Hi Rik,
This code generates a Smatch warning. I normally ignore Smatch warnings
in mm because I'm not clever enough to mess with that, but this one is
pretty recent so I thought I'd ask. Could you take a look?
mm/mmap.c +1980 __split_vma(57) error: we previously assumed 'new->vm_ops' could be null.
1966 if (new->vm_ops && new->vm_ops->open)
^^^^^^^^^^^
We assume new->vm_ops can be NULL here.
1967 new->vm_ops->open(new);
1968
1969 if (new_below)
1970 err = vma_adjust(vma, addr, vma->vm_end, vma->vm_pgoff +
1971 ((addr - new->vm_start) >> PAGE_SHIFT), new);
1972 else
1973 err = vma_adjust(vma, vma->vm_start, addr, vma->vm_pgoff, new);
1974
1975 /* Success. */
1976 if (!err)
1977 return 0;
1978
1979 /* Clean everything up if vma_adjust failed. */
1980 new->vm_ops->close(new);
^^^^^^^^^^^^^^^^^^
But we dereference it unconditionally here.
The dereference was added in 5beb4930: "mm: change anon_vma linking to fix
multi-process server scalability issue".
regards,
dan carpenter
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxxx For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>