| Message ID | 20221102231911.3107438-4-seanjc@google.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | KVM: Rework kvm_init() and hardware enabling | expand |
On Wed, Nov 02, 2022 at 11:18:30PM +0000, Sean Christopherson wrote: > Allocate cpus_hardware_enabled after arch hardware setup so that arch > "init" and "hardware setup" are called back-to-back and thus can be > combined in a future patch. cpus_hardware_enabled is never used before > kvm_create_vm(), i.e. doesn't have a dependency with hardware setup and > only needs to be allocated before /dev/kvm is exposed to userspace. > > Free the object before the arch hooks are invoked to maintain symmetry, > and so that arch code can move away from the hooks without having to > worry about ordering changes. > > Signed-off-by: Sean Christopherson <seanjc@google.com> > --- > virt/kvm/kvm_main.c | 14 +++++++------- > 1 file changed, 7 insertions(+), 7 deletions(-) > > diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c > index e0424af52acc..8b7534cc953b 100644 > --- a/virt/kvm/kvm_main.c > +++ b/virt/kvm/kvm_main.c > @@ -5843,15 +5843,15 @@ int kvm_init(void *opaque, unsigned vcpu_size, unsigned vcpu_align, > if (r) > return r; > > + r = kvm_arch_hardware_setup(opaque); > + if (r < 0) > + goto err_hw_setup; > + > if (!zalloc_cpumask_var(&cpus_hardware_enabled, GFP_KERNEL)) { > r = -ENOMEM; > goto err_hw_enabled; > } > > - r = kvm_arch_hardware_setup(opaque); > - if (r < 0) > - goto out_free_1; > - > c.ret = &r; > c.opaque = opaque; > for_each_online_cpu(cpu) { > @@ -5937,10 +5937,10 @@ int kvm_init(void *opaque, unsigned vcpu_size, unsigned vcpu_align, > unregister_reboot_notifier(&kvm_reboot_notifier); > cpuhp_remove_state_nocalls(CPUHP_AP_KVM_STARTING); > out_free_2: > - kvm_arch_hardware_unsetup(); > -out_free_1: > free_cpumask_var(cpus_hardware_enabled); > err_hw_enabled: > + kvm_arch_hardware_unsetup(); > +err_hw_setup: > kvm_arch_exit(); > return r; > } > @@ -5967,9 +5967,9 @@ void kvm_exit(void) > cpuhp_remove_state_nocalls(CPUHP_AP_KVM_STARTING); > on_each_cpu(hardware_disable_nolock, NULL, 1); > kvm_irqfd_exit(); > + free_cpumask_var(cpus_hardware_enabled); > kvm_arch_hardware_unsetup(); > kvm_arch_exit(); > - free_cpumask_var(cpus_hardware_enabled); > kvm_vfio_ops_exit(); Looks good to me. Reviewed-by: Yuan Yao <yuan.yao@intel.com> > } > EXPORT_SYMBOL_GPL(kvm_exit); > -- > 2.38.1.431.g37b22c650d-goog >
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index e0424af52acc..8b7534cc953b 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c @@ -5843,15 +5843,15 @@ int kvm_init(void *opaque, unsigned vcpu_size, unsigned vcpu_align, if (r) return r; + r = kvm_arch_hardware_setup(opaque); + if (r < 0) + goto err_hw_setup; + if (!zalloc_cpumask_var(&cpus_hardware_enabled, GFP_KERNEL)) { r = -ENOMEM; goto err_hw_enabled; } - r = kvm_arch_hardware_setup(opaque); - if (r < 0) - goto out_free_1; - c.ret = &r; c.opaque = opaque; for_each_online_cpu(cpu) { @@ -5937,10 +5937,10 @@ int kvm_init(void *opaque, unsigned vcpu_size, unsigned vcpu_align, unregister_reboot_notifier(&kvm_reboot_notifier); cpuhp_remove_state_nocalls(CPUHP_AP_KVM_STARTING); out_free_2: - kvm_arch_hardware_unsetup(); -out_free_1: free_cpumask_var(cpus_hardware_enabled); err_hw_enabled: + kvm_arch_hardware_unsetup(); +err_hw_setup: kvm_arch_exit(); return r; } @@ -5967,9 +5967,9 @@ void kvm_exit(void) cpuhp_remove_state_nocalls(CPUHP_AP_KVM_STARTING); on_each_cpu(hardware_disable_nolock, NULL, 1); kvm_irqfd_exit(); + free_cpumask_var(cpus_hardware_enabled); kvm_arch_hardware_unsetup(); kvm_arch_exit(); - free_cpumask_var(cpus_hardware_enabled); kvm_vfio_ops_exit(); } EXPORT_SYMBOL_GPL(kvm_exit);
Allocate cpus_hardware_enabled after arch hardware setup so that arch "init" and "hardware setup" are called back-to-back and thus can be combined in a future patch. cpus_hardware_enabled is never used before kvm_create_vm(), i.e. doesn't have a dependency with hardware setup and only needs to be allocated before /dev/kvm is exposed to userspace. Free the object before the arch hooks are invoked to maintain symmetry, and so that arch code can move away from the hooks without having to worry about ordering changes. Signed-off-by: Sean Christopherson <seanjc@google.com> --- virt/kvm/kvm_main.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-)