On Tue, Dec 15, 2020 at 03:18:49PM +0100, Christian König wrote:
> Am 14.12.20 um 17:01 schrieb Daniel Vetter:
> > On Mon, Dec 14, 2020 at 11:33:10AM +0100, Christian König wrote:
> > > Am 11.12.20 um 16:58 schrieb Daniel Vetter:
> > > > Also try to clarify a bit when dma_buf_begin/end_cpu_access should
> > > > be called.
> > > >
> > > > Signed-off-by: Daniel Vetter <daniel.vetter@xxxxxxxxx>
> > > > Cc: Thomas Zimmermann <tzimmermann@xxxxxxx>
> > > > Cc: Sumit Semwal <sumit.semwal@xxxxxxxxxx>
> > > > Cc: "Christian König" <christian.koenig@xxxxxxx>
> > > > Cc: linux-media@xxxxxxxxxxxxxxx
> > > > Cc: linaro-mm-sig@xxxxxxxxxxxxxxxx
> > > > ---
> > > > drivers/dma-buf/dma-buf.c | 20 ++++++++++++++------
> > > > include/linux/dma-buf.h | 25 +++++++++----------------
> > > > 2 files changed, 23 insertions(+), 22 deletions(-)
> > > >
> > > > diff --git a/drivers/dma-buf/dma-buf.c b/drivers/dma-buf/dma-buf.c
> > > > index e63684d4cd90..a12fdffa130f 100644
> > > > --- a/drivers/dma-buf/dma-buf.c
> > > > +++ b/drivers/dma-buf/dma-buf.c
> > > > @@ -1001,15 +1001,15 @@ EXPORT_SYMBOL_GPL(dma_buf_move_notify);
> > > > * vmalloc space might be limited and result in vmap calls failing.
> > > > *
> > > > * Interfaces::
> > > > + *
> > > > * void \*dma_buf_vmap(struct dma_buf \*dmabuf)
> > > > * void dma_buf_vunmap(struct dma_buf \*dmabuf, void \*vaddr)
> > > > *
> > > > * The vmap call can fail if there is no vmap support in the exporter, or if
> > > > - * it runs out of vmalloc space. Fallback to kmap should be implemented. Note
> > > > - * that the dma-buf layer keeps a reference count for all vmap access and
> > > > - * calls down into the exporter's vmap function only when no vmapping exists,
> > > > - * and only unmaps it once. Protection against concurrent vmap/vunmap calls is
> > > > - * provided by taking the dma_buf->lock mutex.
> > > > + * it runs out of vmalloc space. Note that the dma-buf layer keeps a reference
> > > > + * count for all vmap access and calls down into the exporter's vmap function
> > > > + * only when no vmapping exists, and only unmaps it once. Protection against
> > > > + * concurrent vmap/vunmap calls is provided by taking the &dma_buf.lock mutex.
> > > Who is talking the lock? The caller of the dma_buf_vmap/vunmap() functions,
> > > the functions itself or the callback inside the exporter?
> > That's the part I didn't change at all here, just re-laid out the line
> > breaking. I only removed the outdated kmap section here.
>
> I just wanted to point out that this still isn't described here very very.
>
>
> > Should I do another patch and remove this one sentence here (it's kinda
> > pointless and generally we don't muse about implementation details that
> > callers don't care about)?
>
> Na, works like this for me.
>
> > I did try and do a cursory review of the dma-buf docs, but this is kinda
> > not meant as an all-out revamp. Just a few things I've noticed while
> > reviewing Thomas' vmap_local stuff.
>
>
> Fell free to add an Acked-by: Christian König <christian.koenig@xxxxxxx> to
> the series.
Thanks for taking a look, and yeah I actually want to do a review of all
the dma-buf docs but just haven't found the quiet time for that yet.
Patches pushed to drm-misc-next.
-Daniel
>
> Christian.
>
> > -Daniel
> >
> > > Christian.
> > >
> > > > *
> > > > * - For full compatibility on the importer side with existing userspace
> > > > * interfaces, which might already support mmap'ing buffers. This is needed in
> > > > @@ -1098,6 +1098,11 @@ static int __dma_buf_begin_cpu_access(struct dma_buf *dmabuf,
> > > > * dma_buf_end_cpu_access(). Only when cpu access is braketed by both calls is
> > > > * it guaranteed to be coherent with other DMA access.
> > > > *
> > > > + * This function will also wait for any DMA transactions tracked through
> > > > + * implicit synchronization in &dma_buf.resv. For DMA transactions with explicit
> > > > + * synchronization this function will only ensure cache coherency, callers must
> > > > + * ensure synchronization with such DMA transactions on their own.
> > > > + *
> > > > * Can return negative error values, returns 0 on success.
> > > > */
> > > > int dma_buf_begin_cpu_access(struct dma_buf *dmabuf,
> > > > @@ -1199,7 +1204,10 @@ EXPORT_SYMBOL_GPL(dma_buf_mmap);
> > > > * This call may fail due to lack of virtual mapping address space.
> > > > * These calls are optional in drivers. The intended use for them
> > > > * is for mapping objects linear in kernel space for high use objects.
> > > > - * Please attempt to use kmap/kunmap before thinking about these interfaces.
> > > > + *
> > > > + * To ensure coherency users must call dma_buf_begin_cpu_access() and
> > > > + * dma_buf_end_cpu_access() around any cpu access performed through this
> > > > + * mapping.
> > > > *
> > > > * Returns 0 on success, or a negative errno code otherwise.
> > > > */
> > > > diff --git a/include/linux/dma-buf.h b/include/linux/dma-buf.h
> > > > index cf72699cb2bc..7eca37c8b10c 100644
> > > > --- a/include/linux/dma-buf.h
> > > > +++ b/include/linux/dma-buf.h
> > > > @@ -183,24 +183,19 @@ struct dma_buf_ops {
> > > > * @begin_cpu_access:
> > > > *
> > > > * This is called from dma_buf_begin_cpu_access() and allows the
> > > > - * exporter to ensure that the memory is actually available for cpu
> > > > - * access - the exporter might need to allocate or swap-in and pin the
> > > > - * backing storage. The exporter also needs to ensure that cpu access is
> > > > - * coherent for the access direction. The direction can be used by the
> > > > - * exporter to optimize the cache flushing, i.e. access with a different
> > > > + * exporter to ensure that the memory is actually coherent for cpu
> > > > + * access. The exporter also needs to ensure that cpu access is coherent
> > > > + * for the access direction. The direction can be used by the exporter
> > > > + * to optimize the cache flushing, i.e. access with a different
> > > > * direction (read instead of write) might return stale or even bogus
> > > > * data (e.g. when the exporter needs to copy the data to temporary
> > > > * storage).
> > > > *
> > > > - * This callback is optional.
> > > > + * Note that this is both called through the DMA_BUF_IOCTL_SYNC IOCTL
> > > > + * command for userspace mappings established through @mmap, and also
> > > > + * for kernel mappings established with @vmap.
> > > > *
> > > > - * FIXME: This is both called through the DMA_BUF_IOCTL_SYNC command
> > > > - * from userspace (where storage shouldn't be pinned to avoid handing
> > > > - * de-factor mlock rights to userspace) and for the kernel-internal
> > > > - * users of the various kmap interfaces, where the backing storage must
> > > > - * be pinned to guarantee that the atomic kmap calls can succeed. Since
> > > > - * there's no in-kernel users of the kmap interfaces yet this isn't a
> > > > - * real problem.
> > > > + * This callback is optional.
> > > > *
> > > > * Returns:
> > > > *
> > > > @@ -216,9 +211,7 @@ struct dma_buf_ops {
> > > > *
> > > > * This is called from dma_buf_end_cpu_access() when the importer is
> > > > * done accessing the CPU. The exporter can use this to flush caches and
> > > > - * unpin any resources pinned in @begin_cpu_access.
> > > > - * The result of any dma_buf kmap calls after end_cpu_access is
> > > > - * undefined.
> > > > + * undo anything else done in @begin_cpu_access.
> > > > *
> > > > * This callback is optional.
> > > > *
>
--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
