]> granicus.if.org Git - libvpx/commit
Convert [4][4] matrices to [16] arrays.
authorTimothy B. Terriberry <tterribe@xiph.org>
Fri, 22 Oct 2010 00:04:30 +0000 (17:04 -0700)
committerTimothy B. Terriberry <tterribe@xiph.org>
Fri, 22 Oct 2010 00:04:30 +0000 (17:04 -0700)
commit8f75ea6b5c0a72c233d1ff9a75e16e6fe43fda77
tree476d56cccf6f61b32f3e1db2d5cf7fddd309d625
parent45e64941778058312d72711dbfcf039bb4ba5171
Convert [4][4] matrices to [16] arrays.

Most of the code that actually uses these matrices indexes them as
 if they were a single contiguous array, and coverity produces
 reports about the resulting accesses that overflow the static
 bounds of the first row.
This is perfectly legal in C, but converting them to actual [16]
 arrays should eliminate the report, and removes a good deal of
 extraneous indexing and address operators from the code.

Change-Id: Ibda479e2232b3e51f9edf3b355b8640520fdbf23
13 files changed:
vp8/common/blockd.h
vp8/common/onyxc_int.h
vp8/decoder/arm/dequantize_arm.c
vp8/decoder/decodframe.c
vp8/decoder/dequantize.c
vp8/decoder/threading.c
vp8/encoder/arm/quantize_arm.c
vp8/encoder/block.h
vp8/encoder/encodeframe.c
vp8/encoder/encodemb.c
vp8/encoder/onyx_int.h
vp8/encoder/quantize.c
vp8/encoder/x86/x86_csystemdependent.c