Commit Graph

15739 Commits

Author SHA1 Message Date
Marco Paniconi
204809bfb3 Merge "vp9: Skip computation of best_sad for newmv, unless needed." 2016-06-01 17:37:29 +00:00
Yaowu Xu
6382727dc5 Fix UBSAN/IOC errors
1. test/dct16x16_test.cc
2. test/dct32x32_test.cc
3. test/fdct8x8_test.cc

BUG=webm:1225

Change-Id: I9c9315fbd65ddb3b44f688e01ba265fd22192198
2016-06-01 16:01:18 +00:00
Yaowu Xu
787b38ebb9 Fix VP8 encoder UBSAN/IOC errors
1. vp8/decoder/dboolhuff.c
2. vp8/decoder/dboolhuff.h
3. vp8/encoder/bitstream.c
4. vp8/encoder/boolhuff.h
5. vp8/encoder/rdopt.c

BUG=https://bugs.chromium.org/p/webm/issues/detail?id=1218

Change-Id: I5d315d63fd7aeaee6f3bd79178e593f3db38a6b1
2016-06-01 16:00:56 +00:00
James Zern
e5e2932cb3 ivfdec: tolerate invalid framerates
default invalid framerates to 30, quiets warnings in corrupt / fuzzed
files

Change-Id: Ib10d2b67df83cb6f9ed1cd6ef8e0e637aa7099ff
2016-05-31 17:37:59 -07:00
Yaowu Xu
46ff1072b3 variance_avx2.c: UBSAN/IOC fix
BUG=https://bugs.chromium.org/p/webm/issues/detail?id=1222

Change-Id: Ifb3bedf9b4e1b007b21aebaa4beb9ba50424efef
2016-05-31 16:44:35 -07:00
Marco
bedf1c3af6 vp9: Skip computation of best_sad for newmv, unless needed.
For non-rd pickmode:
best_pred_sad, computed for NEWMV-last, is only used for
skipping golden non-zero modes. Add condition to avoid this
computation if not used (i.e, if golden nonzero modes are not used).

And remove code for computing best_pred_sad for NEWMV-golden,
since that sad is not used.

No change in behavior; small speed gain (~1%) for svc encodes.

Change-Id: Ic2cbdef6c4e9a233a57c0db0eeac8ad5fcead366
2016-05-31 10:29:00 -07:00
Tom Finegan
f80d8011a0 Merge "vpx_ports/mem_ops.h: cast the lhs of bitwise shifts of 24." 2016-05-27 18:52:05 +00:00
James Zern
f6ac6cf5bd Merge "acm_random,Rand9Signed: correct cast" 2016-05-27 18:32:06 +00:00
Linfeng Zhang
2ab7b9a6c9 Merge "Upgrade fwht4x4_mmx() to fwht4x4_sse2() for vp9 and vp10." 2016-05-27 17:51:35 +00:00
James Zern
13d48c4267 acm_random,Rand9Signed: correct cast
convert the random value to int16 before subtracting 256 from it; quiets
a ubsan (sanitize=integer) warning

BUG=webm:1225

Change-Id: Ibc2c5a21f30e112bd6c180f7d6a033327c38d0df
2016-05-27 10:33:56 -07:00
Linfeng Zhang
af7fb17c09 Upgrade fwht4x4_mmx() to fwht4x4_sse2() for vp9 and vp10.
Function level timing test shows about 27% time saving on
a Xeon E5-2680 v2 desktop.

Rename vp9_dct_sse2.c to vp9_dct_intrin_sse2.c for vp9 and
rename dct_sse2.c to dct_intrin_sse2.c for vp10 to avoid
duplicate basenames.

Actually vp9_fwht4x4_mmx/sse2() and vp10_fwht4x4_mmx/sse2()
are identical. TODO: They should be unified later if there is
no intention to keep a duplicate.

Change-Id: I3e537b7bbd9ba417c606cd7c68c4dbbfa583f77d
2016-05-27 09:51:16 -07:00
Tom Finegan
f1de622617 vpx_ports/mem_ops.h: cast the lhs of bitwise shifts of 24.
C does not allow for shifting into the sign bit of a signed
integer, and the two instances here become signed ints via
promotion. Explcitly cast them to unsigned MEM_VALUE_T to
avoid the problem.

BUG=https://bugs.chromium.org/p/chromium/issues/detail?id=614648

Change-Id: I51165361a8c6cbb5c378cf7e4e0f4b80b3ad9a6e
2016-05-27 09:23:11 -07:00
Linfeng Zhang
0ba9b299e9 Merge "Upgrade vpx_lpf_{vertical,horizontal}_4 mmx to sse2" 2016-05-27 15:47:28 +00:00
Linfeng Zhang
4b5e462d08 Upgrade vpx_lpf_{vertical,horizontal}_4 mmx to sse2
Followed the code style of other lpf fuctions.
These 2 functions put 2 rows of data in a single xmm register,
so they have similar but not identical filter operations,
and cannot share the same macros.

Change-Id: I3bab55a5d1a1232926ac8fd1f03251acc38302bc
2016-05-26 14:55:18 -07:00
Yaowu Xu
ff6accf936 Merge "Convert to unsigned int before left shift" 2016-05-26 21:29:50 +00:00
Yaowu Xu
301e345273 Convert to unsigned int before left shift
This is to fix overflow when 128 is left shifted by 24.

Change-Id: Ibb5f6813536d985afa003a9848c0c3dd358955a7
2016-05-26 08:46:01 -07:00
Scott LaVarnway
9d24fe60f1 Merge "Code clean of sub_pixel_variance4xh -- 2" 2016-05-26 13:20:24 +00:00
Marco
75d551783d vp9: Add datarate test for 1 pass VBR mode.
Existing tests are only for CBR mode.

Change-Id: Ie3b2cd46236457748e2650901d1a347a730f38af
2016-05-25 14:20:30 -07:00
Alex Converse
19e0b406c9 Refactor probability savings search.
- Avoid excessive copying

- Don't both searching if no update can possibly offer savings

- Simplify the interface

- Remove the confusing vp9_cost_upd256 macro

Change-Id: Id9d9676a361fd1203b27e930cd29c23b2813ce59
2016-05-25 13:00:09 -07:00
Yaowu Xu
e5b7f14ea7 Merge "Fix comments in build_intra_predictors_high()" 2016-05-25 19:58:18 +00:00
Yaowu Xu
ba8651d474 Fix comments in build_intra_predictors_high()
1. Removed TODOs, no longer applicable to finalized vp9 profiles.
2. Added explanation on assumed values for highbitdepth profiles.

Change-Id: I59e0bebaaab900cc611ed284daa5fa0bdedb8097
2016-05-25 12:18:35 -07:00
Yaowu Xu
75b6cfe1c5 Prevent read to invalid RefBuffer
This commit adds check to validate RefBuffer before reading into the
data structure, to prevent invalid read.

BUG=https://bugs.chromium.org/p/chromium/issues/detail?id=614701

Change-Id: Ie111e95bd18e88fa19d8b25e097cdf52b7139cb6
2016-05-25 09:28:36 -07:00
James Zern
7acd0a59ca Merge "remove vp9_diamond_search_sad_avx.c" 2016-05-25 00:08:38 +00:00
Brion Vibber
35d7e17b03 Move git version extras out of iOS shared framework bundle version
Apple's version format specification is strictly checked on app
store submission, even for embedded frameworks:

http://apple.co/1WgelY1

    The build version number should be a string comprised of
    three non-negative, period-separated integers with the
    first integer being greater than zero. The string should
    only contain numeric (0-9) and period (.) characters.

So that's room for "1.5.0" but not for "1.5.0-906-g656f9c4".

The full version returned from 'version.sh --bare' is now
embedded under a 'VPXFullVersion' custom key in the Info.plist,
so it can still be extracted from the resulting framework.

Change-Id: If34a58d02e407379d1f1859fda533ef7f983170b
2016-05-24 13:08:25 -07:00
James Zern
be12fefa4b remove vp9_diamond_search_sad_avx.c
vp9_diamond_search_sad_avx was disabled in:
057c1c4 disable vp9_diamond_search_sad_avx

this removes a missing prototype warning as the prototype is no longer
included in vp9_rtcd.h. the file can be restored if someone gets around
to fixing the issue.

BUG=https://bugs.chromium.org/p/webm/issues/detail?id=1168

Change-Id: Ia9fda4b81c53dc5fba7c31d780d761f886940b52
2016-05-24 12:02:22 -07:00
Scott LaVarnway
a4f3751be5 Code clean of sub_pixel_variance4xh -- 2
Replace MMX with SSE2.

Change-Id: Id8482d2589131f9427e7f36bc64413f058caf31f
2016-05-24 04:44:05 -07:00
KO Myung-Hun
72e332f767 configure: Add -mstackrealign flags to CFLAGS on OS/2
Many codes require -mstackrealign flags. Although -mstackrealign has
been already added to CFLAGS of some modules, SIGSEGV occurs in other
modules than those modules.

The best way may be to find causes and to fix them. However, we
cannot know those causes until SIGSEGV occur really. In addition, if
SIGSEGV occurs in other programs, it will be fatal.

So adding -mstackrealign flags to CFLAGS unconditionally is
reasonable.

Change-Id: I999ef597a6afe97f5e7cc7bffaa866537c3eedd2
2016-05-22 18:11:59 +09:00
KO Myung-Hun
14e8adea3c vpx: Add OS/2-specific threading codes
With correction of a type of a thread function for new threading
codes.

Change-Id: Ic6dc9f530698800d1cfe2da327848e8f8b62e31f
2016-05-22 18:11:50 +09:00
jackychen
6f397b8a5b vp9: Remove a redundent condition in sub-pixel filter choosing.
Change-Id: I5cbb0f452ec9622437482b3a9496ead1253acfe0
2016-05-20 14:38:45 -07:00
James Zern
4f774ac50e Merge "vp8/error_concealment: remove shift of negative value" 2016-05-20 19:15:53 +00:00
James Zern
e4bdbd3c0b Merge "Revert "Code clean of sub_pixel_variance4xh"" 2016-05-20 19:11:06 +00:00
James Zern
3fb55d24e8 Revert "Code clean of sub_pixel_variance4xh"
This reverts commit 2468163e07.

causes valgrind errors for overread of buffer in SubpelVarianceTest

Change-Id: I448e52c76f815ac199305b71f7d169f2bc167679
2016-05-19 23:37:27 -07:00
James Zern
84e3639454 Revert "Extend the external fb interface to allocate individual planes."
This reverts commit 6dd7f2b50a.

conversion warnings, crashes in 32-bit builds

Change-Id: I529ead34cd93c862dd07c9a29d8542dda2fc20ea
2016-05-19 23:33:51 -07:00
James Zern
971c5a16a9 vp8/error_concealment: remove shift of negative value
fixes:
shifting a negative signed value is undefined [-Wshift-negative-value]

Change-Id: I3a55f2dac7c51c0e264d40081ffce98e2abacb89
2016-05-19 19:19:19 -07:00
Jacky Chen
fa5d54f937 Merge "vp9: Refactor some denoiser logic in vp9_pick_inter_mode." 2016-05-19 23:05:32 +00:00
JackyChen
af0a9fcf70 vp9: Refactor some denoiser logic in vp9_pick_inter_mode.
Move the logic for rechecking zeromv on denoised block out to simplify
the function. To simplify the param passing, add a new structure
VP9_PICKMODE_CTX_DEN which is only used when denoiser is enabled.

Change-Id: Iaa9b4396dfcb8147236c02d4a1868a09103a4476
2016-05-19 15:28:25 -07:00
Daniele Castagna
04fdbdc5ca Merge "Extend the external fb interface to allocate individual planes." 2016-05-19 18:01:59 +00:00
Yaowu Xu
d1f0f4cc63 Merge "Clarify integer value ranges" 2016-05-18 23:55:05 +00:00
Yaowu Xu
489f8b2d88 Merge "Prevent invalid read" 2016-05-18 23:54:56 +00:00
James Zern
146ccd304f Merge "Code clean of sub_pixel_variance4xh" 2016-05-18 23:18:35 +00:00
Alex Converse
284be1c9e0 Merge "Move, rename, and inline high_inter_predictor." 2016-05-18 21:20:04 +00:00
Daniele Castagna
6dd7f2b50a Extend the external fb interface to allocate individual planes.
Change-Id: I73e1b9ea6f4c76ae539e2b3292ee4c751d9c7de4
2016-05-18 16:20:18 -04:00
Johann Koenig
36b610d8c1 Merge "neon hadamard 8x8" 2016-05-18 20:11:16 +00:00
Yaowu Xu
a564b18d7f Clarify integer value ranges
This commit clarifies integer value range for vairables used in
several variance functions, also change to use proper type
conversion to reflect the value ranges.

Change-Id: Ic3234b83a912ce1ad12d1b254f3378763e15cc5c
2016-05-18 10:25:12 -07:00
Alex Converse
a5191f3e60 Move, rename, and inline high_inter_predictor.
The inlining mirrors what was done with the low bit depth
inter_predictor. And the new highbd_inter_predictor name is more
consistent with other high bit depth functions.

Change-Id: I96437f745759aeec6260c6e39a974bf36f1c211c
2016-05-18 09:39:49 -07:00
Yaowu Xu
4f0e4d6cef Prevent invalid read
This commit adds a check before reading into RefBuffer to prevent OOB
read.

BUG=https://bugs.chromium.org/p/chromium/issues/detail?id=612023

Change-Id: I5b02951932e7f457cfbe6b2e650790496b8577ae
2016-05-18 07:40:49 -07:00
Scott LaVarnway
8b8e296019 Merge "VP9: _get_pred_context_switchable_interp()" 2016-05-18 12:42:54 +00:00
Scott LaVarnway
2468163e07 Code clean of sub_pixel_variance4xh
Replace MMX with SSE2.

Change-Id: Ia8fcba755952804e347d7d7736f57d1f90c988a0
2016-05-18 04:24:41 -07:00
Scott LaVarnway
3036fd761a VP9: _get_pred_context_switchable_interp()
Remove unnecessary checks.

Change-Id: Ic7bce8277ac5f4ae88d4ab7d0ae3ab110b2f225b
2016-05-17 15:26:12 -07:00
Yaowu Xu
2240d83d78 Promote to uint32_t before left shift
This commit change to promote uint8_t explicitly to uint32_t before
left shift operation.

BUG=https://bugs.chromium.org/p/chromium/issues/detail?id=612021

Change-Id: Id7059154efb5bdfa45889dabe72aaafd46d79f23
2016-05-17 11:39:57 -07:00