In the Linux kernel, the following vulnerability has been resolved:

mm/hugetlb: do not call vma_add_reservation upon ENOMEM

sysbot reported a splat [1] on __unmap_hugepage_range().  This is because
vma_needs_reservation() can return -ENOMEM if
allocate_file_region_entries() fails to allocate the file_region struct
for the reservation.

Check for that and do not call vma_add_reservation() if that is the case,
otherwise region_abort() and region_del() will see that we do not have any
file_regions.

If we detect that vma_needs_reservation() returned -ENOMEM, we clear the
hugetlb_restore_reserve flag as if this reservation was still consumed, so
free_huge_folio() will not increment the resv count.

[1] https://lore.kernel.org/linux-mm/0000000000004096100617c58d54@google.com/T/#ma5983bc1ab18a54910da83416b3f89f3c7ee43aa

kernel.org https://git.kernel.org/stable/c/8daf9c702ee7f825f0de8600abff764acfedea13 [No types assigned]

kernel.org https://git.kernel.org/stable/c/aa998f9dcb34c28448f86e8f5490f20d5eb0eac7 [No types assigned]