FAILED: patch "[PATCH] RISC-V: KVM: Serialize IMSIC attributes with vCPU migration" failed to apply to 6.6-stable tree

The patch below does not apply to the 6.6-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

To reproduce the conflict and resubmit, you may use the following commands:

git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.6.y
git checkout FETCH_HEAD
git cherry-pick -x 8ae12ccaec6ec74945d8c1ef39f2c1b8df779abc
# <resolve conflicts, build, test, etc.>
git commit -s
git send-email --to '<[email protected]>' --in-reply-to '2026092917-extradite-anagram-9eaf@gregkh' --subject-prefix 'PATCH 6.6.y' 'HEAD^..'

Possible dependencies:

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From 8ae12ccaec6ec74945d8c1ef39f2c1b8df779abc Mon Sep 17 00:00:00 2001
From: Xie Bo <[email protected]>
Date: Mon, 10 Aug 2026 09:21:15 +0800
Subject: [PATCH] RISC-V: KVM: Serialize IMSIC attributes with vCPU migration

KVM device ioctls are not serialized against KVM_RUN. As a result, kvm_riscv_aia_imsic_rw_attr() can snapshot the physical CPU and HGEI of an IMSIC VS-file before a concurrent vCPU migration releases it.

The HGEI can then be allocated to another vCPU before imsic_vsfile_rw() uses the stale tuple. A GET or SET attribute may consequently access the new owner's interrupt file.

Serialize the entire IMSIC attribute operation with the target vCPU mutex. This prevents the VS-file from being migrated and recycled until the attribute access completes. Acquire the mutex killably so that the device ioctl remains interruptible while waiting for KVM_RUN to finish.

index c1af23e79ae0..f17dc559a709 100644
@@ -965,9 +965,14 @@ int kvm_riscv_aia_imsic_rw_attr(struct kvm *kvm, unsigned long type,
 	if (!vcpu)
 		return -ENODEV;
 
+	if (mutex_lock_killable(&vcpu->mutex))
+		return -EINTR;
+
 	imsic = vcpu->arch.aia_context.imsic_state;
-	if (!imsic)
-		return -ENODEV;
+	if (!imsic) {
+		rc = -ENODEV;
+		goto out_unlock;
+	}
 	isel = KVM_DEV_RISCV_AIA_IMSIC_GET_ISEL(type);
 
 	read_lock_irqsave(&imsic->vsfile_lock, flags);

@@ -991,6 +996,8 @@ int kvm_riscv_aia_imsic_rw_attr(struct kvm *kvm, unsigned long type,
 		rc = imsic_vsfile_rw(vsfile_hgei, vsfile_cpu, imsic->nr_eix,
 				     isel, write, val);
 
+out_unlock:
+	mutex_unlock(&vcpu->mutex);
 	return rc;
 }
 

From: Jiakai Xu <[email protected]>

[ Upstream commit aeb1d17d1af5924f7357d7204a293bd8fc06ea13 ]

Add a null pointer check for imsic_state before dereferencing it in kvm_riscv_aia_imsic_rw_attr(). While the function checks that the vcpu exists, it doesn't verify that the vcpu's imsic_state has been initialized, leading to a null pointer dereference when accessed.

The crash manifests as:

  Unable to handle kernel paging request at virtual address
  dfffffff00000006
  ...
  kvm_riscv_aia_imsic_rw_attr+0x2d8/0x854 arch/riscv/kvm/aia_imsic.c:958
  aia_set_attr+0x2ee/0x1726 arch/riscv/kvm/aia_device.c:354
  kvm_device_ioctl_attr virt/kvm/kvm_main.c:4744 [inline]
  kvm_device_ioctl+0x296/0x374 virt/kvm/kvm_main.c:4761
  vfs_ioctl fs/ioctl.c:51 [inline]
  ...

The fix adds a check to return -ENODEV if imsic_state is NULL and moves isel assignment after imsic_state NULL check.

Stable-dep-of: 8ae12ccaec6e ("RISC-V: KVM: Serialize IMSIC attributes with vCPU migration")

---
 arch/riscv/kvm/aia_imsic.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

index c1585444f856e..562890ac5d0b0 100644
@@ -895,8 +895,10 @@ int kvm_riscv_aia_imsic_rw_attr(struct kvm *kvm, unsigned long type,
 	if (!vcpu)
 		return -ENODEV;
 
-	isel = KVM_DEV_RISCV_AIA_IMSIC_GET_ISEL(type);
 	imsic = vcpu->arch.aia_context.imsic_state;
+	if (!imsic)
+		return -ENODEV;
+	isel = KVM_DEV_RISCV_AIA_IMSIC_GET_ISEL(type);
 
 	read_lock_irqsave(&imsic->vsfile_lock, flags);
 

2.53.0

From: Xie Bo <[email protected]>

[ Upstream commit 8ae12ccaec6ec74945d8c1ef39f2c1b8df779abc ]

KVM device ioctls are not serialized against KVM_RUN. As a result, kvm_riscv_aia_imsic_rw_attr() can snapshot the physical CPU and HGEI of an IMSIC VS-file before a concurrent vCPU migration releases it.

The HGEI can then be allocated to another vCPU before imsic_vsfile_rw() uses the stale tuple. A GET or SET attribute may consequently access the new owner's interrupt file.

Serialize the entire IMSIC attribute operation with the target vCPU mutex. This prevents the VS-file from being migrated and recycled until the attribute access completes. Acquire the mutex killably so that the device ioctl remains interruptible while waiting for KVM_RUN to finish.

---
 arch/riscv/kvm/aia_imsic.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

index 562890ac5d0b0..22782b009e370 100644
@@ -895,9 +895,14 @@ int kvm_riscv_aia_imsic_rw_attr(struct kvm *kvm, unsigned long type,
 	if (!vcpu)
 		return -ENODEV;
 
+	if (mutex_lock_killable(&vcpu->mutex))
+		return -EINTR;
+
 	imsic = vcpu->arch.aia_context.imsic_state;
-	if (!imsic)
-		return -ENODEV;
+	if (!imsic) {
+		rc = -ENODEV;
+		goto out_unlock;
+	}
 	isel = KVM_DEV_RISCV_AIA_IMSIC_GET_ISEL(type);
 
 	read_lock_irqsave(&imsic->vsfile_lock, flags);

@@ -921,6 +926,8 @@ int kvm_riscv_aia_imsic_rw_attr(struct kvm *kvm, unsigned long type,
 		rc = imsic_vsfile_rw(vsfile_hgei, vsfile_cpu, imsic->nr_eix,
 				     isel, write, val);
 
+out_unlock:
+	mutex_unlock(&vcpu->mutex);
 	return rc;
 }
 

2.53.0