- 26 Aug, 2020 9 commits
-
-
Santiago Aboy Solanes authored
It has several variants that might need to be simplified but this CL focuses on removing ParameterMode. Bug: v8:9708, v8:6949 Change-Id: I1c300b7abe0b698a9f3d063e0af1ed931dbf4af2 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2376820Reviewed-by:
Dan Elphick <delphick@chromium.org> Commit-Queue: Santiago Aboy Solanes <solanes@chromium.org> Cr-Commit-Position: refs/heads/master@{#69572}
-
Clemens Backes authored
The {priority_} field is being updated in {DefaultJobState::Join}, under {mutex_}. In other places though, it is read unprotected (without holding the mutex), leading to data races. This CL fixes that by reading the field while holding the mutex and using the read priority after releasing the mutex. Note that the {priority_} field is documented to be protected by {mutex_}, so the unprotected read was a bug. R=ulan@chromium.org CC=etiennep@chromium.org Bug: v8:10822 Change-Id: I80079f3cb6689e26116ffeb33755c6938c4a2cf1 Cq-Include-Trybots: luci.v8.try:v8_linux64_tsan_rel_ng Cq-Include-Trybots: luci.v8.try:v8_linux64_tsan_isolates_rel_ng Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2377685Reviewed-by:Ulan Degenbaev <ulan@chromium.org> Commit-Queue: Clemens Backes <clemensb@chromium.org> Cr-Commit-Position: refs/heads/master@{#69571}
-
Clemens Backes authored
The fuzzers were calling the compiled function without passing explicit arguments. Thus all arguments were converted from the "undefined" value, which typically results in a zero value, as expected. For BigInt though, it's not allowed to pass "undefined". We have to pass a proper BigInt. This CL implements this by passing explicit parameter values for all parameters. This effectively unlocks testing BigInt parameters in all fuzzers, thus may increase coverage and find new bugs. R=ahaas@chromium.org Bug: chromium:1120355 Change-Id: I4e451d2418eb73d460fa937d1cf95a1ab6c99cf5 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2377945 Commit-Queue: Clemens Backes <clemensb@chromium.org> Reviewed-by:
Andreas Haas <ahaas@chromium.org> Cr-Commit-Position: refs/heads/master@{#69570}
-
Santiago Aboy Solanes authored
I forgot to remove them when I removed the old API in https://chromium-review.googlesource.com/c/v8/v8/+/2369174. Bug: v8:8116 Change-Id: I74a9670f56d09b7907187d5abcf15d707c8100a6 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2377688 Commit-Queue: Santiago Aboy Solanes <solanes@chromium.org> Commit-Queue: Peter Marshall <petermarshall@chromium.org> Auto-Submit: Santiago Aboy Solanes <solanes@chromium.org> Reviewed-by:
Peter Marshall <petermarshall@chromium.org> Cr-Commit-Position: refs/heads/master@{#69569}
-
Santiago Aboy Solanes authored
Partially TNodify the method just to get ParameterMode out of the way. There is more TNodification needed but method is complicated and we are now focusing on removing ParameterMode entirely. Bug: v8:9708, v8:6949 Change-Id: I705c35e9665488a280111683c275b7292bc47576 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2374547 Commit-Queue: Santiago Aboy Solanes <solanes@chromium.org> Reviewed-by:
Ross McIlroy <rmcilroy@chromium.org> Cr-Commit-Position: refs/heads/master@{#69568}
-
Zeynep Cankara authored
This CL unifies both map-processor and ic-processor's map processing pipeline to the same format. Also, adds the functionality to search for map transitions on map panel by clicking on map property of the ic-panel. Bug: v8:10644 Change-Id: I9f95feebb08620fd489486cf459ebe0babbd32d6 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2367865Reviewed-by:
Sathya Gunasekaran <gsathya@chromium.org> Commit-Queue: Zeynep Cankara <zcankara@google.com> Cr-Commit-Position: refs/heads/master@{#69567}
-
Clemens Backes authored
When we used the tasks API, we distributed the compilation deadlines such that different tasks would finish (and publish) at slightly different times. When moving to the jobs API, this feature was lost, and all tasks now only publish once all compilation is done. This creates major contention and synchronization overhead after compilation, which can take roughly as long as the actual concurrent compilation. This CL reintroduces a mechanism for ensuring regular publishing. We choose a limit which is high enough to ensure that we don't publish too often (because that comes with overhead), but low enough to ensure that not too many units are published at the end, when all compilation is done. At that point, there will still be some contention. Also, choose a slightly different limit per task to ensure proper distribution of publishing over time (different tasks are likely to publish at different times). This removes the last remaining regressions when compiling big modules via the jobs API. In local measurements, it makes us even faster than with the tasks API. R=ahaas@chromium.org Bug: chromium:1101340, chromium:1113234 Change-Id: I504f32606b8ad31a951449709cf407c471fa9b25 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2375196Reviewed-by:
-
Victor Gomes authored
Change-Id: I31e205b696627913584016bb9197e1e719ca0237 Bug: chromium:1120905, v8:10201 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2375191 Commit-Queue: Victor Gomes <victorgomes@chromium.org> Auto-Submit: Victor Gomes <victorgomes@chromium.org> Reviewed-by:
Igor Sheludko <ishell@chromium.org> Reviewed-by:
Jakob Gruber <jgruber@chromium.org> Cr-Commit-Position: refs/heads/master@{#69565}
-
Clemens Backes authored
Decommitting memory ranges is expensive, so try to merge as many of them as possible before decommitting. Minor drive-by: Slightly extend a GC tracing message. R=ahaas@chromium.org Change-Id: I91e44db1212f0c9b70b8c8fccebe8fd33ff37d9c Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2375825Reviewed-by:
-
- 25 Aug, 2020 20 commits
-
-
Clemens Backes authored
This documents the outcome of the discussion with chromium security guts. R=ecmziegler@chromium.org Change-Id: I55835fe5b40e3daf856a8ec1d5ef0d3fed5adb4c Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2375386 Commit-Queue: Clemens Backes <clemensb@chromium.org> Reviewed-by:
Emanuel Ziegler <ecmziegler@chromium.org> Cr-Commit-Position: refs/heads/master@{#69563}
-
Ng Zhi An authored
Some shuffles take have either register or memory operand for second input, but the codegen incorrectly assumes that it is always a register. Bug: v8:10824 Change-Id: Ia2df233dad4ed451e52e57e35cce5c80db0905db Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2373586 Commit-Queue: Zhi An Ng <zhin@chromium.org> Reviewed-by:
Bill Budge <bbudge@chromium.org> Cr-Commit-Position: refs/heads/master@{#69562}
-
Dominik Inführ authored
Experiment with enabling FLAG_local_heaps for a few days to see whether pause times regress and to catch bugs. Bug: v8:10315 Change-Id: Iefc329bbc1b8131796f1db2a729ee358bef88894 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2375190Reviewed-by:
-
Clemens Backes authored
This is a reland of c2ea2047 Original change's description: > [wasm] Move kMaxWasmCodeSpaceSize to wasm directory > > This limit is wasm-internal, and does not need to be exposed via > src/common/globals.h. > This CL moves it into the {WasmCodeAllocator}. > > Drive-by: Minor simplification in jump table stress test. > > R=ecmziegler@chromium.org > > Change-Id: Iff8c4657697ae98123d840a022c5b21c4948fcdf > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2375189 > Reviewed-by: Emanuel Ziegler <ecmziegler@chromium.org> > Commit-Queue: Clemens Backes <clemensb@chromium.org> > Cr-Commit-Position: refs/heads/master@{#69558} Change-Id: I6e0432d14d23978dea599233e620e84d8255caf9 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2375388Reviewed-by:
-
Clemens Backes authored
This reverts commit c2ea2047. Reason for revert: Link failures: https://ci.chromium.org/p/v8/builders/ci/V8%20Linux64%20-%20cfi/26209 Original change's description: > [wasm] Move kMaxWasmCodeSpaceSize to wasm directory > > This limit is wasm-internal, and does not need to be exposed via > src/common/globals.h. > This CL moves it into the {WasmCodeAllocator}. > > Drive-by: Minor simplification in jump table stress test. > > R=ecmziegler@chromium.org > > Change-Id: Iff8c4657697ae98123d840a022c5b21c4948fcdf > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2375189 > Reviewed-by: Emanuel Ziegler <ecmziegler@chromium.org> > Commit-Queue: Clemens Backes <clemensb@chromium.org> > Cr-Commit-Position: refs/heads/master@{#69558} TBR=clemensb@chromium.org,ecmziegler@chromium.org Change-Id: Ic3466eb17f2b3dfa4a0864002b0590fa0f571bb5 No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2375387Reviewed-by:
Clemens Backes <clemensb@chromium.org> Commit-Queue: Clemens Backes <clemensb@chromium.org> Cr-Commit-Position: refs/heads/master@{#69559}
-
Clemens Backes authored
This limit is wasm-internal, and does not need to be exposed via src/common/globals.h. This CL moves it into the {WasmCodeAllocator}. Drive-by: Minor simplification in jump table stress test. R=ecmziegler@chromium.org Change-Id: Iff8c4657697ae98123d840a022c5b21c4948fcdf Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2375189Reviewed-by: -
Dominik Inführ authored
While LocalHeap allow to dereference handles on background threads, this is only possible when FLAG_local_heaps is enabled. Bug: v8:10315 Change-Id: Ia27582cda7e4b4faf4d2c76959de418415e1ae68 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2372147 Commit-Queue: Dominik Inführ <dinfuehr@chromium.org> Reviewed-by:
-
Andreas Haas authored
This CL is the start to implement reference types support in Liftoff. As a first step this CL implements the ref.null instruction, and allows reference types as return values. This allows register allocation to be extended to support reference types, and also adds an easy way to get a reference type value for tests. Additionally with ref.null we don't have to worry about garbage collection because 'null' is an immovable object and therefore does not have to be scanned by the GC. R=thibaudm@chromium.org Bug: v8:7581 Change-Id: I5785dcf522c0d9881e1386f2d8b5d8560a16225c Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2352784 Commit-Queue: Andreas Haas <ahaas@chromium.org> Reviewed-by:
Thibaud Michaud <thibaudm@chromium.org> Cr-Commit-Position: refs/heads/master@{#69556}
-
Jakob Gruber authored
NCI code deopts differently than TF code. Tbr: mythria@chromium.org Bug: v8:8888 Change-Id: Ica1352ed5aba648c95a8a26fd6c2dd703b0f335c Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2375186Reviewed-by:
-
Pierre Langlois authored
We already make sure hints are 16K aligned on the Mac. On linux, it's possible to run with 64K pages enabled so we also need to align the hint there. Bug: v8:10808 Change-Id: I4ae7a56828135f6aa2980f164f25e2eb932b3cc0 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2369176Reviewed-by:
-
Etienne Pierre-doray authored
- JobHandle::IsCompleted() - JobDelegate::GetTaskId() - worker_count passed as argument to GetMaxConcurrency(). Jobs implementation must call the new GetMaxConcurrency(), but Jobs users aren't migrated yet. Bug: chromium:1114823 Change-Id: I0f4295ccaf9eba866dd771f30e2e49aa3eae9551 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2352484 Commit-Queue: Etienne Pierre-Doray <etiennep@chromium.org> Reviewed-by:
-
Santiago Aboy Solanes authored
Also we can do ParameterToIntPtr since this was the last use. Bug: v8:9708, v8:6949 Change-Id: I61747205d9d747c80bd13cd98bc51d8f97946130 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2374546Reviewed-by:
-
Victor Gomes authored
Change-Id: Iee7d2d0d2abbb5e18640d0601203dfb58e5ec474 Bug: v8:10201, v8:10825 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2374527 Commit-Queue: Victor Gomes <victorgomes@chromium.org> Auto-Submit: Victor Gomes <victorgomes@chromium.org> Reviewed-by:
-
Pierre Langlois authored
The cctests need to simulate full pages, however the calculation of the remainging available space is wrong causing an assertion when we change the V8 page size to 512K: $ cctest test-array-buffer-tracker/ArrayBuffer_PagePromotion # # Fatal error in ../../test/cctest/heap/heap-utils.cc, line 94 # Check failed: padding_size <= overall_free_memory || overall_free_memory == 0. The reason is: - On startup, we register a scavenger observer on the new space. The observer is set to trigger when the new space capacity is at 80% by default. - On linux, the initial capacity of the new space is 512K, so the scavenger observer will be placed at 80% of 512K, which will either be in the second page of the space if the page size is 256K, or in the first page if the page size is 512K. - When placing the observer, if the observer hits the first page, we lower the allocation limit (see `ComputeLimit()`). This makes sure the observer isn't skipped by allocations inlined in generated code. However, when we simulate filling the current page, we compute the space left in the current page by comparing the top with the `page_high()` rather than `limit()`. This was done so the tests would also work when inlined allocations are disabled. If we don't look at the `limit()`, we don't take the observer into account and fill more space than is available, triggering the assertion. This can also be reproduced by reducing the % at which the scavenger is triggered so that it hits the first page instead of the second when the page size is 256K, for example passing --scavenge-task-trigger=10. Bug v8:10808, v8:9906 Change-Id: Iad50bb68995de5ee017dcbe069d1fb229c9f5985 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2372545 Commit-Queue: Pierre Langlois <pierre.langlois@arm.com> Reviewed-by: -
Andreas Haas authored
With WebAssembly streaming compilation it is possible that FinishStream indirectly calls AbortStream. This had the effect that while FinishStream fulfilled the promise, AbortStream disposed the promise by removing the compile job from the wasm engine and thereby deallocated AsyncCompileJob and all the state that belongs to it. Without that state, FinishStream could not finish fulfilling the promise correctly. With this CL the streaming decoder remembers that the stream has already been finished. When the stream has been finished, all calls to Abort get ignored. The regression test for this issue requires the Chrome embedding, see https://crrev.com/c/2368359 R=clemensb@chromium.org Bug: chromium:1117258 Change-Id: Ifc28a1ee38c228b051c4d7d85b305fe2a721fa1f Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2367858 Commit-Queue: Andreas Haas <ahaas@chromium.org> Reviewed-by:
-
Clemens Backes authored
This bumps the engine-wide limit for generated machine code from 1GB to 2GB. This will allow compiling (and debugging) bigger modules, as they occur particularly in the debugging use case because the module will be unoptimized then. The limit per allocated code space is left at 1024MB (1GB), so a module will need to allocate two separate code spaces to make use of the full 2GB code space. R=ecmziegler@chromium.org Bug: chromium:1117033, chromium:1114093, chromium:1107649, chromium:1111266 Change-Id: Iab95b3130cbacc21ba078bf0d4ee2d6cd121d675 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2372604Reviewed-by:
-
Dominik Inführ authored
This reverts commit f16d3abf. Reason for revert: register_count() is read from the heap on the background thread. This is only safe when FLAG_local_heaps is enabled (set to true) but this isn't the case on tip-of-tree. Original change's description: > [compiler] Access the heap for BytecodeArray int/Register members > > We can create a new macro to skip the xxxData classes and read directly > from the heap. > > Bug: v8:7790 > Change-Id: I8de9ba0aee78c74d4c3113eb6bc4870a314de552 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2362687 > Commit-Queue: Santiago Aboy Solanes <solanes@chromium.org> > Reviewed-by: Georg Neis <neis@chromium.org> > Cr-Commit-Position: refs/heads/master@{#69471} TBR=neis@chromium.org,solanes@chromium.org Bug: v8:7790 Change-Id: I35bdd44721ce1e9d2f46df7cf5d1f413e22d9acf Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2372602Reviewed-by:
Dominik Inführ <dinfuehr@chromium.org> Reviewed-by:
Nico Hartmann <nicohartmann@chromium.org> Reviewed-by:
Santiago Aboy Solanes <solanes@chromium.org> Commit-Queue: Dominik Inführ <dinfuehr@chromium.org> Cr-Commit-Position: refs/heads/master@{#69547}
-
Marja Hölttä authored
Bug: v8:9237 Change-Id: Iaa1e3485eaecdd5af654177f207a990a2e63396b No-Try: True Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2374545Reviewed-by:
-
Zeynep Cankara authored
Action items: * Replace alls css @import with link tags. * Format all sources in system-analyzer/*. Bug: v8:10644 Change-Id: I6354cf7b0ed2aca8cdab4888cf9d504a1d963b50 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2367869 Commit-Queue: Zeynep Cankara <zcankara@google.com> Reviewed-by:
Camillo Bruni <cbruni@chromium.org> Reviewed-by:
-
v8-ci-autoroll-builder authored
Rolling v8/build: https://chromium.googlesource.com/chromium/src/build/+log/b656b39..7946304 Rolling v8/third_party/catapult: https://chromium.googlesource.com/catapult/+log/e6244cc..a54f10f Rolling v8/third_party/depot_tools: https://chromium.googlesource.com/chromium/tools/depot_tools/+log/4909610..e95b5d6 Rolling v8/tools/clang: https://chromium.googlesource.com/chromium/src/tools/clang/+log/9784c4f..d9986ad TBR=machenbach@chromium.org,tmrts@chromium.org,v8-waterfall-sheriff@grotations.appspotmail.com Change-Id: Ib4c2f5df542ec7aa228c80cf330f6a17ec7ee7e7 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2373478Reviewed-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@{#69544}
-
- 24 Aug, 2020 11 commits
-
-
Shu-yu Guo authored
This is a reland of 8358094d Original change's description: > [atomics] Missing detach check in Atomics.store on MIPS > > There was a missing detach check in the runtime implementation of > Atomics.store on BigInt TypedArrays. This runtime function is only used > on MIPS. > > Bug: v8:10687 > Change-Id: I77946d2ab0abf70c583af514ddaba0d384a21309 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2369663 > Auto-Submit: Shu-yu Guo <syg@chromium.org> > Commit-Queue: Ben Smith <binji@chromium.org> > Reviewed-by: Ben Smith <binji@chromium.org> > Cr-Commit-Position: refs/heads/master@{#69525} Bug: v8:10687 Change-Id: If75e2100712786fc9b39782e24b1457031060169 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2369974Reviewed-by:
Shu-yu Guo <syg@chromium.org> Reviewed-by:
Ben Smith <binji@chromium.org> Commit-Queue: Shu-yu Guo <syg@chromium.org> Cr-Commit-Position: refs/heads/master@{#69543}
-
Junliang Yan authored
Bug: v8:10201 Change-Id: Ia0c89a02840b2b29afe41f8eedd423c4d635c61f Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2319646Reviewed-by:
Victor Gomes <victorgomes@chromium.org> Commit-Queue: Victor Gomes <victorgomes@chromium.org> Cr-Commit-Position: refs/heads/master@{#69542}
-
Santiago Aboy Solanes authored
Also remove ParameterMode since it is not needed. Bug: v8:9708, v8:6949 Change-Id: If82cd5ffb89502e3feed7f5af613182df668080e Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2370639Reviewed-by:
-
Santiago Aboy Solanes authored
Also remove unused macros from the list. Bug: v8:9708, v8:6949 Change-Id: Ib90ce7c2c4996b3f8892106a89884d7e83610ca6 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2370637Reviewed-by:
-
Santiago Aboy Solanes authored
Bug: v8:9708, v8:6949 Change-Id: If2b95333b5821b91296169714359f095bf8b9bd6 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2370636Reviewed-by:
-
Marja Hölttä authored
Bug: v8:9237 No-Try: True Change-Id: Ic1bba1e0f5584350217c3971dfcc769c86beb1fd Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2366694 Commit-Queue: Marja Hölttä <marja@chromium.org> Reviewed-by:
-
Zeynep Cankara authored
This CL highlights the items with clickable functionality. It is implemented based on the received feedback from the V8 Mobile Team. Bug: v8:10644 Change-Id: I0489327681d485eedde59594d36083efdc216575 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2370626Reviewed-by:
-
Santiago Aboy Solanes authored
The new one was created in https://chromium-review.googlesource.com/c/v8/v8/+/1969900, and the old API was deprecated in https://chromium-review.googlesource.com/c/v8/v8/+/2110015, so now we can remove it. Bug: v8:8116 Change-Id: Ia839279609b412c36f1f17368acef23fe07e7c61 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2369174Reviewed-by:
-
Zeynep Cankara authored
This cleanup CL adds a timeline property to ic-panel and directly assigns the logEvents to selectedLogEvents upon data load. Bug: v8:10644 Change-Id: Ic1707ea237abbf57417c0b14e24fc0bf797d9679 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2370627Reviewed-by:
-
Omer Katz authored
Marking on allocation assumes a non-empty stack. Thus marking on allocation will not finalize even if it runs out of work (we prefer to finalize in a non-nestable task with no stack). Instead it will schedule another marking task and delegates finalization to the next marking task. Bug: chromium:1056170 Change-Id: I457090fe0a235707bb9973e330ff24fb25ac66c0 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2366764Reviewed-by:
-
Santiago Aboy Solanes authored
Change-Id: I1df5bbb5f974e09967a7e904556dd109235ecace Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2366780Reviewed-by:
-
