[Xen-devel] [PATCH v2 2/3] tools/libxc: Use an explicit check for PV MSRs in xc_domain_save()

[Andrew Cooper <andrew.cooper3@xxxxxxxxxx>]

Migrating PV domains using MSRs is not supported.  This uses the new
XEN_DOMCTL_get_vcpu_msrs and will fail the migration with an explicit error.

This is an improvement upon the current failure of
  "No extended context for VCPUxx (ENOBUFS)"

Support for migrating PV domains which are using MSRs will be included in the
migration v2 work.

Signed-off-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
CC: Jan Beulich <JBeulich@xxxxxxxx>
CC: Ian Campbell <Ian.Campbell@xxxxxxxxxx>
CC: Ian Jackson <Ian.Jackson@xxxxxxxxxxxxx>

--
v2: Reduce code based on altered Xen implementation
---
 tools/libxc/xc_domain_save.c |   20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/tools/libxc/xc_domain_save.c b/tools/libxc/xc_domain_save.c
index acf3685..778cbde 100644
--- a/tools/libxc/xc_domain_save.c
+++ b/tools/libxc/xc_domain_save.c
@@ -1995,6 +1995,26 @@ int xc_domain_save(xc_interface *xch, int io_fd, 
uint32_t dom, uint32_t max_iter
             goto out;
         }
 
+        /* Check there are no PV MSRs in use. */
+        domctl.cmd = XEN_DOMCTL_get_vcpu_msrs;
+        domctl.domain = dom;
+        memset(&domctl.u, 0, sizeof(domctl.u));
+        domctl.u.vcpu_msrs.vcpu = i;
+        domctl.u.vcpu_msrs.msr_count = 0;
+        set_xen_guest_handle_raw(domctl.u.vcpu_msrs.msrs, (void*)1);
+
+        if ( xc_domctl(xch, &domctl) < 0 )
+        {
+            if ( errno == ENOBUFS )
+            {
+                errno = EOPNOTSUPP;
+                PERROR("Unable to migrate PV guest using MSRs (yet)");
+            }
+            else
+                PERROR("Error querying maximum number of MSRs for VCPU%d", i);
+            goto out;
+        }
+
         /* Start to fetch CPU eXtended States */
         /* Get buffer size first */
         domctl.cmd = XEN_DOMCTL_getvcpuextstate;
-- 
1.7.10.4


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