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;
MIME-Version: 1.0
References: <20190227035448.117169-1-fengc@google.com> <CAO_48GG0sW3AJNVJnydjuWUQHcmWe5aQrB=X0Wo7uBtDnPtoYg@mail.gmail.com>
In-Reply-To: <CAO_48GG0sW3AJNVJnydjuWUQHcmWe5aQrB=X0Wo7uBtDnPtoYg@mail.gmail.com>
From:   Chenbo Feng <fengc@google.com>
Date:   Thu, 14 Mar 2019 18:50:40 -0700
Message-ID: <CAMOXUJmmmvaD7A5taHgsUnt=7RxvbpdnYwqTDun6K42kVHCxhQ@mail.gmail.com>
Subject: Re: [RFC dma-buf 0/3] Improve the dma-buf tracking
To:     Sumit Semwal <sumit.semwal@linaro.org>
Cc:     LKML <linux-kernel@vger.kernel.org>,
        DRI mailing list <dri-devel@lists.freedesktop.org>,
        linux-media@vger.kernel.org, Erick Reyes <erickreyes@google.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Thu, Mar 14, 2019 at 10:49 AM Sumit Semwal <sumit.semwal@linaro.org> wro=
te:
>
> Hello Chenbo,Thank you for your RFC series.
>
> On Wed, 27 Feb 2019 at 09:24, Chenbo Feng <fengc@google.com> wrote:
> >
> > Currently, all dma-bufs share the same anonymous inode. While we can co=
unt
> > how many dma-buf fds or mappings a process has, we can't get the size o=
f
> > the backing buffers or tell if two entries point to the same dma-buf. A=
nd
> > in debugfs, we can get a per-buffer breakdown of size and reference cou=
nt,
> > but can't tell which processes are actually holding the references to e=
ach
> > buffer.
> >
> > To resolve the issue above and provide better method for userspace to t=
rack
> > the dma-buf usage across different processes, the following changes are
> > proposed in dma-buf kernel side. First of all, replace the singleton in=
ode
> > inside the dma-buf subsystem with a mini-filesystem, and assign each
> > dma-buf a unique inode out of this filesystem.  With this change, calli=
ng
> > stat(2) on each entry gives the caller a unique ID (st_ino), the buffer=
's
> > size (st_size), and even the number of pages assigned to each dma-buffe=
r.
> > Secoundly, add the inode information to /sys/kernel/debug/dma_buf/bufin=
fo
> > so in the case where a buffer is mmap()ed into a process=E2=80=99s addr=
ess space
> > but all remaining fds have been closed, we can still get the dma-buf
> > information and try to accociate it with the process by searching the
> > proc/pid/maps and looking for the corresponding inode number exposed in
> > dma-buf debug fs. Thirdly, created an ioctl to assign names to dma-bufs
> > which lets userspace assign short names (e.g., "CAMERA") to buffers. Th=
is
> > information can be extremely helpful for tracking and accounting shared
> > buffers based on their usage and original purpose. Last but not least, =
add
> > dma-buf information to /proc/pid/fdinfo by adding a show_fdinfo() handl=
er
> > to dma_buf_file_operations. The handler will print the file_count and n=
ame
> > of each buffer.
>
> In general, I think I like the idea as it contributes to a much more
> relevant usage analysis of dma-buf backed buffers.
> I will get to doing a more detailed review soon, but immediately, we
> might want to think a bit about the get/set_name IOCTLS - do we need
> to think of disallowing multiple renaming of buffers once they start
> getting used? It could otherwise make the whole metrics a lot
> confused?
Yes, I agree we need to control the dma-buf naming to make sure it
doesn't confuse the user. Ideally the process asked for dma-buf need
to make sure they only set the name once when they are using the
buffer for the same purpose. But I guess we can add check in kernel
side to make sure the name only get initialized once and can never
change after it. Or do we have better way to disallow multiple
renaming of buffers once they start getting used? Can we find a way to
set the name information in dma_buf_export so no one can change it
after that?
>
> >
> > Greg Hackmann (3):
> >   dma-buf: give each buffer a full-fledged inode
> >   dma-buf: add DMA_BUF_{GET,SET}_NAME ioctls
> >   dma-buf: add show_fdinfo handler
> >
> >  drivers/dma-buf/dma-buf.c    | 121 ++++++++++++++++++++++++++++++++---
> >  include/linux/dma-buf.h      |   5 +-
> >  include/uapi/linux/dma-buf.h |   4 ++
> >  include/uapi/linux/magic.h   |   1 +
> >  4 files changed, 122 insertions(+), 9 deletions(-)
> >
> > --
> > 2.21.0.rc2.261.ga7da99ff1b-goog
> >
>
> Best,
> Sumit.