xen/privcmd: fix double free via VMA splitting

privcmd_vm_ops defines .close (privcmd_close), but neither .may_split
nor .open. When userspace does a partial munmap() on a privcmd mapping,
the kernel splits the VMA via __split_vma(). Since may_split is NULL,
the split is allowed. vm_area_dup() copies vm_private_data (a pages
array allocated in alloc_empty_pages()) into the new VMA without any
fixup, because there is no .open callback.

Both VMAs now point to the same pages array. When the unmapped portion
is closed, privcmd_close() calls:
    - xen_unmap_domain_gfn_range()
    - xen_free_unpopulated_pages()
    - kvfree(pages)

The surviving VMA still holds the dangling pointer. When it is later
destroyed, the same sequence runs again, which leads to a double free.

Fix this issue by adding a .may_split callback denying the VMA split.

This is XSA-487 / CVE-2026-31787

Fixes: d71f513985c2 ("xen: privcmd: support autotranslated physmap guests.")
Reported-by: Atharva Vartak <atharva.a.vartak@gmail.com>
Suggested-by: Atharva Vartak <atharva.a.vartak@gmail.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>

1 files changed, 7 insertions, 0 deletions

diff --git a/drivers/xen/privcmd.c b/drivers/xen/privcmd.c
index 15ba592236e8..725a49a0eee7 100644
--- a/drivers/xen/privcmd.c
+++ b/drivers/xen/privcmd.c
@@ -1620,6 +1620,12 @@ static void privcmd_close(struct vm_area_struct *vma)
 	kvfree(pages);
 }
 
+static int privcmd_may_split(struct vm_area_struct *area, unsigned long addr)
+{
+	/* Forbid splitting, avoids double free via privcmd_close(). */
+	return -EINVAL;
+}
+
 static vm_fault_t privcmd_fault(struct vm_fault *vmf)
 {
 	printk(KERN_DEBUG "privcmd_fault: vma=%p %lx-%lx, pgoff=%lx, uv=%p\n",
@@ -1631,6 +1637,7 @@ static vm_fault_t privcmd_fault(struct vm_fault *vmf)
 
 static const struct vm_operations_struct privcmd_vm_ops = {
 	.close = privcmd_close,
+	.may_split = privcmd_may_split,
 	.fault = privcmd_fault
 };