Now all threads need to ack the connection like on real target, dircache is unloaded and playback stops accordingly.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@31009 a1c6a512-1295-4272-9138-f99709370657
When booting with USB inserted, the dircache build can get interrupted by the usb connection, in which case the dircache buffer is freed.
Due to a bug the re-creation of dircache used the old freed buffer and overwrite new allocs (causing screen corruption).
Set allocated_size to 0 to make it not take the code path that assumes an existing buffer,
and bring that and freeing together in the code.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30845 a1c6a512-1295-4272-9138-f99709370657
Also add two dircache function, one of which does what dircache_disable()
did previously as this now also frees the dircache buffer.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30393 a1c6a512-1295-4272-9138-f99709370657
This enables the ability to allocate (and free) memory dynamically
without fragmentation, through compaction. This means allocations can move
and fragmentation be reduced. Most changes are preparing Rockbox for this,
which many times means adding a move callback which can temporarily disable
movement when the corresponding code is in a critical section.
For now, the audio buffer allocation has a central role, because it's the one
having allocated most. This buffer is able to shrink itself, for which it
needs to stop playback for a very short moment. For this,
audio_buffer_available() returns the size of the audio buffer which can
possibly be used by other allocations because the audio buffer can shrink.
lastfm scrobbling and timestretch can now be toggled at runtime without
requiring a reboot.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30381 a1c6a512-1295-4272-9138-f99709370657
The buflib memory allocator is handle based and can free and
compact, move or resize memory on demand. This allows to effeciently
allocate memory dynamically without an MMU, by avoiding fragmentation
through memory compaction.
This patch adds the buflib library to the core, along with
convinience wrappers to omit the context parameter. Compaction is
not yet enabled, but will be in a later patch. Therefore, this acts as a
replacement for buffer_alloc/buffer_get_buffer() with the benifit of a debug
menu.
See buflib.h for some API documentation.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30380 a1c6a512-1295-4272-9138-f99709370657
If the dircache was interrupted during generation (e.g. through USB
insertion), then the allocated buffer was leaked and a new one
was allocated for the second cache generation. This causes a
panic since r30308 since playback holds the control over the
audiobuffer at that time.
The fix is to simply check allocated_size instead of
dircache_size which is reset to 0 upon cancellation.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30321 a1c6a512-1295-4272-9138-f99709370657
dircache_root wasn't initialized at all and the giving allocated_size
passed to buffer_release_buffer() didn't account for alignment padding.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30318 a1c6a512-1295-4272-9138-f99709370657
Namely, introduce buffer_get_buffer() and buffer_release_buffer().
buffer_get_buffer() aquires all available and grabs a lock, attempting to
call buffer_alloc() or buffer_get_buffer() while this lock is locked will cause
a panicf() (doesn't actually happen, but is for debugging purpose).
buffer_release_buffer() unlocks that lock and can additionally increment the
audiobuf buffer to make an allocation. Pass 0 to only unlock if buffer was
used temporarily only.
buffer_available() is a replacement function to query audiobuflen, i.e. what's
left in the buffer.
Buffer init is moved up in the init chain and handles ipodvideo64mb internally.
Further changes happened to mp3data.c and talk.c as to not call the above API
functions, but get the buffer from callers. The caller is the audio system
which has the buffer lock while mp3data.c and talk mess with the buffer.
mpeg.c now implements some buffer related functions of playback.h, especially
audio_get_buffer(), allowing to reduce #ifdef hell a tiny bit.
audiobuf and audiobufend are local to buffer.c now.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30308 a1c6a512-1295-4272-9138-f99709370657
The first is an off-by-one that leads to miscalculation of the dircache size.
The format string size was used but dircache size was incremented by the snprintf() result which is smaller.
The other forgot to update the location of the "." and ".." strings upon compaction,
so that new folders got assigned orphaned pointers for those directory entires.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30224 a1c6a512-1295-4272-9138-f99709370657
prevent d_names data from being overwritten (likely causing garbage in
file browser and other strange symptoms).
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30122 a1c6a512-1295-4272-9138-f99709370657
Only integer IDs are exposed from dircache with this. This way the cache is isolated from other modules.
This is needed for my buflib gsoc project.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30038 a1c6a512-1295-4272-9138-f99709370657
The dircache_entry structs are now allocated subsequently from the front, allowing to treat them as an array. The d_names are allocated from the back (in reverse order, growing downwards).
This allows the cache to be moved around (needed for my buflib gsoc project). It is utilized when loading the cache from disk (on the h100), now the pointer to the cache begin doesn't need to be the same across reboots anymore.
This should save a bit memory usage, since there's no need for aligning padding bytes after d_names anymore.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30036 a1c6a512-1295-4272-9138-f99709370657
Use a recursive helper function with strlcat to build up the path backwards. This way the tree doesn't need to be walked twice and no extraneous size calculation is needed.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30033 a1c6a512-1295-4272-9138-f99709370657
It's reduntant, and enlarges the dircache unnecessarily. Saves 4 byte per file in the whole filesystem.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30032 a1c6a512-1295-4272-9138-f99709370657
Use host's functions for file i/o directly (open(), close() ,etc.), not the sim_* variants.
Some dir functions need to be wrapped still because we need to cache the parents dir's path (host's dirent doesn't let us know).
For the same reason (incompatibility) with host's dirent) detach some members from Rockbox' dirent struct and put it into an extra one,
the values can be retrieved via the new dir_get_info().
Get rid of the sim_ prefix for sleep as well and change the signature to unix sleep().
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@27968 a1c6a512-1295-4272-9138-f99709370657
For RaaA it evaluates user paths at runtime. For everything but codecs/plugins it will give the path under $HOME/.config/rockbox.org if write access is needed or if the file/folder in question exists there (otherwise it gives /usr/local/share/rockbox).
This allows for installing themes under $HOME as well as having config.cfg and other important files there while installing the application (and default themes) under /usr/local.
On the DAPs it's a no-op, returing /.rockbox directly.
Not converted to use get_user_file_path() are plugins themselves, because RaaA doesn't build plugins yet.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@27656 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
- fix check_event_queue to use queue_peek and not mess message order
- change check_dircache_state to a more modest check to avoid messing up dircache state
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@26228 a1c6a512-1295-4272-9138-f99709370657
This is to a) to cleanup firmware/common and firmware/include a bit, but also b) for Rockbox as an application which should use the host system's c library and headers, separating makes it easy to exclude our files from the build.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@25850 a1c6a512-1295-4272-9138-f99709370657
Lots of the changes are due to changing storage_(read|write)sectors() from macros to wrapper functions. This means that they have to be called with IF_MD2(drive,) again.
Flyspray: FS#11167
Author: Frank Gevaerts
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@25459 a1c6a512-1295-4272-9138-f99709370657
-Move fat_dir structure out of dircache stack to RAM. Reduce dircache stack size (max level depth should stay be around 20). This should fix nano2g dircache stkov of FS#10679
-Change the structure returned by readdir_cached to match the one returned by readdir_uncached: remove useless fields to save space and avoid any potential incoherence
-Remove one field from the internal structure used by {opend,read,close}dir_cached because it was mostly redundant.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@24708 a1c6a512-1295-4272-9138-f99709370657
* Move strncpy() from core to the pluginlib
* Introduce strlcpy() and use that instead in most places (use memcpy in a few) in core and some plugins
* Drop strncpy() from the codec api as no codec used it
* Bump codec and plugin api versions
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@21863 a1c6a512-1295-4272-9138-f99709370657
