aboutsummaryrefslogtreecommitdiffstats
path: root/src (follow)
Commit message (Collapse)AuthorAgeFilesLines
* strset: fix duplicate valuesMax Kellermann2008-09-091-1/+1
| | | | | Due to a minor typo, the string set had duplicate values, because strset_add() didn't check the base slot properly.
* added string set libraryMax Kellermann2008-09-093-0/+195
| | | | | | | | "struct strset" is a hashed string set: you can add strings to this library, and it stores them as a set of unique strings. You can get the size of the set, and you can enumerate through all values. This will be used to replace the linear tagTracker library.
* alsa: use blocking instead of non-blocking writeEric Wong2008-09-081-1/+6
| | | | | | | | | | | | The way we used non-blocking mode was HORRIBLE. It was non-blocking to ALSA, but we end up blocking in a busy loop that does absolutely NOTHING but retry. We don't check for playback cancellation (like we do in decoders) or anything. This is seriously broken and I can imagine it affects people on fast CPUs more because we do asynchronous output buffering and our ALSA device will always have data ready.
* alsa: show more debugging informationEric Wong2008-09-081-10/+14
| | | | | | | | | Print out {buffer,period}_{size,time}. Not sure if this is going to help. I've been searching everywhere looking for a possible clue as to what's causing the high CPU usage problems... Also, add device information to some messages I missed earlier.
* alsa: cleanup the mmap logic a bitEric Wong2008-09-081-8/+6
|
* alsa: snd_pcm_sw_params_set_xfer_align is deprecatedEric Wong2008-09-081-3/+0
| | | | | Lets not use deprecated functions. It's apparently possible to not care about the sw_params stuff at all!
* alsa: only run snd_config_update_free_global once atexitEric Wong2008-09-071-3/+7
| | | | | | | | | This is safer than the patch in http://www.musicpd.org/mantis/view.php?id=1542 with multiple audio outputs enabled. Sadly, I only noticed that patch/problem when I googled for "snd_config_update_free_global"
* alsa: move bitformat reading code out of the wayEric Wong2008-09-071-16/+12
|
* alsa: always print the period_time we setEric Wong2008-09-071-2/+1
|
* alsa: avoid unnecessary heap usage if we don't set a device nameEric Wong2008-09-071-11/+12
|
* alsa: get rid of the needless canPause flagEric Wong2008-09-071-3/+0
| | | | | We never use it for anything anyways as we release the device entirely on pause.
* alsa: avoid reassigning ad->writeiEric Wong2008-09-071-7/+4
| | | | This saves me precious terminal space
* alsa: add reasoning for the non-portable macroEric Wong2008-09-071-0/+2
|
* alsa: capitalize "ALSA" consistently in messagesEric Wong2008-09-071-16/+9
| | | | That's the name of this project.
* Merge branch 'ew/alsa'Eric Wong2008-09-071-78/+69
|\ | | | | | | | | | | | | * ew/alsa: alsa: optimistically try resuming from suspend alsa: extra debugging outputs to fix suspend/hibernate alsa: cleanup debug assignment of the "cmd" variable
| * alsa: optimistically try resuming from suspendEric Wong2008-09-071-11/+12
| | | | | | | | | | | | | | | | | | | | | | | | Apparently snd_pcm_hw_params_can_resume() can return false even though my hardware does in fact support resuming. So stop carrying that value in the canResume flag and just try to resume when we're in the suspended state; falling back to snd_pcm_prepare only if resuming fails. libao does something similar on resume, too. While we're at it, use the E() macro which will enable us to have better error reporting.
| * alsa: extra debugging outputs to fix suspend/hibernateEric Wong2008-09-071-3/+9
| | | | | | | | | | | | Hibernating my laptop while MPD is playing results in ugliness about "alsa device foo was suspend" constantly printed to the logs.
| * alsa: cleanup debug assignment of the "cmd" variableEric Wong2008-09-071-66/+50
| | | | | | | | | | Given the length of the ALSA command names, I only want to see them once per-section of code, if at all...
* | playlist: re-randomize when explicitly playing a new songEric Wong2008-09-071-0/+1
|/ | | | | | | | | | | | | | | | | | | | | | | | | | When random is enabled and a user explicitly specifies a certain song on the playlist should be played; we need to re-randomize the internal ordering. To reproduce this, assuming a four song playlist: play <song_a> next => <song_b> next => <song_c> next => <song_d> play <song_a> next => <song_b> next => <song_c> next => <song_d> ... That is, the "next" command restarts song_{b,c,d} the second time "play" starts playing song_a. Thus, the second time "play" is called, the ordering of song_{b,c,d} needs to be reshuffled. Reported-by: Qball
* tag: oops, of course items is now ** and not *Eric Wong2008-09-051-2/+2
| | | | | | Gah, it seems like doing sizeof here either way is error prone. Too easy to leave out a '*' character we can forget.
* audio_format: volatile removalEric Wong2008-09-051-3/+3
| | | | | | | | | | | | | | | | | volatile provides absolutely no guarantee thread-safety in SMP environments. volatile was designed to access memory locations in peripheral hardware directly; not for SMP. If volatile is needed to work properly on SMP, then it is only hiding subtle bugs. volatile only prevents the /compiler/ from making optimizations when accessing variables. CPUs do their own optimizations at runtime so it cannot guarantee registers of CPUs are flushed to memory cache-coherent access on different CPUs. Furthermore, the thread-communication via condition variables between threads sharing audio formats already results in memory barriers.
* tag: lock all accesses to tag_poolEric Wong2008-09-053-6/+15
| | | | | | | | | | The tag pool is a shared global resource that is infrequently modified. However, it can occasionally be modified by several threads, especially by the metadata_pipe for streaming metadata (both reading/writing). The bulk tag_item pool is NOT locked as currently only the update thread uses it.
* tag: introduce handy items_size() functionEric Wong2008-09-051-11/+12
| | | | | | | | | | | Trying to read or remember "tag->numOfItems * sizeof(*tag->items)" requires too much thinking and mental effort on my part. Also, favor "sizeof(struct mpd_tag)" over "sizeof(*tag->items)" because the former is easier to read and follow, even though the latter is easier to modify if the items member changes to a different type.
* Merge branch 'mk/tag'Eric Wong2008-09-0344-1468/+998
|\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * mk/tag: (22 commits) tag: fix segfault on update utf8.h: Fix build (broken os_compat.h #include) tag: optimize tag_dup(), copy item references tag: fix the shout and oggflac plugins const pointers tag: static directory name tag: try not to reallocate tag.items in every add() call song: don't export newNullSong() tag: try not to duplicate the input string tag: pass length to fix_utf8() added "length" parameter to validUtf8String() assert value!=NULL in fix_utf8() tag: converted macro fixUtf8() to an inline function tag: added a pool for tag items tag: converted tag_item.value to a char array removed tree.c tag: converted MpdTag.items to a pointer list tag: moved code to tag_id3.c wavpack: tag_new() cannot fail tag: converted tag_add_item() to an inline function ...
| * tag: fix segfault on updateEric Wong2008-09-032-14/+22
| | | | | | | | | | | | | | | | | | | | | | | | clearMpdTag could be called on a tag that was still in a tag_begin_add transaction before tag_end_add is called. This was causing free() to attempt to operate on bulk.items; which is un-free()-able. Now instead we unmark the bulk.busy to avoid committing the tags to the heap only to be immediately freed. Additionally, we need to remember to call tag_end_add() when a song is updated before we NULL song->tag to avoid tripping an assertion the next time tag_begin_add() is called.
| * utf8.h: Fix build (broken os_compat.h #include)Eric Wong2008-09-021-1/+1
| | | | | | | | This is not a system header
| * tag: optimize tag_dup(), copy item referencesMax Kellermann2008-09-023-7/+42
| | | | | | | | | | Don't call tag_pool_get_item() for duplicating tags, just increase the item's reference counter instead.
| * tag: fix the shout and oggflac pluginsMax Kellermann2008-09-022-6/+8
| | | | | | | | | | | | During the tag library refactoring, the shout plugin was disabled, and I forgot about adapting it to the new API. Apply the same fixes to the oggflac decoder plugin.
| * const pointersMax Kellermann2008-09-024-10/+10
| | | | | | | | Yet another patch which converts pointer arguments to "const".
| * tag: static directory nameMax Kellermann2008-09-021-5/+3
| | | | | | | | | | | | While parsing the tag cache, don't allocate the directory name from the heap, but copy it into a buffer on the stack. This reduces heap fragmentation by 1%.
| * tag: try not to reallocate tag.items in every add() callMax Kellermann2008-09-023-7/+91
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If many tag_items are added at once while the tag cache is being loaded, manage these items in a static fixed list, instead of reallocating the list with every newly created item. This reduces heap fragmentation. Massif results again: mk before: total 12,837,632; useful 10,626,383; extra 2,211,249 mk now: total 12,736,720; useful 10,626,383; extra 2,110,337 The "useful" value is the same since this patch only changes the way we allocate the same amount of memory, but heap fragmentation was reduced by 5%.
| * song: don't export newNullSong()Max Kellermann2008-09-022-3/+1
| | | | | | | | The function newNullSong() is only used internally in song.c.
| * tag: try not to duplicate the input stringMax Kellermann2008-09-021-13/+18
| | | | | | | | | | | | Try to detect if the string needs Latin1-UTF8 conversion, or whitespace cleanup. If not, we don't need to allocate temporary memory, leading to decreased heap fragmentation.
| * tag: pass length to fix_utf8()Max Kellermann2008-09-021-3/+4
| | | | | | | | Same as the previous patch, prepare the function fix_utf8() this time.
| * added "length" parameter to validUtf8String()Max Kellermann2008-09-024-10/+16
| | | | | | | | | | | | | | At several places, we create temporary copies of non-null-terminated strings, just to use them in functions like validUtf8String(). We can save this temporary allocation and avoid heap fragmentation if we add a length parameter instead of expecting a null-terminated string.
| * assert value!=NULL in fix_utf8()Max Kellermann2008-09-021-1/+3
| | | | | | | | | | We must never pass value==NULL to fix_utf(). Replace the run-time check with an assertion.
| * tag: converted macro fixUtf8() to an inline functionMax Kellermann2008-09-021-9/+11
| | | | | | | | | | | | | | Since the inline function cannot modify its caller's variables (which is a good thing for code readability), the new string pointer is the return value. The resulting binary should be the same as with the macro.
| * tag: added a pool for tag itemsMax Kellermann2008-09-025-6/+147
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The new source tag_pool.c manages a pool of reference counted tag_item objects. This is used to merge tag items of the same type and value, saving lots of memory. Formerly, only the value itself was pooled, wasting memory for all the pointers and tag_item structs. The following results were measured with massif. Started MPD on amd64, typed "mpc", no song being played. My music database contains 35k tagged songs. The results are what massif reports as "peak". 0.13.2: total 14,131,392; useful 11,408,972; extra 2,722,420 eric: total 18,370,696; useful 15,648,182; extra 2,722,514 mk f34f694: total 15,833,952; useful 13,111,470; extra 2,722,482 mk now: total 12,837,632; useful 10,626,383; extra 2,211,249 This patch set saves 20% memory, and does a good job in reducing heap fragmentation.
| * tag: converted tag_item.value to a char arrayMax Kellermann2008-09-024-20/+5
| | | | | | | | | | | | | | The value is stored in the same memory allocation as the tag_item struct; this saves memory because we do not store the value pointer anymore. Also remove the getTagItemString()/removeTagItemString() dummies.
| * removed tree.cMax Kellermann2008-09-024-857/+67
| | | | | | | | | | | | | | | | | | This patch makes MPD consume much more memory because string pooling is disabled, but it prepares the next bunch of patches. Replace the code in tagTracker.c with naive algorithms without the tree code. For now, this should do; later we should find better algorithms, especially for getNumberOfTagItems(), which has become wasteful with temporary memory.
| * tag: converted MpdTag.items to a pointer listMax Kellermann2008-09-024-17/+20
| | | | | | | | | | | | This prepares the following patches, which aim to reduce MPD's memory usage: we plan to share tag_item instances, instead of just their values.
| * tag: moved code to tag_id3.cMax Kellermann2008-09-025-352/+402
| | | | | | | | | | The ID3 code uses only the public tag API, but is otherwise unrelated. Move it to a separate source file.
| * wavpack: tag_new() cannot failMax Kellermann2008-09-021-5/+0
| | | | | | | | | | Since tag_new() uses xmalloc(), it cannot fail - if we're really out of memory, the process will abort.
| * tag: converted tag_add_item() to an inline functionMax Kellermann2008-09-021-2/+5
| |
| * tag: renamed functions, no CamelCaseMax Kellermann2008-09-0219-121/+120
| |
| * tag: renamed MpdTag and MpdTagItem to struct mpd_tag, struct tag_itemMax Kellermann2008-09-0226-109/+112
| | | | | | | | | | Getting rid of CamelCase; not having typedefs also allows us to forward-declare the structures.
* | ob_send: avoid void pointer arithmetic warningEric Wong2008-09-021-1/+1
| |
* | volume: include outputBuffer.h for ob_set_sw_volumeEric Wong2008-09-021-0/+1
| |
* | Add missing function prototypesEric Wong2008-09-024-4/+7
|/
* Merge branch 'ew/deconst'Eric Wong2008-09-016-55/+26
|\ | | | | | | | | | | * ew/deconst: use deconst_ptr instead of duplicating deconst logic provide a generic deconst_ptr function