On 7/9/2026 7:49 AM, Yi Xie wrote:
cpu is taken from pgoff & 0xffff. cpu_online() does not reject cpu >=
nr_cpu_ids, and per_cpu_ptr() can then walk off __per_cpu_offset.
Signed-off-by: Yi Xie <[email protected]>
---
drivers/hv/mshv_vtl_main.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/hv/mshv_vtl_main.c b/drivers/hv/mshv_vtl_main.c
index 0d3d4161974f..fc50c44ac1bd 100644
--- a/drivers/hv/mshv_vtl_main.c
+++ b/drivers/hv/mshv_vtl_main.c
@@ -801,7 +801,7 @@ static vm_fault_t mshv_vtl_fault(struct vm_fault *vmf)
int cpu = vmf->pgoff & MSHV_PG_OFF_CPU_MASK;
int real_off = vmf->pgoff >> MSHV_REAL_OFF_SHIFT;
- if (!cpu_online(cpu))
+ if (cpu >= nr_cpu_ids || !cpu_online(cpu))
return VM_FAULT_SIGBUS;
/*
* CPU Hotplug is not supported in VTL2 in OpenHCL, where this kernel
driver exists.
The problem fixed by this patch generally does not happen in practice as
the user space is trusted user space (OpenVMM). Nevertheless, it's good
to have this check.
Nit: subject - s/"mshv:"/"mshv_vtl:"
as this was the agreed upon prefix for changes to mshv_vtl_main driver.
Reviewed-by: Naman Jain <[email protected]>
