qemu-kvm/kvm-i386-do-kvm_put_msr_feature_control-first-thing-when.patch
Miroslav Rezanina 19bc18cc9e * Thu Sep 29 2022 Miroslav Rezanina <mrezanin@redhat.com> - 7.1.0-2
- kvm-vdpa-Skip-the-maps-not-in-the-iova-tree.patch [RHELX-57]
- kvm-vdpa-do-not-save-failed-dma-maps-in-SVQ-iova-tree.patch [RHELX-57]
- kvm-util-accept-iova_tree_remove_parameter-by-value.patch [RHELX-57]
- kvm-vdpa-Remove-SVQ-vring-from-iova_tree-at-shutdown.patch [RHELX-57]
- kvm-vdpa-Make-SVQ-vring-unmapping-return-void.patch [RHELX-57]
- kvm-vhost-Always-store-new-kick-fd-on-vhost_svq_set_svq_.patch [RHELX-57]
- kvm-vdpa-Use-ring-hwaddr-at-vhost_vdpa_svq_unmap_ring.patch [RHELX-57]
- kvm-vhost-stop-transfer-elem-ownership-in-vhost_handle_g.patch [RHELX-57]
- kvm-vhost-use-SVQ-element-ndescs-instead-of-opaque-data-.patch [RHELX-57]
- kvm-vhost-Delete-useless-read-memory-barrier.patch [RHELX-57]
- kvm-vhost-Do-not-depend-on-NULL-VirtQueueElement-on-vhos.patch [RHELX-57]
- kvm-vhost_net-Add-NetClientInfo-start-callback.patch [RHELX-57]
- kvm-vhost_net-Add-NetClientInfo-stop-callback.patch [RHELX-57]
- kvm-vdpa-add-net_vhost_vdpa_cvq_info-NetClientInfo.patch [RHELX-57]
- kvm-vdpa-Move-command-buffers-map-to-start-of-net-device.patch [RHELX-57]
- kvm-vdpa-extract-vhost_vdpa_net_cvq_add-from-vhost_vdpa_.patch [RHELX-57]
- kvm-vhost_net-add-NetClientState-load-callback.patch [RHELX-57]
- kvm-vdpa-Add-virtio-net-mac-address-via-CVQ-at-start.patch [RHELX-57]
- kvm-vdpa-Delete-CVQ-migration-blocker.patch [RHELX-57]
- kvm-vdpa-Make-VhostVDPAState-cvq_cmd_in_buffer-control-a.patch [RHELX-57]
- kvm-vdpa-extract-vhost_vdpa_net_load_mac-from-vhost_vdpa.patch [RHELX-57]
- kvm-vdpa-Add-vhost_vdpa_net_load_mq.patch [RHELX-57]
- kvm-vdpa-validate-MQ-CVQ-commands.patch [RHELX-57]
- kvm-virtio-net-Update-virtio-net-curr_queue_pairs-in-vdp.patch [RHELX-57]
- kvm-vdpa-Allow-MQ-feature-in-SVQ.patch [RHELX-57]
- kvm-i386-reset-KVM-nested-state-upon-CPU-reset.patch [bz#2125281]
- kvm-i386-do-kvm_put_msr_feature_control-first-thing-when.patch [bz#2125281]
- kvm-Revert-Re-enable-capstone-internal-build.patch [bz#2127825]
- kvm-spec-Use-capstone-package.patch [bz#2127825]
- Resolves: RHELX-57
  (vDPA SVQ Multiqueue support )
- Resolves: bz#2125281
  ([RHEL9.1] Guests in VMX root operation fail to reboot with QEMU's 'system_reset' command [rhel-9.2.0])
- Resolves: bz#2127825
  (Use capstone for qemu-kvm build)
2022-09-29 03:59:55 -04:00

67 lines
2.4 KiB
Diff

From aba2a5cb19efa33be871dd951366439cf99c5f13 Mon Sep 17 00:00:00 2001
From: Vitaly Kuznetsov <vkuznets@redhat.com>
Date: Thu, 18 Aug 2022 17:01:13 +0200
Subject: [PATCH 27/29] i386: do kvm_put_msr_feature_control() first thing when
vCPU is reset
RH-Author: Miroslav Rezanina <mrezanin@redhat.com>
RH-MergeRequest: 118: Synchronize qemu-kvm-7.0.0-13.el9
RH-Bugzilla: 2125281
RH-Acked-by: Vitaly Kuznetsov <vkuznets@redhat.com>
RH-Commit: [2/2] 4986c35a04255c8fe2b62a48f5ea489339f3826a (mrezanin/centos-src-qemu-kvm)
kvm_put_sregs2() fails to reset 'locked' CR4/CR0 bits upon vCPU reset when
it is in VMX root operation. Do kvm_put_msr_feature_control() before
kvm_put_sregs2() to (possibly) kick vCPU out of VMX root operation. It also
seems logical to do kvm_put_msr_feature_control() before
kvm_put_nested_state() and not after it, especially when 'real' nested
state is set.
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Message-Id: <20220818150113.479917-3-vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
(cherry picked from commit 45ed68a1a3a19754ade954d75a3c9d13ff560e5c)
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
---
target/i386/kvm/kvm.c | 17 ++++++++++++-----
1 file changed, 12 insertions(+), 5 deletions(-)
diff --git a/target/i386/kvm/kvm.c b/target/i386/kvm/kvm.c
index fd3237310b..a9eba247a5 100644
--- a/target/i386/kvm/kvm.c
+++ b/target/i386/kvm/kvm.c
@@ -4533,6 +4533,18 @@ int kvm_arch_put_registers(CPUState *cpu, int level)
assert(cpu_is_stopped(cpu) || qemu_cpu_is_self(cpu));
+ /*
+ * Put MSR_IA32_FEATURE_CONTROL first, this ensures the VM gets out of VMX
+ * root operation upon vCPU reset. kvm_put_msr_feature_control() should also
+ * preceed kvm_put_nested_state() when 'real' nested state is set.
+ */
+ if (level >= KVM_PUT_RESET_STATE) {
+ ret = kvm_put_msr_feature_control(x86_cpu);
+ if (ret < 0) {
+ return ret;
+ }
+ }
+
/* must be before kvm_put_nested_state so that EFER.SVME is set */
ret = has_sregs2 ? kvm_put_sregs2(x86_cpu) : kvm_put_sregs(x86_cpu);
if (ret < 0) {
@@ -4544,11 +4556,6 @@ int kvm_arch_put_registers(CPUState *cpu, int level)
if (ret < 0) {
return ret;
}
-
- ret = kvm_put_msr_feature_control(x86_cpu);
- if (ret < 0) {
- return ret;
- }
}
if (level == KVM_PUT_FULL_STATE) {
--
2.31.1