Re: fuse, get_user_pages, flush_anon_page, aliasing caches and all that again

[Russell King <rmk+lkml@arm.linux.org.uk>]

On Sun, Dec 31, 2006 at 01:47:56AM -0800, David Miller wrote:
> From: Arjan van de Ven <arjan@infradead.org>
> Date: Sun, 31 Dec 2006 10:27:22 +0100
> > > However, it's not only FUSE which is suffering - direct-IO also doesn't
> > > work. 
> > 
> > for direct-IO the kernel won't touch the data *at all*... (that's the
> > point ;) 
> > 
> > is it still an issue then?
> 
> It can be an issue with virtual caches if the "I/O" is done
> using cpu loads and stores, but we should be handling that
> with explicit flushing anyways.
> 
> The core of the problem is that ARM doesn't look for the user
> mappings for anonymous pages when flush_dcache_page() is invoked.
> I think as a temporary fix it could walk the RMAP list and
> use that to find the user virtual mappings.  Would that work
> Russel?

I'm willing to do that - and I guess this means we can probably do this
instead of walking the list of VMAs for the shared mapping, thereby
hitting both anonymous and shared mappings with the same code?

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of: