for "entries", i.e. indices into the backing store (as opposed to
"public indices" going into the hash function).
This improves consistency and compiler-enforced type safety; no change
in behavior is intended.
Change-Id: I25e57e3ddcf18a406e2dfbd66786b6980c4e9615
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1852768
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64317}
This reverts commit c7c47c68f2.
Reason for revert: breaks TSAN
https://ci.chromium.org/p/v8/builders/ci/V8%20Linux64%20TSAN/28738
Original change's description:
> Reland "[runtime] Move Context::native_context to the map"
>
> This is a reland of f05bae1e0d
>
> Previously I presumed that the context read from a frame in the profiler was
> a valid context. Turns out that on non-intel we're not guaranteed that the
> frame is properly set up. In the case we looked at, the profiler took a
> sample right before writing the frame marker indicating a builtin frame,
> causing the "context" pointer from that frame to be a bytecode array. Since
> we'll read random garbage on the stack as a possible context pointer, I made
> the code reading the native context from it a little more defensive.
>
> Bug: v8:9860
>
> Original change's description:
> > [runtime] Move Context::native_context to the map
> >
> > Remove the native context slot from contexts by making context maps
> > native-context-specific. Now we require 2 loads to go from a context to the
> > native context, but we have 1 field fewer to store when creating contexts.
> >
> > Change-Id: I3c0d7c50c94060c4129db684f46a567de6f30e8d
> > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1859629
> > Commit-Queue: Toon Verwaest <verwaest@chromium.org>
> > Reviewed-by: Igor Sheludko <ishell@chromium.org>
> > Reviewed-by: Peter Marshall <petermarshall@chromium.org>
> > Reviewed-by: Maya Lekova <mslekova@chromium.org>
> > Reviewed-by: Georg Neis <neis@chromium.org>
> > Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> > Reviewed-by: Toon Verwaest <verwaest@chromium.org>
> > Cr-Commit-Position: refs/heads/master@{#64296}
>
> Change-Id: If9461e9b21d35a260d71c79d7f95e518cc429e09
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1864930
> Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> Reviewed-by: Peter Marshall <petermarshall@chromium.org>
> Reviewed-by: Igor Sheludko <ishell@chromium.org>
> Reviewed-by: Georg Neis <neis@chromium.org>
> Commit-Queue: Toon Verwaest <verwaest@chromium.org>
> Auto-Submit: Toon Verwaest <verwaest@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#64314}
TBR=ulan@chromium.org,neis@chromium.org,petermarshall@chromium.org,ishell@chromium.org,verwaest@chromium.org,mslekova@chromium.org,victorgomes@google.com
Change-Id: I4f9edc62ea6f9f5857619ff0ad1a63cab4b33cc3
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: v8:9860
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1864937
Reviewed-by: Sathya Gunasekaran <gsathya@chromium.org>
Commit-Queue: Sathya Gunasekaran <gsathya@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64316}
Port a8cdda9947
Original Commit Message:
WebAssembly locals are specified to be zero on function entry. Liftoff
implements this by just storing the constant 0 in the virtual stack for
integer types, and using one floating point register initialized to
zero for all floating point types.
For big counts of locals this leads to problems (manifesting as huge
blocks of code being generated) once we hit a merge point: All those
constants (for int) and all duplicate register uses (for floats) need to
be fixed up, by using separate registers for the locals or spilling to
the stack if no more registers are available. All this spilling
generates a lot of code, and can even happen multiple times within a
function.
This CL optimizes for such cases by spilling all locals to the stack
initially. All merges within the function body get much smaller then.
The spilled values rarely have to be loaded anyway, because the initial
zero value is usually overwritten before the first use.
To optimize the code size for initializing big numbers of locals on the
stack, this CL also introduces the platform-specific
{FillStackSlotsWithZero} method which uses a loop for bigger local
counts.
This often saves dozens of kilobytes for very big functions, and shows
an overall code size reduction of 4-5 percent for big modules.
R=xwafish@gmail.com
Change-Id: Id65b6d36beadcba0d3f3726bb6559bb316cb212e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1862830
Auto-Submit: Mu Tao <pamilty@gmail.com>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Commit-Queue: Mu Tao <pamilty@gmail.com>
Cr-Commit-Position: refs/heads/master@{#64315}
This is a reland of f05bae1e0d
Previously I presumed that the context read from a frame in the profiler was
a valid context. Turns out that on non-intel we're not guaranteed that the
frame is properly set up. In the case we looked at, the profiler took a
sample right before writing the frame marker indicating a builtin frame,
causing the "context" pointer from that frame to be a bytecode array. Since
we'll read random garbage on the stack as a possible context pointer, I made
the code reading the native context from it a little more defensive.
Bug: v8:9860
Original change's description:
> [runtime] Move Context::native_context to the map
>
> Remove the native context slot from contexts by making context maps
> native-context-specific. Now we require 2 loads to go from a context to the
> native context, but we have 1 field fewer to store when creating contexts.
>
> Change-Id: I3c0d7c50c94060c4129db684f46a567de6f30e8d
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1859629
> Commit-Queue: Toon Verwaest <verwaest@chromium.org>
> Reviewed-by: Igor Sheludko <ishell@chromium.org>
> Reviewed-by: Peter Marshall <petermarshall@chromium.org>
> Reviewed-by: Maya Lekova <mslekova@chromium.org>
> Reviewed-by: Georg Neis <neis@chromium.org>
> Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> Reviewed-by: Toon Verwaest <verwaest@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#64296}
Change-Id: If9461e9b21d35a260d71c79d7f95e518cc429e09
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1864930
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Peter Marshall <petermarshall@chromium.org>
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Commit-Queue: Toon Verwaest <verwaest@chromium.org>
Auto-Submit: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64314}
Symbol and *String classes are now declared on Torque with
generateCppClass, which means they don't use macro accessors anymore. As
such, the gen-postmortem-metadata script is not able to automatically
detect fields for those classes. Define metadata for those fields
manually for now. In the future we might want to generate it from Torque
for consistency.
Also renamed a few *String fields metadata to match the expected format
(className__fieldName__fieldType). For more context:
https://github.com/nodejs/llnode/issues/287#issuecomment-539707117.
R=bmeurer@chromium.org, hpayer@chromium.org, verwaest@chromium.org, yangguo@chromium.org
Change-Id: I82fe8315cdbfd1b8c64c6a8d5dc011b1edaec39e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1847783
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Commit-Queue: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64313}
This prevents the branch table iterator's has_next() method to trigger a
DCHECK when the decoder fails before the end of table decoding.
R=clemensb@chromium.org
Change-Id: I2258886501b77cd4c8fe98bc8a4ed0b66fb23066
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1864931
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Commit-Queue: Thibaud Michaud <thibaudm@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64312}
Using test runner with option --progress=ci can generate
output timeouts in an actual CI environment. To avoid
that we gonna write a timestamp in the standard output
at every minute.
Bug: v8:9146
Change-Id: Id2f05530956b01d9b07809e509cd0cefc0be54b2
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1863196
Commit-Queue: Liviu Rau <liviurau@chromium.org>
Reviewed-by: Tamer Tas <tmrts@chromium.org>
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64311}
DCHECKs are not really special, they just create a non-constexpr path
within an otherwise constexpr function. Since C++14, this is allowed.
Unfortunately, gcc only supports this since version 6, but we still
need to support gcc 5.
R=ulan@chromium.org
Change-Id: If74486144abafa5bbdcdbb9a567ee9295ac4cfc7
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1862568
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64310}
This fixes a bug where coverage for the inline script
<script>function foo() {}<script>
started to get deterministically reported as covered
after crrev.com/c/1771776, while before it, we most of
the time reported it as uncovered (depending on heap
order of SFIs). The correct result is to report `foo`
as uncovered as it is never called.
The problem arose from the fact that v8:9212 needed to
handle extra-wrappers around scripts correctly. Those
wrappers have the same source range as the wrapped
script and a call count of zero even if the wrapped
script is executed. To filter them out, we previously
determined nesting for identical source ranges by
ascending call count. However, in the script case above,
the script has call count one, while `foo` (which has
the same source range) has call count zero. In this
case, nesting is decreasing order of call counts.
This CL is a minimal change that sorts SFIs which are
top-level to the front, only then considers call counts
in descending order. This preserves the behavior that
node's extra wrappers are sorted to the front (and
then filtered out by existing logic), but also ensures
that for the example above, we report the script's
coverage before the coverage for `foo`.
Bug: v8:9857, v9:9212
Change-Id: Id224b0d8f12028b1f586ee5039e126bb5b8d8d36
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1863197
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Sigurd Schneider <sigurds@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64307}
Named capture properties on the groups object should be ordered by the
capture index (and not alpha-sorted). This was accidentally broken in
https://crrev.com/c/1687413.
Bug: v8:9822,v8:9423
Change-Id: Iac6f866f077a1b7ce557ba47e8ba5d7e7014b3ce
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1864829
Auto-Submit: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Peter Marshall <petermarshall@chromium.org>
Commit-Queue: Peter Marshall <petermarshall@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64306}
This is a reland of 1c56974f2a
This is a plain reland of the original CL. The original CL was speculatively
reverted, but ended up not being the cause for bot failures.
Original change's description:
> Unconditionally enable snapshot builds and remove 'v8_use_snapshot'
>
> This CL removes 'v8_use_snapshot' and the usages of the implied
> V8_USE_SNAPSHOT define. One test runner unittest was updated to use the
> "asan" variant instead of the now obsolete "no_snap" variant.
>
> Related chromium CL: https://crrev.com/c/1796325.
>
> Bug: v8:8531
> Change-Id: I5da7c9f8e9110fe7bc0f4e4f821bcb7f7d98f927
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1784282
> Commit-Queue: Simon Zünd <szuend@chromium.org>
> Reviewed-by: Tamer Tas <tmrts@chromium.org>
> Reviewed-by: Michael Achenbach <machenbach@chromium.org>
> Reviewed-by: Nico Weber <thakis@chromium.org>
> Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#64290}
TBR=thakis@chromium.org,machenbach@chromium.org,mstarzinger@chromium.org,jgruber@chromium.org,tmrts@chromium.org,szuend@chromium.org
Bug: v8:8531
Change-Id: Id75a802279238138f7aefec62e0b6425a5acc08d
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1864649
Reviewed-by: Simon Zünd <szuend@chromium.org>
Reviewed-by: Tamer Tas <tmrts@chromium.org>
Commit-Queue: Simon Zünd <szuend@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64305}
Port a8cdda9947
Original Commit Message:
WebAssembly locals are specified to be zero on function entry. Liftoff
implements this by just storing the constant 0 in the virtual stack for
integer types, and using one floating point register initialized to
zero for all floating point types.
For big counts of locals this leads to problems (manifesting as huge
blocks of code being generated) once we hit a merge point: All those
constants (for int) and all duplicate register uses (for floats) need to
be fixed up, by using separate registers for the locals or spilling to
the stack if no more registers are available. All this spilling
generates a lot of code, and can even happen multiple times within a
function.
This CL optimizes for such cases by spilling all locals to the stack
initially. All merges within the function body get much smaller then.
The spilled values rarely have to be loaded anyway, because the initial
zero value is usually overwritten before the first use.
To optimize the code size for initializing big numbers of locals on the
stack, this CL also introduces the platform-specific
{FillStackSlotsWithZero} method which uses a loop for bigger local
counts.
This often saves dozens of kilobytes for very big functions, and shows
an overall code size reduction of 4-5 percent for big modules.
R=clemensb@chromium.org, joransiu@ca.ibm.com, jyan@ca.ibm.com, michael_dawson@ca.ibm.com
BUG=
LOG=N
Change-Id: I2459080a1f6acfdd212e9a93a868d028980c5554
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1863370
Reviewed-by: Junliang Yan <jyan@ca.ibm.com>
Reviewed-by: Milad Farazmand <miladfar@ca.ibm.com>
Commit-Queue: Milad Farazmand <miladfar@ca.ibm.com>
Cr-Commit-Position: refs/heads/master@{#64301}
Previous API owners (adamk, yangguo) are less-involved with the JS API
then ever, so it seems appropriate to have an owner whose responsibility
is squarely in the JS domain.
Change-Id: I4235778fda8f55d0429e875d52ff74049024dad1
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1860328
Reviewed-by: Yang Guo <yangguo@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Commit-Queue: Adam Klein <adamk@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64300}
This reverts commit f05bae1e0d.
Reason for revert: broke arm sim debug
https://ci.chromium.org/p/v8/builders/ci/V8%20Linux%20-%20arm64%20-%20sim%20-%20debug/17714https://logs.chromium.org/logs/v8/buildbucket/cr-buildbucket.appspot.com/8899519852984476944/+/steps/Check_-_trusted/0/logs/FunctionDetailsInlining/0
Original change's description:
> [runtime] Move Context::native_context to the map
>
> Remove the native context slot from contexts by making context maps
> native-context-specific. Now we require 2 loads to go from a context to the
> native context, but we have 1 field fewer to store when creating contexts.
>
> Change-Id: I3c0d7c50c94060c4129db684f46a567de6f30e8d
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1859629
> Commit-Queue: Toon Verwaest <verwaest@chromium.org>
> Reviewed-by: Igor Sheludko <ishell@chromium.org>
> Reviewed-by: Peter Marshall <petermarshall@chromium.org>
> Reviewed-by: Maya Lekova <mslekova@chromium.org>
> Reviewed-by: Georg Neis <neis@chromium.org>
> Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> Reviewed-by: Toon Verwaest <verwaest@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#64296}
TBR=ulan@chromium.org,neis@chromium.org,petermarshall@chromium.org,ishell@chromium.org,verwaest@chromium.org,mslekova@chromium.org,victorgomes@google.com
Change-Id: Ie7b4086c3a9ab2627ecac599da36b20cf8d1f948
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1863200
Reviewed-by: Sathya Gunasekaran <gsathya@chromium.org>
Commit-Queue: Sathya Gunasekaran <gsathya@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64299}
Remove the native context slot from contexts by making context maps
native-context-specific. Now we require 2 loads to go from a context to the
native context, but we have 1 field fewer to store when creating contexts.
Change-Id: I3c0d7c50c94060c4129db684f46a567de6f30e8d
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1859629
Commit-Queue: Toon Verwaest <verwaest@chromium.org>
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Reviewed-by: Peter Marshall <petermarshall@chromium.org>
Reviewed-by: Maya Lekova <mslekova@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64296}
The tests were already passing because they happened to use objects
allocated in the lower half of the heap reservation, but this small
change should make behavior more consistent.
Change-Id: Ib6be3123d347234f4771c213f2209bfe6e19c569
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1860332
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Cr-Commit-Position: refs/heads/master@{#64294}
This reverts commit 1c56974f2a.
Reason for revert: Causes several bots to timeout, e.g. https://ci.chromium.org/p/v8/builders/ci/V8%20Linux%20-%20debug/27945
Original change's description:
> Unconditionally enable snapshot builds and remove 'v8_use_snapshot'
>
> This CL removes 'v8_use_snapshot' and the usages of the implied
> V8_USE_SNAPSHOT define. One test runner unittest was updated to use the
> "asan" variant instead of the now obsolete "no_snap" variant.
>
> Related chromium CL: https://crrev.com/c/1796325.
>
> Bug: v8:8531
> Change-Id: I5da7c9f8e9110fe7bc0f4e4f821bcb7f7d98f927
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1784282
> Commit-Queue: Simon Zünd <szuend@chromium.org>
> Reviewed-by: Tamer Tas <tmrts@chromium.org>
> Reviewed-by: Michael Achenbach <machenbach@chromium.org>
> Reviewed-by: Nico Weber <thakis@chromium.org>
> Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#64290}
TBR=thakis@chromium.org,machenbach@chromium.org,mstarzinger@chromium.org,jgruber@chromium.org,tmrts@chromium.org,szuend@chromium.org
Change-Id: I4024d818877e534b9f7908a2d14f33dca35b5924
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: v8:8531
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1862572
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64293}
We currently generate code for both separately, resulting in five
instructions (three for the i32_eqz, two for the br_if):
test rax,rax
seteq dl
movzxb rdx,rdx
test rdx,rdx
jz <label>
After this CL, we just generate two instructions:
test rax, rax
jnz <label>
This is implemented by a look-ahead in the {kExprI32Eqz} handler. If the
opcode is followed by {kExprBrIf}, no code is emitted. Instead, a flag in
the {LiftoffCompiler} is set to signal to the {kExprBrIf} handler that
the previous instruction was not processed yet.
Note that this mechanism is designed to be reusable for more similar
improvements. For the single instance implemented in this CL, it is not
needed.
Plus some drive-by cleanup.
R=jkummerow@chromium.org
Bug: v8:9831
Change-Id: I47495fe763b7db7cef41aa207c88a2f1b74bf1a9
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1862557
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64292}
AsmJs does not support SharedArrayBuffers. This CL adds a check in
instantiation and reports a proper error.
Bug: chromium:1013920
Change-Id: Id7159f23ddcc2bde139c4c97bdb67ef3dc7f0e22
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1862563
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Ulan Degenbaev <ulan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64291}
This CL removes 'v8_use_snapshot' and the usages of the implied
V8_USE_SNAPSHOT define. One test runner unittest was updated to use the
"asan" variant instead of the now obsolete "no_snap" variant.
Related chromium CL: https://crrev.com/c/1796325.
Bug: v8:8531
Change-Id: I5da7c9f8e9110fe7bc0f4e4f821bcb7f7d98f927
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1784282
Commit-Queue: Simon Zünd <szuend@chromium.org>
Reviewed-by: Tamer Tas <tmrts@chromium.org>
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Reviewed-by: Nico Weber <thakis@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64290}
We are not shipping ptr-compr in M79 on x64 because chromium:1009439
blocks 31-bit Smis on 64-bit architectures, so these's no point in
disabling double fields unboxing.
This CL will be reverted after the M79 branch point.
Bug: v8:9799, chromium:1009439
Change-Id: I28d0013d3ab06ce41d5028ba4f66c9b249de52d7
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1862556
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Commit-Queue: Igor Sheludko <ishell@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64288}
Fix uses of cached descriptors arrays used in loops that map-check
to ensure validity of the cache to also reload the descriptor in
case there are missed in-place representation updates.
As a drive-by, introduce inner HandleScopes for these loops.
Bug: chromium:1012301
Change-Id: I17273caf629a181b846d3c09777b5c08fd8cbb0e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1859621
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64287}
The reason is to unblock M79 blocked by chromium:1009439 while full
solution is not ready yet.
This CL will be reverted after the M79 branch point.
Bug: v8:9767, chromium:1009439
Change-Id: I5302d86fe953ecd94d9a4bba0d29c807b7b9d703
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1862554
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Commit-Queue: Igor Sheludko <ishell@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64286}
This replaces all left-over uses of {WasmGraphBuilder::Buffer} with
proper alternatives (e.g. using {base::SmallVector} instead).
R=clemensb@chromium.org
Change-Id: I2607ce7e2638a1bb35daccbb5b38382d5b62a430
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1859626
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64285}
Use v8::Data as basetype for managed objects that can integrate with
v8::EmbedderHeapTracer.
Bug: v8:9841
Change-Id: Id3e06701207a23870cea89e1d7d334c48fcd3006
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1856002
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64283}
WebAssembly locals are specified to be zero on function entry. Liftoff
implements this by just storing the constant 0 in the virtual stack for
integer types, and using one floating point register initialized to
zero for all floating point types.
For big counts of locals this leads to problems (manifesting as huge
blocks of code being generated) once we hit a merge point: All those
constants (for int) and all duplicate register uses (for floats) need to
be fixed up, by using separate registers for the locals or spilling to
the stack if no more registers are available. All this spilling
generates a lot of code, and can even happen multiple times within a
function.
This CL optimizes for such cases by spilling all locals to the stack
initially. All merges within the function body get much smaller then.
The spilled values rarely have to be loaded anyway, because the initial
zero value is usually overwritten before the first use.
To optimize the code size for initializing big numbers of locals on the
stack, this CL also introduces the platform-specific
{FillStackSlotsWithZero} method which uses a loop for bigger local
counts.
This often saves dozens of kilobytes for very big functions, and shows
an overall code size reduction of 4-5 percent for big modules.
R=jkummerow@chromium.org
Bug: v8:9830
Change-Id: I23fa4145847827420f09e043a11e0e7b606e94cc
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1856004
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64282}
The macros will use AVX variants when available.
Bug: v8:9561
Change-Id: I06872a08184983adc8018ffdf03916feac02016f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1857422
Reviewed-by: Deepti Gandluri <gdeepti@chromium.org>
Commit-Queue: Zhi An Ng <zhin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64280}
This switches the {WasmGraphBuilder} to use {SmallVector} for passing
argument and return values. It deprecates {WasmGraphBuilder::Buffer},
which is a dangerous construct as it makes it easy to accidentally use
the single buffer for two vectors at the same time. This also removes
the by now unused {WasmGraphBuilder::Realloc} method.
R=clemensb@chromium.org
Change-Id: I6bd9be437a99c23bb403a046a75c148ac4a14451
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1859619
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64277}
This switches the {WasmGraphBuilder::GetExceptionValues} to use a proper
vector instead of the deprecated {Buffer} method. This also addresses a
TODO about missing landing pads for the above affected method.
R=clemensb@chromium.org
Change-Id: I33ba7d712a00f2a284ec159a501bcd90e02a3a51
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1859620
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64276}
older gcc compilers (tested on PPC gcc 6) may throw the following error if
a non-copyable element is added to std::map:
src/torque/instance-type-generator.cc:192:76: required from here
/usr/include/c++/6/ext/new_allocator.h:120:4: error: use of deleted
function 'constexpr std::pair<_T1, _T2>::pair(const std::pair<_T1, _T2>&)
[with _T1 = v8::internal::torque::{anonymous}::InstanceTypeTree* const; _T2 =
std::unique_ptr<v8::internal::torque::{anonymous}::InstanceTypeTree>]'
explicitly constructing the std::pair will fix the compilation error.
Bug: v8:9850
Change-Id: I1e69e804be8bb9c16d013a90b532d670a97ed055
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1857552
Commit-Queue: Tobias Tebbi <tebbi@chromium.org>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64275}
With exception handling enabled new call paths open up, which will
perform environment merging while a "call" or "call_indirect" is
currently being emitted. This will lead to double-use of the buffer
returned by calls to {Buffer} or {Realloc}. In general we should
transition away from this optimization to safer constructs such as
{base::SmallVector} to avoid such bugs.
R=clemensb@chromium.org
TEST=mjsunit/regress/regress-9832
BUG=v8:9832
Change-Id: I4c862ac1bc7dc34ad62279c82f6414153e8cbddb
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1856006
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64271}
The assert code gets put into an unreachable block in release builds
to make sure it's type-checked and torque knows the code it contains is
used, but still it doesn't emit actual machine code.
Bug: v8:7793
Change-Id: I580fdd7ac059e0dbe85283fd35c3038634a7228e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1857226
Reviewed-by: Maya Lekova <mslekova@chromium.org>
Commit-Queue: Maya Lekova <mslekova@chromium.org>
Auto-Submit: Tobias Tebbi <tebbi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64270}
