qemu-kvm/kvm-Revert-hw-arm-virt-Use-ACPI-PCI-hotplug-by-default-f.patch
Miroslav Rezanina e880766b3c Reset to RHEL
Resolves: VOYAGER-1263
2026-07-10 10:35:27 +02:00

82 lines
3.3 KiB
Diff

From eee1f8abab9cbcb64ab690737f1a8db293d87c05 Mon Sep 17 00:00:00 2001
From: Eric Auger <eric.auger@redhat.com>
Date: Wed, 11 Feb 2026 09:57:37 -0500
Subject: [PATCH 7/7] Revert "hw/arm/virt: Use ACPI PCI hotplug by default from
10.2 onwards"
RH-Author: Eric Auger <eric.auger@redhat.com>
RH-MergeRequest: 465: Revert "hw/arm/virt: Use ACPI PCI hotplug by default from 10.2 onwards"
RH-Jira: RHEL-134989 RHEL-146584
RH-Acked-by: Sebastian Ott <sebott@redhat.com>
RH-Acked-by: Miroslav Rezanina <mrezanin@redhat.com>
RH-Acked-by: Cornelia Huck <cohuck@redhat.com>
RH-Commit: [1/1] e22612dc81813762f8d7f4bc9f75df0b9f2135e0 (eauger1/centos-qemu-kvm)
JIRA: https://issues.redhat.com/browse/RHEL-134989
JIRA: https://issues.redhat.com/browse/RHEL-146584
UPSTREAM: RHEL-only
This reverts commit 58cba97a715fa3f506234e718191fcc34286f333.
Conflicts: small contextual conflict when reverting changes in
hw/arm/virt.c due to subsequent fix by commit
fa0a758781fc ("arm: fix oob access in compat handling")
Unfortunately the change of the default for the PCI hotplug method
introduced some regressions that cannot be fixed in 10.2 cycle. An
example is hotplugging a virtio-net-pci device with page-per-vq=true.
This induces an increase in the BAR size which is larger than the
default size the FW accomodates. At the moment we do not have any
workaround for those devices with large BARs, ie. we noticed
pcie-root-port pref64-reserve does not work as on x86 and we do not
have any way to opt-in for legacy PCIe hotplug at libvirt
level. So let's revert the change until we get all those stuff
properly fixed.
Signed-off-by: Eric Auger <eric.auger@redhat.com>
---
hw/arm/virt.c | 10 ----------
1 file changed, 10 deletions(-)
diff --git a/hw/arm/virt.c b/hw/arm/virt.c
index 1cfb386f64..752dc08720 100644
--- a/hw/arm/virt.c
+++ b/hw/arm/virt.c
@@ -95,16 +95,9 @@
static GlobalProperty arm_virt_compat[] = {
{ TYPE_VIRTIO_IOMMU_PCI, "aw-bits", "48" },
- { TYPE_ACPI_GED, "acpi-pci-hotplug-with-bridge-support", "on" },
};
static const size_t arm_virt_compat_len = G_N_ELEMENTS(arm_virt_compat);
-GlobalProperty arm_acpi_pci_hp_disabled_compat[] = {
- { TYPE_ACPI_GED, "acpi-pci-hotplug-with-bridge-support", "off" },
-};
-static const size_t arm_acpi_pci_hp_disabled_compat_len =
- G_N_ELEMENTS(arm_acpi_pci_hp_disabled_compat);
-
/*
* RHEL9 kernels have pauth disabled while RHEL10 has it enabled,
* since qemu will setup the VM with pauth when KVM supports it we
@@ -112,7 +105,6 @@ static const size_t arm_acpi_pci_hp_disabled_compat_len =
*/
GlobalProperty arm_rhel9_compat[] = {
{TYPE_ARM_CPU, "pauth", "off", .optional = true},
- {TYPE_ACPI_GED, "acpi-pci-hotplug-with-bridge-support", "off" },
};
const size_t arm_rhel9_compat_len = G_N_ELEMENTS(arm_rhel9_compat);
@@ -3768,8 +3760,6 @@ static void virt_rhel_machine_10_0_0_options(MachineClass *mc)
/* QEMU 9.1 and earlier have only a stage-1 SMMU, not a nested s1+2 one */
vmc->no_nested_smmu = true;
- compat_props_add(mc->compat_props, arm_acpi_pci_hp_disabled_compat,
- arm_acpi_pci_hp_disabled_compat_len);
compat_props_add(mc->compat_props, hw_compat_rhel_10_2, hw_compat_rhel_10_2_len);
compat_props_add(mc->compat_props, hw_compat_rhel_10_1, hw_compat_rhel_10_1_len);
}
--
2.47.3