Revision as of 04:24, 24 April 2016

Request

Index Word	Description
0	Header code [0x00080142]
1	Virtual addr for the input buffer.
2	Physical addr, normally the above vaddr converted to physaddr.
3	Buffer byte-size.
4	Sequence-number / ID. The range normally used for this is 0x0..0x12. This value starts at 0x0, after value 0x12 is used the value that would be used for the next one is reset to zero.
5	Some sort of bool flag. SKATER is only known to use 0x0 here. Patching code so that the value used by MVD here is always 1, breaks video playback under SKATER(or at least in some cases). When non-zero this command will start returning 0x17007 for a lot of frames, this can have a side-affect(?) of video-playback speed-up if the homebrew app doesn't handle the return value properly.
6	Value 0x0
7	KProcess handle

Index Word	Description
0	Header code
1	Resultcode
2	End-address of the processed NAL-unit(internal MVD heap vaddr).
3	End-address of the processed NAL-unit(physaddr following the input physaddr).
4	Total remaining unprocessed input data, if any. Buffer_end_pos=bufsize-<this value>.

This processes the specified NAL-unit. This seems to always start with the 3-byte NAL prefix(00 00 01). Following that is the actual NAL-unit.

This is only used for H.264, not color-conversion.

@@ Line 49: / Line 49: @@
 |-
 | 4
-| Total remaining unprocessed input data. Non-zero means another NAL-unit prefix was detected at buffer_offset=bufsize-<this value>, after the first NAL-unit was processed.
+| Total remaining unprocessed input data, if any. Buffer_end_pos=bufsize-<this value>.
 |}