Re: fuse, get_user_pages, flush_anon_page, aliasing caches and allthat again

[James Bottomley]

On Thu, 2006-12-21 at 16:57 +0000, Russell King wrote:
> I'm not entirely convinced that it can be replaced.  What if the page
> is in the page cache and is shared with other processes?  That quite
> clearly falls under flush_dcache_page()'s remit.

Actually, it should work.  flush_dcache_page() is the big hammer, it
clears the caches of all the user processes and the kernel for that
page.  On most architectures this is not that expensive, but on parisc
it is.  Most of the paths out of fuse have already called
flush_dcache_page() coming in to make the page coherent for the kernel
to use. Then fuse writes data to the page. There's no point calling it
again since there shouldn't be anything in the user cache (the model has
the kernel owning the page), so flush_kernel_dcache_page() is the
shortcut to simply flush out the kernel cache above the page (knowing
the users still have the page uncached).

James


-
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/