Thread (21 messages) 21 messages, 3 authors, 2026-03-17

Re: [PATCH 07/10] KVM: arm64: Use scoped resource management in arm.c

From: Jonathan Cameron <jonathan.cameron@huawei.com>
Date: 2026-03-17 17:53:29
Also in: kvmarm, lkml

On Mon, 16 Mar 2026 17:35:28 +0000
Fuad Tabba [off-list ref] wrote:
Migrate manual spin_lock() calls managing mp_state_lock and manual
mutex_lock() calls managing kvm->arch.config_lock to use the
guard(spinlock) and guard(mutex) macros.

This eliminates manual unlock calls on early return paths and
simplifies the vCPU suspend/resume control flow.

Change-Id: Ifcd8455d08afa5d00fc200daaa3fb13f6736e6ed
Signed-off-by: Fuad Tabba <redacted>
Trivial stuff.
quoted hunk ↗ jump to hunk
---
 arch/arm64/kvm/arm.c | 53 ++++++++++++++++++++--------------------------------
 1 file changed, 20 insertions(+), 33 deletions(-)
diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c
index 410ffd41fd73..017f5bfabe19 100644
--- a/arch/arm64/kvm/arm.c
+++ b/arch/arm64/kvm/arm.c
quoted hunk ↗ jump to hunk
 
 /**
@@ -738,9 +736,8 @@ static void __kvm_arm_vcpu_power_off(struct kvm_vcpu *vcpu)
 
 void kvm_arm_vcpu_power_off(struct kvm_vcpu *vcpu)
 {
-	spin_lock(&vcpu->arch.mp_state_lock);
+	guard(spinlock)(&vcpu->arch.mp_state_lock);
 	__kvm_arm_vcpu_power_off(vcpu);
-	spin_unlock(&vcpu->arch.mp_state_lock);
 }
 
 bool kvm_arm_vcpu_stopped(struct kvm_vcpu *vcpu)
@@ -773,7 +770,7 @@ int kvm_arch_vcpu_ioctl_set_mpstate(struct kvm_vcpu *vcpu,
 {
 	int ret = 0;
 
-	spin_lock(&vcpu->arch.mp_state_lock);
+	guard(spinlock)(&vcpu->arch.mp_state_lock);
 
 	switch (mp_state->mp_state) {
 	case KVM_MP_STATE_RUNNABLE:
@@ -789,8 +786,6 @@ int kvm_arch_vcpu_ioctl_set_mpstate(struct kvm_vcpu *vcpu,
 		ret = -EINVAL;
		return -EINVAL;
	}

	return 0;

Or even return in the switch legs instead of breaking out.
 	}
 
-	spin_unlock(&vcpu->arch.mp_state_lock);
-
 	return ret;
 }
quoted hunk ↗ jump to hunk
 /*
@@ -944,9 +937,8 @@ int kvm_arch_vcpu_run_pid_change(struct kvm_vcpu *vcpu)
 			return ret;
 	}
 
-	mutex_lock(&kvm->arch.config_lock);
+	guard(mutex)(&kvm->arch.config_lock);
 	set_bit(KVM_ARCH_FLAG_HAS_RAN_ONCE, &kvm->arch.flags);
-	mutex_unlock(&kvm->arch.config_lock);
 
 	return ret;
Personally I'd make this return 0 whilst you are here so it's obvious this
path only ever returns success.
 }

Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help