[patch 09/33] holepunch: fix mmap_sem i_mutex deadlock

From: Greg KH
Date: Thu Apr 26 2007 - 13:13:23 EST


-stable review patch. If anyone has any objections, please let us know.

------------------
From: Hugh Dickins <hugh@xxxxxxxxxxx>

sys_madvise has down_write of mmap_sem, then madvise_remove calls
vmtruncate_range which takes i_mutex and i_alloc_sem: no, we can
easily devise deadlocks from that ordering.

madvise_remove drop mmap_sem while calling vmtruncate_range: luckily,
since madvise_remove doesn't split or merge vmas, it's easy to handle
this case with a NULL prev, without restructuring sys_madvise. (Though
sad to retake mmap_sem when it's unlikely to be needed, and certainly
down_read is sufficient for MADV_REMOVE, unlike the other madvices.)

Signed-off-by: Hugh Dickins <hugh@xxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxx>

---
mm/madvise.c | 19 ++++++++++++++-----
1 file changed, 14 insertions(+), 5 deletions(-)

--- a/mm/madvise.c
+++ b/mm/madvise.c
@@ -159,9 +159,10 @@ static long madvise_remove(struct vm_are
unsigned long start, unsigned long end)
{
struct address_space *mapping;
- loff_t offset, endoff;
+ loff_t offset, endoff;
+ int error;

- *prev = vma;
+ *prev = NULL; /* tell sys_madvise we drop mmap_sem */

if (vma->vm_flags & (VM_LOCKED|VM_NONLINEAR|VM_HUGETLB))
return -EINVAL;
@@ -180,7 +181,12 @@ static long madvise_remove(struct vm_are
+ ((loff_t)vma->vm_pgoff << PAGE_SHIFT);
endoff = (loff_t)(end - vma->vm_start - 1)
+ ((loff_t)vma->vm_pgoff << PAGE_SHIFT);
- return vmtruncate_range(mapping->host, offset, endoff);
+
+ /* vmtruncate_range needs to take i_mutex and i_alloc_sem */
+ up_write(&current->mm->mmap_sem);
+ error = vmtruncate_range(mapping->host, offset, endoff);
+ down_write(&current->mm->mmap_sem);
+ return error;
}

static long
@@ -315,12 +321,15 @@ asmlinkage long sys_madvise(unsigned lon
if (error)
goto out;
start = tmp;
- if (start < prev->vm_end)
+ if (prev && start < prev->vm_end)
start = prev->vm_end;
error = unmapped_error;
if (start >= end)
goto out;
- vma = prev->vm_next;
+ if (prev)
+ vma = prev->vm_next;
+ else /* madvise_remove dropped mmap_sem */
+ vma = find_vma(current->mm, start);
}
out:
up_write(&current->mm->mmap_sem);

--
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/