Skip to content

V
V8
Switch branch/tag
  1. 22 Apr, 2020 13 commits
    • Anton Bikineev's avatar
      cppgc: Port FreeList implementation · 308914cc
      Anton Bikineev authored 
      - implemented as a single-linked list with head and tail
  pointers. The tail pointer is needed for freelist appending;
- stores entries in buckets, where bucket[log2(size)] stores
  entries >= size;
- implements worst fit allocation to amortize free list call;
- ported from Blink: https://bit.ly/2yC8XKJ.

Bug: chromium:1056170
Change-Id: I26cf62c948c95a7cbfecd5f7f22ad975e6b8c732
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2157376
Commit-Queue: Anton Bikineev <bikineev@chromium.org>
Reviewed-by: 's avatarMichael Lippautz <mlippautz@chromium.org>
Reviewed-by: 's avatarOmer Katz <omerkatz@chromium.org>
Reviewed-by: 's avatarUlan Degenbaev <ulan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#67310}
      308914cc
    • Leszek Swirski's avatar
      Reland^5 "[parser] Introduce UnoptimizedCompileFlags" · a3228bfc
      Leszek Swirski authored 
      This is a reland of e1b93a4f
which was a reland of 313d4844
which was a reland of 0a59e0cb
which was a reland of 146f5375
which was a reland of d91679bf

Give up on using C++ bitfields, go back to having base::BitField and
getters/setters.

Original change's description:
> [parser] Introduce UnoptimizedCompileFlags
>
> UnoptimizedCompileFlags defines the input flags shared between parse and
> compile (currently parse-only). It is set initially with some values, and
> is immutable after being passed to ParseInfo (ParseInfo still has getters
> for the fields, but no setters).
>
> Since a few of the existing flags were output flags, ParseInfo now has a
> new output_flags field, which will eventually migrate to a ParseOutputs
> structure.
>
> Bug: v8:10314
> Change-Id: If3890a5fad883bca80a97bf9dfe44d91797dc286
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2096580
> Commit-Queue: Leszek Swirski <leszeks@chromium.org>
> Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> Reviewed-by: Simon Zünd <szuend@chromium.org>
> Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#66782}

TBR=ulan@chromium.org,szuend@chromium.org

Bug: v8:10314
Change-Id: I54bcd107a0e85cf1a2ddeef0759100547eb65652
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2157378Reviewed-by: 's avatarLeszek Swirski <leszeks@chromium.org>
Reviewed-by: 's avatarRoss McIlroy <rmcilroy@chromium.org>
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#67309}
      a3228bfc
    • Clemens Backes's avatar
      [wasm] Don't block compilation on publishing · 1262eb02
      Clemens Backes authored 
      If multiple compilation threads want to publish their generated code,
they currently block on each other. This can cause multiple background
threads to be blocked for several hundred milliseconds in the worst
case.

This CL changes this such that instead of blocking, the threads just
put the code in a queue from where it is picked up by the thread that
is currently publishing. Instead of blocking, the threads can then
continue compiling more code already.

This change might produce regressions, because there is now more
TurboFan tier-up compilation happening while Liftoff code is being
published. This might delay the completion of baseline compilation. It
can also happen that we publish (more) TurboFan code before finishing
baseline compilation, which would also regress compile scores.

Let's see what the perf bots have to say about this CL. We might need to
adapt certain things (like delaying TurboFan compilation until all
Liftoff code finished), or we might just accept slight delays in Liftoff
compilation, because tier-up will finish sooner after this CL, giving us
peak performance earlier.

R=ahaas@chromium.org

Bug: v8:10330
Change-Id: I2f5c15810a0a9fc18461f9cbf4e436ab36aa559d
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2154200
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: 's avatarAndreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#67308}
      1262eb02
    • Andreas Haas's avatar
      [wasm][arm] Don't support overlapping LiftoffRegisters in I64Binops · ac518002
      Andreas Haas authored 
      The existing code supports overlapping LiftoffRegisters in I64Binops by
allocating a scratch register. Overlapping LiftoffRegisters means that
the low register of one LiftoffRegister is the high register of another
LiftoffRegister. When I64Binop is used in the implementation of atomic
instructions, no scratch register is available, so overlapping
LiftoffRegisters cannot be supported. As LiftoffRegisters are
not supposed to overlap anyways, this CL removes this special handling.

R=clemensb@chromium.org

Bug: v8:10108
Change-Id: I9ce2143b0930688bc2e6dd56e151f47fba00ec87
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2151351
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Reviewed-by: 's avatarClemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#67307}
      ac518002
    • Dan Elphick's avatar
      [snapshot] Rename partial snapshot to context snapshot · 71bd50e1
      Dan Elphick authored 
      The partial snapshot/serializer/deserializer are renamed to context *,
while the partial snapshot cache is renamed to startup object cache in
line with the read-only object cache (as this better reflects where it
lives and what it does).

To avoid a gap in the file history due to renaming both the files and
identifiers simulataneously, this leaves all the partial-*.* files in
place. They will be renamed in a follow-up CL.

Bug: v8:10416
Change-Id: I5ef41cad751aaa24b35ee2b3c72bd0295832f2c6
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2144115
Commit-Queue: Dan Elphick <delphick@chromium.org>
Commit-Queue: Ulan Degenbaev <ulan@chromium.org>
Auto-Submit: Dan Elphick <delphick@chromium.org>
Reviewed-by: 's avatarUlan Degenbaev <ulan@chromium.org>
Reviewed-by: 's avatarJakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#67306}
      71bd50e1
    • Seth Brenith's avatar
      Reland "[torque] Allow storing to bitfield structs that are stored in Smis" · a9270e2b
      Seth Brenith authored 
      This is a reland of 80843eda

Original change's description:
> [torque] Allow storing to bitfield structs that are stored in Smis
> 
> This change:
> 1. Updates the Torque compiler to allow direct access to bitfields that
>    are packed within Smi values, which previously would have required a
>    separate untagging step,
> 2. Updates JSRegExpStringIterator to represent its flags in Torque,
> 3. Adds reduction cases in MachineOperatorReducer for when the input to
>    a branch or the left-hand side of a Word32Equals is based on a 64-bit
>    shift-and-mask operation which has been truncated to 32 bits, as is
>    the case in the code generated by step 1, and
> 4. Adds a reduction case in MachineOperatorReducer to remove an extra
>    Word64And operation added by step 1.
> 
> Bug: v8:7793
> Change-Id: Ib4ac2def6211b3cae6be25a8b2a644be5c7d6d3f
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2119225
> Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
> Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#67290}

Bug: v8:7793
Change-Id: I783b6ec080042fec0e922927f6675dede458a072
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2159731Reviewed-by: 's avatarTobias Tebbi <tebbi@chromium.org>
Commit-Queue: Tobias Tebbi <tebbi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#67305}
      a9270e2b
    • Tobias Tebbi's avatar
      [torque] improve error message for const-bound this · 5733c063
      Tobias Tebbi authored 
      Bug: v8:7793
Change-Id: Id0478287b2147ea231cd19dab6d08674b632b0e9
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2159494
Commit-Queue: Tobias Tebbi <tebbi@chromium.org>
Commit-Queue: Marja Hölttä <marja@chromium.org>
Auto-Submit: Tobias Tebbi <tebbi@chromium.org>
Reviewed-by: 's avatarMarja Hölttä <marja@chromium.org>
Cr-Commit-Position: refs/heads/master@{#67304}
      5733c063
    • Clemens Backes's avatar
      [wasm] Fix performance bottleneck in DisjointAllocationPool · 7e0279fa
      Clemens Backes authored 
      When compiling modules with many functions, the list of regions in the
{DisjointAllocationPool} can become quite large if the functions die in
a random order (which they typically do, since the order of Liftoff
compilation is different than the order to TurboFan compilation; which
work stealing, both are nondeterministic).
Iterating the list of regions in the {DisjointAllocationPool} was thus
linear in the number of regions, which is linear in the number of
functions of the module. Since we insert new regions one by one, overall
runtime was quadratic.

This CL fixes this by switching from a linked list to a std::set.
Merging a new region is thus logarithmic instead of linear, and overall
we are {n*log(n)} instead of {n^2}.

Note: For {AllocateInRegion} we still need to linearly iterate all
regions that overlap the requested region, but this has not shown to be
a problem so far.

R=ahaas@chromium.org

Bug: v8:10432
Change-Id: I193e56c2abab782e386194fbe64dadfa250916f7
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2154797
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: 's avatarAndreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#67303}
      7e0279fa
    • Andreas Haas's avatar
      [wasm][liftoff][arm] Implement 64-bit atomic binops · 2d45932f
      Andreas Haas authored 
      With this CL we reuse the implementation of AtomicStore to implement
AtomicBinops. The implementation is parametrised by an optional result
Liftoff register, and by an operation that is applied on the value
stored in the memory location.

R=clemensb@chromium.org, v8-arm-ports@googlegroups.com

Bug: v8:10108
Change-Id: I71ff5bf1632d860462d66b43c34515400eae2979
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2151349
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Reviewed-by: 's avatarClemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#67302}
      2d45932f
    • Kong, Fanchen's avatar
      [wasm-simd] [liftoff] Implement fp min/max on ia32 and x64 · 9f41a584
      Kong, Fanchen authored 
      Bug: v8:9909
Change-Id: Ib97bcc7afe516a014cd91128aa3c59f1b8b0b0af
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2151999
Commit-Queue: Fanchen Kong <fanchen.kong@intel.com>
Reviewed-by: 's avatarClemens Backes <clemensb@chromium.org>
Reviewed-by: 's avatarZhi An Ng <zhin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#67301}
      9f41a584
    • Kim-Anh Tran's avatar
      [wasm] Set source url when deserializing wasm module · ec3cadc4
      Kim-Anh Tran authored 
      Bug: v8:1071760
Change-Id: Id2f75aef039574b918768fb323e1a8dbc999606f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2159491Reviewed-by: 's avatarClemens Backes <clemensb@chromium.org>
Commit-Queue: Kim-Anh Tran <kimanh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#67300}
      ec3cadc4
    • Yolanda Chen's avatar
      [regalloc] Fix a regression when enable FindOptimalSpillingPos for phis · 8bfd278c
      Yolanda Chen authored 
      This patch is to avoid spilling the phi at the loop header if there is a back-edge with an input for the phi that interferes with the phi's value, because in case that input gets spilled it might introduce a stack-to-stack move at the back-edge.

Bug: chromium:1063831
Change-Id: Ie7129f10fb573cc799c588e6639b5ad486ea520d
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2152002
Commit-Queue: Yolanda Chen <yolanda.chen@intel.com>
Reviewed-by: 's avatarSigurd Schneider <sigurds@chromium.org>
Reviewed-by: 's avatarThibaud Michaud <thibaudm@chromium.org>
Cr-Commit-Position: refs/heads/master@{#67299}
      8bfd278c
    • Jakob Gruber's avatar
      Reland "[snapshot] Extract more files" · d587f67a
      Jakob Gruber authored 
      This is a reland of 5c4b8056

Original change's description:
> [snapshot] Extract more files
>
> This moves:
>
> - ExternalReferenceEncoder to codegen/external-reference-encoder.h
> - SerializerDeserializer to snapshot/serializer-deserializer.h
> - Checksum() to snapshot/snapshot-utils.h
>
> serializer-common.h and .cc are removed.
>
> Tbr: clemensb@chromium.org,ulan@chromium.org
> Bug: v8:10416
> Change-Id: I36a242dcc1ad8833374aa567f73e0d4a75632c58
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2144118
> Commit-Queue: Jakob Gruber <jgruber@chromium.org>
> Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> Reviewed-by: Clemens Backes <clemensb@chromium.org>
> Reviewed-by: Dan Elphick <delphick@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#67281}

Tbr: delphick@chromium.org,clemensb@chromium.org,ulan@chromium.org
Bug: v8:10416
Change-Id: I6f6a1017435db185778ed931e1ddb13d8d5e920e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2157384Reviewed-by: 's avatarJakob Gruber <jgruber@chromium.org>
Reviewed-by: 's avatarDan Elphick <delphick@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#67298}
      d587f67a
  3. 21 Apr, 2020 27 commits