Andrew McCreight
357b77dcd1
Bug 963844 - Scale ICC budget by time between slices. r=smaug
...
If we exceed the desired length of time between ICC slices, proportionally scale up the slice budget to try to ensure a constant amount of collection work is done per unit of time, even when the browser is under heavy load.
As part of this, track the time the last slice from the current ICC ended.
2014-03-06 10:54:37 -08:00
Luke Wagner
38f3e59b91
Bug 963588 - asmjscache: place cache entries apps that request AOT compilation in persistent storage (r=janv)
2014-03-05 14:47:10 -06:00
Andrew McCreight
b6bf1f0d1f
Bug 979047, part 5 - Make sure we have at least one early forgetSkippable timer fire. r=smaug
...
With ICC, ccDelay can get close to NS_CC_SKIPPABLE_DELAY, but we want to always want to run at least one forgetSkippable before the actual CC.
2014-03-04 16:21:48 -08:00
Andrew McCreight
23dad0ee17
Bug 979047, part 4 - Run ICCSliceTime() after PrepareForCollectionSlice. r=smaug
...
This way we take into account time spent finishing off a GC, or something.
2014-03-04 16:21:48 -08:00
Andrew McCreight
e7cd320eb1
Bug 979047, part 3 - Make sure we trigger a GC after a current ICC when we poke the GC. r=smaug
2014-03-04 16:21:48 -08:00
Andrew McCreight
b3d03a560b
Bug 979047, part 2 - Remove unused endGCTime assignments. r=smaug
2014-03-04 16:21:48 -08:00
Andrew McCreight
a5c88f27d1
Bug 979047, part 1 - Only clear mBeginSliceTime in EndCycleCollectionCallback. r=smaug
...
Run FinishCycleCollectionSlice only once per slice, and only if we've previously run PrepareForCycleCollectionSlice.
2014-03-04 16:21:48 -08:00
Ed Morley
1f3bd2e264
Backed out changeset 508848ad378a (bug 939562) for causing bug 978450
2014-03-03 16:54:26 +00:00
Jan de Mooij
c741c3c67d
Bug 939562 part 3 - Move JIT flags from ContextOptions to RuntimeOptions. r=bent,bholley,luke
2014-02-26 10:25:36 +01:00
Carsten "Tomcat" Book
1d6ff04f98
Backed out changeset 802aa43ae8cc (bug 939562) for PGO mochitest regression on a CLOSED TREE
2014-02-26 16:02:15 +01:00
Jan de Mooij
fc5bf6c865
Bug 939562 part 2 - Move JIT flags from ContextOptions to RuntimeOptions. r=bent,bholley,luke
2014-02-26 10:25:36 +01:00
Luke Wagner
23fb13f948
Bug 976348 - Change DescribeScriptedCaller to not return a JSScript (r=bholley)
2014-02-25 09:43:14 -06:00
Boris Zbarsky
1985f8ead4
Bug 972312 part 2. Get rid of InternalScriptErrorEvent and just use mozilla::dom::ErrorEvent for the cases that used to use it. r=smaug,khuey
2014-02-23 00:01:12 -05:00
Andrew McCreight
21296781d2
Bug 972856 - Use PR_Now() for CC timestamp in JSON log. r=smaug
2014-02-18 09:58:40 -05:00
Julian Seward
a86b0ede10
Bug 960603 - Uninitialised value use relating to nsJSContext::EndCycleCollection. r=mccr8.
2014-02-17 13:16:23 +01:00
Luke Wagner
5ca1cad567
Bug 936236 - Send memory-pressure notification if ArrayBuffer allocation OOMs and retry allocation (r=mccr8)
2014-02-12 22:50:15 -06:00
Phil Ringnalda
64f06206bc
Back out d54433699f2e (bug 936236) for adding 5 rooting hazards
...
CLOSED TREE
2014-02-12 23:26:28 -08:00
Luke Wagner
eb908c8106
Bug 936236 - Notify memory-pressure observers when allocating a large ArrayBuffer fails (r=mccr8)
2014-02-12 22:50:15 -06:00
Jon Coppeard
0df72e25fb
Bug 969812 - Convert JS_NewArrayObject to use HandleValueArray r=terrence r=bz
2014-02-12 10:50:46 +00:00
Jon Coppeard
da8e9bd5e9
Bug 965830 - Remove AutoArrayRooter from the public API r=terrence r=smaug
2014-02-11 10:59:15 +00:00
Andrew McCreight
f5f5f9da17
Bug 965916 - Do an extra GC on memory pressure if it would be productive. r=smaug
2014-02-03 07:31:11 -08:00
Terrence Cole
11a9a45a94
Bug 963738 - Fix some false positives in the Array analysis; r=jonco
2014-01-28 12:05:35 -08:00
Chris Peterson
a16cac4d58
Bug 960416 - Always use JIT hardening. r=luke
2013-12-25 00:21:59 -08:00
Terrence Cole
4a53feb64b
Bug 959787 - Handlify several JSAPI interfaces that can GC, Part 2; r=sfink
2014-01-14 17:19:07 -08:00
Tom Schuster
4a7dd51a60
Bug 939294 - Change xpidl jsval to handles. r=gabor,bz,khuey,bsmedberg,terrence
2014-01-09 18:39:36 +01:00
Boris Zbarsky
98f529fe13
Bug 959926. Fix some miscellaneous unsafe pointer hazards. r=terrence
2014-01-15 14:39:08 -05:00
Till Schneidereit
d05d57c5f9
Bug 958726 - Perform shrinking GC if nsJSContext::GarbageCollectNow is called with aShrinking == ShrinkingGC. r=billm
2014-01-10 23:44:28 +01:00
Andrew McCreight
d1ca98f1f4
Bug 957370 - Make 'total time' measure total time actually running the CC. r=smaug
2014-01-07 17:33:47 -08:00
Andrew McCreight
d721fa69df
Bug 956455, part 1 - WrapNative's holder argument is unnecessary. r=smaug
2014-01-06 11:54:43 -08:00
Ehsan Akhgari
68bfe70f09
Bug 927728 - Part 1: Replace PRUnichar with char16_t; r=roc
...
This patch was automatically generated by the following script:
#!/bin/bash
# Command to convert PRUnichar to char16_t
function convert() {
echo "Converting $1 to $2..."
find . ! -wholename "*nsprpub*" \
! -wholename "*security/nss*" \
! -wholename "*modules/libmar*" \
! -wholename "*/.hg*" \
! -wholename "obj-ff-dbg*" \
! -name prtypes.h \
! -name Char16.h \
-type f \
\( -iname "*.cpp" \
-o -iname "*.h" \
-o -iname "*.c" \
-o -iname "*.cc" \
-o -iname "*.idl" \
-o -iname "*.ipdl" \
-o -iname "*.ipdlh" \
-o -iname "*.mm" \) | \
xargs -n 1 sed -i -e "s/\b$1\b/$2/g"
}
convert PRUnichar char16_t
2014-01-04 10:02:17 -05:00
Andrew McCreight
3af27f9483
Bug 950959 - Rename scheduled cycle collector stuff to slice. r=smaug
2014-01-01 11:00:35 -08:00
Andrew McCreight
b3b5eec196
Bug 950949 - Run forgetSkippable per CC not per slice. r=smaug
2014-01-01 11:00:35 -08:00
Andrew McCreight
bcb3fc886f
Bug 948554, part 3 - Capture max slice time in logs and telemetry. r=smaug
2013-12-31 09:20:21 -08:00
Andrew McCreight
35974ba765
Bug 948554, part 2 - Convert some of nsJSEnvironment to TimeStamp for greater precision. r=smaug
2013-12-31 09:20:21 -08:00
Andrew McCreight
a9e612e653
Bug 948554, part 1 - Rename PrepareForCycleCollection and make it a method. r=smaug
...
This name is more precise, because we must call it once per slice,
not once per CC. Being a method on the CC stats object is more consistent.
2013-12-31 09:20:21 -08:00
Andrew McCreight
32e8958781
Bug 937960, part 5 - Add pref to activate ICC. r=smaug
2013-12-22 10:14:36 -05:00
Andrew McCreight
859bd06975
Bug 937960, part 4 - Reduce time until CC when it is incremental. r=smaug
2013-12-22 09:58:20 -05:00
Andrew McCreight
27829b6a48
Bug 937960, part 3 - Pass in a small time budget with ICC. r=smaug
2013-12-22 09:58:19 -05:00
Andrew McCreight
0ff7ed0343
Bug 937960, part 2 - Add ICC slice timer and scheduling. r=smaug
2013-12-21 09:35:08 -05:00
Andrew McCreight
4711d84a71
Bug 937960, part 1 - Add static var for incremental CC. r=smaug
2013-12-21 09:35:08 -05:00
Ryan VanderMeulen
bc4d191fcb
Backed out changesets ccdc3d4f4571, 4dcc91e771e3, 0ae14946314b, and 94d22ab0b17f (bug 937960) for Windows 7 debug xpcshell failures.
2013-12-19 16:57:55 -05:00
Andrew McCreight
6a7d8e00d3
Bug 937960, part 4 - Reduce time until CC when it is incremental. r=smaug
2013-12-18 21:22:55 -08:00
Andrew McCreight
95a44e0019
Bug 937960, part 3 - Pass in a small time budget with ICC. r=smaug
2013-12-18 21:22:55 -08:00
Andrew McCreight
eec3b7433a
Bug 937960, part 2 - Add ICC slice timer and scheduling. r=smaug
2013-12-18 21:22:55 -08:00
Andrew McCreight
167c393a3d
Bug 937960, part 1 - Add ICC pref to nsJSEnv. r=smaug
2013-12-18 21:22:55 -08:00
Ryan VanderMeulen
8b097825dd
Backed out changesets b03bd1170d1c, 419615a1195d, a1b699b80442, and 524c573fd8a6 (bug 937960) for Win7 debug xpcshell orange.
2013-12-18 19:04:11 -05:00
Ryan VanderMeulen
424df5d1b8
Backed out changesets 472a1be79a9f, 2b43b310858b, and adc20d941bd8 (bug 948554) for Win7 debug xpcshell timeouts.
2013-12-18 19:03:28 -05:00
Andrew McCreight
ffa8faa847
Bug 948554, part 3 - Capture max slice time in logs and telemetry. r=smaug
2013-12-18 11:42:16 -08:00
Andrew McCreight
f64cccb0fa
Bug 948554, part 2 - Convert some of nsJSEnvironment to TimeStamp for greater precision. r=smaug
2013-12-18 11:42:16 -08:00
Andrew McCreight
fa328d2440
Bug 948554, part 1 - Rename PrepareForCycleCollection and make it a method. r=smaug
...
This name is more precise, because we must call it once per slice,
not once per CC. Being a method on the CC stats object is more consistent.
2013-12-18 11:42:16 -08:00