summaryrefslogtreecommitdiff
path: root/src/lib (follow)
AgeCommit message (Collapse)Author
2015-11-01Eo: Partial implementation to remove eo_dodevs/felipealmeida/eo_optimisationsFelipe Magno de Almeida
2015-10-19Eo do: use the __thread directive when available to manage call stack.Tom Hacohen
This is faster in most cases, and to be honest, should be much faster than it is. I don't understand why there's no better directive to mark a variable as *really* important thread storage that is used all the time.
2015-10-19Eo: Remove extra paranoid-never could happen safety checks.Tom Hacohen
I was not able to reach those without arbitrary memory corruption, but in that case we are in trouble anyway.
2015-10-19Eo: Reorganise some of the code and cleanup.Tom Hacohen
This is cleaner and makes more sense.
2015-10-19Eo: Removed weird unneeded condition.Tom Hacohen
2015-10-19Eo: Move op resolve check to where it belongs (out of hot path).Tom Hacohen
It was put in the wrong place. It should abort early if it detects we can't resolve, and shouldn't check it if we already know it's OK.
2015-10-19Eo unref: Decrease amount of checks and hint branch prediction.Tom Hacohen
This may look minor, but this is such a hot path, that this actually speeds things up a bit.
2015-10-19Eo: Split object checking from class checking and simplify.Tom Hacohen
The check there was wrong for objects anyway, and was ultra conservative for classes.
2015-10-19Eo callback call: Remove redundant safety checks.Tom Hacohen
This is already being checked by eo_do, no need to duplicate this here.
2015-10-19Eo callback call: Remove redundant refcounting.Tom Hacohen
We already get refcounting from eo_do, no need to have more refcounting here.
2015-10-19Eo callback call: Simplify callback desc matching code.Tom Hacohen
As part of this change, we also error on adding invalid callbacks.
2015-10-19Eo callbacks: simplify callback call code a bit.Tom Hacohen
We don't need this check because it's already checked in the match function.
2015-10-19Eo: Reduce call stack memory footprint.Tom Hacohen
We don't really need the eo_id most of the time, and when we do, it's very easy to get it. It's better if we just don't save the eo_id on the stack, and just save if it's an object or a class instead.
2015-10-19Eo: Remove useless optimisation that is more harm than good.Tom Hacohen
It seems that the idea behind that optimisation, is to save object data fetching when calling functions implemented by the object's class inside functions implemented by the object's class. This should be rare enough not to worth the upkeep, memory reads and memory writes, especially since for all cases apart of mixins (for which this optimisation won't work for anyway), the upkeep is more costly than fetching the data again.
2015-10-19Eo: use correct mask when checking if an id is a classTom Hacohen
2015-10-19Eo: Optimise object data fetching a bit more.Tom Hacohen
Removed safety check that is not necessary. This may seem small, but this in addition to the previous commit, account for around 2% of CPU usage.
2015-10-19Eo: Optimise object data fetching (minor).Tom Hacohen
Pre-calculate object data offset for improved performance.
2015-10-17eo - another 1.5 percent speedup in eo_bench eo_do by removing err handlCarsten Haitzler (Rasterman)
so we do a bit of error handling like does a stack fail to allocate, does setting the tls var fail, have the stack frames been nulled or not allocated, etc. - these acutally cost every call because they mean some extra compare and branches, but ore because they cause a lot fo extra code to be generated, thus polluting instruction cache with code and cacheline fetches of code that we rarely take - if ever. every if () and DBG, ERR etc. does cost something. in really hotpath code like this, i think it's best we realize that these checks will basically never be triggered, because if a stack fails to grow... we likely alreayd blew our REAL stack for the C/C++ side and that can't allocate anymore and has already just crashed (no magic message there - just segv). so in this case i think this checking is pointless and just costs us rather than gets us anything.
2015-10-17evas - clean if check to be a single return point for clarityCarsten Haitzler (Rasterman)
if (a) return; if (b) return; -> if ((a) || (b)) return;
2015-10-16Revert "ecore-wayland: Redo window animators to not use Custom source animators"Mike Blumenkrantz
This reverts commit bd83d4c03ab9f6f6ae225976b9595a1dfde61237. adding an animator (and then not managing its ticks) causes the animator to fire constantly. in this case, it was causing 100% cpu usage and forcing a compositor re-render for every frame regardless of damages
2015-10-16Eo do: Reuse stack fetching across eo functions.Tom Hacohen
This causes a significant speed up (around 10% here) and is definitely worth it. The way it's done lets the compiler cache the value across different eo_do calls, and across the parts of eo_do. Start and end. This breaks ABI.
2015-10-16Eo do: optimise getting the thread call stack for the main loop thread.Tom Hacohen
This may look like an insignificant change, but it doubles the speed of this function, and since this function is called so often, it actually improves my benchmarks by around 8%.
2015-10-16Eo: Move mainloop checks inside Eo.Tom Hacohen
This breaks ABI in a harmless way, and it will give us the ability to drastically improve Eo in the future without breaking ABI again, thus allowing us to declare Eo stable for this release if we choose to.
2015-10-16Eo events: Add a struct member marking if it's a legacy event or not.Tom Hacohen
My previous patch to this piece of code (37f84b7e966372384e2dfe5d191a6f907a17962e), caused a significant performance regression. This is such a hot path, that even accessing the strings when we don't have to slows things down drastically. It makes more sense to just store it in the structure. This commit breaks ABI (though most people probably won't even need to recompile anything else because of the memory layout). It was discussed on IRC and was decided this is a big enough issue to warrant a fix during the freeze. @fix
2015-10-16evas - add another layer->evas check in addition to layerCarsten Haitzler (Rasterman)
fixes T7215
2015-10-15evas - was missing evlogs for the threaded rendering handling - fixCarsten Haitzler (Rasterman)
2015-10-15edje: load - use evas_object_clipees_has instead of evas_object_clipees_getAmitesh Singh
evas_object_clipees_has is far cheaper than evas_object_clipees_get in case of checking if clipees exist or not. This should improve the performance in case of large set of clipees. @fix
2015-10-15evas: render - use evas_object_clipees_has instead of evas_object_clipees_getAmitesh Singh
evas_object_clipees_has is far cheaper than evas_object_clipees_get in case of checking if clipees exist or not. This should improve the performance in case of large set of clipees. @fix
2015-10-14evas: no need to compute clippees list if we just want to know if there is any.Cedric BAIL
This is a partial fix for a terrible performance scenario where application take forever to close with a large number of edje object. @fix
2015-10-14Eo: Fix confusing indentation and style.Tom Hacohen
2015-10-14ecore-wayland: Redo window animators to not use Custom source animatorsChris Michael
Summary: This moves window animators (for frame callbacks) to not use a custom animator source but rather use a timer-based source. This also moves animators to be per-window based (in that an animator is created per-window). @fix Signed-off-by: Chris Michael <cp.michael@samsung.com>
2015-10-14ecore: fix usage eina_condition_timedwait to use relative time instead of ↵Cedric BAIL
absolute time. This could lead to some very long and unexpected pause as the timeout passed to eina_condition_timedwait was passed as a absolute time instead of relative. Hopefully we don't build rocket. Notes: ref T1636
2015-10-14evas filter: silent coverity on self assignmentAmitesh Singh
This is a false alarm. Just to make coverity happy, use eo_do instead of eo_do_ret CID: 1316017
2015-10-14edje_edit: Fix clang warning of unused variablesChris Michael
Summary: clang reports that these static variables are not being used anywhere (and grep confirms this), so remove them. @fix Signed-off-by: Chris Michael <cp.michael@samsung.com>
2015-10-14ecore-drm: Add missing field initializers for vt_modeChris Michael
Summary: clang reports missing field initializers for usage of vt_mode structure, so add the missing initializers @fix Signed-off-by: Chris Michael <cp.michael@samsung.com>
2015-10-14evas: Fix clang warning for missing field 'async' in default_stateChris Michael
Evas_Object_Filter_Data has a bool for 'async' at the end of the struct. This field was missing from the initialization of 'default_state' @fix Signed-off-by: Chris Michael <cp.michael@samsung.com>
2015-10-14Eina: fix memory leak in eina_file_open()Vincent Torri
@fix
2015-10-14Eina: fix eina_file_current_directory_get()Vincent Torri
the length was not correctly computed and eina_file_path_sanitize() was was writing beyond the limit of the string @fix
2015-10-14Eina: fix spelling in eina_tmpstr documentationVincent Torri
2015-10-14Evil: set EAPI correctly in pwd.hVincent Torri
2015-10-14evas - obj destructrion - handle when obj layer is partly destroyedCarsten Haitzler (Rasterman)
this should fix T2715 and anything similar as during destruction object internals are a bit weird.
2015-10-14evas - fix evlog to push and pop properlyCarsten Haitzler (Rasterman)
2015-10-14eina: util - set tmp to NULL to fix compilation warningAmitesh Singh
lib/eina/eina_util.c: In function 'eina_environment_tmp_get': lib/eina/eina_util.c:96:7: warning: 'tmp' may be used uninitialized in this function [-Wmaybe-uninitialized] if (!tmp) tmp = "/tmp"; ^
2015-10-14Evas: Some coverity fixesJean-Philippe Andre
This are false alarms. But the explicit null check doesn't hurt. Also, Coverity doesn't like eo_do_ret. Neither do I. :) - CID 1327246 - CID 1327245
2015-10-14Evas render: Fix proxy source_clip with source imagesJean-Philippe Andre
In case the source is an Evas_Image, we skip proxy_subrender, and so the clip was not set against the source clip.
2015-10-14Evas render: Fix proxy source_clip logic inversionJean-Philippe Andre
As spotted by @FurryMyad I inverted the logic for source_clip. This should restore the proper behaviour while keeping my previous fixes working. See D2940.
2015-10-13ecore_cocoa: fix a bug where event timestamp was not retrieved correctly.Nicolas Aguirre
Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
2015-10-13ecore_cocoa: use a timer instead of a poller to be more responsive when ↵Nicolas Aguirre
dealing with cocoa Event. As all Mouse and Keyboard events comme from Cocoa, the poll period must be reduced. Backwards of this method is that when no signal are received for a long period, the timer fired anyway and consumes CPU for nothig but it seems there is no easy method to integrate NSApplication mainloop into an existing one. Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
2015-10-13ecore_cocoa: fix warnings, add Debug outputs where needed and remove printfsNicolas Aguirre
Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
2015-10-13eina: Include <unistd.h> for getuid and geteuid functionsChris Michael
Summary: If we do not include unistd.h, we end up with implicit declaration warnings when compiling @fix Signed-off-by: Chris Michael <cp.michael@samsung.com>