Re: [PATCH] libnvdimm/pmem: Bypass CONFIG_HARDENED_USERCOPY overhead
From: Dan Williams
Date: Mon May 20 2019 - 11:43:26 EST
On Mon, May 20, 2019 at 12:52 AM Jan Kara <jack@xxxxxxx> wrote:
> On Sat 18-05-19 21:46:03, Dan Williams wrote:
> > On Fri, May 17, 2019 at 12:25 PM Kees Cook <keescook@xxxxxxxxxxxx> wrote:
> > > On Fri, May 17, 2019 at 10:28:48AM -0700, Dan Williams wrote:
> > > > It seems dax_iomap_actor() is not a path where we'd be worried about
> > > > needing hardened user copy checks.
> > >
> > > I would agree: I think the proposed patch makes sense. :)
> > Sounds like an acked-by to me.
> Yeah, if Kees agrees, I'm fine with skipping the checks as well. I just
> wanted that to be clarified. Also it helped me that you wrote:
> That routine (dax_iomap_actor()) validates that the logical file offset is
> within bounds of the file, then it does a sector-to-pfn translation which
> validates that the physical mapping is within bounds of the block device.
> That is more specific than "dax_iomap_actor() takes care of necessary
> checks" which was in the changelog. And the above paragraph helped me
> clarify which checks in dax_iomap_actor() you think replace those usercopy
> checks. So I think it would be good to add that paragraph to those
> copy_from_pmem() functions as a comment just in case we are wondering in
> the future why we are skipping the checks... Also feel free to add:
> Acked-by: Jan Kara <jack@xxxxxxx>
Will do, thanks Jan.