inspector tests are not yet enabled as a few issues need to be
solved related to endianness and IBM lane numbering.
Change-Id: I3c5c87af45037033d4a112cb7e1da8b7923ce0f3
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3489244
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Commit-Queue: Milad Farazmand <mfarazma@redhat.com>
Cr-Commit-Position: refs/heads/main@{#79306}
Maglev is currently enabled only on x64; BUILD.gn knows about this
through v8_enable_maglev, but v8_check_header_includes doesn't. It
thus tries to compile maglev files on platforms that don't have maglev
support yet.
Add an explicit maglev exclude rule until we support other platforms.
Bug: v8:7700
Change-Id: Iac991741c7d630dc4ed9f9fbf6df7656853cc743
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3494536
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Auto-Submit: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79305}
This was originally introduced to address http://crbug.com/794941,
to make the disassembly generation for WebAssembly modules lazy.
Nowadays we no longer generate a text representation for the Wasm
disassembly in V8, and this method always returns `false`.
Bug: chromium:794941, chromium:1162229
Change-Id: I8b67e451a3657bf732615585577525aeea2b2f55
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3494236
Commit-Queue: Benedikt Meurer <bmeurer@chromium.org>
Auto-Submit: Benedikt Meurer <bmeurer@chromium.org>
Reviewed-by: Simon Zünd <szuend@chromium.org>
Commit-Queue: Simon Zünd <szuend@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79304}
Otherwise std::allocator is used and there is no conversion.
Bug: chromium:819294
Change-Id: Ic93e75a3facef96dc901dda29a6be3b4539b68e7
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3492523
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Stephan Hartmann <stha09@googlemail.com>
Cr-Commit-Position: refs/heads/main@{#79302}
Nowadays these methods allocate a single handle, so there's no point in
having a dedicated handle scope just to close it immediately and escape
the single allocated handle.
Bug: chromium:1162229
Change-Id: I695d8c5577db43b8974b28bdfa6e0600eb41cce9
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3494156
Commit-Queue: Benedikt Meurer <bmeurer@chromium.org>
Auto-Submit: Benedikt Meurer <bmeurer@chromium.org>
Reviewed-by: Simon Zünd <szuend@chromium.org>
Commit-Queue: Simon Zünd <szuend@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79301}
The CL gets rid of the assumption that WeakCallbackItem::parameter
always points to an object with a custom callback, which might not be
the case for weak containers. This fixes failures in cppgc-non-default
bot.
Bug: chromium:1029379
Change-Id: I9ab6cb357153801efaa0d902c9eedff07d12ffe7
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3490932
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Anton Bikineev <bikineev@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79296}
This CL assures scratch registers are passed where needed
and cleans up Simd functions shared between TF and LO.
Change-Id: Ib7633e0d51f3aa92d2bcdfc69d0efe779062af62
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3489239
Reviewed-by: Junliang Yan <junyan@redhat.com>
Commit-Queue: Milad Farazmand <mfarazma@redhat.com>
Cr-Commit-Position: refs/heads/main@{#79295}
This reverts commit 679c373eb2.
Reason for revert: DCHECK failures in post_job.cc when running Chromium gtests, blocking V8 roll into Chromium.
See https://ci.chromium.org/ui/p/chromium/builders/try/mac-rel/928210/overview for example failures/stacks.
Original change's description:
> Reland "heap: Force incremental marking in C++ only workloads"
>
> This is a reland of 4fde332811
>
> Another Blink-related test fix landed, see chromium:1300492.
>
> Original change's description:
> > heap: Force incremental marking in C++ only workloads
> >
> > ... when above a certain minimum threshold. This is to guard against
> > memory running away in scenarios where the JS heap is empty and
> > there's only high throughput C++ allocations that don't allow for a
> > memory reducer GC to kick in.
> >
> > This logic should be revisited after Oilpan's young generation
> > collector is implemented which may allow switching to a more efficient
> > shrinking strategy for initial heap setup.
> >
> > Bug: chromium:1029379, chromium:1300028, chromium:1300492
> > Change-Id: I93924fc2fe77d6226c29358d3afb1cc9d6fbf3b1
> > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3484319
> > Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
> > Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
> > Cr-Commit-Position: refs/heads/main@{#79255}
>
> Bug: chromium:1029379, chromium:1300028, chromium:1300492
> Change-Id: Ida66e0c944094472b4856d5fecef2d199d29549b
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3487991
> Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
> Auto-Submit: Michael Lippautz <mlippautz@chromium.org>
> Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
> Commit-Queue: Dominik Inführ <dinfuehr@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#79281}
Bug: chromium:1029379, chromium:1300028, chromium:1300492
Change-Id: Ifbdcf996a91cbdb5cce3b07059c333a7f1a9444a
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3491187
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Commit-Queue: Adam Klein <adamk@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79294}
The main change is the section name, which is now 'metadata.code.branch_hint'.
The binary format has also a couple of minor changes.
Semantics remain unchanged.
Change-Id: I056c9f672ae494979e8ea55266fa766139b71d38
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3487788
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79292}
After https://crrev.com/c/3484317, passing {nullptr} to the
{CodeSpaceWriteScope} won't work any more. Since the tests do not have a
{NativeModule} to pass instead, make them use
{pthread_jit_write_protect_np} directly.
The jump-table assembler tests have dedicated threads for writing and
executing the code, so we just switch once per thread. The icache test
switches between writing and executing, so we use a little struct for
switching.
R=jkummerow@chromium.org, tebbi@chromium.org
Bug: v8:12644, v8:11974
Change-Id: I116f3ad75454f749cdc4635802a4617ff91548b2
Cq-Include-Trybots: luci.v8.try:v8_mac_arm64_rel_ng
Cq-Include-Trybots: luci.v8.try:v8_mac_arm64_dbg_ng
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3487995
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79290}
Instead of storing a single register, store the entire RegList. This
simplifies a lot of things. We will use RegLists for free registers etc
too later.
Bug: v8:7700
Change-Id: I32146023c7b9bc9e553e3db98fe034e8cef7d09d
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3487994
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Toon Verwaest <verwaest@chromium.org>
Auto-Submit: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79289}
When we can eliminate a branch-on-type instruction based on statically
available type information and replace it with an unconditional branch,
we have to mark the rest of the current block as unreachable.
Change-Id: I9b8cc2f8e76da0b1b7cdf72b150ec675e9aae1a3
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3490931
Reviewed-by: Manos Koukoutos <manoskouk@chromium.org>
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79288}
{Invoke} is the central bottleneck for calling into generated code.
Check that at this point, no {CodeSpaceWriteScope} is open, otherwise
the JS code could write to the code space once it gains access to an
arbitrary-write gadget.
R=jkummerow@chromium.org
Bug: v8:11974
Change-Id: Ie323cea442a5ea355a1c975b300a1cc5a6edf433
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3487787
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79287}
In very rare cases we open CodeSpaceWriteScopes for multiple native
modules at the same time, e.g. for tier down (debugging) via
ExecuteCompilationUnits
--> PublishCompilationResults
--> OnFinishedUnits
--> TriggerCallbacks
--> AsyncCompileJob::CompilationStateCallback::call
--> WasmEngine::UpdateNativeModuleCache
--> RecompileNativeModule [for tier down]
--> InitializeRecompilation
--> FindFunctionsToRecompile
--> CodeSpaceWriteScope
Fixing this would be difficult because we actually want to keep the
CodeSpaceWriteScope open during subsequent publishing. So instead,
remove the assumption that scopes are always only open for one module at
a time.
In order to do this, we remove the {code_space_write_nesting_level_}
counter and instead use the {current_native_module_} field in all
configurations to check whether a scope is currently open, and for which
module.
R=jkummerow@chromium.org
Bug: v8:12644, v8:11974
Change-Id: Idd24c87f5938c43e867c41fa1cd3879def6f3873
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3484317
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79286}
The per-Isolate Symbol tables are implemented using NameDictionary
before, which has additional property details overhead
And NameDictionary is limited to 2^23, which limits the Symbol
tables to be a maximum of 2^23.
- replace NameDictionary with SymbolTable in isolate
Bug: v8:12575
Change-Id: Ica4f05aac3494f7dfa3a074c240d4ba25df814e9
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3476897
Reviewed-by: Shu-yu Guo <syg@chromium.org>
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Commit-Queue: Igor Sheludko <ishell@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79285}
The helper function removes a register from the list and returns it
Bug: v8:7700
Change-Id: I3f9fe9d30113b9e6c7362dc8443e39ae3d1adf07
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3488372
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79284}
When checking whether to merge a region with its surrounding regions in
{InsertIntoWritableRegions}, we did not check first whether the
determined {insert_pos} is within the vector. We were thus accessing
(reading) after the end of the vector.
The bug only happened on MSVC builds, suggesting that clang
deterministically read a value which is never equal to the end of the
new region, whereas for MSVC it sometimes happened that we read exactly
the {region.end()} value, and we tried to merge regions.
R=jkummerow@chromium.org
Bug: v8:12643
Change-Id: If30d910ed6e996f7b0e1d8c5b439c3d842a498f6
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3487988
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79283}
This is a reland of 4fde332811
Another Blink-related test fix landed, see chromium:1300492.
Original change's description:
> heap: Force incremental marking in C++ only workloads
>
> ... when above a certain minimum threshold. This is to guard against
> memory running away in scenarios where the JS heap is empty and
> there's only high throughput C++ allocations that don't allow for a
> memory reducer GC to kick in.
>
> This logic should be revisited after Oilpan's young generation
> collector is implemented which may allow switching to a more efficient
> shrinking strategy for initial heap setup.
>
> Bug: chromium:1029379, chromium:1300028, chromium:1300492
> Change-Id: I93924fc2fe77d6226c29358d3afb1cc9d6fbf3b1
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3484319
> Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
> Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#79255}
Bug: chromium:1029379, chromium:1300028, chromium:1300492
Change-Id: Ida66e0c944094472b4856d5fecef2d199d29549b
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3487991
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Auto-Submit: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Commit-Queue: Dominik Inführ <dinfuehr@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79281}
The {index} argument to {Peek} is unused. Other {Peek} methods use it to
generate the error message, but {Peek} without expected type does not
type check anything so it stays unused.
R=manoskouk@chromium.org
Change-Id: I979063f707f7305987220d6c192db7e679d930ce
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3490930
Reviewed-by: Manos Koukoutos <manoskouk@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79280}
Instead of scanning the array of registers, keep an explicit list of
free registers. Stack slots are equally changed to use an std::vector of
free slots instead of a linked list. Now we only need to scan
- the list of free registers when we want to allocate a specific
register,
- and scan the list of allocated registers to see if the free value is
already in a different register,
- scan the list of allocated registers to free some register if we don't
have enough registers (for input, output, or temp).
Bug: v8:7700
Change-Id: Iff41b06aae656b59e4ed25e9066671a21660a73e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3489487
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Toon Verwaest <verwaest@chromium.org>
Auto-Submit: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79278}
This reverts commit fe822dc984.
Reason for revert: https://ci.chromium.org/ui/p/v8/builders/ci/V8%20Blink%20Linux%20Debug/13306/overview
Original change's description:
> Reland "heap: Force incremental marking in C++ only workloads"
>
> This is a reland of 4fde332811
>
> Two issues in Blink tests have been fixed before this reland.
>
> Original change's description:
> > heap: Force incremental marking in C++ only workloads
> >
> > ... when above a certain minimum threshold. This is to guard against
> > memory running away in scenarios where the JS heap is empty and
> > there's only high throughput C++ allocations that don't allow for a
> > memory reducer GC to kick in.
> >
> > This logic should be revisited after Oilpan's young generation
> > collector is implemented which may allow switching to a more efficient
> > shrinking strategy for initial heap setup.
> >
> > Bug: chromium:1029379, chromium:1300028, chromium:1300492
> > Change-Id: I93924fc2fe77d6226c29358d3afb1cc9d6fbf3b1
> > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3484319
> > Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
> > Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
> > Cr-Commit-Position: refs/heads/main@{#79255}
>
> Bug: chromium:1029379, chromium:1300028, chromium:1300492
> Change-Id: I6cd4a4d358bc1a78f2f001ed50dd9bb3f376f49e
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3488370
> Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
> Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#79274}
Bug: chromium:1029379, chromium:1300028, chromium:1300492
Change-Id: If325d40455f433b7910b68b24bb1cf84337f177a
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3488373
Auto-Submit: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Cr-Commit-Position: refs/heads/main@{#79277}
Instead of returning false and failing in the caller, do fail inside the
PageAllocator directly. Failure to free pages should never happen, and
handling this case in the PageAllocator directly gives us better options
to surface more detailed information in follow-up patches.
R=mlippautz@chromium.org
Bug: v8:12656, chromium:1299735
Change-Id: I6d2aa3a5613c0f1102210fccbccc6ad0e522a6ed
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3484323
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79276}
mjsunit/compiler/deopt-pretenure.js is flaky due to --gc-interval in
some variants.
The flag can cause a variable to be promoted to old space before the
test can force allocation site pretenuring for that variable, which is
essential for the test case.
Bug: v8:12652
Change-Id: If7239deaa3026bb781d3ee96df28a1bbf3a5b6f1
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3488371
Auto-Submit: Patrick Thier <pthier@chromium.org>
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Commit-Queue: Dominik Inführ <dinfuehr@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79275}
This is a reland of 4fde332811
Two issues in Blink tests have been fixed before this reland.
Original change's description:
> heap: Force incremental marking in C++ only workloads
>
> ... when above a certain minimum threshold. This is to guard against
> memory running away in scenarios where the JS heap is empty and
> there's only high throughput C++ allocations that don't allow for a
> memory reducer GC to kick in.
>
> This logic should be revisited after Oilpan's young generation
> collector is implemented which may allow switching to a more efficient
> shrinking strategy for initial heap setup.
>
> Bug: chromium:1029379, chromium:1300028, chromium:1300492
> Change-Id: I93924fc2fe77d6226c29358d3afb1cc9d6fbf3b1
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3484319
> Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
> Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#79255}
Bug: chromium:1029379, chromium:1300028, chromium:1300492
Change-Id: I6cd4a4d358bc1a78f2f001ed50dd9bb3f376f49e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3488370
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79274}
During `WasmDebugBreak` we push the full 128-bit vector register
instead of only the 64-bit FP value. As a result offset calculation
must use kSimd128Size instead of kDoubleSize.
Change-Id: Icaa44d9663024b9740acbbf054f6c334ae349cf8
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3487958
Reviewed-by: Junliang Yan <junyan@redhat.com>
Commit-Queue: Milad Farazmand <mfarazma@redhat.com>
Cr-Commit-Position: refs/heads/main@{#79271}
For creating and unpacking the wrapper objects we currently use
for WasmGC <-> JavaScript interop, we must read the Context from
the "WasmApiFunctionRef" rather than from the "instance_node_".
As a bonus, this patch also slightly improves generated code
for wasm-to-js wrappers by moving conditionally needed code to
labels (previously it was always executed but the results only
conditionally used).
Fixed: v8:12640
Change-Id: I55485cdb6a402f32ddc42ec6316dcbe23f1fcccb
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3489486
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Auto-Submit: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Manos Koukoutos <manoskouk@chromium.org>
Commit-Queue: Manos Koukoutos <manoskouk@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79270}
- First inputs are walked to update next_use and collect dead inputs
- If any dead values were collected, clear them from the registers
- Finally free the LiveNodeInfo from values_.
Bug: v8:7700
Change-Id: I4ae78820d4405470e73d3ec89948e46442286eeb
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3487786
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79268}
This is a reland of a183895687
Now that https://crrev.com/c/3485678 landed and fixed the deadlock
in the linked bug, we can reland this CL without changes.
Original change's description:
> [heap] Allow shared references in WeakMap
>
> Shared references can also be stored in WeakMaps and during marking we
> need to be able to deal with such references. In a client GC shared
> objects are treated as live, so we don't need to update or check mark
> bits for such objects.
>
> Bug: v8:11708
> Change-Id: I0dbf797472c4779f462750dab63cc9b012aad091
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3447365
> Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
> Commit-Queue: Dominik Inführ <dinfuehr@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#79153}
Bug: v8:11708, v8:12642
Change-Id: I5945a16255647c897a1df834267137bf73b6207f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3485679
Auto-Submit: Dominik Inführ <dinfuehr@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79267}
{AtomicSub} on x64 first negates the {value} register, then does an
atomic addition. For that reason, {value} should be a unique register.
So far, we only checked that it's not used in the value stack, but we
should also check for overlap with the destination address or the offset
register.
Drive-by: Remove unneeded handling of non-unique register index on arm,
as that cannot happen (LiftoffCompiler ensures that the result register
is unique).
R=thibaudm@chromium.org
Bug: chromium:1296876
Change-Id: Ie6b97eec8e8dea07b0bcc644d261f47467cc5b8e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3487987
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79265}
This CL introduces an additional verification pass at the end of
SimplifiedLowering. The verification checks consistency of the lowered
graph with respect to node types under the effect of used truncations.
Typing of additional, lower level nodes is required and added in this
CL.
The verification pass can be enabled using --verify-simplified-lowering.
Bug: v8:12619, v8:11682
Change-Id: I21e7ebcf40153e53108ddfad2a871c7cbd61a085
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3452029
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Commit-Queue: Nico Hartmann <nicohartmann@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79264}
This reverts commit 4fde332811.
Reason for revert: causes crashes
Original change's description:
> heap: Force incremental marking in C++ only workloads
>
> ... when above a certain minimum threshold. This is to guard against
> memory running away in scenarios where the JS heap is empty and
> there's only high throughput C++ allocations that don't allow for a
> memory reducer GC to kick in.
>
> This logic should be revisited after Oilpan's young generation
> collector is implemented which may allow switching to a more efficient
> shrinking strategy for initial heap setup.
>
> Bug: chromium:1029379, chromium:1300028, chromium:1300492
> Change-Id: I93924fc2fe77d6226c29358d3afb1cc9d6fbf3b1
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3484319
> Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
> Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#79255}
Bug: chromium:1029379, chromium:1300028, chromium:1300492
Change-Id: I15e8d7b37b9f9b6ef4f72968c262a614618f1863
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3487970
Auto-Submit: Tobias Tebbi <tebbi@chromium.org>
Commit-Queue: Tobias Tebbi <tebbi@chromium.org>
Owners-Override: Tobias Tebbi <tebbi@chromium.org>
Commit-Queue: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Cr-Commit-Position: refs/heads/main@{#79263}
Previously, the interval was max(6, <interval_value>) which was
changed to actually consider the value of the flag.
Change-Id: Iec3cef19b6ec8528f03c36db6239b044ee90cde1
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3487969
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79262}
We need to park the isolate's main thread before blocking in the
semaphore to allow a shared GC to happen in the meantime.
Bug: v8:11708, v8:12647
Change-Id: Ide215d2c811caee84663d8749b7d94a414c44bd8
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3485678
Auto-Submit: Dominik Inführ <dinfuehr@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Commit-Queue: Camillo Bruni <cbruni@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79261}
Avoid going through Heap but rather call it directly on the allocator.
Bug: v8:12615
Change-Id: I395b96d08b685c63c4125245a76c3610acf1643b
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3485677
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79260}
Instead of returning a boolean value, and then failing in the caller via
a CHECK, do fail directly inside OS::Free, OS::Release and similar
functions.
The PageAllocator methods still return a bool (which is always true) to
avoid changing the public API.
R=mlippautz@chromium.org
Bug: v8:12656, chromium:1299735
Cq-Include-Trybots: luci.v8.try:v8_fuchsia_compile_rel
Change-Id: Ide02e7d893e1603326c629797a7defac8bf258ef
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3483671
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79258}
This CL refactors Heap::GCTypeTimer and Heap::GCTypePriorityTimer
and moves them to a GCTracer::RecordGCPhasesInfo class. This is
a necessary change for deprecating counters that are used for
old style GC metrics, like gc_scavenger. When all such counters
are deprecated, GCTracer::RecordGCPhasesInfo will no longer be
necessary and will be removed.
Bug: chromium:1154636
Change-Id: I04504a0f6c7a0955f4300a1c94c969aaeb23b77f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3486556
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Nikolaos Papaspyrou <nikolaos@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79257}
