Re: 2.6.10-rc1-mm3

[Ingo Molnar]

* Andi Kleen <ak@xxxxxxx> wrote:

> > The solution is to clip both 'first' and 'last' to PML4_SIZE boundary.
> > Since when we calculate 'first' we add at least +PML4_SIZE to the value,
> > it is safe to clip 'first'. It is obviously safe to clip 'last'.
> 
> It's a bit tricky because on 3level architectures it clips on PGDs,
> not PML4s, otherwise it would never free any pagetables.  But the
> if()s check that correctly.

indeed ... But in the do_munmap() case we dont even clip to PGD
boundary. So shouldnt we at least do the patch below?

	Ingo

--- linux/mm/mmap.c.orig
+++ linux/mm/mmap.c
@@ -1545,7 +1545,7 @@ no_mmaps:
 	}
 	if (pml4_index(last) > start_pml4_index ||
 	    pgd_index(last) > start_pgd_index) {
-		clear_page_range(tlb, first & PML4_MASK, last & PML4_MASK);
+		clear_page_range(tlb, first & PGDIR_MASK, last & PGDIR_MASK);
 		flush_tlb_pgtables(mm, first & PML4_MASK, last & PML4_MASK);
 	}
 }
-
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/