linux-kernel - Re: [PATCH 2/2] mm: set PG_dma_pinned on get_user

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Mon, 18 Jun 2018 10:12:58 +0200
From:   Christoph Hellwig <hch@....de>
To:     John Hubbard <jhubbard@...dia.com>
Cc:     Dan Williams <dan.j.williams@...el.com>,
        Jason Gunthorpe <jgg@...pe.ca>, john.hubbard@...il.com,
        Matthew Wilcox <willy@...radead.org>,
        Michal Hocko <mhocko@...nel.org>,
        Christopher Lameter <cl@...ux.com>, Jan Kara <jack@...e.cz>,
        Linux MM <linux-mm@...ck.org>,
        LKML <linux-kernel@...r.kernel.org>,
        linux-rdma <linux-rdma@...r.kernel.org>,
        Christoph Hellwig <hch@....de>
Subject: Re: [PATCH 2/2] mm: set PG_dma_pinned on get_user_pages*()

On Sun, Jun 17, 2018 at 01:28:18PM -0700, John Hubbard wrote:
> Yes. However, my thinking was: get_user_pages() can become a way to indicate that 
> these pages are going to be treated specially. In particular, the caller
> does not really want or need to support certain file operations, while the
> page is flagged this way.
> 
> If necessary, we could add a new API call.

That API call is called get_user_pages_longterm.

> But either way, I think we could
> reasonably document that "if you pin these pages (either via get_user_pages,
> or some new, similar-looking API call), you can DMA to/from them, and safely
> mark them as dirty when you're done, and the right things will happen. 
> And in the interim, you can expect that the follow file system API calls
> will not behave predictably: fallocate, truncate, ..."

That is not how get_user_pages(_fast) is used.  We use it all over the
kernel, including for direct I/O.  You'd break a lot of existing use
cases very badly.