diff mbox series

x86/xen: disable CPU idle and frequency drivers for PVH dom0

Message ID 20250407101842.67228-1-roger.pau@citrix.com (mailing list archive)
State Accepted
Commit 64a66e2c3b3113dc78a6124e14825d68ddc2e188
Headers show
Series x86/xen: disable CPU idle and frequency drivers for PVH dom0 | expand

Commit Message

Roger Pau Monné April 7, 2025, 10:18 a.m. UTC
When running as a PVH dom0 the ACPI tables exposed to Linux are (mostly)
the native ones, thus exposing the C and P states, that can lead to
attachment of CPU idle and frequency drivers.  However the entity in
control of the CPU C and P states is Xen, as dom0 doesn't have a full view
of the system load, neither has all CPUs assigned and identity pinned.

Like it's done for classic PV guests, prevent Linux from using idle or
frequency state drivers when running as a PVH dom0.

On an AMD EPYC 7543P system without this fix a Linux PVH dom0 will keep the
host CPUs spinning at 100% even when dom0 is completely idle, as it's
attempting to use the acpi_idle driver.

Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
---
 arch/x86/xen/enlighten_pvh.c | 19 ++++++++++++++++++-
 1 file changed, 18 insertions(+), 1 deletion(-)

Comments

Jason Andryuk April 7, 2025, 2:13 p.m. UTC | #1
On 2025-04-07 06:18, Roger Pau Monne wrote:
> When running as a PVH dom0 the ACPI tables exposed to Linux are (mostly)
> the native ones, thus exposing the C and P states, that can lead to
> attachment of CPU idle and frequency drivers.  However the entity in
> control of the CPU C and P states is Xen, as dom0 doesn't have a full view
> of the system load, neither has all CPUs assigned and identity pinned.
> 
> Like it's done for classic PV guests, prevent Linux from using idle or
> frequency state drivers when running as a PVH dom0.
> 
> On an AMD EPYC 7543P system without this fix a Linux PVH dom0 will keep the
> host CPUs spinning at 100% even when dom0 is completely idle, as it's
> attempting to use the acpi_idle driver.
> 
> Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>

Reviewed-by: Jason Andryuk <jason.andryuk@amd.com>

Thanks,
Jason
diff mbox series

Patch

diff --git a/arch/x86/xen/enlighten_pvh.c b/arch/x86/xen/enlighten_pvh.c
index 0e3d930bcb89..9d25d9373945 100644
--- a/arch/x86/xen/enlighten_pvh.c
+++ b/arch/x86/xen/enlighten_pvh.c
@@ -1,5 +1,7 @@ 
 // SPDX-License-Identifier: GPL-2.0
 #include <linux/acpi.h>
+#include <linux/cpufreq.h>
+#include <linux/cpuidle.h>
 #include <linux/export.h>
 #include <linux/mm.h>
 
@@ -123,8 +125,23 @@  static void __init pvh_arch_setup(void)
 {
 	pvh_reserve_extra_memory();
 
-	if (xen_initial_domain())
+	if (xen_initial_domain()) {
 		xen_add_preferred_consoles();
+
+		/*
+		 * Disable usage of CPU idle and frequency drivers: when
+		 * running as hardware domain the exposed native ACPI tables
+		 * causes idle and/or frequency drivers to attach and
+		 * malfunction.  It's Xen the entity that controls the idle and
+		 * frequency states.
+		 *
+		 * For unprivileged domains the exposed ACPI tables are
+		 * fabricated and don't contain such data.
+		 */
+		disable_cpuidle();
+		disable_cpufreq();
+		WARN_ON(xen_set_default_idle());
+	}
 }
 
 void __init xen_pvh_init(struct boot_params *boot_params)