[PATCH v5 0/7] dma-buf: Performance improvements for system heap & a system-uncached implementation

Fri Nov 20 06:32:45 UTC 2020

Hi Daniel,

On Wed, 18 Nov 2020 at 13:16, Daniel Vetter <daniel at ffwll.ch> wrote:
>
> On Wed, Nov 18, 2020 at 3:40 AM John Stultz <john.stultz at linaro.org> wrote:
> > On Fri, Nov 13, 2020 at 12:39 PM Daniel Vetter <daniel at ffwll.ch> wrote:
> > > On Thu, Nov 12, 2020 at 08:11:02PM -0800, John Stultz wrote:
> > > > On Thu, Nov 12, 2020 at 1:32 AM Daniel Vetter <daniel at ffwll.ch> wrote:
> > > > > On Thu, Nov 12, 2020 at 11:09:04AM +0530, Sumit Semwal wrote:
> > > > > > On Tue, 10 Nov 2020 at 09:19, John Stultz <john.stultz at linaro.org> wrote:
> > > > > > >
> > > > > > > Hey All,
> > > > > > >   So just wanted to send my last revision of my patch series
> > > > > > > of performance optimizations to the dma-buf system heap.
> > > > > >
> > > > > > Thanks very much for your patches - I think the first 5 patches look good to me.
> > > > > >
> > > > > > I know there was a bit of discussion over adding a new system-uncached
> > > > > > heap v/s using a flag to identify that; I think I prefer the separate
> > > > > > heap idea, but lets ask one last time if any one else has any real
> > > > > > objections to it.
> > > > > >
> > > > > > Daniel, Christian: any comments from your side on this?
> > > > >
> > > > > I do wonder a bit where the userspace stack for this all is, since tuning
> > > > > allocators without a full stack is fairly pointless. dma-buf heaps is a
> > > > > bit in a limbo situation here it feels like.
> > > >
> > > > As mentioned in the system-uncached patch:
> > > > Pending opensource users of this code include:
> > > > * AOSP HiKey960 gralloc:
> > > >   - https://android-review.googlesource.com/c/device/linaro/hikey/+/1399519
> > > >   - Visibly improves performance over the system heap
> > > > * AOSP Codec2 (possibly, needs more review):
> > > >   - https://android-review.googlesource.com/c/platform/frameworks/av/+/1360640/17/media/codec2/vndk/C2DmaBufAllocator.cpp#325
> > > >
> > > > Additionally both the HiKey, HiKey960 grallocs  and Codec2 are already
> > > > able to use the current dmabuf heaps instead of ION.
> > > >
> > > > So I'm not sure what you mean by limbo, other than it being in a
> > > > transition state where the interface is upstream and we're working on
> > > > moving vendors to it from ION (which is staged to be dropped in 5.11).
> > > > Part of that work is making sure we don't regress the performance
> > > > expectations.
> > >
> > > The mesa thing below, since if we test this with some downstream kernel
> > > drivers or at least non-mesa userspace I'm somewhat worried we're just
> > > creating a nice split world between the android gfx world and the
> > > mesa/linux desktop gfx world.
> > >
> > > But then that's kinda how android rolls, so *shrug*
> > >
> > > > > Plus I'm vary of anything related to leaking this kind of stuff beyond the
> > > > > dma-api because dma api maintainers don't like us doing that. But
> > > > > personally no concern on that front really, gpus need this. It's just that
> > > > > we do need solid justification I think if we land this. Hence back to
> > > > > first point.
> > > > >
> > > > > Ideally first point comes in the form of benchmarking on android together
> > > > > with a mesa driver (or mesa + some v4l driver or whatever it takes to
> > > > > actually show the benefits, I have no idea).
> > > >
> > > > Tying it with mesa is a little tough as the grallocs for mesa devices
> > > > usually use gbm (gralloc.gbm or gralloc.minigbm). Swapping the
> > > > allocation path for dmabuf heaps there gets a little complex as last I
> > > > tried that (when trying to get HiKey working with Lima graphics, as
> > > > gbm wouldn't allocate the contiguous buffers required by the display),
> > > > I ran into issues with the drm_hwcomposer and mesa expecting the gbm
> > > > private handle metadata in the buffer when it was passed in.
> > > >
> > > > But I might take a look at it again. I got a bit lost digging through
> > > > the mesa gbm allocation paths last time.
> > > >
> > > > I'll also try to see if I can find a benchmark for the codec2 code
> > > > (using dmabuf heaps with and without the uncached heap) on on db845c
> > > > (w/ mesa), as that is already working and I suspect that might be
> > > > close to what you're looking for.
> > >
> > > tbh I think trying to push for this long term is the best we can hope for.
> > >
> > > Media is also a lot more *meh* since it's deeply fragmented and a lot less
> > > of it upstream than on the gles/display side.
> > >
> > > I think confirming that this at least doesn't horrible blow up on a
> > > gralloc/gbm+mesa stack would be useful I think.
> >
> > Sorry, I'm still a little foggy on precisely what you're suggesting here.
> >
> > The patch stack I have has already been used with db845c (mesa +
> > gbm_grallloc), with the codec2 (sw decoders) using dmabuf heaps.
> > So no blowing up there. And I'm working with Hridya to find a
> > benchmark for codec2 so we can try to show the performance delta.
> >
> > However, if you're wanting a dma-buf gralloc implementation with mesa,
> > that may be a little tougher to do, but I guess I can give it a go.
> >
> > Hopefully this will address concerns about the system-uncached heap
> > patch (the last two patches in this series)?
> >
> > In the meantime I hope we can queue the first five patches, as it
> > would be nice to get the code rearranging in as there are others
> > trying to stage their own heaps, and I'd like to avoid dragging that
> > churn out for too long (in addition to improving the allocation
> > performance). Those changes have no ABI implications.
>
> Maybe I'm also misunderstanding what dma-buf heaps is used for in
> Android, at least usually. I thought it's used to allocate all the
> winsys/shared buffers through gralloc (at least in the blobby stacks),
> to handle the allocation constraints problem. In the open stacks we
> don't seem to have a platform with both mesa and v4l (or some other
> codec) with "interesting" allocations constraints, so no one using
> that gralloc+dma-buf heaps combo for what it was meant for. Hence why
> I'm a bit vary that we're creating something here which just misses
> the point a bit when we try to actually use it (in that glorious
> forever-future world where an android platform has enough drivers in
> upstream to do so).
>
> For other "this solves a system problem" we tend to be quite a bit
> more picky with the demonstration use case, to make sure we're
> actually creating something that solves the problem in reality.
>
> But it also looks like Android's just not there yet, so *shrug* ...

For me, looking at the first 5 patches (listed below, for quick
reference), they are only doing code reorganisation and minor updates
for already existing heaps, and no ABI change, I am not able to
clearly see your objection here. To me, these seem to be required
updates that the existing system heap users can benefit from.

dma-buf: system_heap: Rework system heap to use sgtables instead of
    pagelists
  dma-buf: heaps: Move heap-helper logic into the cma_heap
    implementation
  dma-buf: heaps: Remove heap-helpers code
  dma-buf: heaps: Skip sync if not mapped
  dma-buf: system_heap: Allocate higher order pages if available

If we talk about the last two patches - the ones that add system
uncached heap, I somewhat agree that we should be able to show the
performance gains with this approach (which has been in use on ION and
in devices) using dma-buf gralloc or similar.

We can discuss the system-uncached heap when the dma-buf gralloc or
similar demonstration for performance benefits is done, but I am
inclined to push these 5 patches listed above through.

Best,
Sumit.

> -Daniel
> --
> Daniel Vetter
> Software Engineer, Intel Corporation
> http://blog.ffwll.ch