The lossless encoding speed-ups are:
- up to 1% with default parameters
- up to 4% in cruncher mode: -q 100 -m 6
Change-Id: Id92d4bad0b0a2c28c8aa9ff5280eea5717017f30
This reverts commit ee26766a89.
This change also reverts the parent.
Revert "Increase the transform bits if possible."
This reverts commit 7ec51c5916.
These changes result in non-lossless encodes.
Bug: oss-fuzz:69231, oss-fuzz:69109, oss-fuzz:69208
Bug: b:341475869, b:342743143
Change-Id: Ia28f558992e0aa6f024af1ff66da52e0a5e26fa3
A 3 by 1 image would not have its 1st and 3rd lines compared at
the second iteration.
BUG=oss-fuzz:69208
Change-Id: I9213e73995d31907f358310a0b7d5ebb21c1f8b2
This brings minor size improvements because repetitive values in
the transform images are easily explainable through LZ77. Still,
it makes an upcoming pull request a bit more stable.
This is 971a03d820 with a fix to
not forget to analyze the end of the line.
A const has also been added to match VP8LColorSpaceTransform's
signature.
Change-Id: Iae03216fef298c7abc96a766f8a799552b05ade5
This reverts commit 971a03d820.
Reason for revert:
This creates non-lossless encodes.
Original change's description:
> Increase the transform bits if possible.
>
> This brings minor size improvements because repetitive values in
> the transform images are easily explainable through LZ77. Still,
> it makes an upcoming pull request a bit more stable.
>
> Change-Id: I1c7135675cb59b5e27ca960738d74465f10d0deb
Bug: oss-fuzz:69109, b:341475869
Change-Id: I3b9f21a5498735eb3681e62fb35bf9f9c2ed4f9f
This brings minor size improvements because repetitive values in
the transform images are easily explainable through LZ77. Still,
it makes an upcoming pull request a bit more stable.
Change-Id: I1c7135675cb59b5e27ca960738d74465f10d0deb
Replace the 2d histograms with uint32_t 1d versions (to avoid
pointer casting and to use the optimized VP8LAddVectorEq).
Change-Id: I90b0fe98390b49e3fd03e3484289571cf7ae6eca
fixes warnings of the form:
/src/dsp/upsampling_sse41.c:170:1: runtime error: implicit conversion
from type 'int' of value -16 (32-bit, signed) to type 'uintptr_t' (aka
'unsigned long') changed the value to 18446744073709551600 (64-bit,
unsigned)
this is the same change as was done previously in upsampling_sse2.c:
2ee786c7 upsampling_sse2.c: clear int sanitizer warnings
Change-Id: I36064d837ad1a7a118918c16a5551fc732dec2ff
assert idx != IDX_LAST_CHUNK to quiet a static analysis warning
(reported by Coverity):
overrun-local: Overrunning array kChunks of 11 12-byte elements at
element index 11 (byte offset 143) using index idx (which evaluates
to 11).
Change-Id: I85ae159e338d5670c007f76277ffe94e7a1dcafc
assert idx != IDX_LAST_CHUNK to quiet a static analysis warning
(reported by Coverity):
Indexing array kChunks of size 11 with idx.
Change-Id: I5e54b5c09d23c26abbd1a7919b6f828b35fe4f75
With older toolchains, at the least, visibility attributes don't
work and all symbols are exported along with lots of unsupported
warnings.
Change-Id: I6ffe220b7589a6bda00f42c17ead35592de7cc5a
The data mostly must persist until a call to Assemble() is made. If the
data is discarded and subsequent calls are made the behavior is
undefined.
Change-Id: I64b39034ae4663e9ca25b85040b6068d09dfac04
This is available with clang. Clears warnings of the form:
warning: empty expression statement has no effect; remove unnecessary
';' to silence this warning [-Wextra-semi-stmt]
As a side-effect it also clear a few -Wpedantic warnings with gcc:
warning: ISO C does not allow extra ';' outside of a function
[-Wpedantic]
Change-Id: I9295c767aad475c68b1fbbdff855b0d6650a25f5
Initially added to workaround gcc implementation issues that clang
does not have. (gcc hardcodes rbx as the PIC register, clang uses a
virtual register)
Change-Id: I1a3277abf02b1ff437b4aea4d28f4cb1c0176b80
When WebPSafeMalloc fails on VP8LHuffmanTablesAllocate,
next is not initialized to NULL.
VP8LHuffmanTablesDeallocate uses next to know the following nodes.
A patch fixes this issue.
Change-Id: I144ae84cd97e5bca227018ef1afa95361267902c
The first condition is only necessary if we have not read enough
(enough being defined by src_last, not src_end which is the end
of the image).
The second condition now fits the comment below: "if not
incremental, and we are past the end of buffer".
BUG=oss-fuzz:62136
Change-Id: I0700f67c62db8e1c02c2e429a069a71e606a5e4f
First, BuildHuffmanTable is called to check if the data is valid.
If it is and the table is not big enough, more memory is allocated.
This will make sure that valid (but unoptimized because of unbalanced
codes) streams are still decodable.
Bug: chromium:1479274
Change-Id: I31c36dbf3aa78d35ecf38706b50464fd3d375741
(cherry picked from commit 902bc91903)
First, BuildHuffmanTable is called to check if the data is valid.
If it is and the table is not big enough, more memory is allocated.
This will make sure that valid (but unoptimized because of unbalanced
codes) streams are still decodable.
Bug: chromium:1479274
Change-Id: I31c36dbf3aa78d35ecf38706b50464fd3d375741
According to https://gcc.gnu.org/onlinedocs/gcc/extensions-to-the-c-language-family/how-to-use-inline-assembly-language-in-c-code.html
For the C language, the asm keyword is a GNU extension. When
writing C code that can be compiled with -ansi and the -std options
that select C dialects without GNU extensions, use __asm__ instead
of asm (see Alternate Keywords). For the C++ language, asm is a
standard keyword, but __asm__ can be used for code compiled with
-fno-asm.
Change-Id: I4af950e67c857c890290c1e3d9cc886da0748784
This makes it easier to read and copy/paste, though this is just rough
pseudo-code. It also matches the formatting of the advanced decoder
example.
Change-Id: I6a4d0a04a12bfc1f1956ac965ff4869cefc4f3df
4 bits are read ([0, 15]) with 4 added to the result. The check that the
result is not > NUM_CODE_LENGTH_CODES (19) is redundant. This makes the
check an assert for documentation purposes.
webp-lossless-bitstream-spec.txt is updated accordingly.
Bug: webp:611
Change-Id: Ie461017dc9de3da4cfe6856543d3dcf58f53a180
ExpandColorMap() can fail due to a memory allocation failure.
Found by Nallocfuzz (https://github.com/catenacyber/nallocfuzz).
Change-Id: I868c5db2fee18e876f4aa3b63e744d7297d1a91b
in AllocateAlphaPlane() and VP8DecompressAlphaRows().
Found by Nallocfuzz (https://github.com/catenacyber/nallocfuzz).
Change-Id: I39eb305abfe9f4d7915cce5d91e72917f482bc20
This is unnecessary as green will have the largest value; the sizes are
currently: 256 + 24 + color_cache_size, 256, 40, in that order.
Change-Id: Ib9ce3292285a7b9c6089557b68a87184b68d17a5
Users of the encoder (including anim_encode.c) and areas of the encoder
itself rely on the status returned via WebPPicture.
Change-Id: Id786176b8ac3b2329d1e41b9dacbb8dcc5d822e4
VP8LBitWriterFinish() may cause the VP8LBitWriter's buffer to be grown.
If that allocation fails, VP8LBitWriterNumBytes() will return a size
larger than the current allocation resulting in a heap overwrite of the
missing bytes.
==13==ERROR: AddressSanitizer: heap-buffer-overflow on address
0x61900005b880 at pc 0x00000049ffc1 bp 0x7fff144f5b40 sp 0x7fff144f5310
READ of size 1028 at 0x61900005b880 thread T0
#0 0x49ffc0 in __asan_memcpy
#1 0x695861 in VP8BitWriterAppend src/utils/bit_writer_utils.c:186:3
#2 0x65acf9 in EncodeAlphaInternal src/enc/alpha_enc.c:169:14
Found by Nallocfuzz (https://github.com/catenacyber/nallocfuzz).
This is the same issue that was fixed in the non-alpha lossless path in:
d49cfbb3 vp8l_enc,WriteImage: add missing error check
Bug: chromium:1455619
Change-Id: I6bd10de213707d3d6b7ce3d0d2b3942af45d317f
VP8LBitWriterFinish() may cause the VP8LBitWriter's buffer to be grown.
If that allocation fails, VP8LBitWriterNumBytes() will return a size
larger than the current allocation resulting in a heap overwrite of the
missing bytes.
==3531848==ERROR: AddressSanitizer: heap-buffer-overflow on address
0x61d000000880 at pc 0x556eddfa1007 bp 0x7ffe434c7a90 sp 0x7ffe434c7260
READ of size 2052 at 0x61d000000880 thread T0
#0 0x556eddfa1006 in __asan_memcpy
#1 0x556eddfeeccf in WebPMemoryWrite src/enc/picture_enc.c:220:5
#2 0x556ede0f9f87 in WriteImage src/enc/vp8l_enc.c:1454:8
Found by Nallocfuzz (https://github.com/catenacyber/nallocfuzz).
Change-Id: Ib1c9454c2c51849b0ba58c5347e6bd5b02a12fbe
This prevents leaks should an allocation fail.
Spotted by Philippe Antoine (p dot antoine at catenacyber dot fr)
and https://github.com/catenacyber/nallocfuzz.
Change-Id: I07a8a475a282b18bf2f9015cc5b1120772babd48
The buffers are made mandatory to match WebPDecodeYUVInto(), though this
conflicts with WebPIDecGetYUVA().
spotted by Oliver Kunz (okunz at google dot com)
Change-Id: Ic4740c53b75da6b93d4f3462303fb9be0ebfbd48
The histograms count the occurrences of len/dist in entropy images.
Those (at most (1<<14) by (1<<14)) are sub-sampled by at least
MIN_HUFFMAN_BITS == 2, hence at most 24 bits in a histogram value.
At most, we multiply by 19 (because the longest histogram is of
size 40 and we do 40>>1, cf code) for the bit cost. So it all fits
in 32 bits.
Change-Id: Ife24b035f54794851ff31f2fac07901f724c6d7f
and define it to true for __aarch64__ and Win Arm64 + Visual Studio.
Microsoft's compiler (cl.exe) does not define __aarch64__, but relies on
_M_ARM64 & _M_ARM64EC
Bug: b/277254922
Change-Id: I20e4fa07a4031599db69e3d7ba9050345315ef51
This avoids defining a version in each translation unit when using
__declspec(dllexport) which causes failures due to multiply defined
symbols with clang-cl:
lld-link: error: duplicate symbol: VP8GetCPUInfo
>>> defined at CMakeFiles\webpdecode.dir\Debug\src\dec\alpha_dec.c.obj
>>> defined at CMakeFiles\webpdsp.dir\Debug\src\dsp\dec_sse41.c.obj
...
Bug: webp:607
Change-Id: I6cd1ee75b3db984aa513263a05516e867a64925d
this is always defined by default to 0 since:
v0.6.0-158-g663a6d9d unify the ALTERNATE_CODE flag usage
previously the !defined() check would cause a mismatch between C and
assembly.
Change-Id: Idca0b8e39ca90d63785fd4125aeb7af86c5aae61
This avoids a double free should the function fail prior to
VP8BitWriterInit() and a previous trial result's buffer carried over.
Previously in ApplyFiltersAndEncode() trial.bw (with a previous
iteration's buffer) would be freed, followed by best.bw pointing to the
same buffer.
Since:
187d379d add a fallback to ALPHA_NO_COMPRESSION
In addition, check the return value of VP8BitWriterInit() in this
function.
Bug: webp:603
Change-Id: Ic258381ee26c8c16bc211d157c8153831c8c6910
when threading is disabled; fixes:
src/enc/analysis_enc.c:429:15: warning: Value stored to 'split_row'
during its initialization is never read [deadcode.DeadStores]
const int split_row = (9 * last_row + 15) >> 4;
^~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~
Change-Id: I2b02173bb8c930ed54360cce5c7b88ecce098d83
* changes:
picture_csp_enc.c: remove SafeInitSharpYuv
sharpyuv: prefer webp/types.h
sharpyuv,SharpYuvInit: add mutex protection when available
sharyuv_{neon,sse2}.c: merge WEBP_USE_* sections
add a few missing <stddef.h> includes for NULL
sharpyuv.h: remove <inttypes.h>
thread safety was internalized in:
782ed48c sharpyuv,SharpYuvInit: add mutex protection when available
Change-Id: I2877e7bb78bf89ad7504c33d04b22952c0c0a968
move WEBP_EXTERN definition from Makefile.vc to webp/types.h using the
existing WEBP_DLL define to control its declaration to
__declspec(dllexport); the same is done for SHARPYUV_EXTERN
Change-Id: Iecea360305bdd19a878255470a00b9dcc344c730
the ternary used with alpha_offset was removed in:
3b07d327 Import,RGBA: fix for BigEndian import
use the ALPHA_OFFSET directly
Change-Id: Iee9b7c54f1498591e8c5834766dc4d34a321133d
