- 08 Sep, 2020 24 commits
-
-
Leszek Swirski authored
Change FeedbackVector into a Torque-generated class, simplifying its C++ implementation. As a drive-by, also get rid of the lower-cased, integer-indexed slot accessors, and replace the calls that needed this integer index with FeedbackSlot (potentially with a new WithOffset adjustment for extra data slots). Change-Id: Ie4f7e99b462ef95419edab46b4ac7ceecae9a05f Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2398638 Commit-Queue: Leszek Swirski <leszeks@chromium.org> Reviewed-by:
Dominik Inführ <dinfuehr@chromium.org> Reviewed-by:
Igor Sheludko <ishell@chromium.org> Auto-Submit: Leszek Swirski <leszeks@chromium.org> Cr-Commit-Position: refs/heads/master@{#69749}
-
Shu-yu Guo authored
This reverts commit 9e8c54f8. Reason for revert: Data races (https://ci.chromium.org/p/v8/builders/ci/V8%20Linux64%20TSAN/33100) Original change's description: > [Heap] ScavengerCollector use Jobs. > > No yielding is necessary since the main thread Join()s. > > max concurrency is determined based on either > remaining_memory_chunks_ or global pool size > (copied_list_ + promotion_list_) > > Change-Id: Ie30fa86c44d3224b04df5d79569bce126ce7d96b > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2354390 > Commit-Queue: Etienne Pierre-Doray <etiennep@chromium.org> > Reviewed-by: Ulan Degenbaev <ulan@chromium.org> > Cr-Commit-Position: refs/heads/master@{#69746} TBR=ulan@chromium.org,etiennep@chromium.org Change-Id: I206397c6086cf18d5e08cf615bdaaf0f34c6d36b No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2398852Reviewed-by:
Shu-yu Guo <syg@chromium.org> Commit-Queue: Shu-yu Guo <syg@chromium.org> Cr-Commit-Position: refs/heads/master@{#69748}
-
Milad Farazmand authored
Port 10348e8e Original Commit Message: With this CL we emit safepoint maps for externref values on the Liftoff value stack. With that there is support for externref parameters and locals in Liftoff, as well as for intermediate values of type externref. R=ahaas@chromium.org, joransiu@ca.ibm.com, jyan@ca.ibm.com, michael_dawson@ca.ibm.com BUG= LOG=N Change-Id: I369c0e2e7fb1c4681101a149b01e1a9580051d45 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2398845Reviewed-by:
Junliang Yan <jyan@ca.ibm.com> Reviewed-by:
Joran Siu <joransiu@ca.ibm.com> Reviewed-by:
Andreas Haas <ahaas@chromium.org> Commit-Queue: Milad Farazmand <miladfar@ca.ibm.com> Cr-Commit-Position: refs/heads/master@{#69747}
-
Etienne Pierre-doray authored
No yielding is necessary since the main thread Join()s. max concurrency is determined based on either remaining_memory_chunks_ or global pool size (copied_list_ + promotion_list_) Change-Id: Ie30fa86c44d3224b04df5d79569bce126ce7d96b Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2354390 Commit-Queue: Etienne Pierre-Doray <etiennep@chromium.org> Reviewed-by:
Ulan Degenbaev <ulan@chromium.org> Cr-Commit-Position: refs/heads/master@{#69746}
-
Gus Caplan authored
Bug: v8:9891 Change-Id: I04a1eaedc1e3e012a4779671025c8b71d1c6a56e Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2391909 Commit-Queue: Gus Caplan <snek@chromium.org> Reviewed-by:
Jakob Gruber <jgruber@chromium.org> Cr-Commit-Position: refs/heads/master@{#69745}
-
Sathya Gunasekaran authored
This reverts commit e43ec59b. Reason for revert: https://logs.chromium.org/logs/v8/buildbucket/cr-buildbucket.appspot.com/8869719317069732832/+/steps/Check_-_d8/0/logs/externref-globals-liftoff/0 Original change's description: > [wasm][liftoff] Support for most externref globals > > With this CL we add support for all externref globals except for > imported mutable globals. > > R=thibaudm@chromium.org > > Bug: v8:7581 > Change-Id: I63cb26f8ad6f4b8fc1c276e223c5d45745122ebf > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2396086 > Commit-Queue: Andreas Haas <ahaas@chromium.org> > Reviewed-by: Thibaud Michaud <thibaudm@chromium.org> > Cr-Commit-Position: refs/heads/master@{#69741} TBR=ahaas@chromium.org,thibaudm@chromium.org Change-Id: I19d191862aa91a116ea84a4a59a3a0a8036fcb49 No-Presubmit: true No-Tree-Checks: true No-Try: true Bug: v8:7581 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2398643 Commit-Queue: Sathya Gunasekaran <gsathya@chromium.org> Reviewed-by:
Sathya Gunasekaran <gsathya@chromium.org> Cr-Commit-Position: refs/heads/master@{#69744}
-
Santiago Aboy Solanes authored
Bug: v8:7790 Change-Id: I0194178cd73715b2a7e229dd7ce52a67c2280881 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2382312 Commit-Queue: Santiago Aboy Solanes <solanes@chromium.org> Reviewed-by:
Georg Neis <neis@chromium.org> Reviewed-by:
Nico Hartmann <nicohartmann@chromium.org> Cr-Commit-Position: refs/heads/master@{#69743}
-
Santiago Aboy Solanes authored
Original CL by neis@: http://crrev.com/c/v8/v8/+/2362693/1 Bug: v8:7790, v8:10853 Fixed: v8:10853 Change-Id: If0bd45e9dfb00f8ef1a358953dab1f5e1c9ae29e Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2387960Reviewed-by:
Ross McIlroy <rmcilroy@chromium.org> Reviewed-by:
-
Andreas Haas authored
With this CL we add support for all externref globals except for imported mutable globals. R=thibaudm@chromium.org Bug: v8:7581 Change-Id: I63cb26f8ad6f4b8fc1c276e223c5d45745122ebf Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2396086 Commit-Queue: Andreas Haas <ahaas@chromium.org> Reviewed-by:
Thibaud Michaud <thibaudm@chromium.org> Cr-Commit-Position: refs/heads/master@{#69741}
-
Andreas Haas authored
The wasm interpreter is slow on an arm simulator build with asan when it comes to return calls. An infinite return-call recursion therefore caused a timeout on ClusterFuzz. With this CL we increase the costs of return calls, and thereby avoid the timeout. R=clemensb@chromium.org Bug: chromium:1124899 Change-Id: If88e060779fbe2569e289e60170cf487dd31d7db Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2397615Reviewed-by:
-
evih authored
The parameters should be processed and evaluated in an increasing order (starting with the 1st param). Before we started with the last (n-th) parameter which was not correct. Bug: v8:10701, chromium:1124940 Change-Id: I8e0d8b1f0c53832c8f2d09551879c1a4413e1598 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2396085Reviewed-by:
-
Zhao Jiazhong authored
Port a2fd94f0 https://crrev.com/c/2312783 Original Commit Message: This adds the argument count (as intptr) to the standard frame. StandardFrames are now in the same shape as OptimizedFrames. The argument count in the stack will be used to tear down the arguments when we remove the arguments adaptor frame. Change-Id: I5cff9874735c1bdbd685deb0d829e57f1ee07ac2 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2393323Reviewed-by:
Victor Gomes <victorgomes@chromium.org> Reviewed-by:
-
Santiago Aboy Solanes authored
It does a direct access iff the FLAG_turbo_direct_heap_access is enabled. Otherwise, it uses the Data classes as it did before. Bug: v8:7790 Change-Id: I4f42e5734fdb2c91dbe9ef08869aec621c9d04c3 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2382311 Commit-Queue: Santiago Aboy Solanes <solanes@chromium.org> Reviewed-by:
-
Martin Bidlingmaier authored
Cq-Include-Trybots: luci.v8.try:v8_linux64_fyi_rel_ng Bug: v8:10765 Change-Id: I9e23fb5ba8a21aa1ddf96ab7936058f671bcc9ae Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2397895 Commit-Queue: Martin Bidlingmaier <mbid@google.com> Reviewed-by:
-
Marja Hölttä authored
Bug: v8:9237 Change-Id: I06d7e74ba0360334e6fa65c19f24548e220e4c69 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2349297 Commit-Queue: Marja Hölttä <marja@chromium.org> Reviewed-by:
Leszek Swirski <leszeks@chromium.org> Cr-Commit-Position: refs/heads/master@{#69735}
-
Z Nguyen-Huu authored
Cast resource field in ExternalString as v8: :String::ExternalStringResourceBase* would give us more info. Change-Id: Iae97b477f400f58365e2381b7230d2226d490aa7 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2388742 Commit-Queue: Jakob Gruber <jgruber@chromium.org> Reviewed-by:
Seth Brenith <seth.brenith@microsoft.com> Cr-Commit-Position: refs/heads/master@{#69734}
-
Martin Bidlingmaier authored
Sometimes the parser throws away redundant parts of the AST while parsing. For example, the regexp /(?:(?=(f)o))?f/ is (almost) equivalent to just /f/ because the optional block (...)? is zero-length. The parser notices this and returns the same tree as for /f/. However, there is a capture inside the (...)? block (which is never recorded because the quantifier containing it can only match zero-width, which is considered failure), so in this case it doesn't suffice to check that the regexp AST doesn't contain captures. Cq-Include-Trybots: luci.v8.try:v8_linux64_fyi_rel_ng Bug: v8:10765 Change-Id: I6145849d95b3522a397eadd2bae63d1d8e880f28 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2397896Reviewed-by:
-
Sathya Gunasekaran authored
TBR: solanes@chromium.org Bug: v8:10876 Change-Id: Ifd00704ea6328c0f470a2e33f37fb7a7e423435a Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2397897Reviewed-by:
Santiago Aboy Solanes <solanes@chromium.org> Reviewed-by:
-
Thibaud Michaud authored
When the top frame is paused at a breakpoint, and this breakpoint is being removed or was already removed, introduce a "dead breakpoint" in the new code. This ensures that: - The source position for the new frame is correct, otherwise it would just pick the source position of the previous call, - The offset between the source position and return address is the same in the new and old code, which is necessary for OSR to find the correct return address. R=clemensb@chromium.org Bug: v8:10337 Change-Id: I400886ff14846d3973d0634592c05960c05de738 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2377686 Commit-Queue: Thibaud Michaud <thibaudm@chromium.org> Reviewed-by:
Clemens Backes <clemensb@chromium.org> Cr-Commit-Position: refs/heads/master@{#69731}
-
Mythri A authored
An earlier cl: https://chromium-review.googlesource.com/c/v8/v8/+/2354810 fixed loading double fields with dynamic map checks. The fix however didn't update the expected type information to also expect Smi fields. So, in the later optimization phases the CheckHeapObject operation was reduced to a NoP since the expected type was a HeapObject. This cl fixes the bug by updating the type to Type::Any. Bug: chromium:1124696, v8:10582 Change-Id: Ic96dd74c97caf8eaf5289d8e8939f6feb1686a57 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2396088 Commit-Queue: Mythri Alle <mythria@chromium.org> Reviewed-by:
-
Dan Elphick authored
The GetReadOnlyRoots will need to remain on HeapObject due to pointer compression and the Isolate overload of HashSeed is convenient. Change-Id: Id7fa8e57e3b8bfa6f50465c24da1fc5d0b4db0bd Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2397889 Commit-Queue: Dan Elphick <delphick@chromium.org> Commit-Queue: Ross McIlroy <rmcilroy@chromium.org> Auto-Submit: Dan Elphick <delphick@chromium.org> Reviewed-by:
-
Jakob Gruber authored
These flags are later packed into an InstructionCode by the instruction selector, but not all bits are preserved. This CL adds a few assertions and helper methods to clarify and make this a bit safer. Bug: v8:8888 Change-Id: I01adc9f7887cc21c623d790dc204ac1d4b9bc7ac Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2388116 Commit-Queue: Jakob Gruber <jgruber@chromium.org> Reviewed-by:
Tobias Tebbi <tebbi@chromium.org> Cr-Commit-Position: refs/heads/master@{#69728}
-
v8-ci-autoroll-builder authored
Rolling v8/build: https://chromium.googlesource.com/chromium/src/build/+log/fc95b89..2dc8141 Rolling v8/third_party/catapult: https://chromium.googlesource.com/catapult/+log/8584a9d..69e45dc Rolling v8/third_party/depot_tools: https://chromium.googlesource.com/chromium/tools/depot_tools/+log/a070dd1..9218ba2 TBR=machenbach@chromium.org,tmrts@chromium.org,v8-waterfall-sheriff@grotations.appspotmail.com Change-Id: Icccb92296366c45ff1a1a1b33a6310504465861f Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2397199Reviewed-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@{#69727}
-
Yolanda Chen authored
The current implementation in AssembleSwap will generate a push-pop pair to swap between a general register and a stack slot for both x64 and ia32 targets. This is unnecessary for x64 target, as we can use the kScratchRegister to save the general register and swap with the stack slot. Change-Id: I10e0dc360dec22cdf5afa63ece3d5943685d7ecb Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2394177Reviewed-by:
Bill Budge <bbudge@chromium.org> Commit-Queue: Yolanda Chen <yolanda.chen@intel.com> Cr-Commit-Position: refs/heads/master@{#69726}
-
- 07 Sep, 2020 14 commits
-
-
Andreas Haas authored
With this CL we emit safepoint maps for externref values on the Liftoff value stack. With that there is support for externref parameters and locals in Liftoff, as well as for intermediate values of type externref. R=thibaudm@chromium.org Bug: v8:7581 Change-Id: I2df0a8d00b2da33fe06ff474b039cca4c7be726d Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2387571 Commit-Queue: Andreas Haas <ahaas@chromium.org> Reviewed-by:
-
Thibaud Michaud authored
When a WasmGCForegroundTask starts during runMessageLoopOnPause, the stack may contain Wasm code. In this case we need to scan the stack to report live code. This is already caught by some tests when RemoveBreakpoint is implemented correctly, which will be done in a separate CL, so no regression test is added here. R=clemensb@chromium.org Bug: v8:10337 Change-Id: I308145b9c2cf69f0100853be47ec9e30ce79a876 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2379510Reviewed-by:
-
Clemens Backes authored
If we generate more code than allowed by the --wasm-max-code-space flag, the {WasmCodeManager::Commit} method will fail (via fatal error). The DCHECK that was placed before that call checked for a related condition though, and failed before the {Commit} method could fail. By moving it after the call, we fix this issue. Adding a test case for this seems to be involved, since that test would crash the whole process (because of the fatal error). So it would have to be a unittest such that we can use ASSERT_DEATH_IF_SUPPORTED. In unittests though we don't yet have the infrastructure to create a NativeModule (needed for the WasmCodeAllocator). Alternatively we could add special testing methods (and a constructor) to the WasmCodeAllocator, but I do not consider this worth the effort for this little fix. R=thibaudm@chromium.org Bug: chromium:1124646 Change-Id: Ief09dfbd0a5b0b040fe9cf8a34e458d1d01dfacc Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2390754 Auto-Submit: Clemens Backes <clemensb@chromium.org> Commit-Queue: Thibaud Michaud <thibaudm@chromium.org> Reviewed-by: -
Camillo Bruni authored
This avoids race conditions in certain situations detected by TSAN. Bug: v8:10644 Change-Id: Ic3082da4e918890940fcc1cabf0933b0419f41de Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2396083 Commit-Queue: Camillo Bruni <cbruni@chromium.org> Reviewed-by:
Maya Lekova <mslekova@chromium.org> Cr-Commit-Position: refs/heads/master@{#69722}
-
Camillo Bruni authored
memchr is used in the default case but for two-byte strings that contain mostly ascii sequences every other byte is 0. This CL special-cases the 0 byte search in a two-byte string to simply peform a simple for loop which is rouhly 10x faster than the repeated memchr approach. Bug: v8:10737 Change-Id: Icde6b858b6679da69448295616734c46aa95e439 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2364507Reviewed-by:
-
Marcel Laverdet authored
The BigInt constructor has quadratic complexity while parsing strings, and the input is unbounded. Interrupts should be checked during this operation to ensure the embedder has control over runaway execution. since the implicit cast from string may now throw. BigInt: :CompareToString and BigInt::EqualToString now return Maybe<..> Change-Id: Iccb85fafac4df69075a34d1de647cb4f0184cb12 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2392629Reviewed-by:
Jakob Kummerow <jkummerow@chromium.org> Commit-Queue: Jakob Kummerow <jkummerow@chromium.org> Cr-Commit-Position: refs/heads/master@{#69720}
-
Michael Achenbach authored
Guessing the architecture on the testing machine is not required to derive if the simulator is supposed to run. The architecture check doesn't work on some platforms. We derive this now by checking if we have compiled a simulator build: target_cpu != v8_target_cpu. Bug: chromium:1110824 Change-Id: Id30a647f0610f21efb00d68ad1602e62dcd2c65c Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2395563Reviewed-by:
-
Santiago Aboy Solanes authored
This can be done since --local-heaps is also temporarily enabled. This lets us perform a reverse finch trial and start testing out direct heap access. Bug: chromium:1125606, v8:7790 Change-Id: I746e459824aa345dec8184bccc21bbd54070e3b4 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2396077Reviewed-by:
-
Camillo Bruni authored
This reverts commit 95aa697b. Reason for revert: breaks under tsan Original change's description: > [test][d8] Add d8.log.getAndStop helper > > The new helper function allows us to write tests for log parsing > without the need to first generating a log file. This makes it easier > to spot errors when the log format changes. > > - Add d8 global variable > - Add file_name accessor to Logger and Log classes > - Change OS::LogFileOpenMode to w+ / wb+ > - Use separate Log::WriteLogHeader method > - Remove unused logger_ instance variable from Log > > Bug: v8:10644 > Change-Id: Ifc7e35aa4e91b3f01f0847843263946e085944c3 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2387563 > Commit-Queue: Camillo Bruni <cbruni@chromium.org> > Reviewed-by: Michael Lippautz <mlippautz@chromium.org> > Reviewed-by: Toon Verwaest <verwaest@chromium.org> > Reviewed-by: Sathya Gunasekaran <gsathya@chromium.org> > Cr-Commit-Position: refs/heads/master@{#69715} TBR=mlippautz@chromium.org,cbruni@chromium.org,gsathya@chromium.org,verwaest@chromium.org Change-Id: Iad47d2f1e3391cae3c2f8c9e6c904c43925e1671 No-Presubmit: true No-Tree-Checks: true No-Try: true Bug: v8:10644 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2396080Reviewed-by:
Camillo Bruni <cbruni@chromium.org> Commit-Queue: Camillo Bruni <cbruni@chromium.org> Cr-Commit-Position: refs/heads/master@{#69717}
-
Victor Gomes authored
Change-Id: I04bab17c1636e99f7486987bc777bea73997a0c6 Bug: v8:10201 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2387579 Auto-Submit: Victor Gomes <victorgomes@chromium.org> Reviewed-by:
-
Camillo Bruni authored
The new helper function allows us to write tests for log parsing without the need to first generating a log file. This makes it easier to spot errors when the log format changes. - Add d8 global variable - Add file_name accessor to Logger and Log classes - Change OS::LogFileOpenMode to w+ / wb+ - Use separate Log::WriteLogHeader method - Remove unused logger_ instance variable from Log Bug: v8:10644 Change-Id: Ifc7e35aa4e91b3f01f0847843263946e085944c3 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2387563 Commit-Queue: Camillo Bruni <cbruni@chromium.org> Reviewed-by:
Michael Lippautz <mlippautz@chromium.org> Reviewed-by:
Toon Verwaest <verwaest@chromium.org> Reviewed-by:
-
Victor Gomes authored
When the interface descriptor of a builtin uses DEFINE_JS_PARAMETERS, the extra stack arguments must be positioned just above the return address, otherwise we would need to calculate its offset depending on the actual number of the arguments, we currently use a fixed offset to access them in CSA. Therefore, these extra arguments are either the first arguments when V8_REVERSE_JSARGS is enabled or otherwise the last arguments. Change-Id: If38ac7fd7f0079fc0e4fdccdb6cfb26e0425eb84 Bug: v8:10825 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2379854Reviewed-by:
-
Camillo Bruni authored
Normalize the target object if we're guaranteed to end up in dicionary mode after copying over all enumerable properties from the source object. This partially addresses performance issues when using the spread operator on large dict-mode objects. Change-Id: I5be7cbc6e6a7e4b1b59412779db9b651f5f68b45 Bug: v8:10763 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2362961Reviewed-by:
-
v8-ci-autoroll-builder authored
Rolling v8/build: https://chromium.googlesource.com/chromium/src/build/+log/2b8377b..fc95b89 TBR=machenbach@chromium.org,tmrts@chromium.org,v8-waterfall-sheriff@grotations.appspotmail.com Change-Id: Ic59889e29560c4a2d8b91fbe04e4fd1101f4b924 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2394857Reviewed-by:
-
- 05 Sep, 2020 1 commit
-
-
v8-ci-autoroll-builder authored
Rolling v8/build: https://chromium.googlesource.com/chromium/src/build/+log/a9ca728..2b8377b Rolling v8/third_party/depot_tools: https://chromium.googlesource.com/chromium/tools/depot_tools/+log/5dba9c9..a070dd1 Rolling v8/third_party/zlib: https://chromium.googlesource.com/chromium/src/third_party/zlib/+log/59187e1..aec16ef TBR=machenbach@chromium.org,tmrts@chromium.org,v8-waterfall-sheriff@grotations.appspotmail.com Change-Id: I03c0e9bb1562243c059175d4e09cb5b03655a7fa Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2393889Reviewed-by:
-
- 04 Sep, 2020 1 commit
-
-
Milad Farazmand authored
Port 5c0f7219 Original Commit Message: The 64-bit instruction selectors check whether the input value to a ChangeUint32ToUint64 node was produced by a node that sets the upper 32 bits to zero, and if so, they avoid emitting an extra instruction to clear the upper bits. This change: 1. Extends that existing mechanism to also include phi values: if all of the inputs to a phi value guarantee that the upper 32 bits are zero, then the phi value does too. 2. Updates x64 to include non-negative int32 constants in the list of nodes that always clear the upper bits. I didn't add this to any other architecture because I'm less certain of how they resolve moves from constants. This change improves the speed of the Mono interpreter on x64 by about 5%. R=seth.brenith@microsoft.com, joransiu@ca.ibm.com, jyan@ca.ibm.com, michael_dawson@ca.ibm.com BUG= LOG=N Change-Id: Ic515a254025b518c767fb64c5b1282cdf226f5ff Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2393742Reviewed-by:
-
