On Sun, Oct 27, 2013 at 09:42:57AM -0400, Tejun Heo wrote:

On Sun, Oct 27, 2013 at 09:16:53PM +0800, Ming Lei wrote:

quoted

On Sun, Oct 27, 2013 at 8:50 PM, Aaro Koskinen [off-list ref] wrote:

quoted

On ARM v3.9 or older kernels do not trigger this BUG, at seems it only
started to appear with the following commit (bisected):

commit 1bc39742aab09248169ef9d3727c9def3528b3f3
Author: Simon Baatz [off-list ref]
Date:   Mon Jun 10 21:10:12 2013 +0100

    ARM: 7755/1: handle user space mapped pages in flush_kernel_dcache_page

The above commit only starts to implement the helper on ARM,
but according to Documentation/cachetlb.txt, looks caller of
flush_kernel_dcache_page() should make sure the passed
'page' is a user space page.

I don't think PageSlab() is the right test tho.  Wouldn't testing
against user_addr_max() make more sense?

How does that help for a function passed a struct page pointer?