[PATCH 1/6] dma-buf: add dynamic DMA-buf handling v13

Daniel Vetter daniel at ffwll.ch
Wed Jul 31 10:39:51 UTC 2019


On Wed, Jul 31, 2019 at 11:44 AM Christian K├Ânig
<ckoenig.leichtzumerken at gmail.com> wrote:
>
> Am 31.07.19 um 11:12 schrieb Daniel Vetter:
> > [SNIP]
> > I think I brought this up before, but new top-post for a clean start.
> >
> > Use-case I have in mind is something like amdkfd's model, where you have a
> > list of buffers (per context or whatever) that you always need to have
> > present. Idea is to also use this for traditional CS for vk/gl, to cut
> > down on the buffer management overhead, but we'd still allow additional
> > buffers to be listed per-CS on top of that default working set.
> >
> > This of course means no implicit sync anymore on these default buffers
> > (the point is to avoid touching every buffer on every CS, updating fences
> > would defeat that). That's why the CS can still list additional buffers,
> > the only reason for that is to add implicit sync fences. Those buffers
> > would be most likely in the default working set already.
> >
> > Consequence is that I want the amdkfd model of "evict when needed, but
> > keep resident by default", but also working implicit fences. And it must
> > be doable without touching every bo on every CS. Listing possible
> > implementation options:
> >
> > - the amdkfd trick doesn't work because it would break implicit fencing -
> >    any implicit sync would always result in the context getting
> >    preempted/evicted, which isn't great.
>
> I'm actually working on re-working implicit fencing towards better
> supporting this.
>
> Basic idea is that you split up the fences in a reservation object into
> three categories:
> 1. Implicit sync on write.
> 2. Implicit sync on read.
> 3. No implicit sync at all.

Not really sure what you want to do here ... implicit sync is opt-in
(or opt-out flag if you need to keep CS backwards compat) per BO/CS.
At least when we discussed this forever at some XDCs consensus was
that storing the implicit sync mode on the BO is not going to work.

> This should not only help the KFD, but also with amdgpu command
> submission and things like page tables updates.
>
> E.g. we need to fences for page tables updates around in reservation
> objects as well, but you really really really don't want any implicit
> synchronization with them :)

Why do you even try to do implicit sync with your pagetables? How can
your pagetables even get anywhere near where implicit sync matters?
I'm confused ... If it's because ttm doesn't allow you to override the
eviction order because it's a midlayer I think the correct fix is to
demidlayer.

> I think that having a consensus of the meaning of the fences in a
> reservation object will be rather fundamental for what we are planning
> to do here.

Yeah that I can agree on.

> > - we share the resv_obj between all the buffers in the default working set
> >    of a context, with unsharing/resharing the resv_obj if they enter/leave
> >    the default working set. That way there's only one resv_obj to update on
> >    each CS, and we can attach a new shared fence for every CS. Trouble is
> >    that this means a given buffer can only be part of one default working
> >    set, so all shared buffers would need to be listed again separately. Not
> >    so great if userspace has to deal with that fairly arbitrary limitation.
>
> Yeah, that is exactly what we do with the per VM BOs in amdgpu.
>
> The limitation that you have only one working set actually turned out to
> be not a limitation at all, but rather seen as something welcomed by our
> Vulkan guys.

We have per-ctx vms in i915, but I guess even for those sharing will be limited.

> I also don't really see a way to have an implementation with good
> performance where BOs can be in multiple working sets at the same time.
>
> > - we allow the ->move_notify callback to add new fences, which the
> >    exporter needs to wait on before it schedules the pipelined move. This
> >    also avoids the per-BO update on every CS, and it would allow buffers to
> >    be shared and to be in multiple default working sets. The downside is
> >    that ->move_notify needs to be able to cope with added fences, which
> >    means we might need to grow the shared fences array, which might fail
> >    with ENOMEM. Not great. We could fix this with some kind of permanent
> >    shared fence slot reservations (i.e. a reserved slot which outlives
> >    holding the reservation lock), but that might waste quite a bit of
> >    memory. Probably not real problem in the grand scheme of things. I think
> >    the fence itself can be preallocated per context, so that isn't the
> >    problem.
>
> Well the ENOMEM problem is the least of the problems with this approach.
> You can still block for the fence which you wanted to add.
>
> The real problem is that you can't tell if a BO is busy or not by just
> looking at its current fences.
>
> In other words when you stop adding fences you also want to stop moving
> them individually on the LRU.

Well yeah, otherwise you're back a per BO overhead on CS. That's kinda
obvious. But also not sure why this matters, since atm we don't have
any proposed means to pass LRU updates through dma-buf. So exporter
(even with dynamic dma-buf) has to pessimistically assume already that
the exported BO is more busy than what the LRU position suggests, and
evict them later on. If we also need to funnel LRU updates over
dma-buf, then that's another beast entirely (and probably not what we
want to do at all).

Also if you don't want to stall, then just reject after ->move_notify
wherethere the bo is still idle, and give up if so. Or we can add a
counter to indicate a bo needs to be considered busy.

> When the memory management evicts one BO you essentially kick out a
> whole process/working set.
>
> So when you want to kick out the next BO you actually want to do this
> for BOs which now became available anyway.
>
> That approach won't work with the move_notify callback.

So essentially we don't just want move_notify, but also full LRU information?

> > - same as above, but the new fence doesn't get added, but returned to the
> >    caller, and the exporter deals with the ENOMEM mess. Might not work
> >    since an importer could have a lot of contexts using a given object, and
> >    so would have a lot of fences to add.
>
> I don't think that this will work.
>
> See you not only need to be able to add the fence to the BO currently
> evicted, but also to all other BO in your process/working set.
>
> Additional to that moving the ENOMEM handling from the importer to the
> exporter sounds as helpful as adding another layer of abstraction :)

You can go and pick a different victim to evict, or just give up
(there's not much you can do with ENOMEM). Instead of data corruption
because you're not waiting for a fence you should be waiting on.
-Daniel

>
> Regards,
> Christian.
>
> >
> > - something entirely different?
> >
> > Thoughts?
> >
> > Cheers, Daniel
> >
> >> ---
> >>   drivers/dma-buf/dma-buf.c | 183 ++++++++++++++++++++++++++++++++++++--
> >>   include/linux/dma-buf.h   | 108 ++++++++++++++++++++--
> >>   2 files changed, 277 insertions(+), 14 deletions(-)
> >>
> >> diff --git a/drivers/dma-buf/dma-buf.c b/drivers/dma-buf/dma-buf.c
> >> index 6c15deb5d4ad..bd8611fa2cfa 100644
> >> --- a/drivers/dma-buf/dma-buf.c
> >> +++ b/drivers/dma-buf/dma-buf.c
> >> @@ -531,6 +531,9 @@ struct dma_buf *dma_buf_export(const struct dma_buf_export_info *exp_info)
> >>              return ERR_PTR(-EINVAL);
> >>      }
> >>
> >> +    if (WARN_ON(exp_info->ops->cache_sgt_mapping && exp_info->ops->pin))
> >> +            return ERR_PTR(-EINVAL);
> >> +
> >>      if (!try_module_get(exp_info->owner))
> >>              return ERR_PTR(-ENOENT);
> >>
> >> @@ -651,10 +654,12 @@ void dma_buf_put(struct dma_buf *dmabuf)
> >>   EXPORT_SYMBOL_GPL(dma_buf_put);
> >>
> >>   /**
> >> - * dma_buf_attach - Add the device to dma_buf's attachments list; optionally,
> >> + * dma_buf_dynamic_attach - Add the device to dma_buf's attachments list; optionally,
> >>    * calls attach() of dma_buf_ops to allow device-specific attach functionality
> >> - * @dmabuf: [in]    buffer to attach device to.
> >> - * @dev:    [in]    device to be attached.
> >> + * @dmabuf:         [in]    buffer to attach device to.
> >> + * @dev:            [in]    device to be attached.
> >> + * @importer_ops    [in]    importer operations for the attachment
> >> + * @importer_priv   [in]    importer private pointer for the attachment
> >>    *
> >>    * Returns struct dma_buf_attachment pointer for this attachment. Attachments
> >>    * must be cleaned up by calling dma_buf_detach().
> >> @@ -668,8 +673,10 @@ EXPORT_SYMBOL_GPL(dma_buf_put);
> >>    * accessible to @dev, and cannot be moved to a more suitable place. This is
> >>    * indicated with the error code -EBUSY.
> >>    */
> >> -struct dma_buf_attachment *dma_buf_attach(struct dma_buf *dmabuf,
> >> -                                      struct device *dev)
> >> +struct dma_buf_attachment *
> >> +dma_buf_dynamic_attach(struct dma_buf *dmabuf, struct device *dev,
> >> +                   const struct dma_buf_attach_ops *importer_ops,
> >> +                   void *importer_priv)
> >>   {
> >>      struct dma_buf_attachment *attach;
> >>      int ret;
> >> @@ -683,6 +690,8 @@ struct dma_buf_attachment *dma_buf_attach(struct dma_buf *dmabuf,
> >>
> >>      attach->dev = dev;
> >>      attach->dmabuf = dmabuf;
> >> +    attach->importer_ops = importer_ops;
> >> +    attach->importer_priv = importer_priv;
> >>
> >>      mutex_lock(&dmabuf->lock);
> >>
> >> @@ -691,16 +700,72 @@ struct dma_buf_attachment *dma_buf_attach(struct dma_buf *dmabuf,
> >>              if (ret)
> >>                      goto err_attach;
> >>      }
> >> +    reservation_object_lock(dmabuf->resv, NULL);
> >>      list_add(&attach->node, &dmabuf->attachments);
> >> +    reservation_object_unlock(dmabuf->resv);
> >>
> >>      mutex_unlock(&dmabuf->lock);
> >>
> >> +    /* When either the importer or the exporter can't handle dynamic
> >> +     * mappings we cache the mapping here to avoid issues with the
> >> +     * reservation object lock.
> >> +     */
> >> +    if (dma_buf_attachment_is_dynamic(attach) !=
> >> +        dma_buf_is_dynamic(dmabuf)) {
> >> +            struct sg_table *sgt;
> >> +
> >> +            if (dma_buf_is_dynamic(attach->dmabuf)) {
> >> +                    reservation_object_lock(attach->dmabuf->resv, NULL);
> >> +                    ret = dma_buf_pin(attach);
> >> +                    if (ret)
> >> +                            goto err_unlock;
> >> +            }
> >> +
> >> +            sgt = dmabuf->ops->map_dma_buf(attach, DMA_BIDIRECTIONAL);
> >> +            if (!sgt)
> >> +                    sgt = ERR_PTR(-ENOMEM);
> >> +            if (IS_ERR(sgt)) {
> >> +                    ret = PTR_ERR(sgt);
> >> +                    goto err_unpin;
> >> +            }
> >> +            if (dma_buf_is_dynamic(attach->dmabuf))
> >> +                    reservation_object_unlock(attach->dmabuf->resv);
> >> +            attach->sgt = sgt;
> >> +            attach->dir = DMA_BIDIRECTIONAL;
> >> +    }
> >> +
> >>      return attach;
> >>
> >>   err_attach:
> >>      kfree(attach);
> >>      mutex_unlock(&dmabuf->lock);
> >>      return ERR_PTR(ret);
> >> +
> >> +err_unpin:
> >> +    if (dma_buf_is_dynamic(attach->dmabuf))
> >> +            dma_buf_unpin(attach);
> >> +
> >> +err_unlock:
> >> +    if (dma_buf_is_dynamic(attach->dmabuf))
> >> +            reservation_object_unlock(attach->dmabuf->resv);
> >> +
> >> +    dma_buf_detach(dmabuf, attach);
> >> +    return ERR_PTR(ret);
> >> +}
> >> +EXPORT_SYMBOL_GPL(dma_buf_dynamic_attach);
> >> +
> >> +/**
> >> + * dma_buf_attach - Wrapper for dma_buf_dynamic_attach
> >> + * @dmabuf: [in]    buffer to attach device to.
> >> + * @dev:    [in]    device to be attached.
> >> + *
> >> + * Wrapper to call dma_buf_dynamic_attach() for drivers which still use a static
> >> + * mapping.
> >> + */
> >> +struct dma_buf_attachment *dma_buf_attach(struct dma_buf *dmabuf,
> >> +                                      struct device *dev)
> >> +{
> >> +    return dma_buf_dynamic_attach(dmabuf, dev, NULL, NULL);
> >>   }
> >>   EXPORT_SYMBOL_GPL(dma_buf_attach);
> >>
> >> @@ -717,11 +782,22 @@ void dma_buf_detach(struct dma_buf *dmabuf, struct dma_buf_attachment *attach)
> >>      if (WARN_ON(!dmabuf || !attach))
> >>              return;
> >>
> >> -    if (attach->sgt)
> >> +    if (attach->sgt) {
> >> +            if (dma_buf_is_dynamic(attach->dmabuf))
> >> +                    reservation_object_lock(attach->dmabuf->resv, NULL);
> >> +
> >>              dmabuf->ops->unmap_dma_buf(attach, attach->sgt, attach->dir);
> >>
> >> +            if (dma_buf_is_dynamic(attach->dmabuf)) {
> >> +                    dma_buf_unpin(attach);
> >> +                    reservation_object_unlock(attach->dmabuf->resv);
> >> +            }
> >> +    }
> >> +
> >>      mutex_lock(&dmabuf->lock);
> >> +    reservation_object_lock(dmabuf->resv, NULL);
> >>      list_del(&attach->node);
> >> +    reservation_object_unlock(dmabuf->resv);
> >>      if (dmabuf->ops->detach)
> >>              dmabuf->ops->detach(dmabuf, attach);
> >>
> >> @@ -730,6 +806,44 @@ void dma_buf_detach(struct dma_buf *dmabuf, struct dma_buf_attachment *attach)
> >>   }
> >>   EXPORT_SYMBOL_GPL(dma_buf_detach);
> >>
> >> +/**
> >> + * dma_buf_pin - Lock down the DMA-buf
> >> + *
> >> + * @attach: [in]    attachment which should be pinned
> >> + *
> >> + * Returns:
> >> + * 0 on success, negative error code on failure.
> >> + */
> >> +int dma_buf_pin(struct dma_buf_attachment *attach)
> >> +{
> >> +    struct dma_buf *dmabuf = attach->dmabuf;
> >> +    int ret = 0;
> >> +
> >> +    reservation_object_assert_held(dmabuf->resv);
> >> +
> >> +    if (dmabuf->ops->pin)
> >> +            ret = dmabuf->ops->pin(attach);
> >> +
> >> +    return ret;
> >> +}
> >> +EXPORT_SYMBOL_GPL(dma_buf_pin);
> >> +
> >> +/**
> >> + * dma_buf_unpin - Remove lock from DMA-buf
> >> + *
> >> + * @attach: [in]    attachment which should be unpinned
> >> + */
> >> +void dma_buf_unpin(struct dma_buf_attachment *attach)
> >> +{
> >> +    struct dma_buf *dmabuf = attach->dmabuf;
> >> +
> >> +    reservation_object_assert_held(dmabuf->resv);
> >> +
> >> +    if (dmabuf->ops->unpin)
> >> +            dmabuf->ops->unpin(attach);
> >> +}
> >> +EXPORT_SYMBOL_GPL(dma_buf_unpin);
> >> +
> >>   /**
> >>    * dma_buf_map_attachment - Returns the scatterlist table of the attachment;
> >>    * mapped into _device_ address space. Is a wrapper for map_dma_buf() of the
> >> @@ -749,6 +863,7 @@ struct sg_table *dma_buf_map_attachment(struct dma_buf_attachment *attach,
> >>                                      enum dma_data_direction direction)
> >>   {
> >>      struct sg_table *sg_table;
> >> +    int r;
> >>
> >>      might_sleep();
> >>
> >> @@ -767,10 +882,31 @@ struct sg_table *dma_buf_map_attachment(struct dma_buf_attachment *attach,
> >>              return attach->sgt;
> >>      }
> >>
> >> +    if (dma_buf_attachment_is_dynamic(attach)) {
> >> +            reservation_object_assert_held(attach->dmabuf->resv);
> >> +
> >> +    } else if (dma_buf_is_dynamic(attach->dmabuf)) {
> >> +            reservation_object_lock(attach->dmabuf->resv, NULL);
> >> +            r = dma_buf_pin(attach);
> >> +            if (r) {
> >> +                    reservation_object_unlock(attach->dmabuf->resv);
> >> +                    return ERR_PTR(r);
> >> +            }
> >> +    } else {
> >> +            might_lock(&attach->dmabuf->resv->lock.base);
> >> +    }
> >> +
> >>      sg_table = attach->dmabuf->ops->map_dma_buf(attach, direction);
> >>      if (!sg_table)
> >>              sg_table = ERR_PTR(-ENOMEM);
> >>
> >> +    if (!dma_buf_attachment_is_dynamic(attach) &&
> >> +        dma_buf_is_dynamic(attach->dmabuf)) {
> >> +            if (IS_ERR(sg_table))
> >> +                    dma_buf_unpin(attach);
> >> +            reservation_object_unlock(attach->dmabuf->resv);
> >> +    }
> >> +
> >>      if (!IS_ERR(sg_table) && attach->dmabuf->ops->cache_sgt_mapping) {
> >>              attach->sgt = sg_table;
> >>              attach->dir = direction;
> >> @@ -802,10 +938,43 @@ void dma_buf_unmap_attachment(struct dma_buf_attachment *attach,
> >>      if (attach->sgt == sg_table)
> >>              return;
> >>
> >> +    if (dma_buf_attachment_is_dynamic(attach))
> >> +            reservation_object_assert_held(attach->dmabuf->resv);
> >> +    else if (dma_buf_is_dynamic(attach->dmabuf))
> >> +            reservation_object_lock(attach->dmabuf->resv, NULL);
> >> +    else
> >> +            might_lock(&attach->dmabuf->resv->lock.base);
> >> +
> >>      attach->dmabuf->ops->unmap_dma_buf(attach, sg_table, direction);
> >> +
> >> +    if (dma_buf_is_dynamic(attach->dmabuf) &&
> >> +        !dma_buf_attachment_is_dynamic(attach)) {
> >> +            dma_buf_unpin(attach);
> >> +            reservation_object_unlock(attach->dmabuf->resv);
> >> +    }
> >>   }
> >>   EXPORT_SYMBOL_GPL(dma_buf_unmap_attachment);
> >>
> >> +/**
> >> + * dma_buf_move_notify - notify attachments that DMA-buf is moving
> >> + *
> >> + * @dmabuf: [in]    buffer which is moving
> >> + *
> >> + * Informs all attachmenst that they need to destroy and recreated all their
> >> + * mappings.
> >> + */
> >> +void dma_buf_move_notify(struct dma_buf *dmabuf)
> >> +{
> >> +    struct dma_buf_attachment *attach;
> >> +
> >> +    reservation_object_assert_held(dmabuf->resv);
> >> +
> >> +    list_for_each_entry(attach, &dmabuf->attachments, node)
> >> +            if (attach->importer_ops && attach->importer_ops->move_notify)
> >> +                    attach->importer_ops->move_notify(attach);
> >> +}
> >> +EXPORT_SYMBOL_GPL(dma_buf_move_notify);
> >> +
> >>   /**
> >>    * DOC: cpu access
> >>    *
> >> @@ -1225,10 +1394,12 @@ static int dma_buf_debug_show(struct seq_file *s, void *unused)
> >>              seq_puts(s, "\tAttached Devices:\n");
> >>              attach_count = 0;
> >>
> >> +            reservation_object_lock(buf_obj->resv, NULL);
> >>              list_for_each_entry(attach_obj, &buf_obj->attachments, node) {
> >>                      seq_printf(s, "\t%s\n", dev_name(attach_obj->dev));
> >>                      attach_count++;
> >>              }
> >> +            reservation_object_unlock(buf_obj->resv);
> >>
> >>              seq_printf(s, "Total %d devices attached\n\n",
> >>                              attach_count);
> >> diff --git a/include/linux/dma-buf.h b/include/linux/dma-buf.h
> >> index 01ad5b942a6f..ccad2fc1f640 100644
> >> --- a/include/linux/dma-buf.h
> >> +++ b/include/linux/dma-buf.h
> >> @@ -92,14 +92,40 @@ struct dma_buf_ops {
> >>       */
> >>      void (*detach)(struct dma_buf *, struct dma_buf_attachment *);
> >>
> >> +    /**
> >> +     * @pin:
> >> +     *
> >> +     * This is called by dma_buf_pin and lets the exporter know that the
> >> +     * DMA-buf can't be moved any more.
> >> +     *
> >> +     * This is called with the dmabuf->resv object locked.
> >> +     *
> >> +     * This callback is optional.
> >> +     *
> >> +     * Returns:
> >> +     *
> >> +     * 0 on success, negative error code on failure.
> >> +     */
> >> +    int (*pin)(struct dma_buf_attachment *attach);
> >> +
> >> +    /**
> >> +     * @unpin:
> >> +     *
> >> +     * This is called by dma_buf_unpin and lets the exporter know that the
> >> +     * DMA-buf can be moved again.
> >> +     *
> >> +     * This is called with the dmabuf->resv object locked.
> >> +     *
> >> +     * This callback is optional.
> >> +     */
> >> +    void (*unpin)(struct dma_buf_attachment *attach);
> >> +
> >>      /**
> >>       * @map_dma_buf:
> >>       *
> >>       * This is called by dma_buf_map_attachment() and is used to map a
> >>       * shared &dma_buf into device address space, and it is mandatory. It
> >> -     * can only be called if @attach has been called successfully. This
> >> -     * essentially pins the DMA buffer into place, and it cannot be moved
> >> -     * any more
> >> +     * can only be called if @attach has been called successfully.
> >>       *
> >>       * This call may sleep, e.g. when the backing storage first needs to be
> >>       * allocated, or moved to a location suitable for all currently attached
> >> @@ -120,6 +146,9 @@ struct dma_buf_ops {
> >>       * any other kind of sharing that the exporter might wish to make
> >>       * available to buffer-users.
> >>       *
> >> +     * This is always called with the dmabuf->resv object locked when
> >> +     * the pin/unpin callbacks are implemented.
> >> +     *
> >>       * Returns:
> >>       *
> >>       * A &sg_table scatter list of or the backing storage of the DMA buffer,
> >> @@ -137,9 +166,6 @@ struct dma_buf_ops {
> >>       *
> >>       * This is called by dma_buf_unmap_attachment() and should unmap and
> >>       * release the &sg_table allocated in @map_dma_buf, and it is mandatory.
> >> -     * It should also unpin the backing storage if this is the last mapping
> >> -     * of the DMA buffer, it the exporter supports backing storage
> >> -     * migration.
> >>       */
> >>      void (*unmap_dma_buf)(struct dma_buf_attachment *,
> >>                            struct sg_table *,
> >> @@ -330,6 +356,34 @@ struct dma_buf {
> >>      } cb_excl, cb_shared;
> >>   };
> >>
> >> +/**
> >> + * struct dma_buf_attach_ops - importer operations for an attachment
> >> + * @move_notify: [optional] notification that the DMA-buf is moving
> >> + *
> >> + * Attachment operations implemented by the importer.
> >> + */
> >> +struct dma_buf_attach_ops {
> >> +    /**
> >> +     * @move_notify
> >> +     *
> >> +     * If this callback is provided the framework can avoid pinning the
> >> +     * backing store while mappings exists.
> >> +     *
> >> +     * This callback is called with the lock of the reservation object
> >> +     * associated with the dma_buf held and the mapping function must be
> >> +     * called with this lock held as well. This makes sure that no mapping
> >> +     * is created concurrently with an ongoing move operation.
> >> +     *
> >> +     * Mappings stay valid and are not directly affected by this callback.
> >> +     * But the DMA-buf can now be in a different physical location, so all
> >> +     * mappings should be destroyed and re-created as soon as possible.
> >> +     *
> >> +     * New mappings can be created after this callback returns, and will
> >> +     * point to the new location of the DMA-buf.
> >> +     */
> >> +    void (*move_notify)(struct dma_buf_attachment *attach);
> >> +};
> >> +
> >>   /**
> >>    * struct dma_buf_attachment - holds device-buffer attachment data
> >>    * @dmabuf: buffer for this attachment.
> >> @@ -338,6 +392,8 @@ struct dma_buf {
> >>    * @sgt: cached mapping.
> >>    * @dir: direction of cached mapping.
> >>    * @priv: exporter specific attachment data.
> >> + * @importer_ops: importer operations for this attachment.
> >> + * @importer_priv: importer specific attachment data.
> >>    *
> >>    * This structure holds the attachment information between the dma_buf buffer
> >>    * and its user device(s). The list contains one attachment struct per device
> >> @@ -355,6 +411,9 @@ struct dma_buf_attachment {
> >>      struct sg_table *sgt;
> >>      enum dma_data_direction dir;
> >>      void *priv;
> >> +
> >> +    const struct dma_buf_attach_ops *importer_ops;
> >> +    void *importer_priv;
> >>   };
> >>
> >>   /**
> >> @@ -405,10 +464,42 @@ static inline void get_dma_buf(struct dma_buf *dmabuf)
> >>      get_file(dmabuf->file);
> >>   }
> >>
> >> +/**
> >> + * dma_buf_is_dynamic - check if a DMA-buf uses dynamic mappings.
> >> + * @dmabuf: the DMA-buf to check
> >> + *
> >> + * Returns true if a DMA-buf exporter wants to create dynamic sg table mappings
> >> + * for each attachment. False if only a single static sg table should be used.
> >> + */
> >> +static inline bool dma_buf_is_dynamic(struct dma_buf *dmabuf)
> >> +{
> >> +    return !!dmabuf->ops->pin;
> >> +}
> >> +
> >> +/**
> >> + * dma_buf_attachment_is_dynamic - check if a DMA-buf attachment uses dynamic
> >> + * mappinsg
> >> + * @attach: the DMA-buf attachment to check
> >> + *
> >> + * Returns true if a DMA-buf importer wants to use dynamic sg table mappings and
> >> + * calls the map/unmap functions with the reservation object locked.
> >> + */
> >> +static inline bool
> >> +dma_buf_attachment_is_dynamic(struct dma_buf_attachment *attach)
> >> +{
> >> +    return attach->importer_ops && attach->importer_ops->move_notify;
> >> +}
> >> +
> >>   struct dma_buf_attachment *dma_buf_attach(struct dma_buf *dmabuf,
> >> -                                                    struct device *dev);
> >> +                                      struct device *dev);
> >> +struct dma_buf_attachment *
> >> +dma_buf_dynamic_attach(struct dma_buf *dmabuf, struct device *dev,
> >> +                   const struct dma_buf_attach_ops *importer_ops,
> >> +                   void *importer_priv);
> >>   void dma_buf_detach(struct dma_buf *dmabuf,
> >> -                            struct dma_buf_attachment *dmabuf_attach);
> >> +                struct dma_buf_attachment *attach);
> >> +int dma_buf_pin(struct dma_buf_attachment *attach);
> >> +void dma_buf_unpin(struct dma_buf_attachment *attach);
> >>
> >>   struct dma_buf *dma_buf_export(const struct dma_buf_export_info *exp_info);
> >>
> >> @@ -420,6 +511,7 @@ struct sg_table *dma_buf_map_attachment(struct dma_buf_attachment *,
> >>                                      enum dma_data_direction);
> >>   void dma_buf_unmap_attachment(struct dma_buf_attachment *, struct sg_table *,
> >>                              enum dma_data_direction);
> >> +void dma_buf_move_notify(struct dma_buf *dma_buf);
> >>   int dma_buf_begin_cpu_access(struct dma_buf *dma_buf,
> >>                           enum dma_data_direction dir);
> >>   int dma_buf_end_cpu_access(struct dma_buf *dma_buf,
> >> --
> >> 2.17.1
> >>
> >> _______________________________________________
> >> dri-devel mailing list
> >> dri-devel at lists.freedesktop.org
> >> https://lists.freedesktop.org/mailman/listinfo/dri-devel
>


-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch


More information about the amd-gfx mailing list