Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Fri, 8 Mar 2019 03:08:40 +0000
From:   Christopher Lameter <cl@linux.com>
To:     john.hubbard@gmail.com
cc:     Andrew Morton <akpm@linux-foundation.org>, linux-mm@kvack.org,
        Al Viro <viro@zeniv.linux.org.uk>,
        Christian Benvenuti <benve@cisco.com>,
        Christoph Hellwig <hch@infradead.org>,
        Dan Williams <dan.j.williams@intel.com>,
        Dave Chinner <david@fromorbit.com>,
        Dennis Dalessandro <dennis.dalessandro@intel.com>,
        Doug Ledford <dledford@redhat.com>,
        Ira Weiny <ira.weiny@intel.com>, Jan Kara <jack@suse.cz>,
        Jason Gunthorpe <jgg@ziepe.ca>,
        Jerome Glisse <jglisse@redhat.com>,
        Matthew Wilcox <willy@infradead.org>,
        Michal Hocko <mhocko@kernel.org>,
        Mike Rapoport <rppt@linux.ibm.com>,
        Mike Marciniszyn <mike.marciniszyn@intel.com>,
        Ralph Campbell <rcampbell@nvidia.com>,
        Tom Talpey <tom@talpey.com>,
        LKML <linux-kernel@vger.kernel.org>,
        linux-fsdevel@vger.kernel.org, John Hubbard <jhubbard@nvidia.com>
Subject: Re: [PATCH v3 0/1] mm: introduce put_user_page*(), placeholder
 versions
In-Reply-To: <20190306235455.26348-1-jhubbard@nvidia.com>
Message-ID: <010001695b4631cd-f4b8fcbf-a760-4267-afce-fb7969e3ff87-000000@email.amazonses.com>
References: <20190306235455.26348-1-jhubbard@nvidia.com>
User-Agent: Alpine 2.21 (DEB 202 2017-01-01)
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Feedback-ID: 1.us-east-1.fQZZZ0Xtj2+TD7V5apTT/NrT6QKuPgzCT/IC7XYgDKI=:AmazonSES
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Wed, 6 Mar 2019, john.hubbard@gmail.com wrote:


> GUP was first introduced for Direct IO (O_DIRECT), allowing filesystem code
> to get the struct page behind a virtual address and to let storage hardware
> perform a direct copy to or from that page. This is a short-lived access
> pattern, and as such, the window for a concurrent writeback of GUP'd page
> was small enough that there were not (we think) any reported problems.
> Also, userspace was expected to understand and accept that Direct IO was
> not synchronized with memory-mapped access to that data, nor with any
> process address space changes such as munmap(), mremap(), etc.

It would good if that understanding would be enforced somehow given the problems
that we see.

> Interactions with file systems
> ==============================
>
> File systems expect to be able to write back data, both to reclaim pages,

Regular filesystems do that. But usually those are not used with GUP
pinning AFAICT.

> and for data integrity. Allowing other hardware (NICs, GPUs, etc) to gain
> write access to the file memory pages means that such hardware can dirty
> the pages, without the filesystem being aware. This can, in some cases
> (depending on filesystem, filesystem options, block device, block device
> options, and other variables), lead to data corruption, and also to kernel
> bugs of the form:

> Long term GUP
> =============
>
> Long term GUP is an issue when FOLL_WRITE is specified to GUP (so, a
> writeable mapping is created), and the pages are file-backed. That can lead
> to filesystem corruption. What happens is that when a file-backed page is
> being written back, it is first mapped read-only in all of the CPU page
> tables; the file system then assumes that nobody can write to the page, and
> that the page content is therefore stable. Unfortunately, the GUP callers
> generally do not monitor changes to the CPU pages tables; they instead
> assume that the following pattern is safe (it's not):
>
>     get_user_pages()
>
>     Hardware can keep a reference to those pages for a very long time,
>     and write to it at any time. Because "hardware" here means "devices
>     that are not a CPU", this activity occurs without any interaction
>     with the kernel's file system code.
>
>     for each page
>         set_page_dirty
>         put_page()
>
> In fact, the GUP documentation even recommends that pattern.

Isnt that pattern safe for anonymous memory and memory filesystems like
hugetlbfs etc? Which is the common use case.

> Anyway, the file system assumes that the page is stable (nothing is writing
> to the page), and that is a problem: stable page content is necessary for
> many filesystem actions during writeback, such as checksum, encryption,
> RAID striping, etc. Furthermore, filesystem features like COW (copy on
> write) or snapshot also rely on being able to use a new page for as memory
> for that memory range inside the file.
>
> Corruption during write back is clearly possible here. To solve that, one
> idea is to identify pages that have active GUP, so that we can use a bounce
> page to write stable data to the filesystem. The filesystem would work
> on the bounce page, while any of the active GUP might write to the
> original page. This would avoid the stable page violation problem, but note
> that it is only part of the overall solution, because other problems
> remain.

Yes you now have the filesystem as well as the GUP pinner claiming
authority over the contents of a single memory segment. Maybe better not
allow that?

> Direct IO
> =========
>
> Direct IO can cause corruption, if userspace does Direct-IO that writes to
> a range of virtual addresses that are mmap'd to a file.  The pages written
> to are file-backed pages that can be under write back, while the Direct IO
> is taking place.  Here, Direct IO races with a write back: it calls
> GUP before page_mkclean() has replaced the CPU pte with a read-only entry.
> The race window is pretty small, which is probably why years have gone by
> before we noticed this problem: Direct IO is generally very quick, and
> tends to finish up before the filesystem gets around to do anything with
> the page contents.  However, it's still a real problem.  The solution is
> to never let GUP return pages that are under write back, but instead,
> force GUP to take a write fault on those pages.  That way, GUP will
> properly synchronize with the active write back.  This does not change the
> required GUP behavior, it just avoids that race.

Direct IO on a mmapped file backed page doesnt make any sense. The direct
I/O write syscall already specifies one file handle of a filesystem that
the data is to be written onto.  Plus mmap already established another
second filehandle and another filesystem that is also in charge of that
memory segment.

Two filesystem trying to sync one memory segment both believing to have
exclusive access and we want to sort this out. Why? Dont allow this.