This patch redoes the filesystem code from the FAT driver up to the
clipboard code in onplay.c.
Not every aspect of this is finished therefore it is still "WIP". I
don't wish to do too much at once (haha!). What is left to do is get
dircache back in the sim and find an implementation for the dircache
indicies in the tagcache and playlist code or do something else that
has the same benefit. Leaving these out for now does not make anything
unusable. All the basics are done.
Phone app code should probably get vetted (and app path handling
just plain rewritten as environment expansions); the SDL app and
Android run well.
Main things addressed:
1) Thread safety: There is none right now in the trunk code. Most of
what currently works is luck when multiple threads are involved or
multiple descriptors to the same file are open.
2) POSIX compliance: Many of the functions behave nothing like their
counterparts on a host system. This leads to inconsistent code or very
different behavior from native to hosted. One huge offender was
rename(). Going point by point would fill a book.
3) Actual running RAM usage: Many targets will use less RAM and less
stack space (some more RAM because I upped the number of cache buffers
for large memory). There's very little memory lying fallow in rarely-used
areas (see 'Key core changes' below). Also, all targets may open the same
number of directory streams whereas before those with less than 8MB RAM
were limited to 8, not 12 implying those targets will save slightly
less.
4) Performance: The test_disk plugin shows markedly improved performance,
particularly in the area of (uncached) directory scanning, due partly to
more optimal directory reading and to a better sector cache algorithm.
Uncached times tend to be better while there is a bit of a slowdown in
dircache due to it being a bit heavier of an implementation. It's not
noticeable by a human as far as I can say.
Key core changes:
1) Files and directories share core code and data structures.
2) The filesystem code knows which descriptors refer to same file.
This ensures that changes from one stream are appropriately reflected
in every open descriptor for that file (fileobj_mgr.c).
3) File and directory cache buffers are borrowed from the main sector
cache. This means that when they are not in use by a file, they are not
wasted, but used for the cache. Most of the time, only a few of them
are needed. It also means that adding more file and directory handles
is less expensive. All one must do in ensure a large enough cache to
borrow from.
4) Relative path components are supported and the namespace is unified.
It does not support full relative paths to an implied current directory;
what is does support is use of "." and "..". Adding the former would
not be very difficult. The namespace is unified in the sense that
volumes may be specified several times along with relative parts, e.g.:
"/<0>/foo/../../<1>/bar" :<=> "/<1>/bar".
5) Stack usage is down due to sharing of data, static allocation and
less duplication of strings on the stack. This requires more
serialization than I would like but since the number of threads is
limited to a low number, the tradoff in favor of the stack seems
reasonable.
6) Separates and heirarchicalizes (sic) the SIM and APP filesystem
code. SIM path and volume handling is just like the target. Some
aspects of the APP file code get more straightforward (e.g. no path
hashing is needed).
Dircache:
Deserves its own section. Dircache is new but pays homage to the old.
The old one was not compatible and so it, since it got redone, does
all the stuff it always should have done such as:
1) It may be update and used at any time during the build process.
No longer has one to wait for it to finish building to do basic file
management (create, remove, rename, etc.).
2) It does not need to be either fully scanned or completely disabled;
it can be incomplete (i.e. overfilled, missing paths), still be
of benefit and be correct.
3) Handles mounting and dismounting of individual volumes which means
a full rebuild is not needed just because you pop a new SD card in the
slot. Now, because it reuses its freed entry data, may rebuild only
that volume.
4) Much more fundamental to the file code. When it is built, it is
the keeper of the master file list whether enabled or not ("disabled"
is just a state of the cache). Its must always to ready to be started
and bind all streams opened prior to being enabled.
5) Maintains any short filenames in OEM format which means that it does
not need to be rebuilt when changing the default codepage.
Miscellaneous Compatibility:
1) Update any other code that would otherwise not work such as the
hotswap mounting code in various card drivers.
2) File management: Clipboard needed updating because of the behavioral
changes. Still needs a little more work on some finer points.
3) Remove now-obsolete functionality such as the mutex's "no preempt"
flag (which was only for the prior FAT driver).
4) struct dirinfo uses time_t rather than raw FAT directory entry
time fields. I plan to follow up on genericizing everything there
(i.e. no FAT attributes).
5) unicode.c needed some redoing so that the file code does not try
try to load codepages during a scan, which is actually a problem with
the current code. The default codepage, if any is required, is now
kept in RAM separarately (bufalloced) from codepages specified to
iso_decode() (which must not be bufalloced because the conversion
may be done by playback threads).
Brings with it some additional reusable core code:
1) Revised file functions: Reusable code that does things such as
safe path concatenation and parsing without buffer limitations or
data duplication. Variants that copy or alter the input path may be
based off these.
To do:
1) Put dircache functionality back in the sim. Treating it internally
as a different kind of file system seems the best approach at this
time.
2) Restore use of dircache indexes in the playlist and database or
something effectively the same. Since the cache doesn't have to be
complete in order to be used, not getting a hit on the cache doesn't
unambiguously say if the path exists or not.
Change-Id: Ia30f3082a136253e3a0eae0784e3091d138915c8
Reviewed-on: http://gerrit.rockbox.org/566
Reviewed-by: Michael Sevakis <jethead71@rockbox.org>
Tested: Michael Sevakis <jethead71@rockbox.org>
We redefine the top-level macros to our own in order to maintain
compatibility with compound initializers by wrapping the mid or low
level definitions from the OS header.
This allows, hopefully optimized, macros from the host OS's headers to
be used when building any hosted target obviating the need for
NEED_GENERIC_BYTESWAPS unless the target simply doesn't define its
own optimized versions (MIPS!).
Throw in some 64-bit swaps for completeness' sake; they generate no code
if not yet used anyway.
Change-Id: I21b384b55fea46833d01ea3cad1ad8952ea01a11
Abstracts threading from itself a bit, changes the way its queues are
handled and does type hiding for that as well.
Do alot here due to already required major brain surgery.
Threads may now be on a run queue and a wait queue simultaneously so
that the expired timer only has to wake the thread but not remove it
from the wait queue which simplifies the implicit wake handling.
List formats change for wait queues-- doubly-linked, not circular.
Timeout queue is now singly-linked. The run queue is still circular
as before.
Adds a better thread slot allocator that may keep the slot marked as
used regardless of the thread state. Assists in dumping special tasks
that switch_thread was tasked to perform (blocking tasks).
Deletes alot of code yet surprisingly, gets larger than expected.
Well, I'm not not minding that for the time being-- omlettes and break
a few eggs and all that.
Change-Id: I0834d7bb16b2aecb2f63b58886eeda6ae4f29d59
find_first_set_bit() becomes a small inline on ARMv5+ and checkwps now gets
made with -std=gnu99 (it eats all the GCCOPTS) like the rest of things.
Change-Id: Ie6039b17fec057a3dcb0f453d8fd5efac984df89
When the align parameter was a 32bit value (like all default integer literals),
and the to-be-aligned value is a pointer the upper 32bit got corrupted because
the value was casted down to 32bit.
Note: This hasnt been a problem because apparently the sim always gets 32bit
addresses (I found this when compiling Rockbox as a library).
Change-Id: I0d2d3fd8bfa210326b27162bb22c059da97d207a
No code changed, just shuffling stuff around. This should make it easier to
build only select parts kernel and use different implementations.
Change-Id: Ie1f00f93008833ce38419d760afd70062c5e22b5
Creates a standard buffer passing, local data passing and messaging
system for processing stages. Stages can be moved to their own source
files to reduce clutter and ease assimilation of new ones. dsp.c
becomes dsp_core.c which supports an engine and framework for effects.
Formats and change notifications are passed along with the buffer so
that they arrive at the correct time at each stage in the chain
regardless of the internal delays of a particular one.
Removes restrictions on the number of samples that can be processed at
a time and it pays attention to destination buffer size restrictions
without having to limit input count, which also allows pcmbuf to
remain fuller and safely set its own buffer limits as it sees fit.
There is no longer a need to query input/output counts given a certain
number of input samples; just give it the sizes of the source and
destination buffers.
Works in harmony with stages that are not deterministic in terms of
sample input/output ratio (like both resamplers but most notably
the timestretch). As a result it fixes quirks with timestretch hanging
up with certain settings and it now operates properly throughout its
full settings range.
Change-Id: Ib206ec78f6f6c79259c5af9009fe021d68be9734
Reviewed-on: http://gerrit.rockbox.org/200
Reviewed-by: Michael Sevakis <jethead71@rockbox.org>
Tested-by: Michael Sevakis <jethead71@rockbox.org>
* filesize() is not POSIX, so it doesn't need stubbing or redirecting
* make the various directory functions use the sim_ versions for PCTOOL
* PCTOOL needs generic byteswap functions
* fix the database makefile to not use -DSIMULATOR anymore
Change-Id: Ic6abc4f662830b85626c751a472fa4a03e844871
This fixes errornous pointer addition (+ on a short*), which crashed in some situation.
Fixes FS#12317 and should hopefully get the clips booting again.
Thanks to Jonathan Gordon for spotting the bad pointer arithmetic.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30724 a1c6a512-1295-4272-9138-f99709370657
buffer chunks.
* Samples and position indication is closely associated with audio data
instead of compensating by a latency constant. Alleviates problems with
using the elapsed as a track indicator where it could be off by several
steps.
* Timing is accurate throughout track even if resampling for pitch shift,
whereas before it updated during transition latency at the normal 1:1 rate.
* Simpler PCM buffer with a constant chunk size, no linked lists.
In converting crossfade, a minor change was made to not change the WPS until
the fade-in of the incoming track, whereas before it would change upon the
start of the fade-out of the outgoing track possibly having the WPS change
with far too much lead time.
Codec changes are to set elapsed times *before* writing next PCM frame because
time and position data last set are saved in the next committed PCM chunk.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30366 a1c6a512-1295-4272-9138-f99709370657
The NDK includes swap16 and swap32 macros, Rockbox as well. Use the Rockbox
ones and avoid a macro redefined warning.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@29939 a1c6a512-1295-4272-9138-f99709370657
The old cache coherency function names where wrong and misleading.
The new names are (purposely different from vendor manuals)
* commit_* (write-back only)
* discard_* (removing lines from cache only)
* commit_discard_* (write-back and removing lines from cache)
It's suspected the old names have led to wrong uses. The old names still exist
(as aliases) so every call via the old names need to be double checked and changed
to the new name.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@28045 a1c6a512-1295-4272-9138-f99709370657
- Move ATTRIBUTE_PRINTF/ATTRIBUTE_SCANF from _ansi.h
They are not related at all to this file, and this broke compilation
with Code Sourcery GCC which ships its own _ansi.h
- Move LIKELY/UNLIKELY from system.h
There is likely a lot more GCC extensions used everywhere in the source,
conditionally on __GNUC__ or unconditionally
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@27548 a1c6a512-1295-4272-9138-f99709370657
ideally all targets should define CACHEALIGN_BITS, for now we default it
to 16 bytes if it's not specified
Since the buffer is already aligned in playback.c no need to align it
again in buffering.c
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@27073 a1c6a512-1295-4272-9138-f99709370657
The simulator defines PLATFORM_HOSTED, as RaaA will do (RaaA will not define SIMULATOR).
The new define is to (de-)select code to compile on hosted platforms generally.
Should be no functional change to targets or the simulator.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@27019 a1c6a512-1295-4272-9138-f99709370657
- Refactor the program startup. main() is now in main.c like on target, and the implicit application thread will now act as our main thread (previously a separate one was created for this in thread initialization).
This is part of Rockbox as an application and is the first step to make an application port from the uisimulator. In a further step the sim bits from the sdl build will be separated out.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@26065 a1c6a512-1295-4272-9138-f99709370657
Introduce STORAGE_ALIGN_DOWN, STORAGE_PAD (using new CACHE_PAD) and
STORAGE_OVERLAP (using new CACHE_OVERLAP), make them useful only when
PROC_NEEDS_CACHEALIGN and STORAGE_NEEDS_ALIGN are defined
Modify PP and nano2g system-target.h accordingly
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@25336 a1c6a512-1295-4272-9138-f99709370657
