[5/5] memrlimit correct mremap and move_vma accounting

[Balbir Singh]

The memrlimit patches did not account for move_vma() since we account for
address space usage in do_mremap(). The code flow actually increments
total_vm twice (once in do_mremap() and once in move_vma()), the excess
is removed in remove_vma_list() via do_munmap(). Since we did not do the
duplicate accounting, the code was seeing the extra uncharge, causing
our accounting to break. This patch fixes the problem

Signed-off-by: Balbir Singh <balbir@xxxxxxxxxxxxxxxxxx>
---

 mm/mremap.c |    9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff -puN mm/mremap.c~memrlimit-fix-move-vma-accounting mm/mremap.c
--- linux-2.6.26-rc5/mm/mremap.c~memrlimit-fix-move-vma-accounting	2008-06-26 14:48:25.000000000 +0530
+++ linux-2.6.26-rc5-balbir/mm/mremap.c	2008-06-26 14:48:25.000000000 +0530
@@ -177,10 +177,15 @@ static unsigned long move_vma(struct vm_
 	if (mm->map_count >= sysctl_max_map_count - 3)
 		return -ENOMEM;
 
+	if (memrlimit_cgroup_charge_as(mm, new_len >> PAGE_SHIFT))
+		return -ENOMEM;
+
 	new_pgoff = vma->vm_pgoff + ((old_addr - vma->vm_start) >> PAGE_SHIFT);
 	new_vma = copy_vma(&vma, new_addr, new_len, new_pgoff);
-	if (!new_vma)
+	if (!new_vma) {
+		memrlimit_cgroup_uncharge_as(mm, new_len >> PAGE_SHIFT);
 		return -ENOMEM;
+	}
 
 	moved_len = move_page_tables(vma, old_addr, new_vma, new_addr, old_len);
 	if (moved_len < old_len) {
@@ -386,6 +391,8 @@ unsigned long do_mremap(unsigned long ad
 		}
 	}
 
+	memrlimit_cgroup_uncharge_as(mm, (new_len - old_len) >> PAGE_SHIFT);
+
 	/*
 	 * We weren't able to just expand or shrink the area,
 	 * we need to create a new one and move it..
_

-- 
	Warm Regards,
	Balbir Singh
	Linux Technology Center
	IBM, ISTL
--
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/