v8/src/wasm/wasm-code-manager.h

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

747 lines
28 KiB
C
Raw Normal View History

Revert "Revert "[wasm] A simple allocator datastructure for off-the heap"" This reverts commit ee5c31f3354ecd42a99e45a153061bc8b7a0a980. Reason for revert: Fixed compiler failure Original change's description: > Revert "[wasm] A simple allocator datastructure for off-the heap" > > This reverts commit 110d9ab0052f9c70c0c44e9c30397789568bbebf. > > Reason for revert: https://build.chromium.org/p/client.v8/builders/V8%20Linux64%20-%20debug%20builder/builds/26607 > > Surprising we're seeing a failure on Linux 64 *after* CQ. Is the compiler there different? > > Original change's description: > > [wasm] A simple allocator datastructure for off-the heap > > > > We'll use this allocator in a follow-up CL to: > > - allocate speculative sizes of memory for a module that's being > > compiled (e.g. 2*size of wasm code). > > - each module will own such a sub-pool, and then use it to allocate > > contiguous chunks of memory for code. > > > > The underlying assumptions for the chosen allocation strategy is that: > > - the allocation granularity for pools is 1 page, so that no one page > > is owned by more than one wasm module > > - typical pool sizes (given module sizes) are multiple pages. > > - modules and module instances are typically few and long lived. Typically, > > we expect one module and one instance. > > > > This means we shouldn't expect fragmentations that lead to code being > > non-allocatable, or prohibitively many ranges. > > > > The data structure just manages ranges of addresses. Virtual memory management > > will be separate, as part of the responsibility of a "WasmHeap" > > that will be introduced in the future. So will concurrency control. > > > > Bug: > > Change-Id: Id99f46d10c25553b013054d994760f3c2a737c39 > > Reviewed-on: https://chromium-review.googlesource.com/669296 > > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > > Reviewed-by: Eric Holk <eholk@chromium.org> > > Reviewed-by: Brad Nelson <bradnelson@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#48053} > > TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org > > Change-Id: Id82fa341b77624e4971f24c4757a9a666a65930c > No-Presubmit: true > No-Tree-Checks: true > No-Try: true > Reviewed-on: https://chromium-review.googlesource.com/670141 > Reviewed-by: Mircea Trofin <mtrofin@chromium.org> > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > Cr-Commit-Position: refs/heads/master@{#48054} TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org Change-Id: Ib6a7a3e6098d2689e60cdca85ec77e57e5295e48 No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://chromium-review.googlesource.com/670142 Commit-Queue: Mircea Trofin <mtrofin@chromium.org> Reviewed-by: Mircea Trofin <mtrofin@chromium.org> Cr-Commit-Position: refs/heads/master@{#48055}
2017-09-16 05:22:38 +00:00
// Copyright 2017 the V8 project authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
#ifndef V8_WASM_WASM_CODE_MANAGER_H_
#define V8_WASM_WASM_CODE_MANAGER_H_
Revert "Revert "[wasm] A simple allocator datastructure for off-the heap"" This reverts commit ee5c31f3354ecd42a99e45a153061bc8b7a0a980. Reason for revert: Fixed compiler failure Original change's description: > Revert "[wasm] A simple allocator datastructure for off-the heap" > > This reverts commit 110d9ab0052f9c70c0c44e9c30397789568bbebf. > > Reason for revert: https://build.chromium.org/p/client.v8/builders/V8%20Linux64%20-%20debug%20builder/builds/26607 > > Surprising we're seeing a failure on Linux 64 *after* CQ. Is the compiler there different? > > Original change's description: > > [wasm] A simple allocator datastructure for off-the heap > > > > We'll use this allocator in a follow-up CL to: > > - allocate speculative sizes of memory for a module that's being > > compiled (e.g. 2*size of wasm code). > > - each module will own such a sub-pool, and then use it to allocate > > contiguous chunks of memory for code. > > > > The underlying assumptions for the chosen allocation strategy is that: > > - the allocation granularity for pools is 1 page, so that no one page > > is owned by more than one wasm module > > - typical pool sizes (given module sizes) are multiple pages. > > - modules and module instances are typically few and long lived. Typically, > > we expect one module and one instance. > > > > This means we shouldn't expect fragmentations that lead to code being > > non-allocatable, or prohibitively many ranges. > > > > The data structure just manages ranges of addresses. Virtual memory management > > will be separate, as part of the responsibility of a "WasmHeap" > > that will be introduced in the future. So will concurrency control. > > > > Bug: > > Change-Id: Id99f46d10c25553b013054d994760f3c2a737c39 > > Reviewed-on: https://chromium-review.googlesource.com/669296 > > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > > Reviewed-by: Eric Holk <eholk@chromium.org> > > Reviewed-by: Brad Nelson <bradnelson@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#48053} > > TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org > > Change-Id: Id82fa341b77624e4971f24c4757a9a666a65930c > No-Presubmit: true > No-Tree-Checks: true > No-Try: true > Reviewed-on: https://chromium-review.googlesource.com/670141 > Reviewed-by: Mircea Trofin <mtrofin@chromium.org> > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > Cr-Commit-Position: refs/heads/master@{#48054} TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org Change-Id: Ib6a7a3e6098d2689e60cdca85ec77e57e5295e48 No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://chromium-review.googlesource.com/670142 Commit-Queue: Mircea Trofin <mtrofin@chromium.org> Reviewed-by: Mircea Trofin <mtrofin@chromium.org> Cr-Commit-Position: refs/heads/master@{#48055}
2017-09-16 05:22:38 +00:00
#include <atomic>
Revert "Revert "[wasm] A simple allocator datastructure for off-the heap"" This reverts commit ee5c31f3354ecd42a99e45a153061bc8b7a0a980. Reason for revert: Fixed compiler failure Original change's description: > Revert "[wasm] A simple allocator datastructure for off-the heap" > > This reverts commit 110d9ab0052f9c70c0c44e9c30397789568bbebf. > > Reason for revert: https://build.chromium.org/p/client.v8/builders/V8%20Linux64%20-%20debug%20builder/builds/26607 > > Surprising we're seeing a failure on Linux 64 *after* CQ. Is the compiler there different? > > Original change's description: > > [wasm] A simple allocator datastructure for off-the heap > > > > We'll use this allocator in a follow-up CL to: > > - allocate speculative sizes of memory for a module that's being > > compiled (e.g. 2*size of wasm code). > > - each module will own such a sub-pool, and then use it to allocate > > contiguous chunks of memory for code. > > > > The underlying assumptions for the chosen allocation strategy is that: > > - the allocation granularity for pools is 1 page, so that no one page > > is owned by more than one wasm module > > - typical pool sizes (given module sizes) are multiple pages. > > - modules and module instances are typically few and long lived. Typically, > > we expect one module and one instance. > > > > This means we shouldn't expect fragmentations that lead to code being > > non-allocatable, or prohibitively many ranges. > > > > The data structure just manages ranges of addresses. Virtual memory management > > will be separate, as part of the responsibility of a "WasmHeap" > > that will be introduced in the future. So will concurrency control. > > > > Bug: > > Change-Id: Id99f46d10c25553b013054d994760f3c2a737c39 > > Reviewed-on: https://chromium-review.googlesource.com/669296 > > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > > Reviewed-by: Eric Holk <eholk@chromium.org> > > Reviewed-by: Brad Nelson <bradnelson@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#48053} > > TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org > > Change-Id: Id82fa341b77624e4971f24c4757a9a666a65930c > No-Presubmit: true > No-Tree-Checks: true > No-Try: true > Reviewed-on: https://chromium-review.googlesource.com/670141 > Reviewed-by: Mircea Trofin <mtrofin@chromium.org> > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > Cr-Commit-Position: refs/heads/master@{#48054} TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org Change-Id: Ib6a7a3e6098d2689e60cdca85ec77e57e5295e48 No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://chromium-review.googlesource.com/670142 Commit-Queue: Mircea Trofin <mtrofin@chromium.org> Reviewed-by: Mircea Trofin <mtrofin@chromium.org> Cr-Commit-Position: refs/heads/master@{#48055}
2017-09-16 05:22:38 +00:00
#include <list>
2017-11-20 21:34:04 +00:00
#include <map>
#include <memory>
#include <unordered_set>
#include <utility>
#include <vector>
Revert "Revert "[wasm] A simple allocator datastructure for off-the heap"" This reverts commit ee5c31f3354ecd42a99e45a153061bc8b7a0a980. Reason for revert: Fixed compiler failure Original change's description: > Revert "[wasm] A simple allocator datastructure for off-the heap" > > This reverts commit 110d9ab0052f9c70c0c44e9c30397789568bbebf. > > Reason for revert: https://build.chromium.org/p/client.v8/builders/V8%20Linux64%20-%20debug%20builder/builds/26607 > > Surprising we're seeing a failure on Linux 64 *after* CQ. Is the compiler there different? > > Original change's description: > > [wasm] A simple allocator datastructure for off-the heap > > > > We'll use this allocator in a follow-up CL to: > > - allocate speculative sizes of memory for a module that's being > > compiled (e.g. 2*size of wasm code). > > - each module will own such a sub-pool, and then use it to allocate > > contiguous chunks of memory for code. > > > > The underlying assumptions for the chosen allocation strategy is that: > > - the allocation granularity for pools is 1 page, so that no one page > > is owned by more than one wasm module > > - typical pool sizes (given module sizes) are multiple pages. > > - modules and module instances are typically few and long lived. Typically, > > we expect one module and one instance. > > > > This means we shouldn't expect fragmentations that lead to code being > > non-allocatable, or prohibitively many ranges. > > > > The data structure just manages ranges of addresses. Virtual memory management > > will be separate, as part of the responsibility of a "WasmHeap" > > that will be introduced in the future. So will concurrency control. > > > > Bug: > > Change-Id: Id99f46d10c25553b013054d994760f3c2a737c39 > > Reviewed-on: https://chromium-review.googlesource.com/669296 > > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > > Reviewed-by: Eric Holk <eholk@chromium.org> > > Reviewed-by: Brad Nelson <bradnelson@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#48053} > > TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org > > Change-Id: Id82fa341b77624e4971f24c4757a9a666a65930c > No-Presubmit: true > No-Tree-Checks: true > No-Try: true > Reviewed-on: https://chromium-review.googlesource.com/670141 > Reviewed-by: Mircea Trofin <mtrofin@chromium.org> > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > Cr-Commit-Position: refs/heads/master@{#48054} TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org Change-Id: Ib6a7a3e6098d2689e60cdca85ec77e57e5295e48 No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://chromium-review.googlesource.com/670142 Commit-Queue: Mircea Trofin <mtrofin@chromium.org> Reviewed-by: Mircea Trofin <mtrofin@chromium.org> Cr-Commit-Position: refs/heads/master@{#48055}
2017-09-16 05:22:38 +00:00
#include "src/base/address-region.h"
Revert "Revert "[wasm] A simple allocator datastructure for off-the heap"" This reverts commit ee5c31f3354ecd42a99e45a153061bc8b7a0a980. Reason for revert: Fixed compiler failure Original change's description: > Revert "[wasm] A simple allocator datastructure for off-the heap" > > This reverts commit 110d9ab0052f9c70c0c44e9c30397789568bbebf. > > Reason for revert: https://build.chromium.org/p/client.v8/builders/V8%20Linux64%20-%20debug%20builder/builds/26607 > > Surprising we're seeing a failure on Linux 64 *after* CQ. Is the compiler there different? > > Original change's description: > > [wasm] A simple allocator datastructure for off-the heap > > > > We'll use this allocator in a follow-up CL to: > > - allocate speculative sizes of memory for a module that's being > > compiled (e.g. 2*size of wasm code). > > - each module will own such a sub-pool, and then use it to allocate > > contiguous chunks of memory for code. > > > > The underlying assumptions for the chosen allocation strategy is that: > > - the allocation granularity for pools is 1 page, so that no one page > > is owned by more than one wasm module > > - typical pool sizes (given module sizes) are multiple pages. > > - modules and module instances are typically few and long lived. Typically, > > we expect one module and one instance. > > > > This means we shouldn't expect fragmentations that lead to code being > > non-allocatable, or prohibitively many ranges. > > > > The data structure just manages ranges of addresses. Virtual memory management > > will be separate, as part of the responsibility of a "WasmHeap" > > that will be introduced in the future. So will concurrency control. > > > > Bug: > > Change-Id: Id99f46d10c25553b013054d994760f3c2a737c39 > > Reviewed-on: https://chromium-review.googlesource.com/669296 > > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > > Reviewed-by: Eric Holk <eholk@chromium.org> > > Reviewed-by: Brad Nelson <bradnelson@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#48053} > > TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org > > Change-Id: Id82fa341b77624e4971f24c4757a9a666a65930c > No-Presubmit: true > No-Tree-Checks: true > No-Try: true > Reviewed-on: https://chromium-review.googlesource.com/670141 > Reviewed-by: Mircea Trofin <mtrofin@chromium.org> > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > Cr-Commit-Position: refs/heads/master@{#48054} TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org Change-Id: Ib6a7a3e6098d2689e60cdca85ec77e57e5295e48 No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://chromium-review.googlesource.com/670142 Commit-Queue: Mircea Trofin <mtrofin@chromium.org> Reviewed-by: Mircea Trofin <mtrofin@chromium.org> Cr-Commit-Position: refs/heads/master@{#48055}
2017-09-16 05:22:38 +00:00
#include "src/base/macros.h"
#include "src/base/optional.h"
#include "src/builtins/builtins-definitions.h"
#include "src/handles/handles.h"
2017-11-20 21:34:04 +00:00
#include "src/trap-handler/trap-handler.h"
#include "src/utils/vector.h"
#include "src/wasm/compilation-environment.h"
#include "src/wasm/wasm-features.h"
#include "src/wasm/wasm-limits.h"
#include "src/wasm/wasm-tier.h"
Revert "Revert "[wasm] A simple allocator datastructure for off-the heap"" This reverts commit ee5c31f3354ecd42a99e45a153061bc8b7a0a980. Reason for revert: Fixed compiler failure Original change's description: > Revert "[wasm] A simple allocator datastructure for off-the heap" > > This reverts commit 110d9ab0052f9c70c0c44e9c30397789568bbebf. > > Reason for revert: https://build.chromium.org/p/client.v8/builders/V8%20Linux64%20-%20debug%20builder/builds/26607 > > Surprising we're seeing a failure on Linux 64 *after* CQ. Is the compiler there different? > > Original change's description: > > [wasm] A simple allocator datastructure for off-the heap > > > > We'll use this allocator in a follow-up CL to: > > - allocate speculative sizes of memory for a module that's being > > compiled (e.g. 2*size of wasm code). > > - each module will own such a sub-pool, and then use it to allocate > > contiguous chunks of memory for code. > > > > The underlying assumptions for the chosen allocation strategy is that: > > - the allocation granularity for pools is 1 page, so that no one page > > is owned by more than one wasm module > > - typical pool sizes (given module sizes) are multiple pages. > > - modules and module instances are typically few and long lived. Typically, > > we expect one module and one instance. > > > > This means we shouldn't expect fragmentations that lead to code being > > non-allocatable, or prohibitively many ranges. > > > > The data structure just manages ranges of addresses. Virtual memory management > > will be separate, as part of the responsibility of a "WasmHeap" > > that will be introduced in the future. So will concurrency control. > > > > Bug: > > Change-Id: Id99f46d10c25553b013054d994760f3c2a737c39 > > Reviewed-on: https://chromium-review.googlesource.com/669296 > > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > > Reviewed-by: Eric Holk <eholk@chromium.org> > > Reviewed-by: Brad Nelson <bradnelson@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#48053} > > TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org > > Change-Id: Id82fa341b77624e4971f24c4757a9a666a65930c > No-Presubmit: true > No-Tree-Checks: true > No-Try: true > Reviewed-on: https://chromium-review.googlesource.com/670141 > Reviewed-by: Mircea Trofin <mtrofin@chromium.org> > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > Cr-Commit-Position: refs/heads/master@{#48054} TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org Change-Id: Ib6a7a3e6098d2689e60cdca85ec77e57e5295e48 No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://chromium-review.googlesource.com/670142 Commit-Queue: Mircea Trofin <mtrofin@chromium.org> Reviewed-by: Mircea Trofin <mtrofin@chromium.org> Cr-Commit-Position: refs/heads/master@{#48055}
2017-09-16 05:22:38 +00:00
namespace v8 {
namespace internal {
2017-11-20 21:34:04 +00:00
class Code;
class CodeDesc;
class Isolate;
2017-11-20 21:34:04 +00:00
Revert "Revert "[wasm] A simple allocator datastructure for off-the heap"" This reverts commit ee5c31f3354ecd42a99e45a153061bc8b7a0a980. Reason for revert: Fixed compiler failure Original change's description: > Revert "[wasm] A simple allocator datastructure for off-the heap" > > This reverts commit 110d9ab0052f9c70c0c44e9c30397789568bbebf. > > Reason for revert: https://build.chromium.org/p/client.v8/builders/V8%20Linux64%20-%20debug%20builder/builds/26607 > > Surprising we're seeing a failure on Linux 64 *after* CQ. Is the compiler there different? > > Original change's description: > > [wasm] A simple allocator datastructure for off-the heap > > > > We'll use this allocator in a follow-up CL to: > > - allocate speculative sizes of memory for a module that's being > > compiled (e.g. 2*size of wasm code). > > - each module will own such a sub-pool, and then use it to allocate > > contiguous chunks of memory for code. > > > > The underlying assumptions for the chosen allocation strategy is that: > > - the allocation granularity for pools is 1 page, so that no one page > > is owned by more than one wasm module > > - typical pool sizes (given module sizes) are multiple pages. > > - modules and module instances are typically few and long lived. Typically, > > we expect one module and one instance. > > > > This means we shouldn't expect fragmentations that lead to code being > > non-allocatable, or prohibitively many ranges. > > > > The data structure just manages ranges of addresses. Virtual memory management > > will be separate, as part of the responsibility of a "WasmHeap" > > that will be introduced in the future. So will concurrency control. > > > > Bug: > > Change-Id: Id99f46d10c25553b013054d994760f3c2a737c39 > > Reviewed-on: https://chromium-review.googlesource.com/669296 > > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > > Reviewed-by: Eric Holk <eholk@chromium.org> > > Reviewed-by: Brad Nelson <bradnelson@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#48053} > > TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org > > Change-Id: Id82fa341b77624e4971f24c4757a9a666a65930c > No-Presubmit: true > No-Tree-Checks: true > No-Try: true > Reviewed-on: https://chromium-review.googlesource.com/670141 > Reviewed-by: Mircea Trofin <mtrofin@chromium.org> > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > Cr-Commit-Position: refs/heads/master@{#48054} TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org Change-Id: Ib6a7a3e6098d2689e60cdca85ec77e57e5295e48 No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://chromium-review.googlesource.com/670142 Commit-Queue: Mircea Trofin <mtrofin@chromium.org> Reviewed-by: Mircea Trofin <mtrofin@chromium.org> Cr-Commit-Position: refs/heads/master@{#48055}
2017-09-16 05:22:38 +00:00
namespace wasm {
2017-11-20 21:34:04 +00:00
class NativeModule;
class WasmCodeManager;
struct WasmCompilationResult;
class WasmEngine;
Revert ""Reland x3 [arraybuffer] Rearchitect backing store ownership"" This reverts commit df8e61777295ad5568eff27722a99fab765eabbd. Reason for revert: Multiple flakes in apparently related areas: https://logs.chromium.org/logs/v8/buildbucket/cr-buildbucket.appspot.com/8906409837768155568/+/steps/Check__flakes_/0/logs/BackingStoreTest.RacyGrowWasmMem.../0 Original change's description: > "Reland x3 [arraybuffer] Rearchitect backing store ownership" > > This is a reland of bc33f5aeba9ceb13f8bfc401c5ba2521c2207ffb > > Original change's description: > > [arraybuffer] Rearchitect backing store ownership > > > > This CL completely rearchitects the ownership of array buffer backing stores, > > consolidating ownership into a {BackingStore} C++ object that is tracked > > throughout V8 using unique_ptr and shared_ptr where appropriate. > > > > Overall, lifetime management is simpler and more explicit. The numerous > > ways that array buffers were initialized have been streamlined to one > > Attach() method on JSArrayBuffer. The array buffer tracker in the > > GC implementation now manages std::shared_ptr<BackingStore> pointers, > > and the construction and destruction of the BackingStore object itself > > handles the underlying page or embedder-allocated memory. > > > > The embedder API remains unchanged for now. We use the > > v8::ArrayBuffer::Contents struct to hide an additional shared_ptr to > > keep the backing store alive properly, even in the case of aliases > > from live heap objects. Thus the embedder has a lower chance of making > > a mistake. Long-term, we should move the embedder to a model where they > > manage backing stores using shared_ptr to an opaque backing store object. > > R=​mlippautz@chromium.org > BUG=v8:9380,v8:9221,chromium:986318 > TBR=ulan@chromium.org > > Change-Id: I6c49e2425029b5664ef1c68dab8b5146f4ed0ff2 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1719191 > Reviewed-by: Ben Titzer <titzer@chromium.org> > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > Commit-Queue: Ben Titzer <titzer@chromium.org> > Cr-Commit-Position: refs/heads/master@{#63007} TBR=mstarzinger@chromium.org,titzer@chromium.org,mlippautz@chromium.org Change-Id: If0266e5893b1325a332d5986337fa7ece2cb6943 No-Presubmit: true No-Tree-Checks: true No-Try: true Bug: v8:9380, v8:9221, chromium:986318 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1729549 Reviewed-by: Francis McCabe <fgm@chromium.org> Commit-Queue: Francis McCabe <fgm@chromium.org> Cr-Commit-Position: refs/heads/master@{#63011}
2019-07-31 15:50:37 +00:00
class WasmMemoryTracker;
class WasmImportWrapperCache;
2017-11-20 21:34:04 +00:00
struct WasmModule;
// Sorted, disjoint and non-overlapping memory regions. A region is of the
Revert "Revert "[wasm] A simple allocator datastructure for off-the heap"" This reverts commit ee5c31f3354ecd42a99e45a153061bc8b7a0a980. Reason for revert: Fixed compiler failure Original change's description: > Revert "[wasm] A simple allocator datastructure for off-the heap" > > This reverts commit 110d9ab0052f9c70c0c44e9c30397789568bbebf. > > Reason for revert: https://build.chromium.org/p/client.v8/builders/V8%20Linux64%20-%20debug%20builder/builds/26607 > > Surprising we're seeing a failure on Linux 64 *after* CQ. Is the compiler there different? > > Original change's description: > > [wasm] A simple allocator datastructure for off-the heap > > > > We'll use this allocator in a follow-up CL to: > > - allocate speculative sizes of memory for a module that's being > > compiled (e.g. 2*size of wasm code). > > - each module will own such a sub-pool, and then use it to allocate > > contiguous chunks of memory for code. > > > > The underlying assumptions for the chosen allocation strategy is that: > > - the allocation granularity for pools is 1 page, so that no one page > > is owned by more than one wasm module > > - typical pool sizes (given module sizes) are multiple pages. > > - modules and module instances are typically few and long lived. Typically, > > we expect one module and one instance. > > > > This means we shouldn't expect fragmentations that lead to code being > > non-allocatable, or prohibitively many ranges. > > > > The data structure just manages ranges of addresses. Virtual memory management > > will be separate, as part of the responsibility of a "WasmHeap" > > that will be introduced in the future. So will concurrency control. > > > > Bug: > > Change-Id: Id99f46d10c25553b013054d994760f3c2a737c39 > > Reviewed-on: https://chromium-review.googlesource.com/669296 > > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > > Reviewed-by: Eric Holk <eholk@chromium.org> > > Reviewed-by: Brad Nelson <bradnelson@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#48053} > > TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org > > Change-Id: Id82fa341b77624e4971f24c4757a9a666a65930c > No-Presubmit: true > No-Tree-Checks: true > No-Try: true > Reviewed-on: https://chromium-review.googlesource.com/670141 > Reviewed-by: Mircea Trofin <mtrofin@chromium.org> > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > Cr-Commit-Position: refs/heads/master@{#48054} TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org Change-Id: Ib6a7a3e6098d2689e60cdca85ec77e57e5295e48 No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://chromium-review.googlesource.com/670142 Commit-Queue: Mircea Trofin <mtrofin@chromium.org> Reviewed-by: Mircea Trofin <mtrofin@chromium.org> Cr-Commit-Position: refs/heads/master@{#48055}
2017-09-16 05:22:38 +00:00
// form [start, end). So there's no [start, end), [end, other_end),
// because that should have been reduced to [start, other_end).
class V8_EXPORT_PRIVATE DisjointAllocationPool final {
public:
MOVE_ONLY_WITH_DEFAULT_CONSTRUCTORS(DisjointAllocationPool);
explicit DisjointAllocationPool(base::AddressRegion region)
: regions_({region}) {}
Revert "Revert "[wasm] A simple allocator datastructure for off-the heap"" This reverts commit ee5c31f3354ecd42a99e45a153061bc8b7a0a980. Reason for revert: Fixed compiler failure Original change's description: > Revert "[wasm] A simple allocator datastructure for off-the heap" > > This reverts commit 110d9ab0052f9c70c0c44e9c30397789568bbebf. > > Reason for revert: https://build.chromium.org/p/client.v8/builders/V8%20Linux64%20-%20debug%20builder/builds/26607 > > Surprising we're seeing a failure on Linux 64 *after* CQ. Is the compiler there different? > > Original change's description: > > [wasm] A simple allocator datastructure for off-the heap > > > > We'll use this allocator in a follow-up CL to: > > - allocate speculative sizes of memory for a module that's being > > compiled (e.g. 2*size of wasm code). > > - each module will own such a sub-pool, and then use it to allocate > > contiguous chunks of memory for code. > > > > The underlying assumptions for the chosen allocation strategy is that: > > - the allocation granularity for pools is 1 page, so that no one page > > is owned by more than one wasm module > > - typical pool sizes (given module sizes) are multiple pages. > > - modules and module instances are typically few and long lived. Typically, > > we expect one module and one instance. > > > > This means we shouldn't expect fragmentations that lead to code being > > non-allocatable, or prohibitively many ranges. > > > > The data structure just manages ranges of addresses. Virtual memory management > > will be separate, as part of the responsibility of a "WasmHeap" > > that will be introduced in the future. So will concurrency control. > > > > Bug: > > Change-Id: Id99f46d10c25553b013054d994760f3c2a737c39 > > Reviewed-on: https://chromium-review.googlesource.com/669296 > > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > > Reviewed-by: Eric Holk <eholk@chromium.org> > > Reviewed-by: Brad Nelson <bradnelson@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#48053} > > TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org > > Change-Id: Id82fa341b77624e4971f24c4757a9a666a65930c > No-Presubmit: true > No-Tree-Checks: true > No-Try: true > Reviewed-on: https://chromium-review.googlesource.com/670141 > Reviewed-by: Mircea Trofin <mtrofin@chromium.org> > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > Cr-Commit-Position: refs/heads/master@{#48054} TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org Change-Id: Ib6a7a3e6098d2689e60cdca85ec77e57e5295e48 No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://chromium-review.googlesource.com/670142 Commit-Queue: Mircea Trofin <mtrofin@chromium.org> Reviewed-by: Mircea Trofin <mtrofin@chromium.org> Cr-Commit-Position: refs/heads/master@{#48055}
2017-09-16 05:22:38 +00:00
// Merge the parameter region into this object while preserving ordering of
// the regions. The assumption is that the passed parameter is not
// intersecting this object - for example, it was obtained from a previous
// Allocate. Returns the merged region.
base::AddressRegion Merge(base::AddressRegion);
Revert "Revert "[wasm] A simple allocator datastructure for off-the heap"" This reverts commit ee5c31f3354ecd42a99e45a153061bc8b7a0a980. Reason for revert: Fixed compiler failure Original change's description: > Revert "[wasm] A simple allocator datastructure for off-the heap" > > This reverts commit 110d9ab0052f9c70c0c44e9c30397789568bbebf. > > Reason for revert: https://build.chromium.org/p/client.v8/builders/V8%20Linux64%20-%20debug%20builder/builds/26607 > > Surprising we're seeing a failure on Linux 64 *after* CQ. Is the compiler there different? > > Original change's description: > > [wasm] A simple allocator datastructure for off-the heap > > > > We'll use this allocator in a follow-up CL to: > > - allocate speculative sizes of memory for a module that's being > > compiled (e.g. 2*size of wasm code). > > - each module will own such a sub-pool, and then use it to allocate > > contiguous chunks of memory for code. > > > > The underlying assumptions for the chosen allocation strategy is that: > > - the allocation granularity for pools is 1 page, so that no one page > > is owned by more than one wasm module > > - typical pool sizes (given module sizes) are multiple pages. > > - modules and module instances are typically few and long lived. Typically, > > we expect one module and one instance. > > > > This means we shouldn't expect fragmentations that lead to code being > > non-allocatable, or prohibitively many ranges. > > > > The data structure just manages ranges of addresses. Virtual memory management > > will be separate, as part of the responsibility of a "WasmHeap" > > that will be introduced in the future. So will concurrency control. > > > > Bug: > > Change-Id: Id99f46d10c25553b013054d994760f3c2a737c39 > > Reviewed-on: https://chromium-review.googlesource.com/669296 > > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > > Reviewed-by: Eric Holk <eholk@chromium.org> > > Reviewed-by: Brad Nelson <bradnelson@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#48053} > > TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org > > Change-Id: Id82fa341b77624e4971f24c4757a9a666a65930c > No-Presubmit: true > No-Tree-Checks: true > No-Try: true > Reviewed-on: https://chromium-review.googlesource.com/670141 > Reviewed-by: Mircea Trofin <mtrofin@chromium.org> > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > Cr-Commit-Position: refs/heads/master@{#48054} TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org Change-Id: Ib6a7a3e6098d2689e60cdca85ec77e57e5295e48 No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://chromium-review.googlesource.com/670142 Commit-Queue: Mircea Trofin <mtrofin@chromium.org> Reviewed-by: Mircea Trofin <mtrofin@chromium.org> Cr-Commit-Position: refs/heads/master@{#48055}
2017-09-16 05:22:38 +00:00
// Allocate a contiguous region of size {size}. Return an empty pool on
Revert "Revert "[wasm] A simple allocator datastructure for off-the heap"" This reverts commit ee5c31f3354ecd42a99e45a153061bc8b7a0a980. Reason for revert: Fixed compiler failure Original change's description: > Revert "[wasm] A simple allocator datastructure for off-the heap" > > This reverts commit 110d9ab0052f9c70c0c44e9c30397789568bbebf. > > Reason for revert: https://build.chromium.org/p/client.v8/builders/V8%20Linux64%20-%20debug%20builder/builds/26607 > > Surprising we're seeing a failure on Linux 64 *after* CQ. Is the compiler there different? > > Original change's description: > > [wasm] A simple allocator datastructure for off-the heap > > > > We'll use this allocator in a follow-up CL to: > > - allocate speculative sizes of memory for a module that's being > > compiled (e.g. 2*size of wasm code). > > - each module will own such a sub-pool, and then use it to allocate > > contiguous chunks of memory for code. > > > > The underlying assumptions for the chosen allocation strategy is that: > > - the allocation granularity for pools is 1 page, so that no one page > > is owned by more than one wasm module > > - typical pool sizes (given module sizes) are multiple pages. > > - modules and module instances are typically few and long lived. Typically, > > we expect one module and one instance. > > > > This means we shouldn't expect fragmentations that lead to code being > > non-allocatable, or prohibitively many ranges. > > > > The data structure just manages ranges of addresses. Virtual memory management > > will be separate, as part of the responsibility of a "WasmHeap" > > that will be introduced in the future. So will concurrency control. > > > > Bug: > > Change-Id: Id99f46d10c25553b013054d994760f3c2a737c39 > > Reviewed-on: https://chromium-review.googlesource.com/669296 > > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > > Reviewed-by: Eric Holk <eholk@chromium.org> > > Reviewed-by: Brad Nelson <bradnelson@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#48053} > > TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org > > Change-Id: Id82fa341b77624e4971f24c4757a9a666a65930c > No-Presubmit: true > No-Tree-Checks: true > No-Try: true > Reviewed-on: https://chromium-review.googlesource.com/670141 > Reviewed-by: Mircea Trofin <mtrofin@chromium.org> > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > Cr-Commit-Position: refs/heads/master@{#48054} TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org Change-Id: Ib6a7a3e6098d2689e60cdca85ec77e57e5295e48 No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://chromium-review.googlesource.com/670142 Commit-Queue: Mircea Trofin <mtrofin@chromium.org> Reviewed-by: Mircea Trofin <mtrofin@chromium.org> Cr-Commit-Position: refs/heads/master@{#48055}
2017-09-16 05:22:38 +00:00
// failure.
base::AddressRegion Allocate(size_t size);
Revert "Revert "[wasm] A simple allocator datastructure for off-the heap"" This reverts commit ee5c31f3354ecd42a99e45a153061bc8b7a0a980. Reason for revert: Fixed compiler failure Original change's description: > Revert "[wasm] A simple allocator datastructure for off-the heap" > > This reverts commit 110d9ab0052f9c70c0c44e9c30397789568bbebf. > > Reason for revert: https://build.chromium.org/p/client.v8/builders/V8%20Linux64%20-%20debug%20builder/builds/26607 > > Surprising we're seeing a failure on Linux 64 *after* CQ. Is the compiler there different? > > Original change's description: > > [wasm] A simple allocator datastructure for off-the heap > > > > We'll use this allocator in a follow-up CL to: > > - allocate speculative sizes of memory for a module that's being > > compiled (e.g. 2*size of wasm code). > > - each module will own such a sub-pool, and then use it to allocate > > contiguous chunks of memory for code. > > > > The underlying assumptions for the chosen allocation strategy is that: > > - the allocation granularity for pools is 1 page, so that no one page > > is owned by more than one wasm module > > - typical pool sizes (given module sizes) are multiple pages. > > - modules and module instances are typically few and long lived. Typically, > > we expect one module and one instance. > > > > This means we shouldn't expect fragmentations that lead to code being > > non-allocatable, or prohibitively many ranges. > > > > The data structure just manages ranges of addresses. Virtual memory management > > will be separate, as part of the responsibility of a "WasmHeap" > > that will be introduced in the future. So will concurrency control. > > > > Bug: > > Change-Id: Id99f46d10c25553b013054d994760f3c2a737c39 > > Reviewed-on: https://chromium-review.googlesource.com/669296 > > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > > Reviewed-by: Eric Holk <eholk@chromium.org> > > Reviewed-by: Brad Nelson <bradnelson@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#48053} > > TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org > > Change-Id: Id82fa341b77624e4971f24c4757a9a666a65930c > No-Presubmit: true > No-Tree-Checks: true > No-Try: true > Reviewed-on: https://chromium-review.googlesource.com/670141 > Reviewed-by: Mircea Trofin <mtrofin@chromium.org> > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > Cr-Commit-Position: refs/heads/master@{#48054} TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org Change-Id: Ib6a7a3e6098d2689e60cdca85ec77e57e5295e48 No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://chromium-review.googlesource.com/670142 Commit-Queue: Mircea Trofin <mtrofin@chromium.org> Reviewed-by: Mircea Trofin <mtrofin@chromium.org> Cr-Commit-Position: refs/heads/master@{#48055}
2017-09-16 05:22:38 +00:00
bool IsEmpty() const { return regions_.empty(); }
const std::list<base::AddressRegion>& regions() const { return regions_; }
Revert "Revert "[wasm] A simple allocator datastructure for off-the heap"" This reverts commit ee5c31f3354ecd42a99e45a153061bc8b7a0a980. Reason for revert: Fixed compiler failure Original change's description: > Revert "[wasm] A simple allocator datastructure for off-the heap" > > This reverts commit 110d9ab0052f9c70c0c44e9c30397789568bbebf. > > Reason for revert: https://build.chromium.org/p/client.v8/builders/V8%20Linux64%20-%20debug%20builder/builds/26607 > > Surprising we're seeing a failure on Linux 64 *after* CQ. Is the compiler there different? > > Original change's description: > > [wasm] A simple allocator datastructure for off-the heap > > > > We'll use this allocator in a follow-up CL to: > > - allocate speculative sizes of memory for a module that's being > > compiled (e.g. 2*size of wasm code). > > - each module will own such a sub-pool, and then use it to allocate > > contiguous chunks of memory for code. > > > > The underlying assumptions for the chosen allocation strategy is that: > > - the allocation granularity for pools is 1 page, so that no one page > > is owned by more than one wasm module > > - typical pool sizes (given module sizes) are multiple pages. > > - modules and module instances are typically few and long lived. Typically, > > we expect one module and one instance. > > > > This means we shouldn't expect fragmentations that lead to code being > > non-allocatable, or prohibitively many ranges. > > > > The data structure just manages ranges of addresses. Virtual memory management > > will be separate, as part of the responsibility of a "WasmHeap" > > that will be introduced in the future. So will concurrency control. > > > > Bug: > > Change-Id: Id99f46d10c25553b013054d994760f3c2a737c39 > > Reviewed-on: https://chromium-review.googlesource.com/669296 > > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > > Reviewed-by: Eric Holk <eholk@chromium.org> > > Reviewed-by: Brad Nelson <bradnelson@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#48053} > > TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org > > Change-Id: Id82fa341b77624e4971f24c4757a9a666a65930c > No-Presubmit: true > No-Tree-Checks: true > No-Try: true > Reviewed-on: https://chromium-review.googlesource.com/670141 > Reviewed-by: Mircea Trofin <mtrofin@chromium.org> > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > Cr-Commit-Position: refs/heads/master@{#48054} TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org Change-Id: Ib6a7a3e6098d2689e60cdca85ec77e57e5295e48 No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://chromium-review.googlesource.com/670142 Commit-Queue: Mircea Trofin <mtrofin@chromium.org> Reviewed-by: Mircea Trofin <mtrofin@chromium.org> Cr-Commit-Position: refs/heads/master@{#48055}
2017-09-16 05:22:38 +00:00
private:
std::list<base::AddressRegion> regions_;
Revert "Revert "[wasm] A simple allocator datastructure for off-the heap"" This reverts commit ee5c31f3354ecd42a99e45a153061bc8b7a0a980. Reason for revert: Fixed compiler failure Original change's description: > Revert "[wasm] A simple allocator datastructure for off-the heap" > > This reverts commit 110d9ab0052f9c70c0c44e9c30397789568bbebf. > > Reason for revert: https://build.chromium.org/p/client.v8/builders/V8%20Linux64%20-%20debug%20builder/builds/26607 > > Surprising we're seeing a failure on Linux 64 *after* CQ. Is the compiler there different? > > Original change's description: > > [wasm] A simple allocator datastructure for off-the heap > > > > We'll use this allocator in a follow-up CL to: > > - allocate speculative sizes of memory for a module that's being > > compiled (e.g. 2*size of wasm code). > > - each module will own such a sub-pool, and then use it to allocate > > contiguous chunks of memory for code. > > > > The underlying assumptions for the chosen allocation strategy is that: > > - the allocation granularity for pools is 1 page, so that no one page > > is owned by more than one wasm module > > - typical pool sizes (given module sizes) are multiple pages. > > - modules and module instances are typically few and long lived. Typically, > > we expect one module and one instance. > > > > This means we shouldn't expect fragmentations that lead to code being > > non-allocatable, or prohibitively many ranges. > > > > The data structure just manages ranges of addresses. Virtual memory management > > will be separate, as part of the responsibility of a "WasmHeap" > > that will be introduced in the future. So will concurrency control. > > > > Bug: > > Change-Id: Id99f46d10c25553b013054d994760f3c2a737c39 > > Reviewed-on: https://chromium-review.googlesource.com/669296 > > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > > Reviewed-by: Eric Holk <eholk@chromium.org> > > Reviewed-by: Brad Nelson <bradnelson@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#48053} > > TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org > > Change-Id: Id82fa341b77624e4971f24c4757a9a666a65930c > No-Presubmit: true > No-Tree-Checks: true > No-Try: true > Reviewed-on: https://chromium-review.googlesource.com/670141 > Reviewed-by: Mircea Trofin <mtrofin@chromium.org> > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > Cr-Commit-Position: refs/heads/master@{#48054} TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org Change-Id: Ib6a7a3e6098d2689e60cdca85ec77e57e5295e48 No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://chromium-review.googlesource.com/670142 Commit-Queue: Mircea Trofin <mtrofin@chromium.org> Reviewed-by: Mircea Trofin <mtrofin@chromium.org> Cr-Commit-Position: refs/heads/master@{#48055}
2017-09-16 05:22:38 +00:00
};
2017-11-20 21:34:04 +00:00
class V8_EXPORT_PRIVATE WasmCode final {
public:
enum Kind {
kFunction,
kWasmToCapiWrapper,
kWasmToJsWrapper,
kRuntimeStub,
kInterpreterEntry,
2018-06-19 09:47:17 +00:00
kJumpTable
2017-11-20 21:34:04 +00:00
};
// Each runtime stub is identified by an id. This id is used to reference the
// stub via {RelocInfo::WASM_STUB_CALL} and gets resolved during relocation.
enum RuntimeStubId {
#define DEF_ENUM(Name) k##Name,
#define DEF_ENUM_TRAP(Name) kThrowWasm##Name,
WASM_RUNTIME_STUB_LIST(DEF_ENUM, DEF_ENUM_TRAP)
#undef DEF_ENUM_TRAP
#undef DEF_ENUM
kRuntimeStubCount
};
2017-11-20 21:34:04 +00:00
Vector<byte> instructions() const { return instructions_; }
Address instruction_start() const {
return reinterpret_cast<Address>(instructions_.begin());
}
Vector<const byte> reloc_info() const { return reloc_info_.as_vector(); }
Vector<const byte> source_positions() const {
return source_position_table_.as_vector();
}
2017-11-20 21:34:04 +00:00
uint32_t index() const {
DCHECK(!IsAnonymous());
return index_;
}
// Anonymous functions are functions that don't carry an index.
bool IsAnonymous() const { return index_ == kAnonymousFuncIndex; }
2017-11-20 21:34:04 +00:00
Kind kind() const { return kind_; }
NativeModule* native_module() const { return native_module_; }
ExecutionTier tier() const { return tier_; }
2017-11-20 21:34:04 +00:00
Address constant_pool() const;
Address handler_table() const;
uint32_t handler_table_size() const;
Reland "Reland "Reland "[code-comments] Put code comments into the code object""" This is a reland of 9c0a48580bc820d93a16f8914281a7359beb2a7a Original change's description: > Reland "Reland "[code-comments] Put code comments into the code object"" > > This is a reland of ed3d647284538e9d6f013ebf2c460697aa06a5df > > This reland fixes that padding at the end of Wasm instruction streams > triggered asserts in the code printer. > > Original change's description: > > Reland "[code-comments] Put code comments into the code object" > > > > This is a reland of e774cffe2bd3f00332209d4d5695221963888c96 > > > > This reland disables a test as v8:8548 is blocking it, which was > > broken by a recent CL. CQ did not catch this because the merge-base > > CQ used did not yet contain the CL that caused v8:8548. > > > > Original change's description: > > > [code-comments] Put code comments into the code object > > > > > > Code comments in the snapshot can now be enabled with gn > > > arg 'v8_enable_snapshot_code_comments' > > > > > > Bug: v8:7989 > > > Change-Id: I8bd00cafa63132d00d849394c311ba15e6b6daf3 > > > Reviewed-on: https://chromium-review.googlesource.com/c/1329173 > > > Commit-Queue: Sigurd Schneider <sigurds@chromium.org> > > > Reviewed-by: Jakob Gruber <jgruber@chromium.org> > > > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > > > Reviewed-by: Michael Stanton <mvstanton@chromium.org> > > > Cr-Commit-Position: refs/heads/master@{#58020} > > > > TBR=mvstanton@chromium.org,mstarzinger@chromium.org,jgruber@chromium.org,tebbi@chromium.org > > > > Bug: v8:7989, v8:8548 > > Change-Id: I464fc897205fefdf2dfc2eadc54d699c4e08a0e9 > > Reviewed-on: https://chromium-review.googlesource.com/c/1361166 > > Reviewed-by: Sigurd Schneider <sigurds@chromium.org> > > Commit-Queue: Sigurd Schneider <sigurds@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#58028} > > Bug: v8:7989, v8:8548 > Change-Id: I254f55ff687ad049f8d92b09331ed26a2bd05d7d > Reviewed-on: https://chromium-review.googlesource.com/c/1371784 > Commit-Queue: Sigurd Schneider <sigurds@chromium.org> > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > Reviewed-by: Jakob Gruber <jgruber@chromium.org> > Cr-Commit-Position: refs/heads/master@{#58221} TBR=jgruber@chromium.org,mstarzinger@chromium.org Bug: v8:7989, v8:8548, v8:8593 Change-Id: I4f7ffc98e0281c7b744eb4a04ba0763896c7b59b Reviewed-on: https://chromium-review.googlesource.com/c/1375919 Reviewed-by: Sigurd Schneider <sigurds@chromium.org> Commit-Queue: Sigurd Schneider <sigurds@chromium.org> Cr-Commit-Position: refs/heads/master@{#58232}
2018-12-13 19:30:56 +00:00
Address code_comments() const;
uint32_t code_comments_size() const;
2017-11-20 21:34:04 +00:00
size_t constant_pool_offset() const { return constant_pool_offset_; }
size_t safepoint_table_offset() const { return safepoint_table_offset_; }
size_t handler_table_offset() const { return handler_table_offset_; }
Reland "Reland "Reland "[code-comments] Put code comments into the code object""" This is a reland of 9c0a48580bc820d93a16f8914281a7359beb2a7a Original change's description: > Reland "Reland "[code-comments] Put code comments into the code object"" > > This is a reland of ed3d647284538e9d6f013ebf2c460697aa06a5df > > This reland fixes that padding at the end of Wasm instruction streams > triggered asserts in the code printer. > > Original change's description: > > Reland "[code-comments] Put code comments into the code object" > > > > This is a reland of e774cffe2bd3f00332209d4d5695221963888c96 > > > > This reland disables a test as v8:8548 is blocking it, which was > > broken by a recent CL. CQ did not catch this because the merge-base > > CQ used did not yet contain the CL that caused v8:8548. > > > > Original change's description: > > > [code-comments] Put code comments into the code object > > > > > > Code comments in the snapshot can now be enabled with gn > > > arg 'v8_enable_snapshot_code_comments' > > > > > > Bug: v8:7989 > > > Change-Id: I8bd00cafa63132d00d849394c311ba15e6b6daf3 > > > Reviewed-on: https://chromium-review.googlesource.com/c/1329173 > > > Commit-Queue: Sigurd Schneider <sigurds@chromium.org> > > > Reviewed-by: Jakob Gruber <jgruber@chromium.org> > > > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > > > Reviewed-by: Michael Stanton <mvstanton@chromium.org> > > > Cr-Commit-Position: refs/heads/master@{#58020} > > > > TBR=mvstanton@chromium.org,mstarzinger@chromium.org,jgruber@chromium.org,tebbi@chromium.org > > > > Bug: v8:7989, v8:8548 > > Change-Id: I464fc897205fefdf2dfc2eadc54d699c4e08a0e9 > > Reviewed-on: https://chromium-review.googlesource.com/c/1361166 > > Reviewed-by: Sigurd Schneider <sigurds@chromium.org> > > Commit-Queue: Sigurd Schneider <sigurds@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#58028} > > Bug: v8:7989, v8:8548 > Change-Id: I254f55ff687ad049f8d92b09331ed26a2bd05d7d > Reviewed-on: https://chromium-review.googlesource.com/c/1371784 > Commit-Queue: Sigurd Schneider <sigurds@chromium.org> > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > Reviewed-by: Jakob Gruber <jgruber@chromium.org> > Cr-Commit-Position: refs/heads/master@{#58221} TBR=jgruber@chromium.org,mstarzinger@chromium.org Bug: v8:7989, v8:8548, v8:8593 Change-Id: I4f7ffc98e0281c7b744eb4a04ba0763896c7b59b Reviewed-on: https://chromium-review.googlesource.com/c/1375919 Reviewed-by: Sigurd Schneider <sigurds@chromium.org> Commit-Queue: Sigurd Schneider <sigurds@chromium.org> Cr-Commit-Position: refs/heads/master@{#58232}
2018-12-13 19:30:56 +00:00
size_t code_comments_offset() const { return code_comments_offset_; }
size_t unpadded_binary_size() const { return unpadded_binary_size_; }
2017-11-20 21:34:04 +00:00
uint32_t stack_slots() const { return stack_slots_; }
uint32_t tagged_parameter_slots() const { return tagged_parameter_slots_; }
bool is_liftoff() const { return tier_ == ExecutionTier::kLiftoff; }
bool contains(Address pc) const {
return reinterpret_cast<Address>(instructions_.begin()) <= pc &&
pc < reinterpret_cast<Address>(instructions_.end());
}
2017-11-20 21:34:04 +00:00
Vector<trap_handler::ProtectedInstructionData> protected_instructions()
const {
return protected_instructions_.as_vector();
2017-11-20 21:34:04 +00:00
}
void Validate() const;
void Print(const char* name = nullptr) const;
void MaybePrint(const char* name = nullptr) const;
void Disassemble(const char* name, std::ostream& os,
Address current_pc = kNullAddress) const;
2017-11-20 21:34:04 +00:00
static bool ShouldBeLogged(Isolate* isolate);
void LogCode(Isolate* isolate) const;
2017-11-20 21:34:04 +00:00
~WasmCode();
void IncRef() {
int old_val = ref_count_.fetch_add(1, std::memory_order_acq_rel);
DCHECK_LE(1, old_val);
DCHECK_GT(kMaxInt, old_val);
USE(old_val);
}
// Decrement the ref count. Returns whether this code becomes dead and needs
// to be freed.
V8_WARN_UNUSED_RESULT bool DecRef() {
int old_count = ref_count_.load(std::memory_order_acquire);
while (true) {
DCHECK_LE(1, old_count);
if (V8_UNLIKELY(old_count == 1)) return DecRefOnPotentiallyDeadCode();
if (ref_count_.compare_exchange_weak(old_count, old_count - 1,
std::memory_order_acq_rel)) {
return false;
}
}
}
// Decrement the ref count on code that is known to be dead, even though there
// might still be C++ references. Returns whether this drops the last
// reference and the code needs to be freed.
V8_WARN_UNUSED_RESULT bool DecRefOnDeadCode() {
return ref_count_.fetch_sub(1, std::memory_order_acq_rel) == 1;
}
// Decrement the ref count on a set of {WasmCode} objects, potentially
// belonging to different {NativeModule}s. Dead code will be deleted.
static void DecrementRefCount(Vector<WasmCode* const>);
enum FlushICache : bool { kFlushICache = true, kNoFlushICache = false };
STATIC_ASSERT(kAnonymousFuncIndex > kV8MaxWasmFunctions);
2017-11-20 21:34:04 +00:00
private:
friend class NativeModule;
WasmCode(NativeModule* native_module, uint32_t index,
Vector<byte> instructions, uint32_t stack_slots,
uint32_t tagged_parameter_slots, size_t safepoint_table_offset,
size_t handler_table_offset, size_t constant_pool_offset,
size_t code_comments_offset, size_t unpadded_binary_size,
OwnedVector<trap_handler::ProtectedInstructionData>
protected_instructions,
OwnedVector<const byte> reloc_info,
OwnedVector<const byte> source_position_table, Kind kind,
ExecutionTier tier)
2017-11-20 21:34:04 +00:00
: instructions_(instructions),
reloc_info_(std::move(reloc_info)),
source_position_table_(std::move(source_position_table)),
native_module_(native_module),
2017-11-20 21:34:04 +00:00
index_(index),
kind_(kind),
constant_pool_offset_(constant_pool_offset),
stack_slots_(stack_slots),
tagged_parameter_slots_(tagged_parameter_slots),
2017-11-20 21:34:04 +00:00
safepoint_table_offset_(safepoint_table_offset),
handler_table_offset_(handler_table_offset),
Reland "Reland "Reland "[code-comments] Put code comments into the code object""" This is a reland of 9c0a48580bc820d93a16f8914281a7359beb2a7a Original change's description: > Reland "Reland "[code-comments] Put code comments into the code object"" > > This is a reland of ed3d647284538e9d6f013ebf2c460697aa06a5df > > This reland fixes that padding at the end of Wasm instruction streams > triggered asserts in the code printer. > > Original change's description: > > Reland "[code-comments] Put code comments into the code object" > > > > This is a reland of e774cffe2bd3f00332209d4d5695221963888c96 > > > > This reland disables a test as v8:8548 is blocking it, which was > > broken by a recent CL. CQ did not catch this because the merge-base > > CQ used did not yet contain the CL that caused v8:8548. > > > > Original change's description: > > > [code-comments] Put code comments into the code object > > > > > > Code comments in the snapshot can now be enabled with gn > > > arg 'v8_enable_snapshot_code_comments' > > > > > > Bug: v8:7989 > > > Change-Id: I8bd00cafa63132d00d849394c311ba15e6b6daf3 > > > Reviewed-on: https://chromium-review.googlesource.com/c/1329173 > > > Commit-Queue: Sigurd Schneider <sigurds@chromium.org> > > > Reviewed-by: Jakob Gruber <jgruber@chromium.org> > > > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > > > Reviewed-by: Michael Stanton <mvstanton@chromium.org> > > > Cr-Commit-Position: refs/heads/master@{#58020} > > > > TBR=mvstanton@chromium.org,mstarzinger@chromium.org,jgruber@chromium.org,tebbi@chromium.org > > > > Bug: v8:7989, v8:8548 > > Change-Id: I464fc897205fefdf2dfc2eadc54d699c4e08a0e9 > > Reviewed-on: https://chromium-review.googlesource.com/c/1361166 > > Reviewed-by: Sigurd Schneider <sigurds@chromium.org> > > Commit-Queue: Sigurd Schneider <sigurds@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#58028} > > Bug: v8:7989, v8:8548 > Change-Id: I254f55ff687ad049f8d92b09331ed26a2bd05d7d > Reviewed-on: https://chromium-review.googlesource.com/c/1371784 > Commit-Queue: Sigurd Schneider <sigurds@chromium.org> > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > Reviewed-by: Jakob Gruber <jgruber@chromium.org> > Cr-Commit-Position: refs/heads/master@{#58221} TBR=jgruber@chromium.org,mstarzinger@chromium.org Bug: v8:7989, v8:8548, v8:8593 Change-Id: I4f7ffc98e0281c7b744eb4a04ba0763896c7b59b Reviewed-on: https://chromium-review.googlesource.com/c/1375919 Reviewed-by: Sigurd Schneider <sigurds@chromium.org> Commit-Queue: Sigurd Schneider <sigurds@chromium.org> Cr-Commit-Position: refs/heads/master@{#58232}
2018-12-13 19:30:56 +00:00
code_comments_offset_(code_comments_offset),
unpadded_binary_size_(unpadded_binary_size),
protected_instructions_(std::move(protected_instructions)),
tier_(tier) {
Reland "Reland "Reland "[code-comments] Put code comments into the code object""" This is a reland of 9c0a48580bc820d93a16f8914281a7359beb2a7a Original change's description: > Reland "Reland "[code-comments] Put code comments into the code object"" > > This is a reland of ed3d647284538e9d6f013ebf2c460697aa06a5df > > This reland fixes that padding at the end of Wasm instruction streams > triggered asserts in the code printer. > > Original change's description: > > Reland "[code-comments] Put code comments into the code object" > > > > This is a reland of e774cffe2bd3f00332209d4d5695221963888c96 > > > > This reland disables a test as v8:8548 is blocking it, which was > > broken by a recent CL. CQ did not catch this because the merge-base > > CQ used did not yet contain the CL that caused v8:8548. > > > > Original change's description: > > > [code-comments] Put code comments into the code object > > > > > > Code comments in the snapshot can now be enabled with gn > > > arg 'v8_enable_snapshot_code_comments' > > > > > > Bug: v8:7989 > > > Change-Id: I8bd00cafa63132d00d849394c311ba15e6b6daf3 > > > Reviewed-on: https://chromium-review.googlesource.com/c/1329173 > > > Commit-Queue: Sigurd Schneider <sigurds@chromium.org> > > > Reviewed-by: Jakob Gruber <jgruber@chromium.org> > > > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > > > Reviewed-by: Michael Stanton <mvstanton@chromium.org> > > > Cr-Commit-Position: refs/heads/master@{#58020} > > > > TBR=mvstanton@chromium.org,mstarzinger@chromium.org,jgruber@chromium.org,tebbi@chromium.org > > > > Bug: v8:7989, v8:8548 > > Change-Id: I464fc897205fefdf2dfc2eadc54d699c4e08a0e9 > > Reviewed-on: https://chromium-review.googlesource.com/c/1361166 > > Reviewed-by: Sigurd Schneider <sigurds@chromium.org> > > Commit-Queue: Sigurd Schneider <sigurds@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#58028} > > Bug: v8:7989, v8:8548 > Change-Id: I254f55ff687ad049f8d92b09331ed26a2bd05d7d > Reviewed-on: https://chromium-review.googlesource.com/c/1371784 > Commit-Queue: Sigurd Schneider <sigurds@chromium.org> > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > Reviewed-by: Jakob Gruber <jgruber@chromium.org> > Cr-Commit-Position: refs/heads/master@{#58221} TBR=jgruber@chromium.org,mstarzinger@chromium.org Bug: v8:7989, v8:8548, v8:8593 Change-Id: I4f7ffc98e0281c7b744eb4a04ba0763896c7b59b Reviewed-on: https://chromium-review.googlesource.com/c/1375919 Reviewed-by: Sigurd Schneider <sigurds@chromium.org> Commit-Queue: Sigurd Schneider <sigurds@chromium.org> Cr-Commit-Position: refs/heads/master@{#58232}
2018-12-13 19:30:56 +00:00
DCHECK_LE(safepoint_table_offset, unpadded_binary_size);
DCHECK_LE(handler_table_offset, unpadded_binary_size);
DCHECK_LE(code_comments_offset, unpadded_binary_size);
DCHECK_LE(constant_pool_offset, unpadded_binary_size);
}
2017-11-20 21:34:04 +00:00
// Code objects that have been registered with the global trap handler within
// this process, will have a {trap_handler_index} associated with them.
int trap_handler_index() const {
CHECK(has_trap_handler_index());
return trap_handler_index_;
}
void set_trap_handler_index(int value) {
CHECK(!has_trap_handler_index());
trap_handler_index_ = value;
}
bool has_trap_handler_index() const { return trap_handler_index_ >= 0; }
// Register protected instruction information with the trap handler. Sets
// trap_handler_index.
void RegisterTrapHandlerData();
// Slow path for {DecRef}: The code becomes potentially dead.
// Returns whether this code becomes dead and needs to be freed.
V8_NOINLINE bool DecRefOnPotentiallyDeadCode();
2017-11-20 21:34:04 +00:00
Vector<byte> instructions_;
OwnedVector<const byte> reloc_info_;
OwnedVector<const byte> source_position_table_;
NativeModule* native_module_ = nullptr;
uint32_t index_;
2017-11-20 21:34:04 +00:00
Kind kind_;
size_t constant_pool_offset_ = 0;
uint32_t stack_slots_ = 0;
// Number of tagged parameters passed to this function via the stack. This
// value is used by the stack walker (e.g. GC) to find references.
uint32_t tagged_parameter_slots_ = 0;
2017-11-20 21:34:04 +00:00
// we care about safepoint data for wasm-to-js functions,
// since there may be stack/register tagged values for large number
// conversions.
size_t safepoint_table_offset_ = 0;
size_t handler_table_offset_ = 0;
Reland "Reland "Reland "[code-comments] Put code comments into the code object""" This is a reland of 9c0a48580bc820d93a16f8914281a7359beb2a7a Original change's description: > Reland "Reland "[code-comments] Put code comments into the code object"" > > This is a reland of ed3d647284538e9d6f013ebf2c460697aa06a5df > > This reland fixes that padding at the end of Wasm instruction streams > triggered asserts in the code printer. > > Original change's description: > > Reland "[code-comments] Put code comments into the code object" > > > > This is a reland of e774cffe2bd3f00332209d4d5695221963888c96 > > > > This reland disables a test as v8:8548 is blocking it, which was > > broken by a recent CL. CQ did not catch this because the merge-base > > CQ used did not yet contain the CL that caused v8:8548. > > > > Original change's description: > > > [code-comments] Put code comments into the code object > > > > > > Code comments in the snapshot can now be enabled with gn > > > arg 'v8_enable_snapshot_code_comments' > > > > > > Bug: v8:7989 > > > Change-Id: I8bd00cafa63132d00d849394c311ba15e6b6daf3 > > > Reviewed-on: https://chromium-review.googlesource.com/c/1329173 > > > Commit-Queue: Sigurd Schneider <sigurds@chromium.org> > > > Reviewed-by: Jakob Gruber <jgruber@chromium.org> > > > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > > > Reviewed-by: Michael Stanton <mvstanton@chromium.org> > > > Cr-Commit-Position: refs/heads/master@{#58020} > > > > TBR=mvstanton@chromium.org,mstarzinger@chromium.org,jgruber@chromium.org,tebbi@chromium.org > > > > Bug: v8:7989, v8:8548 > > Change-Id: I464fc897205fefdf2dfc2eadc54d699c4e08a0e9 > > Reviewed-on: https://chromium-review.googlesource.com/c/1361166 > > Reviewed-by: Sigurd Schneider <sigurds@chromium.org> > > Commit-Queue: Sigurd Schneider <sigurds@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#58028} > > Bug: v8:7989, v8:8548 > Change-Id: I254f55ff687ad049f8d92b09331ed26a2bd05d7d > Reviewed-on: https://chromium-review.googlesource.com/c/1371784 > Commit-Queue: Sigurd Schneider <sigurds@chromium.org> > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > Reviewed-by: Jakob Gruber <jgruber@chromium.org> > Cr-Commit-Position: refs/heads/master@{#58221} TBR=jgruber@chromium.org,mstarzinger@chromium.org Bug: v8:7989, v8:8548, v8:8593 Change-Id: I4f7ffc98e0281c7b744eb4a04ba0763896c7b59b Reviewed-on: https://chromium-review.googlesource.com/c/1375919 Reviewed-by: Sigurd Schneider <sigurds@chromium.org> Commit-Queue: Sigurd Schneider <sigurds@chromium.org> Cr-Commit-Position: refs/heads/master@{#58232}
2018-12-13 19:30:56 +00:00
size_t code_comments_offset_ = 0;
size_t unpadded_binary_size_ = 0;
int trap_handler_index_ = -1;
OwnedVector<trap_handler::ProtectedInstructionData> protected_instructions_;
ExecutionTier tier_;
// WasmCode is ref counted. Counters are held by:
// 1) The jump table / code table.
// 2) {WasmCodeRefScope}s.
// 3) The set of potentially dead code in the {WasmEngine}.
// If a decrement of (1) would drop the ref count to 0, that code becomes a
// candidate for garbage collection. At that point, we add a ref count for (3)
// *before* decrementing the counter to ensure the code stays alive as long as
// it's being used. Once the ref count drops to zero (i.e. after being removed
// from (3) and all (2)), the code object is deleted and the memory for the
// machine code is freed.
std::atomic<int> ref_count_{1};
DISALLOW_COPY_AND_ASSIGN(WasmCode);
2017-11-20 21:34:04 +00:00
};
WasmCode::Kind GetCodeKind(const WasmCompilationResult& result);
// Return a textual description of the kind.
const char* GetWasmCodeKindAsString(WasmCode::Kind);
// Manages the code reservations and allocations of a single {NativeModule}.
class WasmCodeAllocator {
public:
WasmCodeAllocator(WasmCodeManager*, VirtualMemory code_space,
bool can_request_more,
std::shared_ptr<Counters> async_counters);
~WasmCodeAllocator();
size_t committed_code_space() const {
return committed_code_space_.load(std::memory_order_acquire);
}
size_t generated_code_size() const {
return generated_code_size_.load(std::memory_order_acquire);
}
size_t freed_code_size() const {
return freed_code_size_.load(std::memory_order_acquire);
}
// Allocate code space. Returns a valid buffer or fails with OOM (crash).
Vector<byte> AllocateForCode(NativeModule*, size_t size);
// Sets permissions of all owned code space to executable, or read-write (if
// {executable} is false). Returns true on success.
V8_EXPORT_PRIVATE bool SetExecutable(bool executable);
// Free memory pages of all given code objects. Used for wasm code GC.
void FreeCode(Vector<WasmCode* const>);
private:
// The engine-wide wasm code manager.
WasmCodeManager* const code_manager_;
mutable base::Mutex mutex_;
//////////////////////////////////////////////////////////////////////////////
// Protected by {mutex_}:
// Code space that was reserved and is available for allocations (subset of
// {owned_code_space_}).
DisjointAllocationPool free_code_space_;
// Code space that was allocated for code (subset of {owned_code_space_}).
DisjointAllocationPool allocated_code_space_;
// Code space that was allocated before but is dead now. Full pages within
// this region are discarded. It's still a subset of {owned_code_space_}.
DisjointAllocationPool freed_code_space_;
std::vector<VirtualMemory> owned_code_space_;
// End of fields protected by {mutex_}.
//////////////////////////////////////////////////////////////////////////////
std::atomic<size_t> committed_code_space_{0};
std::atomic<size_t> generated_code_size_{0};
std::atomic<size_t> freed_code_size_{0};
bool is_executable_ = false;
const bool can_request_more_memory_;
std::shared_ptr<Counters> async_counters_;
};
2017-11-20 21:34:04 +00:00
class V8_EXPORT_PRIVATE NativeModule final {
public:
#if V8_TARGET_ARCH_X64 || V8_TARGET_ARCH_S390X || V8_TARGET_ARCH_ARM64
static constexpr bool kCanAllocateMoreMemory = false;
#else
static constexpr bool kCanAllocateMoreMemory = true;
#endif
// {AddCode} is thread safe w.r.t. other calls to {AddCode} or methods adding
// code below, i.e. it can be called concurrently from background threads.
[wasm] Split adding code from publishing it This prepares a refactoring to add and publish compilation results in batches. For this, we need to separate the two phases, so that we can lock the module, allocate all the code space, release the lock, copy the code, lock the module, publish the code, and release the lock again. In particular, this CL does the following: 1) It removes the {AddOwnedCode} method. The functionality of creating the {WasmCode} and memcpy'ing the instruction into that is done in the other {Add*Code} methods. Adding to {owned_code_} is done in {PublishCode}. 2) {PublishInterpreterEntry} is now functionally equivalent to {PublishCode}, so it's removed. 3) After {AddCode}, the caller has to call {PublishCode}. In a follow-up CL, this will be called in batches (first {AddCode} them all, then {PublishCode} them all). 4) {AddCompiledCode} now assumes that the {WasmCompilationResult} succeeded. Otherwise, the caller should directly call {SetError} on the {CompilationState}. 5) {PublishCode} is now the chokepoint for installing code to the code table, the owned code vector, the jump table, and setting interpreter redirections. It replaces previous direct calls to {InstallCode} or explicitly adding to {owned_code_}. 6) Increasing the {generated_code_size_} counter is now done in {AllocateForCode}, which is the chokepoint for allocating space for generated code. This way, we will only increase this counter once once we allocate in batches. R=titzer@chromium.org Bug: v8:8916 Change-Id: I71e02e3a838f21797915cee3ebd373804fb12237 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1530817 Commit-Queue: Clemens Hammacher <clemensh@chromium.org> Reviewed-by: Ben Titzer <titzer@chromium.org> Cr-Commit-Position: refs/heads/master@{#60369}
2019-03-20 13:04:33 +00:00
// The returned code still needs to be published via {PublishCode}.
std::unique_ptr<WasmCode> AddCode(
uint32_t index, const CodeDesc& desc, uint32_t stack_slots,
uint32_t tagged_parameter_slots,
OwnedVector<trap_handler::ProtectedInstructionData>
protected_instructions,
OwnedVector<const byte> source_position_table, WasmCode::Kind kind,
ExecutionTier tier);
[wasm] Split adding code from publishing it This prepares a refactoring to add and publish compilation results in batches. For this, we need to separate the two phases, so that we can lock the module, allocate all the code space, release the lock, copy the code, lock the module, publish the code, and release the lock again. In particular, this CL does the following: 1) It removes the {AddOwnedCode} method. The functionality of creating the {WasmCode} and memcpy'ing the instruction into that is done in the other {Add*Code} methods. Adding to {owned_code_} is done in {PublishCode}. 2) {PublishInterpreterEntry} is now functionally equivalent to {PublishCode}, so it's removed. 3) After {AddCode}, the caller has to call {PublishCode}. In a follow-up CL, this will be called in batches (first {AddCode} them all, then {PublishCode} them all). 4) {AddCompiledCode} now assumes that the {WasmCompilationResult} succeeded. Otherwise, the caller should directly call {SetError} on the {CompilationState}. 5) {PublishCode} is now the chokepoint for installing code to the code table, the owned code vector, the jump table, and setting interpreter redirections. It replaces previous direct calls to {InstallCode} or explicitly adding to {owned_code_}. 6) Increasing the {generated_code_size_} counter is now done in {AllocateForCode}, which is the chokepoint for allocating space for generated code. This way, we will only increase this counter once once we allocate in batches. R=titzer@chromium.org Bug: v8:8916 Change-Id: I71e02e3a838f21797915cee3ebd373804fb12237 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1530817 Commit-Queue: Clemens Hammacher <clemensh@chromium.org> Reviewed-by: Ben Titzer <titzer@chromium.org> Cr-Commit-Position: refs/heads/master@{#60369}
2019-03-20 13:04:33 +00:00
// {PublishCode} makes the code available to the system by entering it into
// the code table and patching the jump table. It returns a raw pointer to the
// given {WasmCode} object.
WasmCode* PublishCode(std::unique_ptr<WasmCode>);
// Hold the {allocation_mutex_} when calling {PublishCodeLocked}.
WasmCode* PublishCodeLocked(std::unique_ptr<WasmCode>);
2017-11-20 21:34:04 +00:00
WasmCode* AddDeserializedCode(
uint32_t index, Vector<const byte> instructions, uint32_t stack_slots,
uint32_t tagged_parameter_slots, size_t safepoint_table_offset,
size_t handler_table_offset, size_t constant_pool_offset,
size_t code_comments_offset, size_t unpadded_binary_size,
OwnedVector<trap_handler::ProtectedInstructionData>
protected_instructions,
OwnedVector<const byte> reloc_info,
OwnedVector<const byte> source_position_table, WasmCode::Kind kind,
ExecutionTier tier);
// Adds anonymous code for testing purposes.
WasmCode* AddCodeForTesting(Handle<Code> code);
// Use {UseLazyStub} to setup lazy compilation per function. It will use the
// existing {WasmCode::kWasmCompileLazy} runtime stub and populate the jump
// table with trampolines accordingly.
void UseLazyStub(uint32_t func_index);
2017-11-20 21:34:04 +00:00
// Initializes all runtime stubs by setting up entry addresses in the runtime
// stub table. It must be called exactly once per native module before adding
// other WasmCode so that runtime stub ids can be resolved during relocation.
void SetRuntimeStubs(Isolate* isolate);
// Creates a snapshot of the current state of the code table. This is useful
// to get a consistent view of the table (e.g. used by the serializer).
std::vector<WasmCode*> SnapshotCodeTable() const;
WasmCode* GetCode(uint32_t index) const;
bool HasCode(uint32_t index) const;
Address runtime_stub_entry(WasmCode::RuntimeStubId index) const {
DCHECK_LT(index, WasmCode::kRuntimeStubCount);
Address entry_address = runtime_stub_entries_[index];
DCHECK_NE(kNullAddress, entry_address);
return entry_address;
}
Address jump_table_start() const {
return jump_table_ ? jump_table_->instruction_start() : kNullAddress;
}
uint32_t GetJumpTableOffset(uint32_t func_index) const;
2018-06-19 09:47:17 +00:00
bool is_jump_table_slot(Address address) const {
return jump_table_->contains(address);
}
// Returns the target to call for the given function (returns a jump table
// slot within {jump_table_}).
Address GetCallTargetForFunction(uint32_t func_index) const;
2017-11-20 21:34:04 +00:00
// Reverse lookup from a given call target (i.e. a jump table slot as the
// above {GetCallTargetForFunction} returns) to a function index.
uint32_t GetFunctionIndexFromJumpTableSlot(Address slot_address) const;
bool SetExecutable(bool executable) {
return code_allocator_.SetExecutable(executable);
}
2017-11-20 21:34:04 +00:00
// For cctests, where we build both WasmModule and the runtime objects
// on the fly, and bypass the instance builder pipeline.
void ReserveCodeTableForTesting(uint32_t max_functions);
2017-11-20 21:34:04 +00:00
void LogWasmCodes(Isolate* isolate);
CompilationState* compilation_state() { return compilation_state_.get(); }
Reland "[wasm] Decouple background compile jobs from NativeModule" This is a reland of 92d9b09c0e68382d1427ad25c70db016bb9e5b80. Patch unchanged, errors fixed by https://crrev.com/c/1430059. Original change's description: > [wasm] Decouple background compile jobs from NativeModule > > Background compile jobs should not keep the NativeModule alive, for two > reasons: > 1) We sometimes have to wait for background compilation to finish (from > a foreground task!). This introduces unnecessary latency. > 2) Giving the background compile tasks shared ownership of the > NativeModule causes the NativeModule (and the CompilationState) to > be freed from background tasks, which is error-prone (see > https://crrev.com/c/1400420). > > Instead, this CL introduces a BackgroundCompileToken which is held > alive by the NativeModule and all background compile jobs. The initial > and the final phase of compilation (getting and submitting work) > synchronize on this token to check and ensure that the NativeModule is > and stays alive. During compilation itself, the mutex is released, such > that the NativeModule can die. > The destructor of the NativeModule cancels the BackgroundCompileToken. > Immediately afterwards, the NativeModule and the CompilationState can > die. > > This change allows to remove two hacks introduced previously: The atomic > {aborted_} flag and the {FreeCallbacksTask}. > > R=mstarzinger@chromium.org > CC=titzer@chromium.org > > Bug: v8:8689, v8:7921 > Change-Id: I42e06eab3c944b0988286f2ce18e3c294535dfb6 > Cq-Include-Trybots: luci.v8.try:v8_linux64_tsan_rel > Reviewed-on: https://chromium-review.googlesource.com/c/1421364 > Commit-Queue: Clemens Hammacher <clemensh@chromium.org> > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > Cr-Commit-Position: refs/heads/master@{#59020} TBR=mstarzinger@chromium.org Bug: v8:8689, v8:7921 Change-Id: Iead972ef77c8503da7246cab48e7693b176d8f02 Cq-Include-Trybots: luci.v8.try:v8_linux64_tsan_rel Reviewed-on: https://chromium-review.googlesource.com/c/1429862 Reviewed-by: Clemens Hammacher <clemensh@chromium.org> Commit-Queue: Clemens Hammacher <clemensh@chromium.org> Cr-Commit-Position: refs/heads/master@{#59035}
2019-01-23 11:49:36 +00:00
// Create a {CompilationEnv} object for compilation. The caller has to ensure
// that the {WasmModule} pointer stays valid while the {CompilationEnv} is
// being used.
CompilationEnv CreateCompilationEnv() const;
uint32_t num_functions() const {
return module_->num_declared_functions + module_->num_imported_functions;
}
uint32_t num_imported_functions() const {
return module_->num_imported_functions;
}
UseTrapHandler use_trap_handler() const { return use_trap_handler_; }
void set_lazy_compile_frozen(bool frozen) { lazy_compile_frozen_ = frozen; }
bool lazy_compile_frozen() const { return lazy_compile_frozen_; }
Vector<const uint8_t> wire_bytes() const { return wire_bytes_->as_vector(); }
const WasmModule* module() const { return module_.get(); }
Reland "[wasm] Decouple background compile jobs from NativeModule" This is a reland of 92d9b09c0e68382d1427ad25c70db016bb9e5b80. Patch unchanged, errors fixed by https://crrev.com/c/1430059. Original change's description: > [wasm] Decouple background compile jobs from NativeModule > > Background compile jobs should not keep the NativeModule alive, for two > reasons: > 1) We sometimes have to wait for background compilation to finish (from > a foreground task!). This introduces unnecessary latency. > 2) Giving the background compile tasks shared ownership of the > NativeModule causes the NativeModule (and the CompilationState) to > be freed from background tasks, which is error-prone (see > https://crrev.com/c/1400420). > > Instead, this CL introduces a BackgroundCompileToken which is held > alive by the NativeModule and all background compile jobs. The initial > and the final phase of compilation (getting and submitting work) > synchronize on this token to check and ensure that the NativeModule is > and stays alive. During compilation itself, the mutex is released, such > that the NativeModule can die. > The destructor of the NativeModule cancels the BackgroundCompileToken. > Immediately afterwards, the NativeModule and the CompilationState can > die. > > This change allows to remove two hacks introduced previously: The atomic > {aborted_} flag and the {FreeCallbacksTask}. > > R=mstarzinger@chromium.org > CC=titzer@chromium.org > > Bug: v8:8689, v8:7921 > Change-Id: I42e06eab3c944b0988286f2ce18e3c294535dfb6 > Cq-Include-Trybots: luci.v8.try:v8_linux64_tsan_rel > Reviewed-on: https://chromium-review.googlesource.com/c/1421364 > Commit-Queue: Clemens Hammacher <clemensh@chromium.org> > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > Cr-Commit-Position: refs/heads/master@{#59020} TBR=mstarzinger@chromium.org Bug: v8:8689, v8:7921 Change-Id: Iead972ef77c8503da7246cab48e7693b176d8f02 Cq-Include-Trybots: luci.v8.try:v8_linux64_tsan_rel Reviewed-on: https://chromium-review.googlesource.com/c/1429862 Reviewed-by: Clemens Hammacher <clemensh@chromium.org> Commit-Queue: Clemens Hammacher <clemensh@chromium.org> Cr-Commit-Position: refs/heads/master@{#59035}
2019-01-23 11:49:36 +00:00
std::shared_ptr<const WasmModule> shared_module() const { return module_; }
size_t committed_code_space() const {
return code_allocator_.committed_code_space();
}
WasmEngine* engine() const { return engine_; }
void SetWireBytes(OwnedVector<const uint8_t> wire_bytes);
2018-06-19 09:47:17 +00:00
WasmCode* Lookup(Address) const;
WasmImportWrapperCache* import_wrapper_cache() const {
return import_wrapper_cache_.get();
}
2017-11-20 21:34:04 +00:00
~NativeModule();
const WasmFeatures& enabled_features() const { return enabled_features_; }
const char* GetRuntimeStubName(Address runtime_stub_entry) const;
// Sample the current code size of this modules to the given counters.
enum CodeSamplingTime : int8_t { kAfterBaseline, kAfterTopTier, kSampling };
void SampleCodeSize(Counters*, CodeSamplingTime) const;
WasmCode* AddCompiledCode(WasmCompilationResult);
std::vector<WasmCode*> AddCompiledCode(Vector<WasmCompilationResult>);
// Allows to check whether a function has been redirected to the interpreter
// by publishing an entry stub with the {Kind::kInterpreterEntry} code kind.
bool IsRedirectedToInterpreter(uint32_t func_index);
// Free a set of functions of this module. Uncommits whole pages if possible.
// The given vector must be ordered by the instruction start address, and all
// {WasmCode} objects must not be used any more.
// Should only be called via {WasmEngine::FreeDeadCode}, so the engine can do
// its accounting.
void FreeCode(Vector<WasmCode* const>);
2017-11-20 21:34:04 +00:00
private:
friend class WasmCode;
2017-11-20 21:34:04 +00:00
friend class WasmCodeManager;
friend class NativeModuleModificationScope;
2017-11-20 21:34:04 +00:00
// Private constructor, called via {WasmCodeManager::NewNativeModule()}.
NativeModule(WasmEngine* engine, const WasmFeatures& enabled_features,
bool can_request_more, VirtualMemory code_space,
std::shared_ptr<const WasmModule> module,
std::shared_ptr<Counters> async_counters,
std::shared_ptr<NativeModule>* shared_this);
2017-11-20 21:34:04 +00:00
std::unique_ptr<WasmCode> AddCodeWithCodeSpace(
uint32_t index, const CodeDesc& desc, uint32_t stack_slots,
uint32_t tagged_parameter_slots,
OwnedVector<trap_handler::ProtectedInstructionData>
protected_instructions,
OwnedVector<const byte> source_position_table, WasmCode::Kind kind,
ExecutionTier tier, Vector<uint8_t> code_space);
[wasm] Split adding code from publishing it This prepares a refactoring to add and publish compilation results in batches. For this, we need to separate the two phases, so that we can lock the module, allocate all the code space, release the lock, copy the code, lock the module, publish the code, and release the lock again. In particular, this CL does the following: 1) It removes the {AddOwnedCode} method. The functionality of creating the {WasmCode} and memcpy'ing the instruction into that is done in the other {Add*Code} methods. Adding to {owned_code_} is done in {PublishCode}. 2) {PublishInterpreterEntry} is now functionally equivalent to {PublishCode}, so it's removed. 3) After {AddCode}, the caller has to call {PublishCode}. In a follow-up CL, this will be called in batches (first {AddCode} them all, then {PublishCode} them all). 4) {AddCompiledCode} now assumes that the {WasmCompilationResult} succeeded. Otherwise, the caller should directly call {SetError} on the {CompilationState}. 5) {PublishCode} is now the chokepoint for installing code to the code table, the owned code vector, the jump table, and setting interpreter redirections. It replaces previous direct calls to {InstallCode} or explicitly adding to {owned_code_}. 6) Increasing the {generated_code_size_} counter is now done in {AllocateForCode}, which is the chokepoint for allocating space for generated code. This way, we will only increase this counter once once we allocate in batches. R=titzer@chromium.org Bug: v8:8916 Change-Id: I71e02e3a838f21797915cee3ebd373804fb12237 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1530817 Commit-Queue: Clemens Hammacher <clemensh@chromium.org> Reviewed-by: Ben Titzer <titzer@chromium.org> Cr-Commit-Position: refs/heads/master@{#60369}
2019-03-20 13:04:33 +00:00
// Add and publish anonymous code.
WasmCode* AddAndPublishAnonymousCode(Handle<Code>, WasmCode::Kind kind,
const char* name = nullptr);
2017-11-20 21:34:04 +00:00
WasmCode* CreateEmptyJumpTable(uint32_t jump_table_size);
2018-06-19 09:47:17 +00:00
// Hold the {allocation_mutex_} when calling this method.
bool has_interpreter_redirection(uint32_t func_index) {
DCHECK_LT(func_index, num_functions());
DCHECK_LE(module_->num_imported_functions, func_index);
if (!interpreter_redirections_) return false;
uint32_t bitset_idx = func_index - module_->num_imported_functions;
uint8_t byte = interpreter_redirections_[bitset_idx / kBitsPerByte];
return byte & (1 << (bitset_idx % kBitsPerByte));
}
// Hold the {allocation_mutex_} when calling this method.
void SetInterpreterRedirection(uint32_t func_index) {
DCHECK_LT(func_index, num_functions());
DCHECK_LE(module_->num_imported_functions, func_index);
if (!interpreter_redirections_) {
interpreter_redirections_.reset(
new uint8_t[RoundUp<kBitsPerByte>(module_->num_declared_functions) /
kBitsPerByte]{});
}
uint32_t bitset_idx = func_index - module_->num_imported_functions;
uint8_t& byte = interpreter_redirections_[bitset_idx / kBitsPerByte];
byte |= 1 << (bitset_idx % kBitsPerByte);
}
// {WasmCodeAllocator} manages all code reservations and allocations for this
// {NativeModule}.
WasmCodeAllocator code_allocator_;
// Features enabled for this module. We keep a copy of the features that
// were enabled at the time of the creation of this native module,
// to be consistent across asynchronous compilations later.
const WasmFeatures enabled_features_;
Reland "[wasm] Decouple background compile jobs from NativeModule" This is a reland of 92d9b09c0e68382d1427ad25c70db016bb9e5b80. Patch unchanged, errors fixed by https://crrev.com/c/1430059. Original change's description: > [wasm] Decouple background compile jobs from NativeModule > > Background compile jobs should not keep the NativeModule alive, for two > reasons: > 1) We sometimes have to wait for background compilation to finish (from > a foreground task!). This introduces unnecessary latency. > 2) Giving the background compile tasks shared ownership of the > NativeModule causes the NativeModule (and the CompilationState) to > be freed from background tasks, which is error-prone (see > https://crrev.com/c/1400420). > > Instead, this CL introduces a BackgroundCompileToken which is held > alive by the NativeModule and all background compile jobs. The initial > and the final phase of compilation (getting and submitting work) > synchronize on this token to check and ensure that the NativeModule is > and stays alive. During compilation itself, the mutex is released, such > that the NativeModule can die. > The destructor of the NativeModule cancels the BackgroundCompileToken. > Immediately afterwards, the NativeModule and the CompilationState can > die. > > This change allows to remove two hacks introduced previously: The atomic > {aborted_} flag and the {FreeCallbacksTask}. > > R=mstarzinger@chromium.org > CC=titzer@chromium.org > > Bug: v8:8689, v8:7921 > Change-Id: I42e06eab3c944b0988286f2ce18e3c294535dfb6 > Cq-Include-Trybots: luci.v8.try:v8_linux64_tsan_rel > Reviewed-on: https://chromium-review.googlesource.com/c/1421364 > Commit-Queue: Clemens Hammacher <clemensh@chromium.org> > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > Cr-Commit-Position: refs/heads/master@{#59020} TBR=mstarzinger@chromium.org Bug: v8:8689, v8:7921 Change-Id: Iead972ef77c8503da7246cab48e7693b176d8f02 Cq-Include-Trybots: luci.v8.try:v8_linux64_tsan_rel Reviewed-on: https://chromium-review.googlesource.com/c/1429862 Reviewed-by: Clemens Hammacher <clemensh@chromium.org> Commit-Queue: Clemens Hammacher <clemensh@chromium.org> Cr-Commit-Position: refs/heads/master@{#59035}
2019-01-23 11:49:36 +00:00
// The decoded module, stored in a shared_ptr such that background compile
// tasks can keep this alive.
std::shared_ptr<const WasmModule> module_;
// Wire bytes, held in a shared_ptr so they can be kept alive by the
// {WireBytesStorage}, held by background compile tasks.
std::shared_ptr<OwnedVector<const uint8_t>> wire_bytes_;
// Contains entry points for runtime stub calls via {WASM_STUB_CALL}.
Address runtime_stub_entries_[WasmCode::kRuntimeStubCount] = {kNullAddress};
// Jump table used for runtime stubs (i.e. trampolines to embedded builtins).
WasmCode* runtime_stub_table_ = nullptr;
2017-11-20 21:34:04 +00:00
2018-06-19 09:47:17 +00:00
// Jump table used to easily redirect wasm function calls.
WasmCode* jump_table_ = nullptr;
// Lazy compile stub table, containing entries to jump to the
// {WasmCompileLazy} builtin, passing the function index.
WasmCode* lazy_compile_table_ = nullptr;
// The compilation state keeps track of compilation tasks for this module.
// Note that its destructor blocks until all tasks are finished/aborted and
// hence needs to be destructed first when this native module dies.
std::unique_ptr<CompilationState> compilation_state_;
// A cache of the import wrappers, keyed on the kind and signature.
std::unique_ptr<WasmImportWrapperCache> import_wrapper_cache_;
// This mutex protects concurrent calls to {AddCode} and friends.
mutable base::Mutex allocation_mutex_;
//////////////////////////////////////////////////////////////////////////////
// Protected by {allocation_mutex_}:
// Holds all allocated code objects. For lookup based on pc, the key is the
// instruction start address of the value.
std::map<Address, std::unique_ptr<WasmCode>> owned_code_;
std::unique_ptr<WasmCode* []> code_table_;
// Null if no redirections exist, otherwise a bitset over all functions in
// this module marking those functions that have been redirected.
std::unique_ptr<uint8_t[]> interpreter_redirections_;
// End of fields protected by {allocation_mutex_}.
//////////////////////////////////////////////////////////////////////////////
WasmEngine* const engine_;
int modification_scope_depth_ = 0;
UseTrapHandler use_trap_handler_ = kNoTrapHandler;
bool lazy_compile_frozen_ = false;
DISALLOW_COPY_AND_ASSIGN(NativeModule);
2017-11-20 21:34:04 +00:00
};
class V8_EXPORT_PRIVATE WasmCodeManager final {
public:
Revert ""Reland x3 [arraybuffer] Rearchitect backing store ownership"" This reverts commit df8e61777295ad5568eff27722a99fab765eabbd. Reason for revert: Multiple flakes in apparently related areas: https://logs.chromium.org/logs/v8/buildbucket/cr-buildbucket.appspot.com/8906409837768155568/+/steps/Check__flakes_/0/logs/BackingStoreTest.RacyGrowWasmMem.../0 Original change's description: > "Reland x3 [arraybuffer] Rearchitect backing store ownership" > > This is a reland of bc33f5aeba9ceb13f8bfc401c5ba2521c2207ffb > > Original change's description: > > [arraybuffer] Rearchitect backing store ownership > > > > This CL completely rearchitects the ownership of array buffer backing stores, > > consolidating ownership into a {BackingStore} C++ object that is tracked > > throughout V8 using unique_ptr and shared_ptr where appropriate. > > > > Overall, lifetime management is simpler and more explicit. The numerous > > ways that array buffers were initialized have been streamlined to one > > Attach() method on JSArrayBuffer. The array buffer tracker in the > > GC implementation now manages std::shared_ptr<BackingStore> pointers, > > and the construction and destruction of the BackingStore object itself > > handles the underlying page or embedder-allocated memory. > > > > The embedder API remains unchanged for now. We use the > > v8::ArrayBuffer::Contents struct to hide an additional shared_ptr to > > keep the backing store alive properly, even in the case of aliases > > from live heap objects. Thus the embedder has a lower chance of making > > a mistake. Long-term, we should move the embedder to a model where they > > manage backing stores using shared_ptr to an opaque backing store object. > > R=​mlippautz@chromium.org > BUG=v8:9380,v8:9221,chromium:986318 > TBR=ulan@chromium.org > > Change-Id: I6c49e2425029b5664ef1c68dab8b5146f4ed0ff2 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1719191 > Reviewed-by: Ben Titzer <titzer@chromium.org> > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > Commit-Queue: Ben Titzer <titzer@chromium.org> > Cr-Commit-Position: refs/heads/master@{#63007} TBR=mstarzinger@chromium.org,titzer@chromium.org,mlippautz@chromium.org Change-Id: If0266e5893b1325a332d5986337fa7ece2cb6943 No-Presubmit: true No-Tree-Checks: true No-Try: true Bug: v8:9380, v8:9221, chromium:986318 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1729549 Reviewed-by: Francis McCabe <fgm@chromium.org> Commit-Queue: Francis McCabe <fgm@chromium.org> Cr-Commit-Position: refs/heads/master@{#63011}
2019-07-31 15:50:37 +00:00
explicit WasmCodeManager(WasmMemoryTracker* memory_tracker,
size_t max_committed);
2017-11-20 21:34:04 +00:00
#ifdef DEBUG
~WasmCodeManager() {
// No more committed code space.
DCHECK_EQ(0, total_committed_code_space_.load());
}
#endif
Unwind V8 frames correctly on Windows ARM64 On Windows ARM64, OS stack walking does not work because the V8 ARM64 backend doesn't emit unwinding info and also because it doesn't emit ABI compliant stack frames. This was fixed for Windows X64 (https://crrev.com/c/1469329) and documented below: https://docs.google.com/document/d/1-wf50jFlii0c_Pr52lm2ZU-49m220nhYMrHDi3vXnh0 This problem can be fixed similarly for Windows ARM64 by observing that V8 frames usually all have the same prolog which maintains a chain via frame pointer (fp or x29 register). stp fp, lr, [sp, ...] One exception is JSEntry which stops fp pointer chain and needs to be handled specially. So it is possible to define XDATA with UNWIND_CODE which specify how Windows should walk through V8 dynamic frames. The same as X64, since V8 Code objects are all allocated in the same code-range for an Isolate, it is possible to register at most 2 XDATA and a group of PDATA entries to cover stack walking for all the code generated inside that code-range. This is more than 1 PDATA/XDATA because according to the Windows ARM64 exeption handling document, 1 PDATA can cover less than 1MB code range (see below doc). https://docs.microsoft.com/en-us/cpp/build/arm64-exception-handling This PR implements stackwalk for Windows ARM64 to be on par with X64, including embedded builtins, jitted code and wasm jitted code, but not including register handler for handling exception only, because there is no backward compatibility to maintain for Windows ARM64 which was released since 1709 windows build. Bug: chromium:893460 Change-Id: Ic74cbdad8af5cf342185030a4c53796f12ea5429 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1701133 Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> Reviewed-by: Jakob Gruber <jgruber@chromium.org> Commit-Queue: Jakob Gruber <jgruber@chromium.org> Cr-Commit-Position: refs/heads/master@{#63002}
2019-07-27 06:36:52 +00:00
#if defined(V8_OS_WIN64)
Reland "Generate unwind info on Win/x64 by default" The original CL title was updated to reflect CL contents. The --win64-unwinding-info flag still exists but it is set by default. This is a reland of efd8c2d9752c4206966dfd72e4794e025b9843e1 Original change's description: > Remove --win64-unwinding-info flag and always generate unwind info on Win/x64 > > The generation of unwind info to enable stack walking on Windows/x64 > (https://chromium-review.googlesource.com/c/v8/v8/+/1469329) was implemented > behind a temporary flag, in order to coordinate these changes with the > corresponding changes in Chromium. > > The required changes to Chromium > (https://chromium-review.googlesource.com/c/chromium/src/+/1474703) have also > been merged, so we can now remove the flag and enable the generation of stack > unwinding info by default on Windows/x64. > > Bug: v8:3598 > Change-Id: I88814aaeabecc007f5262227aa0681a1d16156d5 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1573138 > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > Reviewed-by: Jakob Gruber <jgruber@chromium.org> > Reviewed-by: Ulan Degenbaev <ulan@chromium.org> > Commit-Queue: Paolo Severini <paolosev@microsoft.com> > Cr-Commit-Position: refs/heads/master@{#61020} Bug: v8:3598, chromium:958035 Change-Id: Ie53b39f3bb31567797a61e5110685284c266c1f9 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1599596 Commit-Queue: Jakob Gruber <jgruber@chromium.org> Reviewed-by: Jakob Gruber <jgruber@chromium.org> Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> Reviewed-by: Ulan Degenbaev <ulan@chromium.org> Cr-Commit-Position: refs/heads/master@{#61368}
2019-04-24 22:01:00 +00:00
bool CanRegisterUnwindInfoForNonABICompliantCodeRange() const;
Unwind V8 frames correctly on Windows ARM64 On Windows ARM64, OS stack walking does not work because the V8 ARM64 backend doesn't emit unwinding info and also because it doesn't emit ABI compliant stack frames. This was fixed for Windows X64 (https://crrev.com/c/1469329) and documented below: https://docs.google.com/document/d/1-wf50jFlii0c_Pr52lm2ZU-49m220nhYMrHDi3vXnh0 This problem can be fixed similarly for Windows ARM64 by observing that V8 frames usually all have the same prolog which maintains a chain via frame pointer (fp or x29 register). stp fp, lr, [sp, ...] One exception is JSEntry which stops fp pointer chain and needs to be handled specially. So it is possible to define XDATA with UNWIND_CODE which specify how Windows should walk through V8 dynamic frames. The same as X64, since V8 Code objects are all allocated in the same code-range for an Isolate, it is possible to register at most 2 XDATA and a group of PDATA entries to cover stack walking for all the code generated inside that code-range. This is more than 1 PDATA/XDATA because according to the Windows ARM64 exeption handling document, 1 PDATA can cover less than 1MB code range (see below doc). https://docs.microsoft.com/en-us/cpp/build/arm64-exception-handling This PR implements stackwalk for Windows ARM64 to be on par with X64, including embedded builtins, jitted code and wasm jitted code, but not including register handler for handling exception only, because there is no backward compatibility to maintain for Windows ARM64 which was released since 1709 windows build. Bug: chromium:893460 Change-Id: Ic74cbdad8af5cf342185030a4c53796f12ea5429 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1701133 Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> Reviewed-by: Jakob Gruber <jgruber@chromium.org> Commit-Queue: Jakob Gruber <jgruber@chromium.org> Cr-Commit-Position: refs/heads/master@{#63002}
2019-07-27 06:36:52 +00:00
#endif // V8_OS_WIN64
Reland "Generate unwind info on Win/x64 by default" The original CL title was updated to reflect CL contents. The --win64-unwinding-info flag still exists but it is set by default. This is a reland of efd8c2d9752c4206966dfd72e4794e025b9843e1 Original change's description: > Remove --win64-unwinding-info flag and always generate unwind info on Win/x64 > > The generation of unwind info to enable stack walking on Windows/x64 > (https://chromium-review.googlesource.com/c/v8/v8/+/1469329) was implemented > behind a temporary flag, in order to coordinate these changes with the > corresponding changes in Chromium. > > The required changes to Chromium > (https://chromium-review.googlesource.com/c/chromium/src/+/1474703) have also > been merged, so we can now remove the flag and enable the generation of stack > unwinding info by default on Windows/x64. > > Bug: v8:3598 > Change-Id: I88814aaeabecc007f5262227aa0681a1d16156d5 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1573138 > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > Reviewed-by: Jakob Gruber <jgruber@chromium.org> > Reviewed-by: Ulan Degenbaev <ulan@chromium.org> > Commit-Queue: Paolo Severini <paolosev@microsoft.com> > Cr-Commit-Position: refs/heads/master@{#61020} Bug: v8:3598, chromium:958035 Change-Id: Ie53b39f3bb31567797a61e5110685284c266c1f9 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1599596 Commit-Queue: Jakob Gruber <jgruber@chromium.org> Reviewed-by: Jakob Gruber <jgruber@chromium.org> Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> Reviewed-by: Ulan Degenbaev <ulan@chromium.org> Cr-Commit-Position: refs/heads/master@{#61368}
2019-04-24 22:01:00 +00:00
2018-06-19 09:47:17 +00:00
NativeModule* LookupNativeModule(Address pc) const;
2017-11-20 21:34:04 +00:00
WasmCode* LookupCode(Address pc) const;
size_t committed_code_space() const {
return total_committed_code_space_.load();
}
2017-11-20 21:34:04 +00:00
void SetMaxCommittedMemoryForTesting(size_t limit);
Unwind V8 frames correctly on Windows ARM64 On Windows ARM64, OS stack walking does not work because the V8 ARM64 backend doesn't emit unwinding info and also because it doesn't emit ABI compliant stack frames. This was fixed for Windows X64 (https://crrev.com/c/1469329) and documented below: https://docs.google.com/document/d/1-wf50jFlii0c_Pr52lm2ZU-49m220nhYMrHDi3vXnh0 This problem can be fixed similarly for Windows ARM64 by observing that V8 frames usually all have the same prolog which maintains a chain via frame pointer (fp or x29 register). stp fp, lr, [sp, ...] One exception is JSEntry which stops fp pointer chain and needs to be handled specially. So it is possible to define XDATA with UNWIND_CODE which specify how Windows should walk through V8 dynamic frames. The same as X64, since V8 Code objects are all allocated in the same code-range for an Isolate, it is possible to register at most 2 XDATA and a group of PDATA entries to cover stack walking for all the code generated inside that code-range. This is more than 1 PDATA/XDATA because according to the Windows ARM64 exeption handling document, 1 PDATA can cover less than 1MB code range (see below doc). https://docs.microsoft.com/en-us/cpp/build/arm64-exception-handling This PR implements stackwalk for Windows ARM64 to be on par with X64, including embedded builtins, jitted code and wasm jitted code, but not including register handler for handling exception only, because there is no backward compatibility to maintain for Windows ARM64 which was released since 1709 windows build. Bug: chromium:893460 Change-Id: Ic74cbdad8af5cf342185030a4c53796f12ea5429 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1701133 Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> Reviewed-by: Jakob Gruber <jgruber@chromium.org> Commit-Queue: Jakob Gruber <jgruber@chromium.org> Cr-Commit-Position: refs/heads/master@{#63002}
2019-07-27 06:36:52 +00:00
#if defined(V8_OS_WIN64)
Reland "Generate unwind info on Win/x64 by default" The original CL title was updated to reflect CL contents. The --win64-unwinding-info flag still exists but it is set by default. This is a reland of efd8c2d9752c4206966dfd72e4794e025b9843e1 Original change's description: > Remove --win64-unwinding-info flag and always generate unwind info on Win/x64 > > The generation of unwind info to enable stack walking on Windows/x64 > (https://chromium-review.googlesource.com/c/v8/v8/+/1469329) was implemented > behind a temporary flag, in order to coordinate these changes with the > corresponding changes in Chromium. > > The required changes to Chromium > (https://chromium-review.googlesource.com/c/chromium/src/+/1474703) have also > been merged, so we can now remove the flag and enable the generation of stack > unwinding info by default on Windows/x64. > > Bug: v8:3598 > Change-Id: I88814aaeabecc007f5262227aa0681a1d16156d5 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1573138 > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > Reviewed-by: Jakob Gruber <jgruber@chromium.org> > Reviewed-by: Ulan Degenbaev <ulan@chromium.org> > Commit-Queue: Paolo Severini <paolosev@microsoft.com> > Cr-Commit-Position: refs/heads/master@{#61020} Bug: v8:3598, chromium:958035 Change-Id: Ie53b39f3bb31567797a61e5110685284c266c1f9 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1599596 Commit-Queue: Jakob Gruber <jgruber@chromium.org> Reviewed-by: Jakob Gruber <jgruber@chromium.org> Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> Reviewed-by: Ulan Degenbaev <ulan@chromium.org> Cr-Commit-Position: refs/heads/master@{#61368}
2019-04-24 22:01:00 +00:00
void DisableWin64UnwindInfoForTesting() {
is_win64_unwind_info_disabled_for_testing_ = true;
}
Unwind V8 frames correctly on Windows ARM64 On Windows ARM64, OS stack walking does not work because the V8 ARM64 backend doesn't emit unwinding info and also because it doesn't emit ABI compliant stack frames. This was fixed for Windows X64 (https://crrev.com/c/1469329) and documented below: https://docs.google.com/document/d/1-wf50jFlii0c_Pr52lm2ZU-49m220nhYMrHDi3vXnh0 This problem can be fixed similarly for Windows ARM64 by observing that V8 frames usually all have the same prolog which maintains a chain via frame pointer (fp or x29 register). stp fp, lr, [sp, ...] One exception is JSEntry which stops fp pointer chain and needs to be handled specially. So it is possible to define XDATA with UNWIND_CODE which specify how Windows should walk through V8 dynamic frames. The same as X64, since V8 Code objects are all allocated in the same code-range for an Isolate, it is possible to register at most 2 XDATA and a group of PDATA entries to cover stack walking for all the code generated inside that code-range. This is more than 1 PDATA/XDATA because according to the Windows ARM64 exeption handling document, 1 PDATA can cover less than 1MB code range (see below doc). https://docs.microsoft.com/en-us/cpp/build/arm64-exception-handling This PR implements stackwalk for Windows ARM64 to be on par with X64, including embedded builtins, jitted code and wasm jitted code, but not including register handler for handling exception only, because there is no backward compatibility to maintain for Windows ARM64 which was released since 1709 windows build. Bug: chromium:893460 Change-Id: Ic74cbdad8af5cf342185030a4c53796f12ea5429 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1701133 Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> Reviewed-by: Jakob Gruber <jgruber@chromium.org> Commit-Queue: Jakob Gruber <jgruber@chromium.org> Cr-Commit-Position: refs/heads/master@{#63002}
2019-07-27 06:36:52 +00:00
#endif // V8_OS_WIN64
Reland "Generate unwind info on Win/x64 by default" The original CL title was updated to reflect CL contents. The --win64-unwinding-info flag still exists but it is set by default. This is a reland of efd8c2d9752c4206966dfd72e4794e025b9843e1 Original change's description: > Remove --win64-unwinding-info flag and always generate unwind info on Win/x64 > > The generation of unwind info to enable stack walking on Windows/x64 > (https://chromium-review.googlesource.com/c/v8/v8/+/1469329) was implemented > behind a temporary flag, in order to coordinate these changes with the > corresponding changes in Chromium. > > The required changes to Chromium > (https://chromium-review.googlesource.com/c/chromium/src/+/1474703) have also > been merged, so we can now remove the flag and enable the generation of stack > unwinding info by default on Windows/x64. > > Bug: v8:3598 > Change-Id: I88814aaeabecc007f5262227aa0681a1d16156d5 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1573138 > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > Reviewed-by: Jakob Gruber <jgruber@chromium.org> > Reviewed-by: Ulan Degenbaev <ulan@chromium.org> > Commit-Queue: Paolo Severini <paolosev@microsoft.com> > Cr-Commit-Position: refs/heads/master@{#61020} Bug: v8:3598, chromium:958035 Change-Id: Ie53b39f3bb31567797a61e5110685284c266c1f9 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1599596 Commit-Queue: Jakob Gruber <jgruber@chromium.org> Reviewed-by: Jakob Gruber <jgruber@chromium.org> Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> Reviewed-by: Ulan Degenbaev <ulan@chromium.org> Cr-Commit-Position: refs/heads/master@{#61368}
2019-04-24 22:01:00 +00:00
static size_t EstimateNativeModuleCodeSize(const WasmModule* module);
static size_t EstimateNativeModuleNonCodeSize(const WasmModule* module);
2017-11-20 21:34:04 +00:00
private:
friend class WasmCodeAllocator;
friend class WasmEngine;
std::shared_ptr<NativeModule> NewNativeModule(
WasmEngine* engine, Isolate* isolate,
const WasmFeatures& enabled_features, size_t code_size_estimate,
bool can_request_more, std::shared_ptr<const WasmModule> module);
2017-11-20 21:34:04 +00:00
V8_WARN_UNUSED_RESULT VirtualMemory TryAllocate(size_t size,
void* hint = nullptr);
bool Commit(base::AddressRegion);
void Decommit(base::AddressRegion);
2017-11-20 21:34:04 +00:00
void FreeNativeModule(Vector<VirtualMemory> owned_code,
size_t committed_size);
void AssignRange(base::AddressRegion, NativeModule*);
2017-11-20 21:34:04 +00:00
Revert ""Reland x3 [arraybuffer] Rearchitect backing store ownership"" This reverts commit df8e61777295ad5568eff27722a99fab765eabbd. Reason for revert: Multiple flakes in apparently related areas: https://logs.chromium.org/logs/v8/buildbucket/cr-buildbucket.appspot.com/8906409837768155568/+/steps/Check__flakes_/0/logs/BackingStoreTest.RacyGrowWasmMem.../0 Original change's description: > "Reland x3 [arraybuffer] Rearchitect backing store ownership" > > This is a reland of bc33f5aeba9ceb13f8bfc401c5ba2521c2207ffb > > Original change's description: > > [arraybuffer] Rearchitect backing store ownership > > > > This CL completely rearchitects the ownership of array buffer backing stores, > > consolidating ownership into a {BackingStore} C++ object that is tracked > > throughout V8 using unique_ptr and shared_ptr where appropriate. > > > > Overall, lifetime management is simpler and more explicit. The numerous > > ways that array buffers were initialized have been streamlined to one > > Attach() method on JSArrayBuffer. The array buffer tracker in the > > GC implementation now manages std::shared_ptr<BackingStore> pointers, > > and the construction and destruction of the BackingStore object itself > > handles the underlying page or embedder-allocated memory. > > > > The embedder API remains unchanged for now. We use the > > v8::ArrayBuffer::Contents struct to hide an additional shared_ptr to > > keep the backing store alive properly, even in the case of aliases > > from live heap objects. Thus the embedder has a lower chance of making > > a mistake. Long-term, we should move the embedder to a model where they > > manage backing stores using shared_ptr to an opaque backing store object. > > R=​mlippautz@chromium.org > BUG=v8:9380,v8:9221,chromium:986318 > TBR=ulan@chromium.org > > Change-Id: I6c49e2425029b5664ef1c68dab8b5146f4ed0ff2 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1719191 > Reviewed-by: Ben Titzer <titzer@chromium.org> > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > Commit-Queue: Ben Titzer <titzer@chromium.org> > Cr-Commit-Position: refs/heads/master@{#63007} TBR=mstarzinger@chromium.org,titzer@chromium.org,mlippautz@chromium.org Change-Id: If0266e5893b1325a332d5986337fa7ece2cb6943 No-Presubmit: true No-Tree-Checks: true No-Try: true Bug: v8:9380, v8:9221, chromium:986318 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1729549 Reviewed-by: Francis McCabe <fgm@chromium.org> Commit-Queue: Francis McCabe <fgm@chromium.org> Cr-Commit-Position: refs/heads/master@{#63011}
2019-07-31 15:50:37 +00:00
WasmMemoryTracker* const memory_tracker_;
size_t max_committed_code_space_;
Unwind V8 frames correctly on Windows ARM64 On Windows ARM64, OS stack walking does not work because the V8 ARM64 backend doesn't emit unwinding info and also because it doesn't emit ABI compliant stack frames. This was fixed for Windows X64 (https://crrev.com/c/1469329) and documented below: https://docs.google.com/document/d/1-wf50jFlii0c_Pr52lm2ZU-49m220nhYMrHDi3vXnh0 This problem can be fixed similarly for Windows ARM64 by observing that V8 frames usually all have the same prolog which maintains a chain via frame pointer (fp or x29 register). stp fp, lr, [sp, ...] One exception is JSEntry which stops fp pointer chain and needs to be handled specially. So it is possible to define XDATA with UNWIND_CODE which specify how Windows should walk through V8 dynamic frames. The same as X64, since V8 Code objects are all allocated in the same code-range for an Isolate, it is possible to register at most 2 XDATA and a group of PDATA entries to cover stack walking for all the code generated inside that code-range. This is more than 1 PDATA/XDATA because according to the Windows ARM64 exeption handling document, 1 PDATA can cover less than 1MB code range (see below doc). https://docs.microsoft.com/en-us/cpp/build/arm64-exception-handling This PR implements stackwalk for Windows ARM64 to be on par with X64, including embedded builtins, jitted code and wasm jitted code, but not including register handler for handling exception only, because there is no backward compatibility to maintain for Windows ARM64 which was released since 1709 windows build. Bug: chromium:893460 Change-Id: Ic74cbdad8af5cf342185030a4c53796f12ea5429 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1701133 Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> Reviewed-by: Jakob Gruber <jgruber@chromium.org> Commit-Queue: Jakob Gruber <jgruber@chromium.org> Cr-Commit-Position: refs/heads/master@{#63002}
2019-07-27 06:36:52 +00:00
#if defined(V8_OS_WIN64)
Reland "Generate unwind info on Win/x64 by default" The original CL title was updated to reflect CL contents. The --win64-unwinding-info flag still exists but it is set by default. This is a reland of efd8c2d9752c4206966dfd72e4794e025b9843e1 Original change's description: > Remove --win64-unwinding-info flag and always generate unwind info on Win/x64 > > The generation of unwind info to enable stack walking on Windows/x64 > (https://chromium-review.googlesource.com/c/v8/v8/+/1469329) was implemented > behind a temporary flag, in order to coordinate these changes with the > corresponding changes in Chromium. > > The required changes to Chromium > (https://chromium-review.googlesource.com/c/chromium/src/+/1474703) have also > been merged, so we can now remove the flag and enable the generation of stack > unwinding info by default on Windows/x64. > > Bug: v8:3598 > Change-Id: I88814aaeabecc007f5262227aa0681a1d16156d5 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1573138 > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > Reviewed-by: Jakob Gruber <jgruber@chromium.org> > Reviewed-by: Ulan Degenbaev <ulan@chromium.org> > Commit-Queue: Paolo Severini <paolosev@microsoft.com> > Cr-Commit-Position: refs/heads/master@{#61020} Bug: v8:3598, chromium:958035 Change-Id: Ie53b39f3bb31567797a61e5110685284c266c1f9 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1599596 Commit-Queue: Jakob Gruber <jgruber@chromium.org> Reviewed-by: Jakob Gruber <jgruber@chromium.org> Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> Reviewed-by: Ulan Degenbaev <ulan@chromium.org> Cr-Commit-Position: refs/heads/master@{#61368}
2019-04-24 22:01:00 +00:00
bool is_win64_unwind_info_disabled_for_testing_;
Unwind V8 frames correctly on Windows ARM64 On Windows ARM64, OS stack walking does not work because the V8 ARM64 backend doesn't emit unwinding info and also because it doesn't emit ABI compliant stack frames. This was fixed for Windows X64 (https://crrev.com/c/1469329) and documented below: https://docs.google.com/document/d/1-wf50jFlii0c_Pr52lm2ZU-49m220nhYMrHDi3vXnh0 This problem can be fixed similarly for Windows ARM64 by observing that V8 frames usually all have the same prolog which maintains a chain via frame pointer (fp or x29 register). stp fp, lr, [sp, ...] One exception is JSEntry which stops fp pointer chain and needs to be handled specially. So it is possible to define XDATA with UNWIND_CODE which specify how Windows should walk through V8 dynamic frames. The same as X64, since V8 Code objects are all allocated in the same code-range for an Isolate, it is possible to register at most 2 XDATA and a group of PDATA entries to cover stack walking for all the code generated inside that code-range. This is more than 1 PDATA/XDATA because according to the Windows ARM64 exeption handling document, 1 PDATA can cover less than 1MB code range (see below doc). https://docs.microsoft.com/en-us/cpp/build/arm64-exception-handling This PR implements stackwalk for Windows ARM64 to be on par with X64, including embedded builtins, jitted code and wasm jitted code, but not including register handler for handling exception only, because there is no backward compatibility to maintain for Windows ARM64 which was released since 1709 windows build. Bug: chromium:893460 Change-Id: Ic74cbdad8af5cf342185030a4c53796f12ea5429 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1701133 Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> Reviewed-by: Jakob Gruber <jgruber@chromium.org> Commit-Queue: Jakob Gruber <jgruber@chromium.org> Cr-Commit-Position: refs/heads/master@{#63002}
2019-07-27 06:36:52 +00:00
#endif // V8_OS_WIN64
Reland "Generate unwind info on Win/x64 by default" The original CL title was updated to reflect CL contents. The --win64-unwinding-info flag still exists but it is set by default. This is a reland of efd8c2d9752c4206966dfd72e4794e025b9843e1 Original change's description: > Remove --win64-unwinding-info flag and always generate unwind info on Win/x64 > > The generation of unwind info to enable stack walking on Windows/x64 > (https://chromium-review.googlesource.com/c/v8/v8/+/1469329) was implemented > behind a temporary flag, in order to coordinate these changes with the > corresponding changes in Chromium. > > The required changes to Chromium > (https://chromium-review.googlesource.com/c/chromium/src/+/1474703) have also > been merged, so we can now remove the flag and enable the generation of stack > unwinding info by default on Windows/x64. > > Bug: v8:3598 > Change-Id: I88814aaeabecc007f5262227aa0681a1d16156d5 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1573138 > Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> > Reviewed-by: Jakob Gruber <jgruber@chromium.org> > Reviewed-by: Ulan Degenbaev <ulan@chromium.org> > Commit-Queue: Paolo Severini <paolosev@microsoft.com> > Cr-Commit-Position: refs/heads/master@{#61020} Bug: v8:3598, chromium:958035 Change-Id: Ie53b39f3bb31567797a61e5110685284c266c1f9 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1599596 Commit-Queue: Jakob Gruber <jgruber@chromium.org> Reviewed-by: Jakob Gruber <jgruber@chromium.org> Reviewed-by: Michael Starzinger <mstarzinger@chromium.org> Reviewed-by: Ulan Degenbaev <ulan@chromium.org> Cr-Commit-Position: refs/heads/master@{#61368}
2019-04-24 22:01:00 +00:00
std::atomic<size_t> total_committed_code_space_;
// If the committed code space exceeds {critical_committed_code_space_}, then
// we trigger a GC before creating the next module. This value is set to the
// currently committed space plus 50% of the available code space on creation
// and updated after each GC.
std::atomic<size_t> critical_committed_code_space_;
mutable base::Mutex native_modules_mutex_;
//////////////////////////////////////////////////////////////////////////////
// Protected by {native_modules_mutex_}:
2017-11-20 21:34:04 +00:00
std::map<Address, std::pair<Address, NativeModule*>> lookup_map_;
// End of fields protected by {native_modules_mutex_}.
//////////////////////////////////////////////////////////////////////////////
DISALLOW_COPY_AND_ASSIGN(WasmCodeManager);
2017-11-20 21:34:04 +00:00
};
// Within the scope, the native_module is writable and not executable.
// At the scope's destruction, the native_module is executable and not writable.
// The states inside the scope and at the scope termination are irrespective of
// native_module's state when entering the scope.
// We currently mark the entire module's memory W^X:
// - for AOT, that's as efficient as it can be.
// - for Lazy, we don't have a heuristic for functions that may need patching,
// and even if we did, the resulting set of pages may be fragmented.
// Currently, we try and keep the number of syscalls low.
// - similar argument for debug time.
class NativeModuleModificationScope final {
public:
explicit NativeModuleModificationScope(NativeModule* native_module);
~NativeModuleModificationScope();
private:
NativeModule* native_module_;
};
// {WasmCodeRefScope}s form a perfect stack. New {WasmCode} pointers generated
// by e.g. creating new code or looking up code by its address are added to the
// top-most {WasmCodeRefScope}.
class V8_EXPORT_PRIVATE WasmCodeRefScope {
public:
WasmCodeRefScope();
~WasmCodeRefScope();
// Register a {WasmCode} reference in the current {WasmCodeRefScope}. Fails if
// there is no current scope.
static void AddRef(WasmCode*);
private:
WasmCodeRefScope* const previous_scope_;
std::unordered_set<WasmCode*> code_ptrs_;
DISALLOW_COPY_AND_ASSIGN(WasmCodeRefScope);
};
// Similarly to a global handle, a {GlobalWasmCodeRef} stores a single
// ref-counted pointer to a {WasmCode} object.
class GlobalWasmCodeRef {
public:
explicit GlobalWasmCodeRef(WasmCode* code,
std::shared_ptr<NativeModule> native_module)
: code_(code), native_module_(std::move(native_module)) {
code_->IncRef();
}
~GlobalWasmCodeRef() { WasmCode::DecrementRefCount({&code_, 1}); }
// Get a pointer to the contained {WasmCode} object. This is only guaranteed
// to exist as long as this {GlobalWasmCodeRef} exists.
WasmCode* code() const { return code_; }
private:
WasmCode* const code_;
// Also keep the {NativeModule} alive.
const std::shared_ptr<NativeModule> native_module_;
DISALLOW_COPY_AND_ASSIGN(GlobalWasmCodeRef);
};
Revert "Revert "[wasm] A simple allocator datastructure for off-the heap"" This reverts commit ee5c31f3354ecd42a99e45a153061bc8b7a0a980. Reason for revert: Fixed compiler failure Original change's description: > Revert "[wasm] A simple allocator datastructure for off-the heap" > > This reverts commit 110d9ab0052f9c70c0c44e9c30397789568bbebf. > > Reason for revert: https://build.chromium.org/p/client.v8/builders/V8%20Linux64%20-%20debug%20builder/builds/26607 > > Surprising we're seeing a failure on Linux 64 *after* CQ. Is the compiler there different? > > Original change's description: > > [wasm] A simple allocator datastructure for off-the heap > > > > We'll use this allocator in a follow-up CL to: > > - allocate speculative sizes of memory for a module that's being > > compiled (e.g. 2*size of wasm code). > > - each module will own such a sub-pool, and then use it to allocate > > contiguous chunks of memory for code. > > > > The underlying assumptions for the chosen allocation strategy is that: > > - the allocation granularity for pools is 1 page, so that no one page > > is owned by more than one wasm module > > - typical pool sizes (given module sizes) are multiple pages. > > - modules and module instances are typically few and long lived. Typically, > > we expect one module and one instance. > > > > This means we shouldn't expect fragmentations that lead to code being > > non-allocatable, or prohibitively many ranges. > > > > The data structure just manages ranges of addresses. Virtual memory management > > will be separate, as part of the responsibility of a "WasmHeap" > > that will be introduced in the future. So will concurrency control. > > > > Bug: > > Change-Id: Id99f46d10c25553b013054d994760f3c2a737c39 > > Reviewed-on: https://chromium-review.googlesource.com/669296 > > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > > Reviewed-by: Eric Holk <eholk@chromium.org> > > Reviewed-by: Brad Nelson <bradnelson@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#48053} > > TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org > > Change-Id: Id82fa341b77624e4971f24c4757a9a666a65930c > No-Presubmit: true > No-Tree-Checks: true > No-Try: true > Reviewed-on: https://chromium-review.googlesource.com/670141 > Reviewed-by: Mircea Trofin <mtrofin@chromium.org> > Commit-Queue: Mircea Trofin <mtrofin@chromium.org> > Cr-Commit-Position: refs/heads/master@{#48054} TBR=bradnelson@chromium.org,mtrofin@chromium.org,eholk@chromium.org Change-Id: Ib6a7a3e6098d2689e60cdca85ec77e57e5295e48 No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://chromium-review.googlesource.com/670142 Commit-Queue: Mircea Trofin <mtrofin@chromium.org> Reviewed-by: Mircea Trofin <mtrofin@chromium.org> Cr-Commit-Position: refs/heads/master@{#48055}
2017-09-16 05:22:38 +00:00
} // namespace wasm
} // namespace internal
} // namespace v8
#endif // V8_WASM_WASM_CODE_MANAGER_H_