It can cause excessively long interrupt outages if moving a larger
buffer and disrupt audio where DMA is not at a higher interrupt priority
such as FIQ.
Some targets, like Gigabeat S, have very low audio interrupt latency
requirements and will even channel swap if they are missed. Pictureflow
will make the issue very obvious. Even then, moves could take
milliseconds or more depending on the buffer size which is far too long
for any target.
Change-Id: I8e7817213e901da67c36b7eb25d7cb1c1e3ba802
Reviewed-on: http://gerrit.rockbox.org/472
Reviewed-by: Michael Sevakis <jethead71@rockbox.org>
Tested-by: Michael Sevakis <jethead71@rockbox.org>
Letting go of the button before the host is detected allows storage mode
to be entered even though a button was down when plugging. Sometimes the
host would try several times and the button would have to be held down
for several seconds to avoid a storage connection.
The adjustment wasn't made when switching to setup request host
detection.
Change-Id: Iab3bced5bfb1478d4d9a7baab2a2a1144afaf437
Boost only for tests and unboost immediately afterwards. It was leaving
the CPU boosted after exiting.
Change-Id: I010c4b975bd27e8990dc057ddd244dab233411ca
Most SoCs are these days are fast enough for realtime BRR, gaussian
interpolation and echo processing.
Change-Id: I180ce8ad45242c67b5e573a406b9522098a3f12b
Affected BRR cached waveforms but not realtime BRR decode as far as
I could ascertain. BRR cached waves required loop points to be inside
the initial waveform but this change removes that restriction.
Change-Id: I0ef4db720e5c28bd7b2fb9ae255d27c0a7213f79
CPU optimization gets its own files in which to fill-in optimizable
routines.
Some pointless #if 0's for profiling need removal. Those macros are
empty if not profiling.
Force some functions that are undesirable to be force-inlined by the
compiler to be not inlined.
Change-Id: Ia7b7e45380d7efb20c9b1a4d52e05db3ef6bbaab
Use the tlsf malloc and friends instead of the silly
codec_malloc to get actually working free and saner
realloc that doesn't leak memory.
Makes files with moderately sized embedded AA play
on targets with large enough codec buffers and files
with too large AA are now skipped rather than crashing.
Fixes crash when playing example file in FS#12842.
Change-Id: I06562955c4d9a95bd90f55738214fba462092b71
Uses the Catmull-Rom case of Hermite cubic splines.
Vastly improves the quality and accuracy of audio resampling with a
rather minor additional overhead compared to the previous linear
implementation.
ARM and Coldfire assembly implementations included.
Change-Id: Ic45d84bc66c5b312ef373198297a952167a4be26
Reviewed-on: http://gerrit.rockbox.org/304
Reviewed-by: Michael Sevakis <jethead71@rockbox.org>
Tested-by: Michael Sevakis <jethead71@rockbox.org>
Split the ISR into two parts and alllow quick return from first half.
Introduces a uevent() API to have a callback happen in a specified
number of microseconds. Right now only one event is supported.
Change-Id: Ib1666165be2f6082e5275d64961f083cab104f9f
Similar as done in d2b8f91 change dependency generation to use one file per
source file. This removes the need to have external tools installed on Windows.
Previously Cygwin or msys tools needed to be in the system PATH which was
problematic at times.
This means that now building on Windows (using MinGW) doesn't require
additional tools anymore.
Change-Id: I4c0675e99c3cc3a729b91beefd58320db498ae0a
Automatically generate dependency files for all sources so changes in headers
are picked up automatically. Use one dependency file for each source file,
since that allows to create them without using external tools (except the
compiler of course).
Change-Id: I246c1ceb525692547af22a2e32c4bab0c11507e1
Similar as done for libtools.make ask the compiler for the binary it creates
instead of using uname to figure the target to avoid special handling when
cross compiling.
Change-Id: Icb4654616e4339bf9e6e3be5177b35e0bb313bcd
Windows mkdir doesn't know about the -p option and requires paths to use \ as
path separator. Try to detect when building on Windows and use the Windows
internal function instead of relying on a compatible mkdir in the path.
Change-Id: I47d47d45edeb38c672321f77d6e91268bf744dba
Create the object - source dependency internally and use a more generic rule
for compiling. Removes the need for explicit rules for source files located in
a different folder.
This adds the limitation for SOURCES and LIBSOURCES to only hold C files.
Change-Id: I56f6a4b1e7df36347cd2f54051e072251f456092
This speeds up partial updates quite a bit but what is more
important it opens up a way to efficiently implement
lcd_blit_yuv() using hw colorspace conversion.
Tested on rk27generic, hm60x v1 and v2 and on ma9.
Benchmark for hm60x v1 (by mortalis):
HEAD patched
1/1 141fps 138fps
1/4 315fps 395fps
Change-Id: I4cc115786c3139000fc14c49a7290e289cfd6c42
Make sure its value quotes are properly escaped. Otherwise make on Windows
causes problems. Also, expand CPPDEFINES only once.
Change-Id: I52e91f17626b2596f389811b22abc12d94f721d1
Setup AR before checking for ccache to avoid adding ccache to ar. Setup CC and
CFLAGS in the beginning as well. Quote ar and system speex options to fix
building with ccache enabled.
Change-Id: I21714f0849e990abeac72dcda52c81474e090934
Conversion from centibels to register values was left out of
audiohw_set_lineout_volume().
Credit goes to Dave Marsh for providing the fix.
Just another little oversight from sound/audiohw rework.
Change-Id: I7f8b1c5fb575ac95b5d595c2f4824fbb5ebbd3c8
We need this only for cross compiling. Trying to figure the compiler prefix
fails if the compiler isn't gcc (i.e. clang), and we only support gcc when
cross compiling.
Change-Id: I822e1b47ec7e9d7e9fdeae8ccf3b8d417a472ef9
DSP_SWITCH_FREQUENCY has been deprecated and the same enumerated value
as DSP_SET_FREQUENCY since major DSP revisions were committed. This
task should have been performed much earlier but, oh well, do it now.
Change-Id: I3f30d651b894136a07c7e17f78fc16a7d98631ff
Just one more SNAFU after redoing sound.c. Software bass/treble controls
in the DSP expect .1 dB gain increments but the gains were being set as
1 dB increments because that's what the AUDIOHW_SETTING specifies. Just
x10 the gains given to audiohw_set_bass/treble.
Change-Id: Id5296f93908ec0036a5605d3a60a2cb5eec91bb5
95e23de made -Wl,--gc-sections appear after -T by combining it with the global
linker options. Moving the global linker options in front of -T seems to fix
the crash observed on mini2g after that change -- as it seems the linker
behaves differently if --gc-sections is passed after -T.
Fixes FS#12857.
Change-Id: Id64dc6b91726760d75a1a0cd85bb7c06208d2f45
Previously several additional rules were added by qmake to build the required
libraries. This has a couple of problems on Windows, requiring the use of a
msys shell.
Additional variables are passed as environment variables. Windows doesn't
understand this. Passing the variables via make command line argument works but
make won't be able to change those variables anymore, breaking things. Use
intermediate variables passed via make command line that are then exported by
the called Makefile. This doesn't work in the generated Makefile since the
export lines must not be part of a rule.
Also cleans up the project file a bit.
Change-Id: Iaffb059ce8bafe17a35bd1e0a74ae7acd966c138
Similar as done in ea0bfe7 ask the compiler for the binary it creates instead
of using uname to figure the target.
Simplify source file preprocessing to avoid using grep to improve compatibility
with Windows.
Change-Id: I39ecfe690c9bb70b07b0af5a7804d7e5ab512d71
While it made the mini2g not crash during startup anymore further tests showed
that other mini2g devices still exhibit the crash, or end up with a "No
partition found" error; furthermore the device tested first still crashes on
USB disconnect. Therefore the change doesn't really help with the problem, and
at the expense of increasing binary size for all other targets there is no
point in keeping it for now.
This reverts commit 850491a043.
Input type can only change once per call because the DSP parameters
are only copied at the start and input is always taken from the src
buffer which means sample input format switching can be once per call
instead of once per loop.
Change-Id: Ifa3521753428fb0e6997e4934f24a3b915628cc7
Some things can just be a bit simpler in handling the list of stages
and some things, especially format change handling, can be simplified
for each stage implementation. Format changes are sent through the
configure() callback.
Hide some internal details and variables from processing stages and
let the core deal with it.
Do some miscellaneous cleanup and keep things a bit better factored.
Change-Id: I19dd8ce1d0b792ba914d426013088a49a52ecb7e
