Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1572886 > unrolled thread

Re: [PATCH 00/10] android: binder: support for domains and scatter-gather.

Started byJohn Stultz <john.stultz@linaro.org>
First post2017-02-03 06:00 +0100
Last post2017-02-03 08:20 +0100
Articles 2 — 2 participants

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.


Contents

  Re: [PATCH 00/10] android: binder: support for domains and scatter-gather. John Stultz <john.stultz@linaro.org> - 2017-02-03 06:00 +0100
    Re: [PATCH 00/10] android: binder: support for domains and  scatter-gather. Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-02-03 08:20 +0100

#1572886 — Re: [PATCH 00/10] android: binder: support for domains and scatter-gather.

FromJohn Stultz <john.stultz@linaro.org>
Date2017-02-03 06:00 +0100
SubjectRe: [PATCH 00/10] android: binder: support for domains and scatter-gather.
Message-ID<t6E8y-5gf-5@gated-at.bofh.it>
On Mon, Oct 24, 2016 at 6:20 AM, Martijn Coenen <maco@android.com> wrote:
> Android userspace will start using binder IPC for communication with HAL
> modules. To clearly separate this IPC domain from the existing framework
> IPC domain, this patch series adds support for multiple "binder domains".
> This is implemented by each domain having its own binder context manager,
> and then having a separate device node per domain.
>
> The other change introduced by this series is scatter-gather; currently
> all objects passed through binder IPC must be serialized in userspace, with
> with the exception of binder objects and file descriptors. By adding scatter-
> gather support for memory buffers, we can avoid the serialization copy,
> thereby increasing performance for larger transaction sizes.
>
> The two patches from Arve are security patches that we've already applied
> in android-common. I included them in front of the series because my changes
> touch quite some of that code.

Hey Greg,
   Curious what the status of these patches are? Did you have any
outstanding objection, or were you just waiting for them to be resent?

AOSP is now making use of this feature, so mainline kernels have
trouble booting w/o it, so it would be good to get merged. :)

thanks
-john

[toc] | [next] | [standalone]


#1572911 — Re: [PATCH 00/10] android: binder: support for domains and scatter-gather.

FromGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Date2017-02-03 08:20 +0100
SubjectRe: [PATCH 00/10] android: binder: support for domains and scatter-gather.
Message-ID<t6Gk1-6Pp-13@gated-at.bofh.it>
In reply to#1572886
On Thu, Feb 02, 2017 at 08:56:04PM -0800, John Stultz wrote:
> On Mon, Oct 24, 2016 at 6:20 AM, Martijn Coenen <maco@android.com> wrote:
> > Android userspace will start using binder IPC for communication with HAL
> > modules. To clearly separate this IPC domain from the existing framework
> > IPC domain, this patch series adds support for multiple "binder domains".
> > This is implemented by each domain having its own binder context manager,
> > and then having a separate device node per domain.
> >
> > The other change introduced by this series is scatter-gather; currently
> > all objects passed through binder IPC must be serialized in userspace, with
> > with the exception of binder objects and file descriptors. By adding scatter-
> > gather support for memory buffers, we can avoid the serialization copy,
> > thereby increasing performance for larger transaction sizes.
> >
> > The two patches from Arve are security patches that we've already applied
> > in android-common. I included them in front of the series because my changes
> > touch quite some of that code.
> 
> Hey Greg,
>    Curious what the status of these patches are? Did you have any
> outstanding objection, or were you just waiting for them to be resent?
> 
> AOSP is now making use of this feature, so mainline kernels have
> trouble booting w/o it, so it would be good to get merged. :)

The patches that ended up in AOSP are a bit different from this series
from what I remember, which is why I didn't end up applying this series
at the time (it needed more work.)  If you could dig up the latest
versions, I'll be glad to review/merge them as well.

There's also the 'hwbinder' work that should be in AOSP also, that
should get merged too.

thanks,

greg k-h

[toc] | [prev] | [standalone]


Back to top | Article view | linux.kernel


csiph-web