[QEMU][PATCH v5 06/10] hw/xen/xen-hvm-common: skip ioreq creation on ioreq registration failure

Vikram Garhwal posted 10 patches 1 year, 7 months ago
There is a newer version of this series
[QEMU][PATCH v5 06/10] hw/xen/xen-hvm-common: skip ioreq creation on ioreq registration failure
Posted by Vikram Garhwal 1 year, 7 months ago
From: Stefano Stabellini <stefano.stabellini@amd.com>

On ARM it is possible to have a functioning xenpv machine with only the
PV backends and no IOREQ server. If the IOREQ server creation fails continue
to the PV backends initialization.

Also, moved the IOREQ registration and mapping subroutine to new function
xen_do_ioreq_register().

Signed-off-by: Stefano Stabellini <stefano.stabellini@amd.com>
Signed-off-by: Vikram Garhwal <vikram.garhwal@amd.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
---
 hw/xen/xen-hvm-common.c | 53 ++++++++++++++++++++++++++++-------------
 1 file changed, 36 insertions(+), 17 deletions(-)

diff --git a/hw/xen/xen-hvm-common.c b/hw/xen/xen-hvm-common.c
index c2e1e08124..5e3c7b073f 100644
--- a/hw/xen/xen-hvm-common.c
+++ b/hw/xen/xen-hvm-common.c
@@ -781,25 +781,12 @@ err:
     exit(1);
 }
 
-void xen_register_ioreq(XenIOState *state, unsigned int max_cpus,
-                        MemoryListener xen_memory_listener)
+static void xen_do_ioreq_register(XenIOState *state,
+                                           unsigned int max_cpus,
+                                           MemoryListener xen_memory_listener)
 {
     int i, rc;
 
-    state->xce_handle = xenevtchn_open(NULL, 0);
-    if (state->xce_handle == NULL) {
-        perror("xen: event channel open");
-        goto err;
-    }
-
-    state->xenstore = xs_daemon_open();
-    if (state->xenstore == NULL) {
-        perror("xen: xenstore open");
-        goto err;
-    }
-
-    xen_create_ioreq_server(xen_domid, &state->ioservid);
-
     state->exit.notify = xen_exit_notifier;
     qemu_add_exit_notifier(&state->exit);
 
@@ -863,12 +850,44 @@ void xen_register_ioreq(XenIOState *state, unsigned int max_cpus,
     QLIST_INIT(&state->dev_list);
     device_listener_register(&state->device_listener);
 
+    return;
+
+err:
+    error_report("xen hardware virtual machine initialisation failed");
+    exit(1);
+}
+
+void xen_register_ioreq(XenIOState *state, unsigned int max_cpus,
+                        MemoryListener xen_memory_listener)
+{
+    int rc;
+
+    state->xce_handle = xenevtchn_open(NULL, 0);
+    if (state->xce_handle == NULL) {
+        perror("xen: event channel open");
+        goto err;
+    }
+
+    state->xenstore = xs_daemon_open();
+    if (state->xenstore == NULL) {
+        perror("xen: xenstore open");
+        goto err;
+    }
+
+    rc = xen_create_ioreq_server(xen_domid, &state->ioservid);
+    if (!rc) {
+        xen_do_ioreq_register(state, max_cpus, xen_memory_listener);
+    } else {
+        warn_report("xen: failed to create ioreq server");
+    }
+
     xen_bus_init();
 
     xen_register_backend(state);
 
     return;
+
 err:
-    error_report("xen hardware virtual machine initialisation failed");
+    error_report("xen hardware virtual machine backend registration failed");
     exit(1);
 }
-- 
2.17.0
Re: [QEMU][PATCH v5 06/10] hw/xen/xen-hvm-common: skip ioreq creation on ioreq registration failure
Posted by Paul Durrant 1 year, 7 months ago
On 31/01/2023 22:51, Vikram Garhwal wrote:
> From: Stefano Stabellini <stefano.stabellini@amd.com>
> 
> On ARM it is possible to have a functioning xenpv machine with only the
> PV backends and no IOREQ server. If the IOREQ server creation fails continue
> to the PV backends initialization.
> 
> Also, moved the IOREQ registration and mapping subroutine to new function
> xen_do_ioreq_register().
> 
> Signed-off-by: Stefano Stabellini <stefano.stabellini@amd.com>
> Signed-off-by: Vikram Garhwal <vikram.garhwal@amd.com>
> Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
> ---
>   hw/xen/xen-hvm-common.c | 53 ++++++++++++++++++++++++++++-------------
>   1 file changed, 36 insertions(+), 17 deletions(-)

Reviewed-by: Paul Durrant <paul@xen.org>