Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit 32952371 authored by Kyle Yan's avatar Kyle Yan Committed by Gerrit - the friendly Code Review server
Browse files

Merge "msm: vidc: Update documentation for VIDC DTSI" into msm-4.8

parents 78be33e7 3fddf480
Loading
Loading
Loading
Loading
+0 −48
Original line number Diff line number Diff line
@@ -10,7 +10,6 @@ Required properties:
  of macroblocks per second. The load is a reflection of hardware capability
  rather than a performance guarantee. Performance is guaranteed only up to
  advertised capability of the chipset.
- qcom,firmware-name : firmware file name to be downloaded.

Optional properties:
- reg : offset and length of the register set for the device.
@@ -24,22 +23,10 @@ Optional properties:
    supports mvc decoder = 0x00000003
    supports h264 encoder = 0x00000004
    supports h264 decoder = 0x0000000c
    supports h263 encoder = 0x00000010
    supports h263 decoder = 0x00000030
    supports mpeg1 encoder = 0x00000040
    supports mpeg1 decoder = 0x000000c0
    supports mpeg2 encoder = 0x00000100
    supports mpeg2 decoder = 0x00000300
    supports mpeg4 encoder = 0x00000400
    supports mpeg4 decoder = 0x00000c00
    supports divx_311 encoder = 0x00001000
    supports divx_311 decoder = 0x00003000
    supports divx encoder = 0x00004000
    supports divx decoder = 0x0000c000
    supports vc1 encoder = 0x00010000
    supports vc1 decoder = 0x00030000
    supports spark encoder = 0x00040000
    supports spark decoder = 0x000c0000
    supports vp6 encoder = 0x00100000
    supports vp6 decoder = 0x00300000
    supports vp7 encoder = 0x00400000
@@ -48,25 +35,12 @@ Optional properties:
    supports vp8 decoder = 0x03000000
    supports hevc encoder = 0x04000000
    supports hevc decoder = 0x0c000000
    supports hevc_hybrid encoder = 0x10000000
    supports hevc_hybrid decoder = 0x30000000
- qcom,dcvs-tbl : specifies the parameter to configure DCVS algorithm. Video
  instance load (in mbs/sec) and corresponding low and high threshold DCVS
  load for supported codec formats.
- qcom,dcvs-limit : specifies the minimum specifications required to kick in
  DCVS algorithm. Min MBs per frame and the fps for encoder and decoder to
  kick in DCVS.
- qcom,imem-ab-tbl : video core frequency in Hz and corresponding imem ab value
  in kbps. The imem ab value corresponds to the clock frequency at which imem
  should operate for a particular video core frequency.
- qcom,reg-presets : list of offset-value pairs for registers to be written.
  The offsets are from the base offset specified in 'reg'. This is mainly
  used for QoS, VBIF, etc. presets for video.
- qcom,qdss-presets : list of physical address and memory allocation size pairs.
  when fw_debug_mode is set as HFI_DEBUG_MODE_QDSS, all firmware messages will be
  written to QDSS memory.
- qcom,imem-size: imem size required by hardware for optimum performance.
  If imem is not present then there is no need to specify this property.
- *-supply: A phandle pointing to the appropriate regulator. Number of
  regulators vary across targets.
- clock-names: an array of clocks that the driver is supposed to be
@@ -90,8 +64,6 @@ Optional properties:
       represented in Q16 format.
- qcom,sw-power-collapse = A bool indicating if video hardware core can be
  power collapsed in idle state.
- qcom,slave-side-cp = A bool indicating the content protection mode for an
  ongoing video session. It is true for targets where the mode is slave side.
- qcom,never-unload-fw = A bool indicating if video firmware should be not be
  unloaded after all active sessions have closed.  Once a new session starts up
  after this, the firmware will be ready to go.  This should be set on platforms
@@ -99,8 +71,6 @@ Optional properties:
- qcom,use-non-secure-pil = A bool indicating which type of pil to use to load
  the fw.
- qcom,fw-bias = The address at which venus fw is loaded (manually).
- qcom,enable-idle-indicator = A bool to enable video hardware to generate
  idle message when it is in idle state.
- qcom,enable-thermal-mitigation = A bool to enable thermal mitigation when
  thermal run away occurs.
- qcom,hfi-version = The hfi packetization version supported by venus firmware.
@@ -204,22 +174,6 @@ Example:
			<72000 133330000 0x0c000000>, /* HEVC decoder VGA 60fps   */
			<36000 133330000 0x0c000000>, /* HEVC VGA 30 fps  */
			<36000 133330000 0x01000414>; /* Legacy encoder VGA 30 fps   */
		qcom,dcvs-tbl =
			<972000 972000 19944000 0xffffffff>, /* Legacy decoder UHD 30fps */
			<489600 489600   972000 0xffffffff>, /* Legacy decoder 1080p 60fps */
			<244800 244800   489600 0xffffffff>, /* Legacy decoder 1080p 30fps */
			<829440 489600   972000 0x55555555>; /* Legacy encoder DCI 24fps
		qcom,dcvs-limit =
			<32400 30>, /* Encoder UHD */
			<14400 30>; /* Decoder WQHD */
		qcom,imem-ab-tbl =
			<75000000  1500000>, /* imem clk @ 75 Mhz  */
			<150000000 1500000>, /* imem clk @ 75 Mhz  */
			<355333333 2500000>, /* imem clk @ 170 Mhz */
			<533000000 6000000>; /* imem clk @ 320 Mhz */

		qcom,imem-size = <4096>;
		qcom,firmware-name = "venus";
		qcom,hfi-version = "3xx";
		qcom,reg-presets = <0x80004 0x1>,
			<0x80178 0x00001FFF>;
@@ -230,12 +184,10 @@ Example:
		clock-names = "foo_clk", "bar_clk", "baz_clk";
		qcom,clock-configs = <0x3 0x1 0x0>;
		qcom,sw-power-collapse;
		qcom,enable-idle-indicator;
		qcom,buffer-type-tz-usage-table = <0x1 0x1>,
						<0x1fe 0x2>;
		qcom,enable-thermal-mitigation;
		qcom,use-non-secure-pil;
		qcom,slave-side-cp;
		qcom,use_dynamic_bw_update;
		qcom,fw-bias = <0xe000000>;
		msm_vidc_cb1: msm_vidc_cb1 {