- 20 Mar, 2018 32 commits
-
-
Eric Holk authored
This was a bug found while working on https://crrev.com/c/850550 Change-Id: I92b04d1814e75f0a1817f8409496612a3597cd20 Reviewed-on: https://chromium-review.googlesource.com/966997 Commit-Queue: Eric Holk <eholk@chromium.org> Reviewed-by:
Yang Guo <yangguo@chromium.org> Reviewed-by:
Michael Lippautz <mlippautz@chromium.org> Cr-Commit-Position: refs/heads/master@{#52078}
-
Clemens Hammacher authored
Avoid loading to a general purpose register if the bit pattern consists of a consecutive block of 1 bits. Drive-by: Change a parameter from int8_t to byte such that the AVX_OP macro works on these methods. R=mstarzinger@chromium.org Change-Id: Ib469ddd29d92ddeabe98460d2951b01159a6548a Reviewed-on: https://chromium-review.googlesource.com/969123 Commit-Queue: Clemens Hammacher <clemensh@chromium.org> Reviewed-by:
Michael Starzinger <mstarzinger@chromium.org> Cr-Commit-Position: refs/heads/master@{#52077}
-
Clemens Hammacher authored
If FLAG_enable_embedded_constant_pool is false, the field for the builtin index would alias the field for the constant pool offset in the code object. This makes constant_pool() return the builtin index, but it also makes set_constant_pool() override the builtin index (and vice versa). This CL fixes this by making all constant_pool accessors honor that flag. R=mstarzinger@chromium.org Change-Id: I88803a4f28bd5a2fe85a310708c7a365cc457339 Reviewed-on: https://chromium-review.googlesource.com/970586Reviewed-by:
-
Igor Sheludko authored
Bug: v8:7206, chromium:576312 Change-Id: I1628e7966a40fa7d6f628bd972b0ad0142a5d349 Reviewed-on: https://chromium-review.googlesource.com/970825Reviewed-by:
Mythri Alle <mythria@google.com> Commit-Queue: Igor Sheludko <ishell@chromium.org> Cr-Commit-Position: refs/heads/master@{#52075}
-
jgruber authored
TBR=jarin@chromium.org Bug: v8:6666 Change-Id: I9f4654955f297c54e67b98afb42125b659e35fd3 Reviewed-on: https://chromium-review.googlesource.com/971042Reviewed-by:
Jakob Gruber <jgruber@chromium.org> Commit-Queue: Jakob Gruber <jgruber@chromium.org> Cr-Commit-Position: refs/heads/master@{#52074}
-
jgruber authored
Off-heap code cannot embed external references. With this CL, we load from the external reference table (reached through the root pointer) instead. In a follow-up, the table could be stored within the isolate itself, removing one more level of indirection. Bug: v8:6666 Change-Id: I4c612ad3d4112ec03c3b389f5bfb9cdc3dc8a671 Reviewed-on: https://chromium-review.googlesource.com/970468 Commit-Queue: Jakob Gruber <jgruber@chromium.org> Reviewed-by:
-
Peter Marshall authored
Part of ongoing work to remove the construct_stub field of the SFI. Generate_InternalArrayConstructor was actually incorrect for packed internal arrays, where it would instead create a regular internal array because it loaded the constructor function from the context every time. Ultimately InternalArray should be removed, or the constructor ported to CSA in the meantime. But for now, it is off the critical path for the construct_stub removal. Also fix a bug: Runtime_NewArray expects a type_info parameter, which should be in rbx (on x64). Because we now go through JSBuiltinsConstructStubHelper first, rbx is loaded with a value that doesn't look like a heap object, which causes a crash in NewArray. Fix that by first loading undefined explicitly (which is what the ArrayConstructor builtin does already). Bug: v8:7503 Change-Id: Ic92fa8864b0af2d32200eb0176ba55ccff03b114 Reviewed-on: https://chromium-review.googlesource.com/970823 Commit-Queue: Peter Marshall <petermarshall@chromium.org> Reviewed-by:
Benedikt Meurer <bmeurer@chromium.org> Cr-Commit-Position: refs/heads/master@{#52072}
-
Vasili Skurydzin authored
src/base/debug/stack_trace_posix.cc: suppressed unused function warnings for functions DemangleSymbols, OutputPointer(in order to compile with -Werror flag) test/cctest/test-isolate-independent-builtins.cc: corrections to make ByteInText test case compatible with aix. (affects aix only) Change-Id: I49e45e63545404c77aaed3f51b26557f6f03455e Reviewed-on: https://chromium-review.googlesource.com/927484Reviewed-by:
Michael Achenbach <machenbach@chromium.org> Commit-Queue: Jakob Gruber <jgruber@chromium.org> Cr-Commit-Position: refs/heads/master@{#52071}
-
Camillo Bruni authored
We currently never shrink the StringTable which causes excessive memory usage on certain websites. This CL tries to mitigate this by shrinking the StringTable if it is very empty (nof_elements * 16 < capacity) hopefully avoiding costly reallocations. Bug: chromium:818642, v8:5443 Change-Id: I4e6a95b3a6992b499fa6dd59ae159c51f089965a Reviewed-on: https://chromium-review.googlesource.com/970465Reviewed-by:
Toon Verwaest <verwaest@chromium.org> Reviewed-by:
Hannes Payer <hpayer@chromium.org> Commit-Queue: Camillo Bruni <cbruni@chromium.org> Cr-Commit-Position: refs/heads/master@{#52070}
-
jgruber authored
Loading external references from off-heap builtins will be root-pointer-relative. At least initially, these loads will happen in CSA and thus need access to the root pointer value. Bug: v8:6666 Change-Id: Iae4c89061df442f5afd03f93e5ba35c4e125b850 Reviewed-on: https://chromium-review.googlesource.com/970264Reviewed-by:
Jaroslav Sevcik <jarin@chromium.org> Commit-Queue: Jakob Gruber <jgruber@chromium.org> Cr-Commit-Position: refs/heads/master@{#52069}
-
Yang Guo authored
Win64 MSVC build seems to be fixed. Bug: v8:178 Change-Id: I6d3ec22511a7a3c25467c486d3d05df0939c7c87 Reviewed-on: https://chromium-review.googlesource.com/970469 Commit-Queue: Yang Guo <yangguo@chromium.org> Reviewed-by:
-
Ulan Degenbaev authored
The embedders should use the EmbedderGraph API. The similar structure can be created with the following steps: 1) Create a root node for each retainer info group. 2) Iterate all handles using Isolate::VisitHandlesWithClassIds. 3) Add an edge from the retainer info node to the v8 wrapper node. 4) Add an edge from the v8 wrapper node to the retainer info node. See how HeapSnapshotRetainedObjectInfo is converted to the new API. Bug: chromium:749490 Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng Change-Id: I124ae3853354863b4f888e6aa2ea13777dcaa37d Reviewed-on: https://chromium-review.googlesource.com/948842 Commit-Queue: Ulan Degenbaev <ulan@chromium.org> Reviewed-by:
-
Leszek Swirski authored
This reverts commit 520b025f. Reason for revert: Breaks internal-snapshot build: https://build.chromium.org/p/client.v8/builders/V8%20Linux64%20-%20internal%20snapshot/builds/14603 Original change's description: > Remove SFI code field > > Remove the SharedFunctionInfo code field, inferring the code object > from the function_data field instead. In some cases, the function_data > field can now hold a Code object (e.g. some WASM cases). > > Bug: chromium:783853 > Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng > Change-Id: I1219a4d6aa5abaa9fee54dda883da7a3186e347a > Reviewed-on: https://chromium-review.googlesource.com/952452 > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > Reviewed-by: Jakob Gruber <jgruber@chromium.org> > Reviewed-by: Yang Guo <yangguo@chromium.org> > Commit-Queue: Leszek Swirski <leszeks@chromium.org> > Cr-Commit-Position: refs/heads/master@{#52064} TBR=yangguo@chromium.org,mstarzinger@chromium.org,jgruber@chromium.org,leszeks@chromium.org,bmeurer@chromium.org,verwaest@chromium.org Change-Id: Iaf464ed29ba4317bbbb255c1aec76fb65cddba84 No-Presubmit: true No-Tree-Checks: true No-Try: true Bug: chromium:783853 Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng Reviewed-on: https://chromium-review.googlesource.com/970647Reviewed-by:
Leszek Swirski <leszeks@chromium.org> Commit-Queue: Leszek Swirski <leszeks@chromium.org> Cr-Commit-Position: refs/heads/master@{#52066}
-
sreten.kovacevic authored
Port CallIndirect and CallRuntime methods. Also, implement methods for allocation and deallocation of stack slots, which are used in trace memory operations. Bug: v8:6600 Change-Id: I99e0115dcf6d971229892b27b4b4f01d0c5441e8 Reviewed-on: https://chromium-review.googlesource.com/970262 Commit-Queue: Sreten Kovacevic <sreten.kovacevic@mips.com> Reviewed-by:
Ivica Bogosavljevic <ivica.bogosavljevic@mips.com> Cr-Commit-Position: refs/heads/master@{#52065}
-
Leszek Swirski authored
Remove the SharedFunctionInfo code field, inferring the code object from the function_data field instead. In some cases, the function_data field can now hold a Code object (e.g. some WASM cases). Bug: chromium:783853 Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng Change-Id: I1219a4d6aa5abaa9fee54dda883da7a3186e347a Reviewed-on: https://chromium-review.googlesource.com/952452Reviewed-by:
-
Clemens Hammacher authored
This adds support for i64.eqz, and all binary comparisons (i64.eq, i64.ne, i64.lt_s, i64.le_s, i64.lt_u, i64.le_u, i64.gt_s, i64.ge_s, i64.gt_u, i64.and i64.ge_u). R=titzer@chromium.org Bug: v8:6600 Change-Id: Ic6c59529b007220698e09d5959394bcfb6590173 Reviewed-on: https://chromium-review.googlesource.com/969125 Commit-Queue: Clemens Hammacher <clemensh@chromium.org> Reviewed-by:
Ben Titzer <titzer@chromium.org> Cr-Commit-Position: refs/heads/master@{#52063}
-
Predrag Rudic authored
Embedded builtins cause many failures therefore we'll disable them. Bug: v8:7571 Change-Id: I20fae241b2fab4ed41339de3a469d514e16c3894 Reviewed-on: https://chromium-review.googlesource.com/970542 Commit-Queue: Jakob Gruber <jgruber@chromium.org> Reviewed-by:
-
Ulan Degenbaev authored
This was removed in one of the branches during weak-refs work. Bug: v8:7574 Change-Id: Id2a1af22b1150d8c888c117c023e8c78f532b9f2 Reviewed-on: https://chromium-review.googlesource.com/970702Reviewed-by:
-
Michael Achenbach authored
NOTRY=true Change-Id: I30bb7c1935db33e49729ea66a8535809531c7130 Reviewed-on: https://chromium-review.googlesource.com/970642Reviewed-by:
Sergiy Byelozyorov <sergiyb@chromium.org> Commit-Queue: Michael Achenbach <machenbach@chromium.org> Cr-Commit-Position: refs/heads/master@{#52060}
-
Ulan Degenbaev authored
The embedders should use the new EmbedderGraph API to provide retainer info. Bug: chromium:749490 Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng Change-Id: Iac8dc1e749ef14277b027f43e799357c5bd413ea Reviewed-on: https://chromium-review.googlesource.com/948489 Commit-Queue: Ulan Degenbaev <ulan@chromium.org> Reviewed-by:
-
Yang Guo authored
R=jgruber@chromium.org Bug: v8:178 Change-Id: Id00ae14cb5a50e560f93249eed4e4b20aa0a6ce7 Reviewed-on: https://chromium-review.googlesource.com/970467 Commit-Queue: Yang Guo <yangguo@chromium.org> Reviewed-by:
-
Clemens Hammacher authored
The MessageLoopBehavior might change over time: Sometimes we want to wait because wasm background compilation is going on, sometimes we don't. This makes the semaphore go out of sync with the task queue (we always notify it when a new task is scheduled, but we only sometimes wait on it). Using a condition variable instead of a semaphore avoids this problem. R=ahaas@chromium.org Change-Id: Ib9850efc634f5988d3f824895b6566bd76475985 Reviewed-on: https://chromium-review.googlesource.com/969122Reviewed-by:
Andreas Haas <ahaas@chromium.org> Commit-Queue: Clemens Hammacher <clemensh@chromium.org> Cr-Commit-Position: refs/heads/master@{#52057}
-
Daniel Clifford authored
In general, TurboFan doesn't encounter phi nodes with only a single data input in the backend. However, CSA-based builtins (especially auto-generated ones, e.g. from Torque), may contain single-input phi nodes, although outside the auto-generated case this doesn't happen much in practice. Single input phi nodes (i.e. phis in blocks with one predecessor) don't have any side effects and are essentially useless and harmless, but to avoid problems in the backend of TurboFan (whose SSA deconstruction disallows control flow splits that continue to blocks with phis), this CL tweaks the existing CSA-only control flow and graph sanitization in the CSA path to ensure no no-op phis. Change-Id: I109f4dc6cde5ad1794585a09609a230b1848e0d5 Reviewed-on: https://chromium-review.googlesource.com/963711Reviewed-by:
-
Michael Achenbach authored
This is a reland of 40d66d8b The fix disambiguates duplicate symbols in the generated embedded builtins file. Original change's description: > [build] Make separate snapshot for trusted variant > > This enables side-by-side snapshots with and without untrusted-code > mitigations. It'll be the default in all V8 stand-alone builds > with external startup data. Internal snapshots are not supported. > > The files snapshot_blob.bin and snapshot_blob_trusted.bin will be > bundled with V8 on swarming and the correct file is loaded dependent > on the --untrusted-code-mitigations runtime flag. > > Likewise we embed two snapshots for builtins. > > Side-by-side snapshots won't be supported in Chromium. > > Bug: v8:7441 > Change-Id: I2949ddfd5773649946b1c8e74751d48ad1d9c524 > Reviewed-on: https://chromium-review.googlesource.com/960004 > Commit-Queue: Michael Achenbach <machenbach@chromium.org> > Reviewed-by: Jakob Gruber <jgruber@chromium.org> > Reviewed-by: Yang Guo <yangguo@chromium.org> > Reviewed-by: Ross McIlroy <rmcilroy@chromium.org> > Cr-Commit-Position: refs/heads/master@{#52028} Bug: v8:7441 Change-Id: I626171d4e07389f0453b4d0a698e2772fd37e8c5 Reviewed-on: https://chromium-review.googlesource.com/968623 Commit-Queue: Michael Achenbach <machenbach@chromium.org> Reviewed-by:
-
Clemens Hammacher authored
This adds support to get or set globals of all the standard types (i32, i64, f32, f64). R=titzer@chromium.org Bug: v8:6600 Change-Id: Ie8d14d3d964e2abe3f19945a0e80b0e8462e9485 Reviewed-on: https://chromium-review.googlesource.com/969262 Commit-Queue: Clemens Hammacher <clemensh@chromium.org> Reviewed-by:
-
Dan Elphick authored
Adds a new space RO_SPACE and modifies the serializer and other machinery to support it. Currently RO_SPACE has nothing in it, but will eventually contain all the immovable immutable objects, so the GC can ignore it. Bug: v8:7464 Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng Change-Id: Ib2ff474699196c138df8c24f7a2248471e30fbac Reviewed-on: https://chromium-review.googlesource.com/925703 Commit-Queue: Dan Elphick <delphick@chromium.org> Reviewed-by:
-
Clemens Hammacher authored
It was stored in a shared_ptr so far, which makes it more difficult to reason about life times. Since there is always exactly one owner of the AsyncCompileJob, a unique_ptr actually suffices. R=ahaas@chromium.org Change-Id: If94c9091889ad05325c559a97e9a9ffeee8d450c Reviewed-on: https://chromium-review.googlesource.com/968604Reviewed-by:
-
Clemens Hammacher authored
This adds support for the f32.abs and f64.abs opcodes. R=titzer@chromium.org Bug: v8:6600 Change-Id: I05a16bb4301d492ba8d22a6326c7b2ce0f9f2faa Reviewed-on: https://chromium-review.googlesource.com/968502Reviewed-by:
-
Yang Guo authored
Also unskip test that has already been fixed in f1b1ec70. R=jgruber@chromium.org Bug: v8:178 Change-Id: I9cd2156ef41146b0dd58a974088726f5cbda8058 Reviewed-on: https://chromium-review.googlesource.com/970243Reviewed-by:
-
Jaroslav Sevcik authored
This prevent forgotten field initializations. Bug: v8:5267 Change-Id: Iaef5f79af360e2400b29afe878f80870c8a64e71 Reviewed-on: https://chromium-review.googlesource.com/970241 Commit-Queue: Jaroslav Sevcik <jarin@chromium.org> Reviewed-by:
-
Michael Achenbach authored
TBR=sergiyb@chromium.org NOTRY=true NOTREECHECKS=true Bug: chromium:822593 Change-Id: I86291a5039fc368c0fddea4d46e0d6c72c56a226 Reviewed-on: https://chromium-review.googlesource.com/968921 Commit-Queue: Michael Achenbach <machenbach@chromium.org> Reviewed-by:
-
v8-autoroll authored
Rolling v8/build: https://chromium.googlesource.com/chromium/src/build/+log/76da9f5..a7e9a83 Rolling v8/buildtools: https://chromium.googlesource.com/chromium/buildtools/+log/2888931..3748a2a Rolling v8/third_party/catapult: https://chromium.googlesource.com/catapult/+log/6c4a8ca..734f737 Rolling v8/tools/clang: https://chromium.googlesource.com/chromium/src/tools/clang/+log/0a9737b..7d56ff9 TBR=machenbach@chromium.org,hablich@chromium.org,sergiyb@chromium.org Change-Id: I60e4196d8ed8d39e88bbf3688569bbd0aadef919 Reviewed-on: https://chromium-review.googlesource.com/969956Reviewed-by:
v8 autoroll <v8-autoroll@chromium.org> Commit-Queue: v8 autoroll <v8-autoroll@chromium.org> Cr-Commit-Position: refs/heads/master@{#52047}
-
- 19 Mar, 2018 8 commits
-
-
Peter Marshall authored
In order to remove the construct_stub field of the SFI we need all construct stubs to be the same, and do any branching at runtime instead. For builtins we don't need to set the construct stub because the builtins construct stub will call into it for us. There should only be two builtins left without the builtins construct stub: Array and InternalArray, which are special cases that need to be dealt with in another CL. Bug: v8:7503 Change-Id: If0d419399a9ee22c09cf2a5a3d3dbea7a04dee77 Reviewed-on: https://chromium-review.googlesource.com/968862Reviewed-by:
-
Michael Stanton authored
This reverts commit 46a3c772. Reason for revert: This is actually not quite ready. What we need is a speculation free poisoning, and if we do another branch, then I think that won't happen. Original change's description: > [turbofan] Masking/poisoning in codegen (optimized code, mips & mips64) > > This introduces masking of loads with speculation bit during code generation. > At the moment, this is done only under the > --branch-load-poisoning flag, and this CL enlarges the set of supported > platforms from {x64, arm, arm64} to {x64, arm, arm64, mips, mips64}. > > Overview of changes: > - new register configuration configuration with one register reserved for > the speculation poison/mask (kSpeculationPoisonRegister). > - in codegen, we introduce an update to the poison register at the starts > of all successors of branches (and deopts) that are marked as safety > branches (deopts). > - in memory optimizer, we lower all field and element loads to PoisonedLoads. > - poisoned loads are then masked in codegen with the poison register. > * only integer loads are masked at the moment. > > Bug: chromium:798964 > Change-Id: I211395b8305ed0ad9288d6da48fa159fa970c827 > Reviewed-on: https://chromium-review.googlesource.com/951382 > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > Reviewed-by: Ivica Bogosavljevic <ivica.bogosavljevic@mips.com> > Commit-Queue: Ivica Bogosavljevic <ivica.bogosavljevic@mips.com> > Cr-Commit-Position: refs/heads/master@{#52042} TBR=mvstanton@chromium.org,mstarzinger@chromium.org,ivica.bogosavljevic@mips.com Change-Id: Ief4d9ef56d918172f0b545d321a64b1ab5b46915 No-Presubmit: true No-Tree-Checks: true No-Try: true Bug: chromium:798964 Reviewed-on: https://chromium-review.googlesource.com/969041Reviewed-by:
Michael Stanton <mvstanton@chromium.org> Commit-Queue: Michael Stanton <mvstanton@chromium.org> Cr-Commit-Position: refs/heads/master@{#52045}
-
Kim-Anh Tran authored
Synchronous compilation currently continues creating new tasks even though compilation has already failed. This stops the creation of new background tasks and makes sure that the background task manager in the CompilationState is not canceled twice. Change-Id: Ic4c55275ff70e7eca901ad357253f81aa8e2e8e1 Reviewed-on: https://chromium-review.googlesource.com/968781 Commit-Queue: Kim-Anh Tran <kimanh@google.com> Reviewed-by:
Clemens Hammacher <clemensh@chromium.org> Cr-Commit-Position: refs/heads/master@{#52044}
-
Clemens Hammacher authored
EmitMonomorphicBinOp and EmitBinOpWithDifferentResultType were nearly identical, they just differ in one argument to GetUnusedRegister. This CL merges them. R=titzer@chromium.org Bug: v8:6600 Change-Id: Ief75beb410c8ba248b43cd382693f25bd9153d74 Reviewed-on: https://chromium-review.googlesource.com/968501 Commit-Queue: Clemens Hammacher <clemensh@chromium.org> Reviewed-by:
-
Mike Stanton authored
This introduces masking of loads with speculation bit during code generation. At the moment, this is done only under the --branch-load-poisoning flag, and this CL enlarges the set of supported platforms from {x64, arm, arm64} to {x64, arm, arm64, mips, mips64}. Overview of changes: - new register configuration configuration with one register reserved for the speculation poison/mask (kSpeculationPoisonRegister). - in codegen, we introduce an update to the poison register at the starts of all successors of branches (and deopts) that are marked as safety branches (deopts). - in memory optimizer, we lower all field and element loads to PoisonedLoads. - poisoned loads are then masked in codegen with the poison register. * only integer loads are masked at the moment. Bug: chromium:798964 Change-Id: I211395b8305ed0ad9288d6da48fa159fa970c827 Reviewed-on: https://chromium-review.googlesource.com/951382Reviewed-by: -
Marja Hölttä authored
For problems with the current approach, see crbug.com/v8/7564. We can instead gather all weak references in code in VisitEmbeddedPointer. BUG=v8:7564, v8:7308 Change-Id: Ib369e7ab9efd62c90bdac69835318929c58217f2 Reviewed-on: https://chromium-review.googlesource.com/968250 Commit-Queue: Marja Hölttä <marja@chromium.org> Reviewed-by:
Ulan Degenbaev <ulan@chromium.org> Cr-Commit-Position: refs/heads/master@{#52041}
-
Georg Neis authored
This type is dangerous because it can become smaller over time (as strings get internalized). Bug: v8:6521 Change-Id: Iea650789ab52c13a0519f46999edc8a7959ccc71 Reviewed-on: https://chromium-review.googlesource.com/968525 Commit-Queue: Georg Neis <neis@chromium.org> Reviewed-by:
-
jgruber authored
This is another step towards efficient external reference access from off-heap builtins. Bug: v8:6666 Change-Id: I6f128a8a64724acaa092f28b1fceae01e3c05487 Reviewed-on: https://chromium-review.googlesource.com/968481Reviewed-by:
Peter Marshall <petermarshall@chromium.org> Commit-Queue: Jakob Gruber <jgruber@chromium.org> Cr-Commit-Position: refs/heads/master@{#52039}
-
