Commit 3063bae6 authored by Sachin Shah's avatar Sachin Shah Committed by Stanimir Varbanov
Browse files

media: vidc: v4l2 based video driver for Qualcomm chipsets



Compatible for B-family onward targets only.

- fix stop_streaming method signature -
  stop_streaming now returns void. Update enc/dec definitions
  accordingly.

- make firmware-name optional - to maintain the backward
  compatibility with older targets.

- move resource initialization from _get_device() into _load_fw() -
  venus_hfi_get_device() is generally called during probe.
  However not all hardware resources are available at probe time
  (e.g. buses, which are represented as distinct devices that
  might get probed separately later on).  Hence push the resource
  initialization to a point where we know that the resources will
  be known and when we actually need it. Also, don't vote for bus
  or enable clocks in msm_comm_load_fw.  It's already taken care
  of in hfi->load_fw.  Besides, only the HFI layer is supposed to
  worry about hardware specific things.

- don't use resource managed devm_* calls in acquiring resources
  We no longer acquire resources in probe, hence we shouldn't be
  using devm_* variants of regulator_get or clock_get.  Also
  improve the error handling when init_clocks() or
  init_regulators() fail.

- misc changes to get Video driver compiling
  * Add MSM bus driver related header files.
  * Add OCMEM header file.
  * Add msm_vidc specific v4l2 extensions and
    other header files.
  * Remove ION from defconfig and feature out ION
    dependencies in video driver
  * Enable ARM SMMU config option.

- SMEM related changes to use DMA instead of ION
- add support for old SMMU API and switch to DMA buf for internal buffers
  * Switch to DMA contig buffers for internal and scratch buffers.
  * Move to vb2_dma_contig_memops for all v4l2 memops.
  * Comment out clock/regulators/CP related stuff.

Signed-off-by: default avatarSachin Shah <sachins@codeaurora.org>
parent b54fd555
Loading
Loading
Loading
Loading
0% Loading or .
You are about to add 0 people to the discussion. Proceed with caution.
Please to comment