Jon Coppeard
2b287fb36b
Bug 1125101 - Check budget in incremental compaction r=terrence
2015-02-26 12:35:59 +00:00
Jon Coppeard
00824bdb94
Bug 1125101 - Incrementalise compacting GC by zones r=terrence
2015-02-26 12:35:59 +00:00
Jon Coppeard
97111629f8
Bug 1130439 - Trigger compacting GCs after the user becomes inactive r=smaug r=terrence
2015-02-24 09:40:02 +00:00
Brian Hackett
f89e0a41c5
Bug 1136267 - Handle inlining when folding &&/||/?: operations in Ion, r=sunfish.
2015-02-26 06:34:04 -06:00
Nicolas B. Pierron
59f8375283
Bug 1112164 part 19 - LSRA should spill aliased registers around call-sites. r=bhackett
2015-02-26 12:18:31 +01:00
Nicolas B. Pierron
91f884d4d5
Bug 1112164 part 18 - PushRegsInMask: Do not spill SIMD register if there is no support. r=bbouvier
2015-02-26 12:18:30 +01:00
Nicolas B. Pierron
6067e57284
Bug 1112164 part 17 - StupidAllocator: Fix issues with SIMD spill and aliased registers. r=mjrosenb
2015-02-26 12:18:30 +01:00
Nicolas B. Pierron
3070e6abff
Bug 1112164 part 16 - Clean-up: Remove FloatRegister::GetSizeInBytes. r=bbouvier
2015-02-26 12:18:29 +01:00
Nicolas B. Pierron
b4c6aa82f7
Bug 1112164 part 15 - x86/x64: Add register type in the register allocator spew. r=bbouvier
2015-02-26 12:18:28 +01:00
Nicolas B. Pierron
d578d5f79f
Bug 1112164 part 14 - Add types to x86/x64 float registers. r=bbouvier,jandem
2015-02-26 12:18:28 +01:00
Nicolas B. Pierron
fa375effb5
Bug 1112164 part 13 - HandleRegisterDump operations should support all register type. r=jandem
2015-02-26 12:18:27 +01:00
Nicolas B. Pierron
8a914a468b
Bug 1112164 part 12 - Use RegisterDump size for bailout spills. r=jandem
2015-02-26 12:18:27 +01:00
Nicolas B. Pierron
dd5b6b3130
Bug 1112164 part 11 - Add an union to declare how data can overlap in one register. r=mjrosenb
2015-02-26 12:18:26 +01:00
Nicolas B. Pierron
b72eb39785
Bug 1112164 part 10 - Clean-up: Use SetType for register mask declarations. r=bbouvier
2015-02-26 12:18:25 +01:00
Nicolas B. Pierron
ef7ab2c9b8
Bug 1112164 part 9 - PushRegsInMask no longer assumes that any FloatRegister is 8 bytes. r=bbouvier
2015-02-26 12:18:25 +01:00
Nicolas B. Pierron
8c7c5e6671
Bug 1112164 part 8 - Use architecture specific SetSize. r=bbouvier
2015-02-26 12:18:24 +01:00
Nicolas B. Pierron
602486a0c0
Bug 1112164 part 7 - Add common architecture functions to query/convert a register type. r=mjrosenb
2015-02-26 12:18:23 +01:00
Nicolas B. Pierron
5aaec30202
Bug 1112164 part 6 - RegisterSets: takeAny should take one register and all aliases of it. r=mjrosenb
2015-02-26 12:18:23 +01:00
Nicolas B. Pierron
80a06dbe94
Bug 1112164 part 5 - Distinguish between the FloatRegister code and the encoding on x86 & x64. r=jandem
2015-02-26 12:18:22 +01:00
Nicolas B. Pierron
037b9b36fd
Bug 1112164 part 4 - SimdReinterpretCast: Do not use redefine as the MIRType of the input & output are different. r=sunfish
2015-02-26 12:18:21 +01:00
Nicolas B. Pierron
99b10983e5
Bug 1112164 part 3 - Add {Float,}Register::Encoding to represent what can be decoded from the assembly. r=jandem
2015-02-26 12:18:21 +01:00
Nicolas B. Pierron
1edfae06a0
Bug 1112164 part 2 - Rename ReturnSimdReg to ReturnInt32x4Reg and ReturnFloat32x4Reg. r=bbouvier
2015-02-26 12:18:20 +01:00
Nicolas B. Pierron
924048ae6f
Bug 1112164 part 1 - x86/x64 lowerForFPU: Do not reuse the input register if the MIRType are different. r=sunfish
2015-02-26 12:18:19 +01:00
Nicolas B. Pierron
52a5714abf
Bug 1112164 part 0 - PrintUse: Use AnyRegister::name method for printing the name of Fixed register uses. r=bhackett
2015-02-26 12:18:18 +01:00
Nathan Froyd
5c4535f109
Bug 1122021 - make number_constants internal to js_InitNumberClass; r=jorendorff
...
ThreadSanitizer reports write-after-write conflicts to number_constants
while initializing JSRuntimes on different threads. While an argument
could be made that these writes are entirely safe (they are, after all,
writing the same value every time), it seems better to move to a more
obviously not-racy solution.
This patch implements such a solution, by moving number_constants prior
to its only use in js_InitNumberClass, and forcing the initialization of
the "problematic" constants (infinities, NaNs, etc.) to take place the
first time js_InitNumberClass is executed.
In passing, this patch also removes the unused nc_slot enum.
2015-02-25 11:18:28 -05:00
Jarda
a4d6261b39
Bug 1127932 - IonMonkey: Inline SIMD.float32x4.add/sub/mul calls; r=bbouvier
2015-02-24 11:42:48 +01:00
Jon Coppeard
6685089a51
Bug 1136768 - Tidy detection of possibly relocated types r=terrence
2015-02-26 09:04:46 +00:00
Jon Coppeard
7c8c621e4e
Bug 1136046 - Increase maximum capacity of js::HashMap and HashSet r=luke
2015-02-26 09:02:13 +00:00
Jon Coppeard
52a1742276
Bug 1133759 - Always shrink GC buffers at the end of a shrinking GC, even if we don't compact r=terrence
2015-02-26 09:01:40 +00:00
Nicholas Nethercote
d341543d7d
Bug 979293 - Add a FrozenAtomSet to clarify how |permanentAtoms| works. r=bhackett.
...
This clarifies the two phases -- (a) initialization and (b) read-only use
-- that |permanentAtoms| goes through. It also gives some type-based protection
against potential misuse.
2015-02-25 19:11:28 -08:00
Nicholas Nethercote
af301d189b
Bug 979293 - Don't write collision bits in HashTable unnecessarily. r=luke.
...
This avoids no-op writes to the keyHash of entries when doing a no-add-lookup.
This removes a genuine data race in JSRuntime::permanentAtoms, which receives
frequent no-add-lookups from multiple threads after JSRuntime initialization
without any kind of locking.
2015-02-25 10:39:46 -08:00
Eric Faust
2958c3cabc
Bug 1066229 - Follow up: Enable |let| in ecma_6/Class/ in browser JS reftests. (rs=bz)
2015-02-25 20:15:22 -08:00
Eric Faust
dfe4980cd3
Bug 1066229 - Follow up: Push args to callVM in correct, reversed, order. (r=me) on a CLOSED TREE
2015-02-25 18:43:09 -08:00
Eric Faust
21eedc2939
Backed out changeset f32c20fcbff0 for landing with incorrect bug number on a CLOSED TREE
2015-02-25 18:41:46 -08:00
Eric Faust
abd3eab36b
Bug 106229 - Follow up: Push args to callVM in correct, reversed, order. (r=me) on a CLOSED TREE
2015-02-25 18:33:28 -08:00
Eric Faust
dbf4ddd6f6
Bug 1066229 - Tests. (r=jorendorff)
2015-02-25 17:30:08 -08:00
Eric Faust
c3812ee127
Bug 1066229 - Part 5: Emitter support for basic ES6 ClassStatements. (r=jorendorff)
2015-02-25 17:30:08 -08:00
Eric Faust
834f675127
Bug 1066229 - Part 4: Create JSOP_INITHIDDENPROP, which adds non-enumerable properties. (r=jorendorff)
2015-02-25 17:30:08 -08:00
Eric Faust
8da201c364
Bug 1066229 - Part 3: Create JSOP_INITLOCKEDDPROP, which adds non-configurable non-writable non-enumerable properties. (r=jorendorff)
2015-02-25 17:30:08 -08:00
Eric Faust
6e3a09f93d
Bug 1066229 - Part 2: Factor EmitPropertyList() out of EmitObject(). (r=jorendorff)
2015-02-25 17:30:08 -08:00
Eric Faust
10b7fa1160
Bug 1066229 - Part 1: Create a clean way to emit lexical initializers. (r=jorendorff)
2015-02-25 17:30:08 -08:00
Eric Faust
49559095cd
Bug 1066227 - Tests. (r=jorendorff)
2015-02-25 17:30:08 -08:00
Eric Faust
62d7c0fbf8
Bug 1066227 - Part 4: Reflect.parse support for ClassStatements. (r=jorendorff)
2015-02-25 17:30:08 -08:00
Eric Faust
6ef5512719
Bug 1066227 - Part 3: Parser support for basic ES6 ClassStatements (Nightly Only). (r=jorendorff)
2015-02-25 17:30:08 -08:00
Eric Faust
716c14b5e4
Bug 1066227 - Part 2: Rename objectLiteral() propertyList() in preparation for classes. (r=jorendorff)
2015-02-25 17:30:07 -08:00
Eric Faust
4bd836fdab
Bug 1066227 - Part 1: Create a clean way to create lexical bindings at initalizer sites. (r=jorendorff)
2015-02-25 17:30:07 -08:00
Eric Faust
fbd64a8199
Bug 1101903 - Part 2: Allow parsing and emitting strict mode code in smaller than script-sized units. (r=Waldo)
2015-02-25 17:30:07 -08:00
Eric Faust
19d475634f
Bug 1101903 - Part 1: Convert SharedContext::strict to a method. (r=Waldo)
2015-02-25 17:30:07 -08:00
Eric Faust
b9884814c7
Bug 1124362 - Allow strict-reserved names to be method names. (r=jorendorff)
2015-02-25 17:30:07 -08:00
Nick Fitzgerald
79562d4566
Bug 1031152 - Part 2: Define a JSAPI test for the SavedFrame public API; r=jandem
2015-02-25 17:07:05 -08:00
