arch/riscv/kvm/aia_imsic.c | 3 +++ 1 file changed, 3 insertions(+)
Add a null pointer check for imsic_state before dereferencing it in
kvm_riscv_aia_imsic_has_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.
This issue was discovered during fuzzing of RISC-V KVM code. The
crash occurs when userspace calls KVM_HAS_DEVICE_ATTR ioctl on an
AIA IMSIC device before the IMSIC state has been fully initialized
for a vcpu.
The crash manifests as:
Unable to handle kernel paging request at virtual address
dfffffff00000001
...
epc : kvm_riscv_aia_imsic_has_attr+0x464/0x50e
arch/riscv/kvm/aia_imsic.c:998
...
kvm_riscv_aia_imsic_has_attr+0x464/0x50e arch/riscv/kvm/aia_imsic.c:998
aia_has_attr+0x128/0x2bc arch/riscv/kvm/aia_device.c:471
kvm_device_ioctl_attr virt/kvm/kvm_main.c:4722 [inline]
kvm_device_ioctl+0x296/0x374 virt/kvm/kvm_main.c:4739
...
The fix adds a check to return -ENODEV if imsic_state is NULL, which
is consistent with other error handling in the function and prevents
the null pointer dereference.
Reproducer and detailed analysis available at:
https://github.com/j1akai/temp/tree/main/20260125
Signed-off-by: Jiakai Xu <xujiakai2025@iscas.ac.cn>
---
arch/riscv/kvm/aia_imsic.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/arch/riscv/kvm/aia_imsic.c b/arch/riscv/kvm/aia_imsic.c
index e597e86491c3b..9c58a66068447 100644
--- a/arch/riscv/kvm/aia_imsic.c
+++ b/arch/riscv/kvm/aia_imsic.c
@@ -995,6 +995,9 @@ int kvm_riscv_aia_imsic_has_attr(struct kvm *kvm, unsigned long type)
isel = KVM_DEV_RISCV_AIA_IMSIC_GET_ISEL(type);
imsic = vcpu->arch.aia_context.imsic_state;
+ if (!imsic)
+ return -ENODEV;
+
return imsic_mrif_isel_check(imsic->nr_eix, isel);
}
--
2.34.1
On Sun, Jan 25, 2026 at 11:34 AM Jiakai Xu <jiakaipeanut@gmail.com> wrote: > > Add a null pointer check for imsic_state before dereferencing it in > kvm_riscv_aia_imsic_has_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. > > This issue was discovered during fuzzing of RISC-V KVM code. The > crash occurs when userspace calls KVM_HAS_DEVICE_ATTR ioctl on an > AIA IMSIC device before the IMSIC state has been fully initialized > for a vcpu. > > The crash manifests as: > Unable to handle kernel paging request at virtual address > dfffffff00000001 > ... > epc : kvm_riscv_aia_imsic_has_attr+0x464/0x50e > arch/riscv/kvm/aia_imsic.c:998 > ... > kvm_riscv_aia_imsic_has_attr+0x464/0x50e arch/riscv/kvm/aia_imsic.c:998 > aia_has_attr+0x128/0x2bc arch/riscv/kvm/aia_device.c:471 > kvm_device_ioctl_attr virt/kvm/kvm_main.c:4722 [inline] > kvm_device_ioctl+0x296/0x374 virt/kvm/kvm_main.c:4739 > ... > > The fix adds a check to return -ENODEV if imsic_state is NULL, which > is consistent with other error handling in the function and prevents > the null pointer dereference. > > Reproducer and detailed analysis available at: > https://github.com/j1akai/temp/tree/main/20260125 No need for a link to your personal repo over here. Please add Fixes tag here. > > Signed-off-by: Jiakai Xu <xujiakai2025@iscas.ac.cn> Otherwise, LGTM. Reviewed-by: Anup Patel <anup@brainfault.org> Regards, Anup > --- > arch/riscv/kvm/aia_imsic.c | 3 +++ > 1 file changed, 3 insertions(+) > > diff --git a/arch/riscv/kvm/aia_imsic.c b/arch/riscv/kvm/aia_imsic.c > index e597e86491c3b..9c58a66068447 100644 > --- a/arch/riscv/kvm/aia_imsic.c > +++ b/arch/riscv/kvm/aia_imsic.c > @@ -995,6 +995,9 @@ int kvm_riscv_aia_imsic_has_attr(struct kvm *kvm, unsigned long type) > > isel = KVM_DEV_RISCV_AIA_IMSIC_GET_ISEL(type); > imsic = vcpu->arch.aia_context.imsic_state; > + if (!imsic) > + return -ENODEV; > + > return imsic_mrif_isel_check(imsic->nr_eix, isel); > } > > -- > 2.34.1 >
© 2016 - 2026 Red Hat, Inc.