On 2025-07-17 02:24, Juergen Gross wrote:
On 16.07.25 23:14, Jason Andryuk wrote:
Expose a domain's capabilities - control, hardware or xenstore - through
stable get domain state hypercall.

The xenstore domain can use this information to assign appropriate
permissions on connections.

Repurpose the 16bit pad field for this purpose.

Signed-off-by: Jason Andryuk <jason.andr...@amd.com>
---
  xen/common/domain.c         | 8 +++++++-
  xen/include/public/domctl.h | 7 +++++--
  2 files changed, 12 insertions(+), 3 deletions(-)

diff --git a/xen/common/domain.c b/xen/common/domain.c
index c347de4335..8b0151a8f7 100644
--- a/xen/common/domain.c
+++ b/xen/common/domain.c
@@ -195,6 +195,12 @@ static void set_domain_state_info(struct xen_domctl_get_domain_state *info,
          info->state |= XEN_DOMCTL_GETDOMSTATE_STATE_DYING;
      if ( d->is_dying == DOMDYING_dead )
          info->state |= XEN_DOMCTL_GETDOMSTATE_STATE_DEAD;

I'll add:

+
+    info->caps = 0

here, to ensure all the caps bits start off cleared.

+    if ( is_control_domain(d) )
+        info->caps |= XEN_DOMCTL_GETDOMSTATE_CAP_CONTROL;
+    if ( is_hardware_domain(d) )
+        info->caps |= XEN_DOMCTL_GETDOMSTATE_CAP_HARDWARE;
+    if ( is_xenstore_domain(d) )
+        info->caps |= XEN_DOMCTL_GETDOMSTATE_CAP_XENSTORE;
      info->unique_id = d->unique_id;
  }

diff --git a/xen/include/public/domctl.h b/xen/include/public/domctl.h
index f1f6f96bc2..879343b21d 100644
--- a/xen/include/public/domctl.h
+++ b/xen/include/public/domctl.h
@@ -1269,8 +1269,11 @@ struct xen_domctl_get_domain_state {
  #define XEN_DOMCTL_GETDOMSTATE_STATE_SHUTDOWN  0x0002  /* Shutdown finished. */   #define XEN_DOMCTL_GETDOMSTATE_STATE_DYING     0x0004  /* Domain dying. */   #define XEN_DOMCTL_GETDOMSTATE_STATE_DEAD      0x0008  /* Domain dead. */
-    uint16_t pad0;           /* Must be 0 on input, returned as 0. */
-    uint32_t pad1;           /* Must be 0 on input, returned as 0. */
+    uint16_t caps;           /* Must be 0 on input, returned as 0. */

I'd drop the "returned as 0" in the comment. :-)

:)

I'll drop the whole comment since the input value is not checked.

With that fixed:

Reviewed-by: Juergen Gross <jgr...@suse.com>

Does this stand with the above changes?

Thanks,
Jason

Reply via email to