.com>
> >
> > Acked-by: Ian Campbell <ian.campbell@citrix.com>
> >
> >> Cc: Daniel De Graaf <dgdegra@tycho.nsa.gov>
> >
> > Daniel, any comments?
> >
> 
> Yes, the backend_domname conversion needs to happen earlier or the checks
> against LIBXL_TOOLSTACK_DOMID won''t work right - I actually had
run into
> this and was bypassing it by specifying a script, so this fix is needed.
> 
> Reviewed-by: Daniel De Graaf <dgdegra@tycho.nsa.gov>
Thanks, applied.

Eric, The patch was whitespace wrapped, I fixed it by hand this time but
for next time please consult some of the links from
http://wiki.xen.org/wiki/Submitting_Xen_Patches about beating various
MUAs into submission or use git send-email or something.

Thanks,
Ian.