The number of arguments for the LiftoffCompiler has grown significantly
since its initial implementation, and it becomes hard to keep track of
all options at the call sites.
This CL refactors all optional parameters into a {LiftoffOptions} struct
which has a factory-like interface.
This will allow us to add more options in the future, e.g. for dynamic
tiering.
R=thibaudm@chromium.org
Change-Id: I66697bb2f99b676a84c158304cc3a285e1b077d0
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3069148
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Cr-Commit-Position: refs/heads/master@{#76098}
There is exactly one WasmEngine per process, hence we do not need to
store or pass a pointer to it. We just use {GetWasmEngine} (which just
reads a global variable) whenever we need it.
R=jkummerow@chromium.org
Bug: v8:11879
Change-Id: I7e0e86e326f4cafe5a894af0ff6d35803c0340a9
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2972725
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Maya Lekova <mslekova@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Cr-Commit-Position: refs/heads/master@{#75266}
The adding of base:: was mostly prepared using git grep and sed:
git grep -l <pattern> | grep -v base/vector.h | \
xargs sed -i 's/\b<pattern>\b/base::<pattern>/
with lots of manual clean-ups due to the resulting
v8::internal::base::Vectors.
#includes were fixed using:
git grep -l "src/utils/vector.h" | \
axargs sed -i 's!src/utils/vector.h!src/base/vector.h!'
Bug: v8:11879
Change-Id: I3e6d622987fee4478089c40539724c19735bd625
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2968412
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Hannes Payer <hpayer@chromium.org>
Commit-Queue: Dan Elphick <delphick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#75243}
The "num_exceptions" field of the control block must be updated before
entering the stack check, because it is used in
"GetCurrentDebugSideTable" to compute the correct indices for the debug
side table.
R=clemensb@chromium.org
Bug: chromium:1199526
Change-Id: I54f1e4244bf84d0a78b47a764fedc83b54758d01
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2831483
Commit-Queue: Thibaud Michaud <thibaudm@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#74003}
Take locals into account when computing the stack index for the next
control. Also include unwind in the list of blocks that have an implicit
exception reference on the stack.
R=ahaas@chromium.org
Bug: chromium:1183774
Change-Id: I29c67d286f1ec5efa9f2f80e13d083d6eff5836e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2794421
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Commit-Queue: Thibaud Michaud <thibaudm@chromium.org>
Cr-Commit-Position: refs/heads/master@{#73722}
Push the caught exception on the value stack, so that we can access it
from an inner catch block and rethrow it.
R=clemensb@chromium.org
Bug: v8:11453
Change-Id: Ibc5e653a07c3e4436e252c001b53bc2d3402abc9
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2739974
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Commit-Queue: Thibaud Michaud <thibaudm@chromium.org>
Cr-Commit-Position: refs/heads/master@{#73295}
In https://crrev.com/c/2707170, Liftoff was changed to only store the
ValueKind instead of the ValueType, because we only need to know kind
for code emission. For debugging though, the whole type is useful.
This CL changes the debug sidetable back to store the full type, and
retrieves this information from the decoder.
R=jkummerow@chromium.org
Bug: v8:11477
Change-Id: I08a512d24cdf0955c95f3b9261d68a02a39b9b4e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2720302
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Cr-Commit-Position: refs/heads/master@{#73068}
The precise type is only used for validation. For code generation,
knowing the kind is more than enough. Hence, only store and pass the
ValueKind in Liftoff, and not the full ValueType.
R=manoskouk@chromium.org
Bug: v8:11477
Change-Id: Ia42c0fa419f75b508bd2f210c767b631e93d3398
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2707170
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Manos Koukoutos <manoskouk@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72997}
For functions with a very large stack, the debug side table repeats a
lot of information: Most values will be spilled to the stack, still
every single entry in the debug side table repeats information about
them (type, stack offset). This leads to the size of the debug side
table to be quadratic in the size of the function.
In the linked bug, the generation of the debug side table took ~400ms,
whereas Liftoff compilation alone just took 16ms.
This CL optimized the debug side table by delta-encoding the entries,
i.e. only storing stack slots that changed. This reduces the size of the
table significantly, at the cost of making lookup slower, since that now
has to search the table backwards for the last entry that had
information about a specific slot. For now, this seems like a good
compromise. If it turns out to be a problem, we could speed up the
lookup by either forcing a full dump of the stack state after N entries,
or by dynamically inserting new entries during lookup, whenever we find
that we had to search backwards more than N entries. That would speed up
subsequent lookups then.
On the reproducer in the linked bug, this change reduces the time to
generate the debug side table from ~400ms to ~120ms.
Before this CL, the debug side table has 13,314 entries with a total of
38,599,606 stack value entries. After this CL, it shrinks to 20,037
stack value entries in the 13,314 entries (average of ~1.5 instead of
~2,899).
R=thibaudm@chromium.org
Bug: chromium:1172299
Change-Id: Ie726bb82d4c6648cc9ebd130115ee7ab3d1d551b
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2676636
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72558}
Instead of passing a bunch of objects and pointers to
{GenerateLiftoffDebugSideTable}, just pass the WasmCode pointer for
which the debug sidetable should be created.
This requires changing the corresponding cctests to actually compile
code, such that we can get a WasmCode pointer.
R=thibaudm@chromium.org
Bug: chromium:1172299
Change-Id: If42f06a545feb590f9c2377ce95e6214bbc6f566
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2674006
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72526}
We are often stepping multiple times without inspecting the state
in-between. Hence, the generated debug side table is often not being
used. Instead of always generating it, we can generate it lazily on
demand, which can avoid the need to generate it at all.
R=thibaudm@chromium.org
TEST=inspector/debugger/wasm-stepping
Bug: chromium:1172299
Change-Id: I9b9ff4485d65d720d23585856b3d672925460667
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2664446
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72484}
The opcodes were renamed long ago, but the macros were still using the
old syntax.
This CL was created using the following command (for WASM_GET_LOCAL,
WASM_SET_LOCAL, and WASM_TEE_LOCAL):
ag -l WASM_GET_LOCAL | xargs -L 1 sed -i 's/\bWASM_SET_LOCAL\b/WASM_LOCAL_SET/g'
R=ahaas@chromium.org
Bug: v8:11074
Change-Id: I0018bea185030be29344e66e59706fed183cc2f1
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2595446
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71835}
The interpreter is not an execution tier in production any more. It's
only used in tests.
Thus, remove {ExecutionTier::kInterpreter} and instead add a
{TestExecutionTier} that still has {kInterpreter}.
If needed (in {TestingModuleBuilder::execution_tier()}), we translate
back from {TestExecutionTier} to {ExecutionTier} (for {kLiftoff} and
{kTurboFan} only).
The {TraceMemoryOperation} method, which is shared between interpreter
and production code, now receives a {base::Optional<ExecutionTier>}, and
we will just pass en empty optional if called from the interpreter.
R=thibaudm@chromium.org
Bug: v8:10389
Change-Id: Ibe133b91e8dca6d6edbfaee5ffa0d7fe72ed6d64
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2335186
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#69260}
On desktop systems, we use a very basic tiering strategy: Everything is
initially compiled with Liftoff, and once that is done, the module can
start being used. Concurrently to the execution, we re-compile all code
with TurboFan, and hot-swap each function once TurboFan finishes.
We should start using a more dynamic strategy where each function is
tiered-up when judged necessary. This change will then tier-up each
liftoff function once it has been called 5 times.
I then added a counter in the native module, that is updated directly
from Liftoff code, and a runtime call is then made when the counter
reaches the goal.
R=clemensb@chromium.org
CC=thibaudm@chromium.org
Bug: v8:10728
Change-Id: I8dc2b02fdff8d97781bb1cf496886594b3d7f644
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2306803
Commit-Queue: Arnaud Robin <arobin@google.com>
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#68971}
... by migrating old-style code
MyObject* obj = new (zone) MyObject(...)
to the new style
MyObject* obj = zone->New<MyObject>(...)
Bug: v8:10689
Change-Id: I2fc4a44ea05e4d087565811f343893f0e97dc660
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2288857
Commit-Queue: Igor Sheludko <ishell@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#68789}
Before the "debug" flag was stored on the {CompilationEnv}. But each
background compilation task only gets the {CompilationEnv} once when
starting compilation, so by the time it picks up the "Liftoff for
debugging" compilation jobs, it might still compile them without the
debug flag being set. This leads to flakes in the "debug-step-into-wasm"
test, because we won't stop in the function prologue when stepping in
(because the function prologue does not check the "hook on function
call" flag if debug mode was not enabled).
This CL does not increase the size of a compilation unit, since both the
tier and the debug flag only need a single byte each.
As a nice side effect, this change allows us to remove the lock in
{CreateCompilationEnv}, because no modifyable flag is read any more.
R=thibaudm@chromium.org
Bug: v8:10410
Change-Id: Ic296ea0c4dd1d4dedde119f0536e87e5d301b5a1
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2144116
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#67115}
In production, the debug side tables will always be produced with
Liftoff in debug mode ({CompilationEnv::debug} being set).
Thus, this CL switches the tests to also test this configuration.
R=thibaudm@chromium.org
Bug: v8:10410
Change-Id: I3fa16f099d2bb612c7abf8c3ef518e1446564306
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2144119
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#67090}
In preparation for adding reference types, which need an additional
parameter to indicate the referenced type.
Bug: v8:7748
Change-Id: If4023f3d9c7f42ed603b69c43356d2e8b81a0daa
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2091471
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#66687}
This extends the debug side table to also store register locations in
addition to constants and stack values.
Previously, every value that was not constant was assumed to be spilled
to the stack. This made sense, because without breakpoints we would only
emit debug side table entries at call sites, where all registers are
spilled.
With breakpoints, this changes. At break locations, values might be live
in registers.
The logic to decide whether a value will live in the register or on the
stack is extended, because we sometimes generate the debug side table
entry at a point where the registers are not spilled yet. The debug side
table entry creation needs to account for that, and assume that these
registers will still be spilled.
R=thibaudm@chromium.org
Bug: v8:10147, v8:10222
Change-Id: I3b020dfaa29fc007047663706ee286180a996bfd
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2066960
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Cr-Commit-Position: refs/heads/master@{#66407}
This extends the debug side table test to check that the two methods
{ExecuteLiftoffCompilation} and {GenerateLiftoffDebugSideTable} return
the same debug side table.
This is important, because for code without breakpoints, we generate the
debug side table lazily via the {GenerateLiftoffDebugSideTable} method,
and it needs to match the code generated previously via
{ExecuteLiftoffCompilation}.
R=thibaudm@chromium.org
Bug: v8:10147
Change-Id: I267f599beb3fe39a1ccf22840a9d0a7f9bc5143e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2066957
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#66384}
This refactors the debug side table such that we can easily add
register information later.
In particular
- vectors for types and stack offsets are combined into one;
- constants are stored in the same vector;
- locals and operand stack values are stored in the same vector.
A follow-up CL will extend the DebugSideTable to also encode locals
or operand stack values held in registers.
R=thibaudm@chromium.org
Bug: v8:10147, v8:10222
Change-Id: I97adb56b31afdb22896530c7ba2e8a24b5d31da9
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2062405
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Cr-Commit-Position: refs/heads/master@{#66374}
The debug side table is indexed by pc offset. Offsets change if
breakpoints are added or removed, hence we cannot reuse the debug side
table when compiling another version of the function (with a different
set of breakpoints). Thus store the debug side table per code object
instead of per function.
R=thibaudm@chromium.org
Bug: v8:10147
Change-Id: Ifd77dd8f43c9b80bc4715ffe5ca8f0adca2aaf42
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2030922
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#66110}
Store the types of locals in the {DebugSideTable}, and the type of all
stack values on each entry.
Especially the stack value types would be difficult to reconstruct later
on.
R=jkummerow@chromium.org
Bug: v8:10019
Change-Id: I9b945b4e0a51166460420099908442703d3d486a
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1975759
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Cr-Commit-Position: refs/heads/master@{#65670}
For out-of-line code, we need to generate the debug side table
information at the point where the out-of-line code is being triggered,
not when it is emitted (at the end of the function).
This CL also adds more tests to check the actual content of the debug
side table in different scenarios.
R=jkummerow@chromium.org
Bug: v8:10019
Change-Id: I7714c86ee7edc4918b5ecc97cbded84c27b00e09
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1967388
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#65481}
This adds a method to generate the debug side table via Liftoff, and
adds first tests that check that the number of entries is as expected.
These tests will be extended in a follow-up CL to test the actual
content of the debug side table.
R=mstarzinger@chromium.org
Bug: v8:10019
Change-Id: I393ffabed3408463ffba232a66e2dffd7dd74f15
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1954390
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#65370}
Fix a TODO from Ben to change the macro argument order to match the
actual order in wasm code.
After this fix, we can remove the individual {WASM_CALL_INDIRECT[0-5]}
macros and implement them via a common variadic macro.
Also, rename {WASM_CALL_INDIRECT_TABLE0} to {WASM_CALL_INDIRECT_TABLE}.
The name was confusing, because this macro explictly allows to set a
table index different from 0. Thus, just drop the "0" in the name.
The individual test changes were done via a vim macro, to avoid manual
errors.
R=mstarzinger@chromium.org
Bug: v8:10021
Change-Id: I9f0f31511c5c6e20a0b07524bf75fe9cf1598eba
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1940265
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#65242}
This adds a few tests that test that the output of Liftoff is
deterministic.
These tests will be extended to test the debug side table in follow-up
CLs.
R=mstarzinger@chromium.org
Bug: v8:10019
Change-Id: Ia7992e7a889bf6dad963e5abe5b50507735996a1
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1932371
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#65190}
