[BUG] mlocked page counter mismatch

[MinChan Kim]

After executing following program, 'cat /proc/meminfo' shows
following result. 

--
# cat /proc/meminfo 
..
Unevictable:           8 kB
Mlocked:               8 kB
..

-- test program --

#include <stdio.h>
#include <sys/mman.h>
int main()
{
        char buf[64] = {0,};
        char *ptr;
        int k;
        int i,j;
        int x,y;
        mlockall(MCL_CURRENT);
        sprintf(buf, "cat /proc/%d/maps", getpid());
        system(buf);
        return 0;
}

--

It seems mlocked page counter have a problem.
After I diged in source, I found that try_to_unmap_file called 
try_to_mlock_page about shared mapping pages 
since other vma had VM_LOCKED flag.
After all, try_to_mlock_page called mlock_vma_page. 
so, mlocked counter increased

But, After I called munlockall intentionally, the counter work well. 
In case of munlockall, we already had a mmap_sem about write. 
Such a case, try_to_mlock_page can't call mlock_vma_page.
so, mlocked counter didn't increased. 
As a result, the counter seems to be work well but I think 
it also have a problem.



--
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/