dma-buf/sync_file: refactor fence storage in struct sync_file
authorGustavo Padovan <gustavo.padovan@collabora.co.uk>
Fri, 5 Aug 2016 13:39:35 +0000 (10:39 -0300)
committerSumit Semwal <sumit.semwal@linaro.org>
Thu, 11 Aug 2016 10:03:24 +0000 (15:33 +0530)
commita02b9dc90d844cc7df7b63264e7920cc425052d9
treec6e8cf1d7d2f7f6c0e3c205f4af5ae2880f924d6
parente1aaf311dbe82221910cc0e0809c988de210cc3c
dma-buf/sync_file: refactor fence storage in struct sync_file

Create sync_file->fence to abstract the type of fence we are using for
each sync_file. If only one fence is present we use a normal struct fence
but if there is more fences to be added to the sync_file a fence_array
is created.

This change cleans up sync_file a bit. We don't need to have sync_file_cb
array anymore. Instead, as we always have  one fence, only one fence
callback is registered per sync_file.

v2: Comments from Chris Wilson and Christian König
- Not using fence_ops anymore
- fence_is_array() was created to differentiate fence from fence_array
- fence_array_teardown() is now exported and used under fence_is_array()
- struct sync_file lost num_fences member

v3: Comments from Chris Wilson and Christian König
- struct sync_file lost status member in favor of fence_is_signaled()
- drop use of fence_array_teardown()
- use sizeof(*fence) to allocate only an array on fence pointers

v4: Comments from Chris Wilson
- use sizeof(*fence) to reallocate array
- fix typo in comments
- protect num_fences sum against overflows
- use array->base instead of casting the to struct fence

v5: fixes checkpatch warnings

v6: fix case where all fences are signaled.

Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Christian König <christian.koenig@amd.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sumit Semwal <sumit.semwal@linaro.org>
drivers/dma-buf/sync_file.c
drivers/staging/android/sync_debug.c
include/linux/sync_file.h