[PATCH 5/5] qemuProcessUpdateCPU: do not change 'fallback' to ALLOW for pSeries guests

[Daniel Henrique Barboza]

Commit v3.10.0-182-g237f045d9a ("qemu: Ignore fallback CPU attribute
on reconnect") forced CPU 'fallback' to ALLOW, regardless of user
choice. This fixed a situation in which guests created with older
Libvirt versions, which used CPU mode 'host-model' in runtime, would
fail to launch in a newer Libvirt if the fallback was set to FORBID.
This would lead to a scenario where the CPU was translated to 'host-model'
to 'custom', but then the FORBID setting would make the translation
process fail.

This fix has a side effect for PSeries guests. PSeries can operate
with 'host-model' in runtime due to specific PPC64 mechanics regarding
compatibility mode. In fact, the update() implementation of the
cpuDriverPPC64 driver is a NO-OP if CPU mode is 'host-model', and
the driver does not implement translate(). The result is that PSeries
guests aren't affected by the problem, but they are being affected by
the fix - users are seeing 'fallback' mode being changed without
necessity during daemon restart.

All other cpuArchDrivers implements update() and changes guest mode
to VIR_CPU_MODE_CUSTOM, meaning that PSeries is currently the only
exception to this logic. Let's make it official.

https://bugzilla.redhat.com/show_bug.cgi?id=1660711

CC: Jiri Denemark <jdenemar at redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413 at gmail.com>
---
 src/qemu/qemu_process.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c
index a1ef1d42b0..fec1720f33 100644
--- a/src/qemu/qemu_process.c
+++ b/src/qemu/qemu_process.c
@@ -4347,8 +4347,14 @@ qemuProcessUpdateCPU(virQEMUDriverPtr driver,
 
     /* The host CPU model comes from host caps rather than QEMU caps so
      * fallback must be allowed no matter what the user specified in the XML.
+     *
+     * Note: PSeries domains are able to run with host-model CPU by design,
+     * even on Libvirt newer than 2.3, never replacing host-model with
+     * custom in the virCPUUpdate() call prior to this function. It is not
+     * needed to change the user defined 'fallback' attribute in this case.
      */
-    vm->def->cpu->fallback = VIR_CPU_FALLBACK_ALLOW;
+    if (!qemuDomainIsPSeries(vm->def))
+        vm->def->cpu->fallback = VIR_CPU_FALLBACK_ALLOW;
 
     if (qemuProcessFetchGuestCPU(driver, vm, asyncJob, &cpu, &disabled) < 0)
         return -1;
-- 
2.26.2