- 25 Sep, 2019 8 commits
-
-
Yu Yin authored
when wasm have multiple returns but not aligned such as return {kWasmI32, kWasmI64, kWasmI64, kWasmI32 } like test/wasm-api-tests/multi-return.cc do, wasm compiler will generate store instructions but not unaligned store instructions to store the return values, this will cause check failed on mips simulator, the test maybe will successful on mips native machine if the host kernel can handle the unaligend stores. This patch also fix the return address offset. Change-Id: I7de93fdbef3341e7d0057f6ecbc95a9d2f86c943 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1824309 Commit-Queue: Clemens Backes [né Hammacher] <clemensh@chromium.org> Reviewed-by:Clemens Backes [né Hammacher] <clemensh@chromium.org> Reviewed-by:
Jakob Kummerow <jkummerow@chromium.org> Cr-Commit-Position: refs/heads/master@{#63963}
-
Dan Elphick authored
In preparation for Code objects that aren't executable and can be in RO_SPACE, pass an Isolate into Code::Disassemble. Bug: v8:7464 Change-Id: I99f5faf23dd4709a48925e1ae44d1ce595ea6f5f Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1822043Reviewed-by:
Jakob Gruber <jgruber@chromium.org> Reviewed-by:
Ross McIlroy <rmcilroy@chromium.org> Commit-Queue: Dan Elphick <delphick@chromium.org> Cr-Commit-Position: refs/heads/master@{#63962}
-
Clemens Hammacher authored
With the far jump table, we need to distinguish the maximum size of a single code space from the maximum total code size per module. On arm64, they differ, because we now support 1GB of code space, but each code space is still limited to 128MB. Bug: v8:9477 R=mstarzinger@chromium.org Change-Id: I7b9aaec56a1d9d1f70573b6b895216d5b3f38346 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1815253 Commit-Queue: Clemens Backes [né Hammacher] <clemensh@chromium.org> Reviewed-by:
Michael Starzinger <mstarzinger@chromium.org> Cr-Commit-Position: refs/heads/master@{#63961}
-
Yu Yin authored
Port 12a9ee3a Original Commit Message: 32 bit Smis are incompatible with pointer compression so we land disable them before enabling pointer compression in order to separate memory and performance regressions caused by 31 bit Smis from pointer compression change. Change-Id: I19647f18e7feafc6ca5b922c73785839ab528421 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1824306Reviewed-by:
-
Dan Elphick authored
Moves the following functions from CodeStubAssembler to BuiltinsStringAssembler: SubString CopyStringCharacters AllocAndCopyStringCharacters Bug: v8:9396 Change-Id: Ieb534b7fa7e72db9b05cdc2a34bd88b7a52ee985 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1822040Reviewed-by:
-
Dan Elphick authored
Move CodeStubAssembler::FindOrderedHashTableEntry into CollectionsBuiltinsAssembler and remove the specializations since they can now be generated directly by the call sites. Bug: v8:9396 Change-Id: I23e84167dafc2fec1886a842a8649f0500c4f447 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1822041 Auto-Submit: Dan Elphick <delphick@chromium.org> Reviewed-by:
-
Dan Elphick authored
Last use of BasicStoreNumberDictionaryElement was removed in https://chromium-review.googlesource.com/c/v8/v8/+/1627977 Bug: v8:9396 Change-Id: I08c1da67393c3549a8c756abd2a5ca4e35fa3b2a Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1822042 Auto-Submit: Dan Elphick <delphick@chromium.org> Commit-Queue: Jakob Gruber <jgruber@chromium.org> Reviewed-by:
-
v8-ci-autoroll-builder authored
Rolling v8/build: https://chromium.googlesource.com/chromium/src/build/+log/04fbb76..7f88e8b Rolling v8/third_party/catapult: https://chromium.googlesource.com/catapult/+log/8669b37..6bb5733 Rolling v8/third_party/depot_tools: https://chromium.googlesource.com/chromium/tools/depot_tools/+log/ba5bc99..7211cf6 Rolling v8/tools/clang: https://chromium.googlesource.com/chromium/src/tools/clang/+log/e514199..d0c91c0 TBR=machenbach@chromium.org,tmrts@chromium.org Change-Id: I8139c9c1b3d322cf0cc5d49d190a20d55ff35a4f Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1823582Reviewed-by:
v8-ci-autoroll-builder <v8-ci-autoroll-builder@chops-service-accounts.iam.gserviceaccount.com> Commit-Queue: v8-ci-autoroll-builder <v8-ci-autoroll-builder@chops-service-accounts.iam.gserviceaccount.com> Cr-Commit-Position: refs/heads/master@{#63956}
-
- 24 Sep, 2019 19 commits
-
-
Ng Zhi An authored
Bug: v8:9728 Change-Id: I8d993368fc23ab9e8cc08e31f4405678ec4ce824 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1803790Reviewed-by:
Deepti Gandluri <gdeepti@chromium.org> Commit-Queue: Zhi An Ng <zhin@chromium.org> Cr-Commit-Position: refs/heads/master@{#63955}
-
Ng Zhi An authored
This gives nicer error messages. Was Fatal error in ../../src/compiler/backend/instruction.h, line 778 Debug check failed: i < InputCount(). Now Fatal error in ../../src/compiler/backend/instruction.h, line 778 Debug check failed: i < InputCount() (5 vs. 3). Bug: v8:9396 Change-Id: Iab2aea49245a42397bf07f4abf9e3de501c758b9 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1817258Reviewed-by:
-
Andreas Haas authored
This CL fixes a memory leak in the interpreter. The leak was caused by a cycle the object graph that was rooted with a global object. The cycle was the following: A global handle, owned by the interpreter -> reference stack of the Interpreter -> ref.func element (WasmExportedFunction) -> WasmInstanceObject -> WasmDebugInfo -> InterpreterHandle -> Interpreter With this CL we get rid of the global handle. Instead we store the stack in the WasmDebugInfo. We then have to load the reference stack every time we enter the Interpreter and want access the reference stack. R=mstarzinger@chromium.org Bug: chromium:1000610 Change-Id: If8995725f7ec35862b2f99a07582c861027daaf1 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1800582 Commit-Queue: Andreas Haas <ahaas@chromium.org> Reviewed-by:
-
Santiago Aboy Solanes authored
We are going to change again Tagged to not mean 'Full Tagged'. Then, we have to reintroduce this operator, which was removed in https://chromium-review.googlesource.com/c/v8/v8/+/1594439. Bug: v8:7703 Change-Id: I8df65b76eed67613297b04d67c791cb6580bb49e Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1817613 Commit-Queue: Santiago Aboy Solanes <solanes@chromium.org> Reviewed-by:
Georg Neis <neis@chromium.org> Cr-Commit-Position: refs/heads/master@{#63952}
-
Milad Farazmand authored
Port 12a9ee3a Original Commit Message: 32 bit Smis are incompatible with pointer compression so we land disable them before enabling pointer compression in order to separate memory and performance regressions caused by 31 bit Smis from pointer compression change. R=ishell@chromium.org, joransiu@ca.ibm.com, jyan@ca.ibm.com, michael_dawson@ca.ibm.com BUG= LOG=N Change-Id: I1a353f2d6a682ba27e579a7de1bf7ea2240a6bbd Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1822117Reviewed-by:
Junliang Yan <jyan@ca.ibm.com> Reviewed-by:
Igor Sheludko <ishell@chromium.org> Commit-Queue: Milad Farazmand <miladfar@ca.ibm.com> Cr-Commit-Position: refs/heads/master@{#63951}
-
Georg Neis authored
R=mslekova@chromium.org Bug: v8:7790 Change-Id: Ie6cc694ac3d11bdd454c2f6d23e24b4fdd59b00e Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1822039Reviewed-by:
Maya Lekova <mslekova@chromium.org> Commit-Queue: Georg Neis <neis@chromium.org> Cr-Commit-Position: refs/heads/master@{#63950}
-
Santiago Aboy Solanes authored
We are going to change again Tagged to not mean 'Full Tagged'. Then, we have to reintroduce these operators. These allow to write a tagged uncompressed pointer even if pointer compression is enabled. This CL is basically a revert of https://chromium-review.googlesource.com/c/v8/v8/+/1593301. Bug: v8:7703 Change-Id: Ic1c66690b6bc74f697c9f5bf6d3ff45bb764c268 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1817612Reviewed-by:
-
Jakob Gruber authored
This extends suppressions to also handle excess zone allocations in the parser. Bug: chromium:1002510 Change-Id: Id418564d244ed618703e62260de31a6b1a2842f8 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1821459 Auto-Submit: Jakob Gruber <jgruber@chromium.org> Commit-Queue: Michael Achenbach <machenbach@chromium.org> Reviewed-by:
Michael Achenbach <machenbach@chromium.org> Cr-Commit-Position: refs/heads/master@{#63948}
-
Dan Elphick authored
Always unmark arrowhead parameters as assigned directly after their initialization as the parser doesn't know when it first sees the "assignment" that it may be in an arrowhead. Bug: chromium:1003403, v8:8510 Change-Id: Iad5a4136d5ec06331fc43b81a809fd72cee2dd65 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1815131 Commit-Queue: Dan Elphick <delphick@chromium.org> Reviewed-by:
Toon Verwaest <verwaest@chromium.org> Cr-Commit-Position: refs/heads/master@{#63947}
-
Joshua Litt authored
Adds support for parsing top level await to V8, as well as many tests. This is the final cl in the series to add support for top level await to v8. Spec is here: https://tc39.es/proposal-top-level-await/#sec-execute-async-module Bug: v8:9344 Change-Id: Ie8f17ad8c7c60d1f6996d134ae154416cc1f31e3 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1703878Reviewed-by:
Adam Klein <adamk@chromium.org> Commit-Queue: Joshua Litt <joshualitt@chromium.org> Cr-Commit-Position: refs/heads/master@{#63946}
-
Jakob Gruber authored
This introduces a limit for the interpreter's BacktrackStack to match the limit used by generated code (RegExpStack::kMaximumStackSize). Bug: chromium:1006670 Change-Id: I0b7613698e61257aecca89535ad9109c7e454692 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1821458 Commit-Queue: Jakob Gruber <jgruber@chromium.org> Commit-Queue: Peter Marshall <petermarshall@chromium.org> Auto-Submit: Jakob Gruber <jgruber@chromium.org> Reviewed-by:
Peter Marshall <petermarshall@chromium.org> Cr-Commit-Position: refs/heads/master@{#63945}
-
Michael Starzinger authored
This reduces the number of label indices accepted by {br_table} from the full function body size to specifically 65520 labels. Note that TurboFan already had a similar limitation on switches, but caused a crash during compilation up until now. This change just makes the limit explicit and avoids the crash during compilation. R=clemensh@chromium.org TEST=mjsunit/regress/wasm/regress-9759 BUG=v8:9759 Change-Id: I3a9a4406b19a7f98fc36707b3b946be846170a15 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1821457 Commit-Queue: Michael Starzinger <mstarzinger@chromium.org> Reviewed-by: -
Matheus Marchini authored
Also report code relocation events through the public CodeEventHandler API, so that embedders can expose this information to external profilers, allowing those profilers to correctly translate function names even after they were relocated. R=bmeurer@chromium.org, jgruber@chromium.org, yangguo@chromium.org Change-Id: I8795186f5f5c58ede9054e4b83e2d290d92b6e00 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1802657Reviewed-by:
Yang Guo <yangguo@chromium.org> Reviewed-by:
Benedikt Meurer <bmeurer@chromium.org> Reviewed-by:
-
Clemens Hammacher authored
In the {Fixed} variant, the {WasmCodeManagerTest} always reserves 1GB of memory. This makes the test run OOM on many 32-bit platforms. Instead of skipping it selectively, this CL just removes the whole test. It caused a lot of trouble in the past, and needs two test-only methods in the WasmCodeManager. Also, the {Fixed} variant will not be needed any more with the wasm far jump table, since modules can always grow then. Drive-by: Clean up the unittests status file a bit. R=mstarzinger@chromium.org Bug: v8:9477 Change-Id: I5b6f8ed9f800863575c69d49d5df82f21fd23030 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1815251Reviewed-by: -
Ulan Degenbaev authored
If a BackingStore is marked as !free_on_destruct, then we don't have to guarantee that there is only one such BackingStore pointing to the underlying buffer. So we can skip costly registration in process-global table of backing stores. Bug: v8:9380,chromium:1002693 Change-Id: Iad1ec5c4811d6c52a9a9d78dd700acf69170db60 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1815136 Commit-Queue: Ulan Degenbaev <ulan@chromium.org> Reviewed-by:
-
Igor Sheludko authored
This is a reland of 12a9ee3a Fixed arm64 disasm test. Original change's description: > [ptr-compr] Switch to 31 bit Smis on 64-bit architectures > > 32 bit Smis are incompatible with pointer compression so we land disable > them before enabling pointer compression in order to separate memory and > performance regressions caused by 31 bit Smis from pointer compression > change. > > Bug: v8:9767 > Change-Id: I3d4a675df4208f808b1ba6e7816be545eae0dc24 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1815249 > Reviewed-by: Toon Verwaest <verwaest@chromium.org> > Commit-Queue: Igor Sheludko <ishell@chromium.org> > Cr-Commit-Position: refs/heads/master@{#63934} Bug: v8:9767 Change-Id: Ife46a4240141dd89d841eac152032ad6ca471810 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1820939Reviewed-by:
-
Ulan Degenbaev authored
This moves the initialization to JSArrayBuffer::SetupEmpty, which is the proper bottleneck for all paths constructing array buffers. Bug: chromium:1006600,v8:9380 Change-Id: I1887cb867627d69ade20654e5bc372b1ba1ac4e3 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1815132Reviewed-by:
-
v8-ci-autoroll-builder authored
Rolling v8/build: https://chromium.googlesource.com/chromium/src/build/+log/c05af3e..04fbb76 Rolling v8/third_party/catapult: https://chromium.googlesource.com/catapult/+log/fa002e6..8669b37 Rolling v8/third_party/depot_tools: https://chromium.googlesource.com/chromium/tools/depot_tools/+log/7cb60e8..ba5bc99 Rolling v8/tools/clang: https://chromium.googlesource.com/chromium/src/tools/clang/+log/235f673..e514199 Rolling v8/tools/swarming_client: https://chromium.googlesource.com/infra/luci/client-py/+log/96f1257..885b3fe TBR=machenbach@chromium.org,tmrts@chromium.org Change-Id: I755fb986c321cffd5ea4e6a14d31d50bec82f9fe Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1817576Reviewed-by:
-
Yu Yin authored
port f5ab7d38 https://crrev.com/c/1627539. Change-Id: Ibe721144301bff6d843b6d1861b5a20890a057bb Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1818004 Commit-Queue: Yu Yin <xwafish@gmail.com> Reviewed-by:
-
- 23 Sep, 2019 13 commits
-
-
Bill Budge authored
This reverts commit 12a9ee3a. Reason for revert: Breaks arm64 sim bot https://ci.chromium.org/p/v8/builders/ci/V8%20Linux%20-%20arm64%20-%20sim%20-%20debug/17467 Original change's description: > [ptr-compr] Switch to 31 bit Smis on 64-bit architectures > > 32 bit Smis are incompatible with pointer compression so we land disable > them before enabling pointer compression in order to separate memory and > performance regressions caused by 31 bit Smis from pointer compression > change. > > Bug: v8:9767 > Change-Id: I3d4a675df4208f808b1ba6e7816be545eae0dc24 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1815249 > Reviewed-by: Toon Verwaest <verwaest@chromium.org> > Commit-Queue: Igor Sheludko <ishell@chromium.org> > Cr-Commit-Position: refs/heads/master@{#63934} TBR=ishell@chromium.org,verwaest@chromium.org Change-Id: Icc5a5b676a7f7d3fdb3fe3cf84007d4f9bce6eaf No-Presubmit: true No-Tree-Checks: true No-Try: true Bug: v8:9767 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1815842Reviewed-by:
Bill Budge <bbudge@chromium.org> Commit-Queue: Bill Budge <bbudge@chromium.org> Cr-Commit-Position: refs/heads/master@{#63936}
-
Santiago Aboy Solanes authored
Since we were already loading the undefined value a few lines above, we could avoid loading it again. Bug: v8:9771 Change-Id: Id05a77494c06a212612a365b5c0dcba4803ba756 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1817609 Commit-Queue: Santiago Aboy Solanes <solanes@chromium.org> Reviewed-by:
Dan Elphick <delphick@chromium.org> Cr-Commit-Position: refs/heads/master@{#63935}
-
Igor Sheludko authored
32 bit Smis are incompatible with pointer compression so we land disable them before enabling pointer compression in order to separate memory and performance regressions caused by 31 bit Smis from pointer compression change. Bug: v8:9767 Change-Id: I3d4a675df4208f808b1ba6e7816be545eae0dc24 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1815249Reviewed-by:
-
Joshua Litt authored
Bug: v8:9463 Change-Id: Ie0e04e102b56ffdfb636e94ef293bb0d46e5f4a9 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1808485Reviewed-by:
Michael Stanton <mvstanton@chromium.org> Reviewed-by:
Ulan Degenbaev <ulan@chromium.org> Commit-Queue: Joshua Litt <joshualitt@chromium.org> Cr-Commit-Position: refs/heads/master@{#63933}
-
Dan Elphick authored
Moves the following functions only used in string builtins out of CodeStubAssembler: StringAdd AllocateConsString StringFromSingleUTF16EncodedCodePoint BranchIfCanDerefIndirectString DerefIndirectString MaybeDerefIndirectString MaybeDerefIndirectStrings Bug: v8:9396 Change-Id: Ib89966b9c170ca23dd7535a0f550c69966a6e21c Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1817608 Commit-Queue: Dan Elphick <delphick@chromium.org> Reviewed-by:
-
Georg Neis authored
Change-Id: I4ddb49bb4cb2cda52247354c53386369ad0e0844 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1817610Reviewed-by:
-
Dan Elphick authored
Move CodeStubAssembler::StoreJSTypedArrayElementFromTagged, SetJSTypedArrayOnHeapDataPtr and SetJSTypedArrayOffHeapDataPtr into TypedArrayBuiltinsAssembler. Bug: v8:9396 Change-Id: I16e6aa82d0de9bd1d29a9e2c095032f2c6dd5e29 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1817611Reviewed-by:
-
Dominik Inführ authored
Since invalidation is disabled at the moment, slots still need to be cleared explicitly. This was not the case in MigrateFastToFast in case a slot transitioned from tagged to untagged. Bug: chromium:1006630 Change-Id: I8d32ab9196603dabad9a382ec49b39144624c30d Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1815248Reviewed-by:
-
Ulan Degenbaev authored
Bug: chromium:1006629,v8:9380 Change-Id: I8e45759fe3ad1b0ef8f1ebdb33919c84e1e8a044 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1815244Reviewed-by:
-
Dominik Inführ authored
This reverts commit c271cb74. Reason for revert: CL is not needed anymore. Original change's description: > [heap] Emit debugging data before crashing > > We see crashes in the wild, however the minidump does not contain enough > information for properly investigating this. This is temporary code > that should be deleted as soon as possible. > > Bug: chromium:1004365 > Change-Id: I25e119d891e03f980d63f5d559fcb4429b574cb1 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1809373 > Commit-Queue: Dominik Inführ <dinfuehr@chromium.org> > Reviewed-by: Ulan Degenbaev <ulan@chromium.org> > Cr-Commit-Position: refs/heads/master@{#63873} TBR=ulan@chromium.org,dinfuehr@chromium.org # Not skipping CQ checks because original CL landed > 1 day ago. Bug: chromium:1004365 Change-Id: Ic8765790bfc96ec144ad02dbdf84e53c6f3e6571 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1815252Reviewed-by:
Dominik Inführ <dinfuehr@chromium.org> Commit-Queue: Dominik Inführ <dinfuehr@chromium.org> Cr-Commit-Position: refs/heads/master@{#63927}
-
Michael Starzinger authored
This fixes how arguments of a call to {fround} are being parsed. It now accepts a single "AssignmentExpression" only instead of an "Expression" which could potentially be a whole comma-separated list of expressions. R=clemensh@chromium.org TEST=mjsunit/regress/regress-crbug-1006592 BUG=chromium:1006592 Change-Id: Ifaf0c2b048e4ec18429cc6039c0e7dcdecc1d0bb Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1815255Reviewed-by: -
Clemens Hammacher authored
This is a very minor cleanup, which makes the output of --trace-wasm-compiler more compact. R=mstarzinger@chromium.org Change-Id: I6b941f32e2ec6c3963e14dd376f8a1d65cf46ec6 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1815254Reviewed-by:
-
Michael Starzinger authored
This loads the call builtin from the Isolate root instead of embedding it into the instruction stream. This can be more efficient, but more importantly it fixes an issue with tracing and eventually allows for background compilation of these wrappers. R=clemensh@chromium.org TEST=mjsunit/regress/wasm/regress-crbug-1006631 BUG=chromium:1006631 Change-Id: Ife1bc513340d233a3c01789c7b56126fe3b87f6f Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1815245Reviewed-by:
-
