Commit Graph

1732 Commits

Author SHA1 Message Date
ulan@chromium.org
6a6012c0e7 Emit RSB for subtraction with constant lhs on ARM
Instead of generating the following for "100 - x":

  mov r1, #100
  subs r0, r1, r0

this patch makes Crankshaft generate this:

  rsbs r0, r0, #100

thus saving a register and a mov.

BUG=

Review URL: https://chromiumcodereview.appspot.com/11308040
Patch from Hans Wennborg <hans@chromium.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13017 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-20 14:46:16 +00:00
jkummerow@chromium.org
d9ba8ed8a1 Fix register confusion in non-VFP3 BinaryOpStubs on ARM
Review URL: https://codereview.chromium.org/11413014

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12980 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-16 08:54:01 +00:00
svenpanne@chromium.org
d5f589808d Removed a bunch of GetExistingThreadLocal calls by threading the Isolate.
For Octane, the number of calls go down from 7341629 to 1947880, i.e. they are
reduced by more than 73%. TLS access is not especially cheap, so this exercise
seems worthwhile.

Review URL: https://codereview.chromium.org/11412007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12979 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-16 08:38:11 +00:00
yangguo@chromium.org
3699616609 Rename SeqAsciiString
This is a straight rename:

NewRawAsciiString -> NewRawOneByteString
SeqAscii -> SeqOneByte

SeqOneByteString cannot yet take non-ascii data.

R=yangguo@chromium.org,
BUG=

Review URL: https://chromiumcodereview.appspot.com/11411005
Patch from Dan Carney <dcarney@google.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12972 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-15 13:31:27 +00:00
verwaest@chromium.org
a08194c83a Support all fast elements kinds in the major array operations.
Currently missing support for unshift.

BUG=

Review URL: https://chromiumcodereview.appspot.com/11377132

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12969 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-15 12:19:14 +00:00
yangguo@chromium.org
7fc1751821 Make kAsciiDataHintTag have correct semantics for all both 1 and 2 byte string types.
BUG=

Review URL: https://chromiumcodereview.appspot.com/11360251
Patch from Dan Carney <dcarney@google.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12966 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-15 10:29:11 +00:00
jkummerow@chromium.org
1c086d1202 Lattice-based representation inference, powered by left/right specific type feedback for BinaryOps and comparisons
Review URL: https://chromiumcodereview.appspot.com/10837165

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12961 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-14 15:59:45 +00:00
ulan@chromium.org
74492ab2d4 Emit VMLA for multiply-add on ARM
Review URL: https://chromiumcodereview.appspot.com/11293061
Patch from Hans Wennborg <hans@chromium.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12958 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-14 11:01:18 +00:00
mmassi@chromium.org
ce682a2489 Allow property indexes to refer to slots inside the object header.
BUG=

Review URL: https://chromiumcodereview.appspot.com/11365221

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12944 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-13 11:07:04 +00:00
mvstanton@chromium.org
50bd6e9092 Revert change r12912 (http://code.google.com/p/v8/source/detail?r=12912) due to x64 perf issue
R=verwaest@chromium.org
BUG=

Review URL: https://codereview.chromium.org/11312206

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12943 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-13 10:53:34 +00:00
yangguo@chromium.org
827f8919f4 Reverting r12937.
R=verwaest@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/11293246

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12942 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-13 10:36:20 +00:00
yangguo@chromium.org
ce884e13e3 Make kAsciiDataHintTag have correct semantics for all both 1 and 2 byte string types.
BUG=

Review URL: https://chromiumcodereview.appspot.com/11361171

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12937 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-12 14:57:59 +00:00
svenpanne@chromium.org
237684fa45 Refactoring: Make emit_debug_code flag handling architecture-independent.
Review URL: https://codereview.chromium.org/11312165

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12921 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-09 13:43:48 +00:00
svenpanne@chromium.org
a28bc2fd97 Refactoring: Make predictable code flag handling architecture-independent.
Review URL: https://codereview.chromium.org/11359127

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12920 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-09 13:10:10 +00:00
mvstanton@chromium.org
803d8eada1 Some improvements in register usage in lithium compilation of LoadKeyed/StoreKeyed operations.
BUG=

Review URL: https://codereview.chromium.org/11365084

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12912 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-09 10:21:07 +00:00
danno@chromium.org
bd4e114b8e Add code again to allow reclaiming old unexecuted functions.
When code objects in the heap for FUNCTIONs and OPTIMIZED_FUNCTIONs are marked by the GC, their prologue is patched with a call to a stub that removes the patch. This allows the collector to quickly identify code objects that haven't been executed since the last full collection (they are the ones that sill contain the patch). The functionality is currently disabled, but can be activated by specifying the "--age-code".

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/10837037

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12898 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-08 12:18:11 +00:00
yangguo@chromium.org
943c10bb87 Rename kAsciiStringTag to kOneByteStringTag
This is just a rename. After commit, I'll begin with the semantic changes.
Until those are complete, kOneByteStringTag will have the same meaning as
kAsciiStringTag.

BUG=

Review URL: https://chromiumcodereview.appspot.com/11293168
Patch from Dan Carney <dcarney@google.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12897 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-08 12:14:29 +00:00
mstarzinger@chromium.org
a31889e2de Fix slack tracking when instance prototype changes.
This fixes a corner case when the instance prototype of a function is
changed while inobject slack tracking is still in progress. This caused
the intial map to be unrelated for functions with the same shared info
and hence the shared construct stub is no longer generic enough to work
for all those functions.

R=danno@chromium.org
BUG=chromium:157019
TEST=mjsunit/regress/regress-crbug-157019

Review URL: https://codereview.chromium.org/11293059

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12896 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-08 11:56:44 +00:00
danno@chromium.org
bbcfa17134 Use movw/movt on ARM to load constant roots
BUG=none
TEST=none

Review URL: https://codereview.chromium.org/11307012
Patch from Jay Conrod <dconrod@codeaurora.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12895 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-08 10:32:39 +00:00
danno@chromium.org
63ada1da1b Move vldr in DoLoadKeyedFastDoubleElement before the hole check
This reschedules instructions to better absorb VFP load latency.
BUG=none
TEST=none

Review URL: https://codereview.chromium.org/11363086
Patch from Rajeev Krithivasan <rkrithiv@codeaurora.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12890 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-07 21:38:46 +00:00
ulan@chromium.org
101d64c1a6 Simplify some of the startup code for SubStringStub::Generate.
R=ulan@chromium.org

Review URL: https://chromiumcodereview.appspot.com/11098043
Patch from JF Bastien <jfb@chromium.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12887 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-07 13:22:03 +00:00
ulan@chromium.org
f106c9c9f7 Add rotate-right instruction to hydrogen and use it instead of bitwise operations
of the form ((x >>> i) | (x << (32 - i))).

This CL is based on https://chromiumcodereview.appspot.com/10984057/
by Jay Conrod <dconrod@codeaurora.org>.

R=danno@chromium.org,mstarzinger@chromium.org,dconrod@codeaurora.org

Review URL: https://chromiumcodereview.appspot.com/11033005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12855 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-05 13:28:10 +00:00
svenpanne@chromium.org
9de1d40d28 Improve handling of property loads on the proto chain.
Previously Crankshaft emitted a generic load for these, now we emit a load of a
named field, guarded by a proto chain check.

LCheckPrototypeMaps now returns the holder, which is for free, because it
already had to check its map as the last step, anyway. This is in sync with what
StubCompiler::CheckPrototype does.

Review URL: https://codereview.chromium.org/11338030

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12847 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-05 08:53:54 +00:00
mvstanton@chromium.org
a85fd03caa Consolidated all the key store/load classes in the Hydrogen and Lithium
space into just two:
HLoadKeyed/HLoadKeyedGeneric and HStoreKeyed/HStoreKeyedGeneric
LLoadKeyed/LLoadKeyedGeneric and LStoreKeyed/LStoreKeyedGeneric
BUG=

Review URL: https://codereview.chromium.org/11238016

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12839 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-02 09:18:53 +00:00
danno@chromium.org
5da0bbf390 Loosen aligned code target requirement on ARM
Fixes crashes when V8 is built as Thumb code.

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/11343014

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12832 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-29 16:27:54 +00:00
danno@chromium.org
e96d49047b Remove redundant jump to deoptimization (the jump already exists in DoCheckMapCommon)
BUG=none
TEST=none

Review URL: https://codereview.chromium.org/11265044
Patch from Rajeev Krithivasan <rkrithiv@codeaurora.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12828 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-29 11:45:40 +00:00
ulan@chromium.org
d247c5e6c7 Change constant pool marker to be the unconditional, permanently undefined instruction.
The previously-used instruction isn't guaranteed to always be undefined,
and the encoding used was conditional (failing the condition on an
undefined instruction is itself undefined and not guaranteed to
fault!). I would have like to use a more clever encoding (see bug 2963),
but we need the extra bits to encode the size of the constant pool.

BUG=security
R=ulan@chromium.org

Review URL: https://chromiumcodereview.appspot.com/11242002
Patch from JF Bastien <jfb@chromium.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12791 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-22 15:25:17 +00:00
ulan@chromium.org
18ff15ca8c Speed up function deoptimization by avoiding quadratic pass over optimized function list.
R=danno@chromium.org
BUG=155270

Review URL: https://chromiumcodereview.appspot.com/11189091

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12780 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-22 09:48:56 +00:00
verwaest@chromium.org
c9e86f4b69 Remove scratch register requirement from LoadInstanceDescriptors on arm and mips.
Review URL: https://chromiumcodereview.appspot.com/11193022

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12776 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-19 12:39:59 +00:00
danno@chromium.org
ace5bf3f94 Fix ARM dissambler test problems with movw/movt.
R=ulan@chromium.org

Review URL: https://codereview.chromium.org/11198061

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12771 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-19 10:44:56 +00:00
danno@chromium.org
4431e0c54e Put more constants in movw/movt instructions
R=ulan@chromium.org

Review URL: https://codereview.chromium.org/11185052

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12759 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-18 14:37:04 +00:00
danno@chromium.org
43e248d2e5 Use movw/movt instead of constant pool on ARMv7.
Some ARM architectures load 32-bit immediate constants more efficiently using movw/movt pairs rather than constant pool loads. This patch allows the assembler to generate one or the other load form at runtime depending on what is faster.

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/11037023

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12755 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-18 12:21:42 +00:00
verwaest@chromium.org
72424b3987 Move DescriptorArray into the map.
Review URL: https://chromiumcodereview.appspot.com/11188031

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12746 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-17 13:04:49 +00:00
ulan@chromium.org
a86e897924 Simulate and disasm NOP on ARM
R=ulan@chromium.org

Review URL: https://chromiumcodereview.appspot.com/11116011
Patch from JF Bastien <jfb@chromium.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12737 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-15 15:42:09 +00:00
ulan@chromium.org
86eedc26b5 Allow unaligned accesses for ARMv7.
R=mstarzinger@chromium.org,jfb@chromium.org

Review URL: https://chromiumcodereview.appspot.com/11087047

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12735 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-15 15:19:36 +00:00
svenpanne@chromium.org
5a0176222b Consistently make the bounds check for AccessArgumentsAt explicit.
This has the advantage that AccessArgumentsAt itself can't deopt anymore and the
bounds check is visible for the elimination phase. Furthermore, things are
simply more consistent now, a good thing in itself. :-)

Review URL: https://codereview.chromium.org/11106012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12721 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-15 07:25:20 +00:00
danno@chromium.org
bf83c5fe32 Use immediate add when possible in space allocator
Save one instruction in allocating new space by using an immediate add if
possible to calculate the new top of heap.

BUG=

Review URL: https://chromiumcodereview.appspot.com/11091068
Patch from Anthony Berent <aberent@chromium.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12718 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-12 14:06:03 +00:00
svenpanne@chromium.org
1f3bdabd80 Ensure that we do not accidentally emit debug code.
The call sites are more concise, the naming is better and at least one
case of superfluous code in release mode was removed.

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/11108011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12712 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-12 11:09:14 +00:00
danno@chromium.org
0457a3be0a DoNumberTagD performance improvement
Allocate heap entry untagged and tag at end to avoid having to subtract off
the tag offset before storing the value.

BUG=

Review URL: https://codereview.chromium.org/11028115
Patch from Anthony Berent <aberent@chromium.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12711 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-12 10:58:25 +00:00
ulan@chromium.org
ede725d2f7 Fix vstr.cc S, [mem] to emit vstr instead of vldr.
BUG=emits wrong instruction
R=ulan@chromium.org

Review URL: https://chromiumcodereview.appspot.com/11118015
Patch from JF Bastien <jfb@chromium.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12710 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-12 09:27:12 +00:00
verwaest@chromium.org
ad6a21c0d0 Remove descriptors pointer.
Secondary changes:
- don't transfer ownership back on CNLT
- turned debugging checks back into ASSERT

Review URL: https://chromiumcodereview.appspot.com/11099064

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12699 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-11 12:01:19 +00:00
mstarzinger@chromium.org
270316e953 Improve page flag checking sequence on ARM.
R=ulan@chromium.org

Review URL: https://codereview.chromium.org/11090021

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12682 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-09 15:12:39 +00:00
danno@chromium.org
0a41d2731c ARM: Fast path for integer inputs to EmitVFPTruncate
BUG=none
TEST=none

Review URL: https://codereview.chromium.org/11049025
Patch from Martyn Capewell <m.m.capewell@googlemail.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12676 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-08 12:50:15 +00:00
ulan@chromium.org
a80f7f42dd Make sure all labels are bound and linked in ARM DoModI.
BUG=v8:2354

TBR=danno@chromium.org,subratokde@codeaurora.org

Review URL: https://chromiumcodereview.appspot.com/11038017

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12656 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-02 16:48:45 +00:00
danno@chromium.org
08747c3a16 Support for SDIV and MLS ARM instructions, and implement DoModI using them.
Also added support for the runtime detection to check if hardware supports SDIV/UDIV
Other new opportunities to exploit SDIV/UDIV will be done in separate issues.

Review URL: https://chromiumcodereview.appspot.com/10977051
Patch from Subrato K De <subratokde@codeaurora.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12646 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-10-01 21:27:33 +00:00
erik.corry@gmail.com
fbcc4a408e Add the VFP-ness to the minor number of the keyed store elements
IC so that the version from the snapshot is not used if we have
a more capable CPU at runtime.
Review URL: https://chromiumcodereview.appspot.com/10984065

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12624 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-09-27 11:31:26 +00:00
erik.corry@gmail.com
dc57f54548 Fix ARM bug introduced in r12604 that caused crashes on ARM
on crypto-md5 from SunSpider.  Bug=152402
Review URL: https://chromiumcodereview.appspot.com/10991045

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12623 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-09-27 07:45:49 +00:00
erik.corry@gmail.com
e8ffc2bebd Make the speed of incremental marking depend also on the rate
at which we are hitting expensive write barrier operations,
not just on the rate of allocation.
Review URL: https://chromiumcodereview.appspot.com/10974003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12618 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-09-26 11:35:42 +00:00
danno@chromium.org
c467b2659f ARM: Small optimisation of VFP immediate creation
Save an instruction in VFP immediate creation by passing a scratch register.

BUG=none
TEST=none

Review URL: https://chromiumcodereview.appspot.com/10990024
Patch from Martyn Capewell <m.m.capewell@googlemail.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12605 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-09-25 14:32:07 +00:00
erik.corry@gmail.com
72e9f1bea1 x64 and ARM: Fix issue 2346 (order of operations in keyed store
on arrays) and turn get-own-property-descriptor.js test into
a regression test.
Review URL: https://chromiumcodereview.appspot.com/10985017

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12604 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-09-25 13:35:42 +00:00