This simplifies testing arbitrary code fragments within a function
body.
Signed-off-by: Mans Rullgard <mans@mansr.com>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
Instead of inlining everything into ff_h264_hl_decode_mb(), use
explicit templating to create versions of the called functions
with constant parameters filled in. This greatly speeds up
compilation of h264.c and reduces the code size without any
measurable impact on performance.
Compilation time for h264.c on an i7 goes from 30s to 5.5s.
Code size is reduced by 430kB.
Signed-off-by: Mans Rullgard <mans@mansr.com>
This tool uses lavfi internal symbols not accessible in shared
libraries. TESTPROGS are linked statically to allow them use of
library internals not normally exported.
Signed-off-by: Mans Rullgard <mans@mansr.com>
The buffers are only allocated once, although it can happen from
any of a few different places, so there is no need to use realloc.
Using av_malloc() ensures they are aligned suitably for SIMD
optimisations.
Signed-off-by: Mans Rullgard <mans@mansr.com>
Currently, --enable-small turns av_always_inline into plain inline,
which is more or less ignored by the compiler. While the intent of
this is probably to reduce code size by avoiding some inlining, it
has more far-reaching effects.
We use av_always_inline in two situations:
1. The body of a function is smaller than the call overhead.
Instances of these are abundant in libavutil, the bswap.h
functions being good examples.
2. The function is a template relying on constant propagation
through inlined calls for sane code generation. These are
often found in motion compensation code.
Both of these types of functions should be inlined even if targeting
small code size.
Although GCC has heuristics for detecting the first of these types,
it is not always reliable, especially when the function uses inline
assembler, which is often the reason for having those functions in
the first place, so making it explicit is generally a good idea.
The size increase from inlining template-type functions is usually
much smaller than it seems due to different branches being mutually
exclusive between the different invocations. The dead branches can,
however, only be removed after inlining and constant propagation have
been performed, which means the initial cost estimate for inlining
these is much higher than is actually the case, resulting in GCC
often making bad choices if left to its own devices.
Furthermore, the GCC inliner limits how much it allows a function to
grow due to automatic inlining of calls, and this appears to not take
call overhead into account. When nested inlining is used, the limit
may be hit before the innermost level is reached. In some cases, this
has prevented inlining of type 1 functions as defined above, resulting
in significant performance loss.
Signed-off-by: Mans Rullgard <mans@mansr.com>
This adds macros for accessing the EFLAGS register and uses
these instead of coding the entire check in inline asm.
Signed-off-by: Mans Rullgard <mans@mansr.com>
This creates proper position independent code when accessing
data symbols if CONFIG_PIC is set.
References to external symbols should now use the movrelx macro.
Some additional code changes are required since this macro may
need a register to hold the GOT pointer.
Signed-off-by: Mans Rullgard <mans@mansr.com>
The problem is that the ssse3 psign instruction does the wrong
thing here. Commit ea60dfe incorrectly removed a macro emulating
this instruction for pre-ssse3 code. However, the emulation is
incorrect, and the code relies on the behaviour of the macro.
Specifically, the psign sets destination elements to zero where
the corresponding source element is zero, whereas the emulation
only negates destination elements where the source is negative.
Furthermore, the PSIGNW_MMX macro in x86util.asm is totally bogus,
which is why the original VC-1 code had an additional right shift
when using it. Since the psign instruction cannot be used here,
skip all the macro hell and use the working instruction sequence
directly.
None of this was noticed due a stray return statement in
ff_vc1dsp_init_mmx() which meant that only the mmx version of the
loop filter was ever used (before being removed in ea60dfe).
Signed-off-by: Mans Rullgard <mans@mansr.com>
This adds a fallback for cbrtf() using powf(x, 1/3). Since
powf() with a non-integer exponent requires a non-negative
base, special handling of negative inputs is needed.
Signed-off-by: Mans Rullgard <mans@mansr.com>
This adds whitespace around operators, aligns line continuation
backslashes, and breaks long lines. Also fixes an ifdef halfway
through a statement. The one line of duplication this saved is
not worth the ugliness.
Signed-off-by: Mans Rullgard <mans@mansr.com>
This removes a dependency on implementation details from generic
code and allows easy addition of the equivalent optimisation for
other architectures than x86.
Signed-off-by: Mans Rullgard <mans@mansr.com>
This function implements a delay using the first available
of the following functions:
- nanosleep()
- usleep()
- Sleep() (Windows)
The conditional #includes in time.c are simplified by including
unistd.h and windows.h whenever they are available rather than
having these lines triggered by specific functions.
Signed-off-by: Mans Rullgard <mans@mansr.com>
The unistd.h header is only needed for the close() declaration.
If this header is not available, the close() declaration may be
provided by another header, e.g. io.h.
Signed-off-by: Mans Rullgard <mans@mansr.com>
The only symbol this file uses from unistd.h is isatty(). By
including the header only when this function is used, the file
can be built on systems without unistd.h (which presumably also
lack isatty).
Signed-off-by: Mans Rullgard <mans@mansr.com>
Override the frame size from the SPS with AVCodecContext values
if the latter specify a size smaller by less than one macroblock.
This is required for correct cropping of MOV files from Canon cameras.
Signed-off-by: Mans Rullgard <mans@mansr.com>
This ensures that these functions are inlined into the per-position
entry points, allowing constant propagation as needed for proper
optimisation.
18% faster VC1 decoding on Cortex-A9.
Signed-off-by: Mans Rullgard <mans@mansr.com>
A number of systems do not implement freopen() with a NULL filename
correctly. This changes these programs to output individual images
if opening a named output argument as a file fails, in this case
assuming it is a directory.
Signed-off-by: Mans Rullgard <mans@mansr.com>
On systems where the 'md5' command is used, there is a conflict
with the md5() shell function in fate-run.sh. Using the 'command'
keyword bypasses the shell function for correct behaviour.
Signed-off-by: Mans Rullgard <mans@mansr.com>
GNU make 3.81 applies pattern rules in declaration order rather than
by stem length as in 3.82. This moves the more generic patterns above
the more specific ones such that they work with either make version.
Some of the vsynth patterns are also simplified a little.
Signed-off-by: Mans Rullgard <mans@mansr.com>
The encode/decode tests should all depend on avconv. Since
avconv requires libavfilter, there is no need to enable those
tests selectively.
Signed-off-by: Mans Rullgard <mans@mansr.com>
On some systems, the wc command prints spaces before the first
number causing mismatches with the test references. Using the
output of wc as arguments to echo removes any extra whitespace.
Signed-off-by: Mans Rullgard <mans@mansr.com>
diff -w is not a standard option. This fixes the reference files
to match what the tests actually output and switches to using the
standard diff -b which is sufficient to handle different line ending
styles.
Signed-off-by: Mans Rullgard <mans@mansr.com>
The codec (adpcm-ima-ws) is tested elsewhere. Using framecrc output
provides more information than a single md5 if something goes wrong.
Signed-off-by: Mans Rullgard <mans@mansr.com>
Mans RullgardandLuca Barbato
