summaryrefslogtreecommitdiff
path: root/firmware/kernel/include/thread.h (unfollow)
Commit message (Collapse)Author
2014-08-30Remove I/O priority. It is harmful when used with the new file code.Michael Sevakis
HAVE_IO_PRIORITY was defined for native targets with dircache. It is already effectively disabled for the most part since dircache no longer lowers its thread's I/O priority. It existed primarily for the aforementioned configuration. Change-Id: Ia04935305397ba14df34647c8ea29c2acaea92aa
2014-08-16Base scheduler queues off linked lists and do cleanup/consolidationMichael Sevakis
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
2014-08-08Do some kernel cleanupMichael Sevakis
* Seal away private thread and kernel definitions and declarations into the internal headers in order to better hide internal structure. * Add a thread-common.c file that keeps shared functions together. List functions aren't messed with since that's about to be changed to different ones. * It is necessary to modify some ARM/PP stuff since GCC was complaining about constant pool distance and I would rather not force dump it. Just bl the cache calls in the startup and exit code and let it use veneers if it must. * Clean up redundant #includes in relevant areas and reorganize them. * Expunge useless and dangerous stuff like remove_thread(). Change-Id: I6e22932fad61a9fac30fd1363c071074ee7ab382
2014-08-06Fix some stuff for no priority andMichael Sevakis
thread_queue_wake() doesn't need the 2nd parameter. The original purpose for it never came to be. Non priority version mrsw_writer_wakeup_readers was left improperly finished. Get that back into line. Change-Id: Ic613a2479f3cc14dc7c761517670eb15178da9f5
2014-08-06Add multi-reader, single-writer locks to kernel.Michael Sevakis
Any number of readers may be in the critical section at a time and writers are mutually exclusive to all other threads. They are a better choice when data is rarely modified but often read and multiple threads can safely access it for reading. Priority inheritance is fully implemented along with other changes to the kernel to fully support it on multiowner objects. This also cleans up priority code in the kernel and updates some associated structures in existing objects to the cleaner form. Currently doesn't add the mrsw_lock.[ch] files since they're not yet needed by anything but the supporting improvements are still useful. This includes a typed bitarray API (bitarray.h) which is pretty basic for now. Change-Id: Idbe43dcd9170358e06d48d00f1c69728ff45b0e3 Reviewed-on: http://gerrit.rockbox.org/801 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested: Michael Sevakis <jethead71@rockbox.org>
2014-05-09Make errno a thread-local variable.Michael Sevakis
errno is supposed to be thread local and we do that here in a rather makeshift way by implementing a function that returns a pointer to the __errno variable in the thread block. If more serious TLS is required, perhaps it would be worthwhile implementing it using the linker and the "__thread" storage attribute. That's a bit overkill just for this. I'm also not liking what I saw the compiler producing. Change-Id: I03bc0bd6a89f6e3d6bae7653284ee01054614f9a Reviewed-on: http://gerrit.rockbox.org/803 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested: Michael Sevakis <jethead71@rockbox.org>
2014-03-03kernel: Break out kernel primitives into separate files and move to separate ↵Thomas Martitz
dir. 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
2013-05-31Refactor audio thread to run both recording and playback.Michael Sevakis
Eliminates the pcmrec thread and keeps playback and recording engine operation mutually-exclusive. audio_thread.c contains the audio thread which branches to the correct engine depending upon the request. It also handles the main audio initialization. Moves pcm_init into main.c just before dsp_init because I don't want that one in audio_init in the new file. (Also makes revision df6e1bc pointless ;) Change-Id: Ifc1db24404e6d8dd9ac42d9f4dfbc207aa9a26e1
2012-01-22Fix checkwps.Thomas Martitz
Change-Id: Ia530de560c17b9b3a71c3d9117c7cdc72ec7569d
2012-01-22Create fimrware/asm directory for assembly optimized stuff.Thomas Martitz
This dir is suitable for stuff that doesn't fit the target tree, e.g. because it also builds on hosted or otherwise. It also has a generic subfolder for fallback C implementations so that not all archs need to provide asm files. SOURCES should only contain "foo.c" where foo.c includes the specific <arch>/foo.c files from the subdirs using the preprocessor. This way automatic selection of asm versions or generic C verion is possible. For the start, the thread support files are moved, since ASM threads can be used on hosted platforms as well. Since core_sleep() remains platform specific it's moved to the corresponding system.h headers. Change-Id: Iebff272f3407a6eaafeb7656ceb0ae9eca3f7cb9
2011-11-16Finally commit FS#5111 - piezo clicker for ipods!Jonathan Gordon
Origional implementation by Robert Keevil with contributions from Frederik Vestre, Stoyan Stratev, Craig Elliott, Michael Sparmann, Thomas Schott, Rosso Maltese, and syncs from a bunch of other people! git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30995 a1c6a512-1295-4272-9138-f99709370657
2011-03-05Change the thread api a bit.Thomas Martitz
* Remove THREAD_ID_CURRENT macro in favor of a thread_self() function, this allows thread functions to be simpler. * thread_self_entry() shortcut for kernel.c. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@29521 a1c6a512-1295-4272-9138-f99709370657
2011-03-02Merge functionality of wakeups and semaphores-- fewer APIs and object types. ↵Michael Sevakis
semaphore_wait takes a timeout now so codecs and plugins have to be made incompatible. Don't make semaphores for targets not using them. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@29492 a1c6a512-1295-4272-9138-f99709370657
2011-02-19Cleanup preprocessor around corelock usage and move its definition outside ↵Thomas Martitz
#ifdef ASSEMBLER_THREADS git-svn-id: svn://svn.rockbox.org/rockbox/trunk@29330 a1c6a512-1295-4272-9138-f99709370657
2011-02-18Fix yellows and checkwps.Thomas Martitz
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@29328 a1c6a512-1295-4272-9138-f99709370657
2011-02-18Implement cooperative threads on hosted platforms using C code.Thomas Martitz
This replaces SDL threads with real cooperative threads, which are less cpu intensive and allow priority scheduling. The backend for context switching is dependant on the host (sigaltstack/longjmp on Unix, Fibers on Windows). configure has options to force or disallow SDL threads. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@29327 a1c6a512-1295-4272-9138-f99709370657
2010-12-29Just a technical matter even though all is well anyway: struct ↵Michael Sevakis
blocker::thread should be volatile for > 1 core and the change should make no difference to anything for 1 core. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@28923 a1c6a512-1295-4272-9138-f99709370657
2010-12-27Certain data accesses in the kernel should have volatile semantics to be ↵Michael Sevakis
correct and not rely on the whims of the compiler. Change queue clearing to simply catch read up to write rather than reset both to 0 to ensure sane results for queue_count and queue_empty with concurrency. Binsize may or may not increase a bit depending upon whether the output was as intended in all places; wrong stuff was already unlikely to cause any issue. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@28909 a1c6a512-1295-4272-9138-f99709370657
2010-12-22Fix yellowThomas Martitz
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@28878 a1c6a512-1295-4272-9138-f99709370657
2010-08-12Introduce NORETURN_ATTR wrapper for __attribute__((noreturn)), using this ↵Thomas Martitz
and a bit further cleanup in main gets rid of a warning when compiling for android. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@27788 a1c6a512-1295-4272-9138-f99709370657
2010-08-02Rockbox as an application: Commit current Android port progress.Thomas Martitz
General state is: Rockbox is usable (plays music, saves configuration, touchscreen works too). Problems: - Playing music in the background (i.e. when switching to another app) doesn't work reliably, but I'm working on that now. - no cabbiev2 (only some preliminary files for it), no other default theme. - screen flickers sometimes if the updates are too frequent - no multi screen apk/package - strange behavior when a phone call comes in The java files (and the eclipse project) resides in android/, which is also supposed to be the build folder. I've put a small README in there for instructions. There are some steps needed after the make part, which are described there, and which eclipse mostly handles. But there ought to be some script/makefile rules which do that instead in the future. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@27668 a1c6a512-1295-4272-9138-f99709370657
2010-06-21Rockbox as an application: Replace many occurences of #ifdef SIMULATOR with ↵Thomas Martitz
#if (CONFIG_PLATFORM & PLATFORM_HOSTED) (or equivalently). 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
2010-06-17Switch iPod 3G to use EABI toolchain. Make necessary threading changes to ↵Michael Sevakis
avoid use of stack after switching to idle stack. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@26898 a1c6a512-1295-4272-9138-f99709370657
2010-06-10Clean unused stuff out of thread.h and config.h and reorganize thread-pp.c ↵Michael Sevakis
to simplify the preprocessor blocks. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@26743 a1c6a512-1295-4272-9138-f99709370657
2010-06-08CPU_PP implies CPU_ARM is definedRafaël Carré
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@26690 a1c6a512-1295-4272-9138-f99709370657
2010-06-07Remove CORELOCK_SWAP (unused since r16981)Rafaël Carré
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@26664 a1c6a512-1295-4272-9138-f99709370657
2010-05-27SDL Simulator: Get thread shutdown and properly handled and fix a minor ↵Michael Sevakis
memory leak that happens when threads exit. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@26336 a1c6a512-1295-4272-9138-f99709370657
2010-04-17Use a more natural guard for the callback definitionAlexander Levin
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@25664 a1c6a512-1295-4272-9138-f99709370657
2010-04-16Cleanup thread_entry definition a bit. Reorder members so that related ones ↵Thomas Martitz
are next to each other and move a pretty hidden define into config.h. Doesn't make a binsize difference (the comment says the order should be optimized for size). git-svn-id: svn://svn.rockbox.org/rockbox/trunk@25659 a1c6a512-1295-4272-9138-f99709370657
2010-04-03Add IO priority handling. Currently all IO has equal priority, except the ↵Frank Gevaerts
dircache scanning thread which is lower. This fixes the slow boot problem for me, with the added benefit that actual audio playback also starts faster. 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
2010-03-03FS#10756 - Free unused init codeThomas Martitz
Introduce a new .init section for initialisation code, so that it can be copied to an area which is later overwritten before calling. The stack/bss can then overwrite that code, effectively freeing the code size that the initialisation routines need. Gives a few kB ram usage back. Only implemented for PP and as3525 so far. More targets could be added, as well as more functions. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@25013 a1c6a512-1295-4272-9138-f99709370657
2009-05-06MIPS: don't save gp register when switching threadsMaurus Cuelenaere
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@20863 a1c6a512-1295-4272-9138-f99709370657
2009-01-05Unless I'm severely mistaken, we can drop a thread slot now since mpegplayer ↵Michael Sevakis
creates one less thread. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@19686 a1c6a512-1295-4272-9138-f99709370657
2008-12-10Use cookies for thread identification instead of pointers directly which ↵Michael Sevakis
gives a buffer against wrongly identifying a thread when the slot is recycled (which has been nagging me for awhile). A slot gets 255 uses before it repeats. Everything gets incompatible so a full update is required. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@19377 a1c6a512-1295-4272-9138-f99709370657
2008-09-14Onda VX747:Maurus Cuelenaere
* Get rid of bug when interrupts are enabled * Get threading to work (although with some weirdness) * Other fixes/optimizations git-svn-id: svn://svn.rockbox.org/rockbox/trunk@18512 a1c6a512-1295-4272-9138-f99709370657
2008-07-14Add preliminary support for the Onda VX747 (MIPS target)Maurus Cuelenaere
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@18032 a1c6a512-1295-4272-9138-f99709370657
2008-06-28Updated our source code header to explicitly mention that we are GPL v2 orDaniel Stenberg
later. We still need to hunt down snippets used that are not. 1324 modified files... http://www.rockbox.org/mail/archive/rockbox-dev-archive-2008-06/0060.shtml git-svn-id: svn://svn.rockbox.org/rockbox/trunk@17847 a1c6a512-1295-4272-9138-f99709370657
2008-05-18Spelling fixes.Bertrik Sikken
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@17574 a1c6a512-1295-4272-9138-f99709370657
2008-05-04The presence of the run list corelocks are dependent on core number and not ↵Michael Sevakis
building with priority scheduling. Fix building for multiprocessor without HAVE_PRIORITY_SCHEDULING defined. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@17362 a1c6a512-1295-4272-9138-f99709370657
2008-04-12Add GPIO manager. Get the PMIC interrupt handling working (along with power ↵Michael Sevakis
button and HP detect). Add some reg field defined instead of using raw numbers. Add PMIC info to debug ports screen. Refine PMIC driver ops a little bit. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@17086 a1c6a512-1295-4272-9138-f99709370657
2008-03-29Just change PRIORITY_BUFFERING to 15. Seems a teeny tiny bit helpful to ↵Michael Sevakis
mpegplayer as well. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@16884 a1c6a512-1295-4272-9138-f99709370657
2008-03-25Add a complete priority inheritance implementation to the scheduler (all ↵Michael Sevakis
mutex ownership and queue_send calls are inheritable). Priorities are differential so that dispatch depends on the runnable range of priorities. Codec priority can therefore be raised in small steps (pcmbuf updated to enable). Simplify the kernel functions to ease implementation and use the same kernel.c for both sim and target (I'm tired of maintaining two ;_). 1) Not sure if a minor audio break at first buffering issue will exist on large-sector disks (the main mutex speed issue was genuinely resolved earlier). At this point it's best dealt with at the buffering level. It seems a larger filechunk could be used again. 2) Perhaps 64-bit sims will have some minor issues (finicky) but a backroll of the code of concern there is a 5-minute job. All kernel objects become incompatible so a full rebuild and update is needed. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@16791 a1c6a512-1295-4272-9138-f99709370657
2008-01-19Do some simplification in switch_thread regarding interrupt enabling which ↵Michael Sevakis
simplifies kernel objects as well. Simply doing unconditional disable/enable should be fine in the firmware. Not sure about the case in the bootloader for all targets and so will evaluate but that should be fine too if everything is masked. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@16107 a1c6a512-1295-4272-9138-f99709370657
2007-12-29mpegplayer: Make playback engine fully seekable and frame-accurate and split ↵Michael Sevakis
into logical parts. Be sure to have all current features work. Actual UI for seeking will be added soon. Recommended GOP size is about 15-30 frames depending on target or seeking can be slow with really long GOPs (nature of MPEG video). More refined encoding recommendations for a particular player should be posted soon. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@15977 a1c6a512-1295-4272-9138-f99709370657
2007-11-20Make threads responsible for explicit cancellation of their own boosted ↵Michael Sevakis
status. Sleeping and timeouts will no longer cancel it. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@15709 a1c6a512-1295-4272-9138-f99709370657
2007-10-25Finally commit Metadata on Buffer!Nicolas Pennequin
buffering.c and buffering.h implement the new buffering API. playback.c is translated to that API. For more information about the whole concept, see http://www.rockbox.org/wiki/MetadataOnBuffer. There should be no major visible changes, but most existing bugs remain (though fixing them should be easier now that playback.c is a bit less complex) and there probably will be new ones. Please report any problem! Next step is to adapt cuesheet support, which is partly disabled here, and of course fix as much bugs as possible. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@15306 a1c6a512-1295-4272-9138-f99709370657
2007-10-17Apply Akio Idehara's fix for FS#7972 - Fix ARM's swp (xchg) inline assembly ↵Michael Sevakis
for gcc 4.2. Also avoids UNPREDICTABLE behavior that GCC should have always warned about. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@15157 a1c6a512-1295-4272-9138-f99709370657
2007-10-16Oops made a mistake that messed up test_and_set selection.Michael Sevakis
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@15139 a1c6a512-1295-4272-9138-f99709370657
2007-10-16Finally full multicore support for PortalPlayer 502x targets with an eye ↵Michael Sevakis
towards the possibility of other types. All SVN targets the low-lag code to speed up blocking operations. Most files are modified here simple due to a name change to actually support a real event object and a param change to create_thread. Add some use of new features but just sit on things for a bit and leave full integration for later. Work will continue on to address size on sensitive targets and simplify things if possible. Any PP target having problems with SWP can easily be changed to sw corelocks with one #define change in config.h though only PP5020 has shown an issue and seems to work without any difficulties. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@15134 a1c6a512-1295-4272-9138-f99709370657
2007-10-06Discontinue any use of 'swp(b)' on PP5020. While clocking is stable, some ↵Michael Sevakis
testing revealed this instruction can still cause problems without concurrent access. Make sure mpegplayer is safe while not using spinlock (no longer atomic) between cores to protect the stream byte counters - use nonwrapping head and tail pointers. git-svn-id: svn://svn.rockbox.org/rockbox/trunk@15005 a1c6a512-1295-4272-9138-f99709370657
generated by cgit v1.1 at 2026-04-25 09:17:45 +0000