v8/test/unittests/wasm/wasm-code-manager-unittest.cc

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

393 lines
14 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.
#include "test/unittests/test-utils.h"
#include "testing/gmock/include/gmock/gmock.h"
#include "src/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
namespace v8 {
namespace internal {
namespace wasm {
namespace wasm_heap_unittest {
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
class DisjointAllocationPoolTest : public ::testing::Test {
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
public:
Address A(size_t n) { return reinterpret_cast<Address>(n); }
void CheckLooksLike(const DisjointAllocationPool& mem,
std::vector<std::pair<size_t, size_t>> expectation);
DisjointAllocationPool Make(std::vector<std::pair<size_t, size_t>> model);
};
void DisjointAllocationPoolTest::CheckLooksLike(
const DisjointAllocationPool& mem,
std::vector<std::pair<size_t, size_t>> expectation) {
const auto& ranges = mem.ranges();
CHECK_EQ(ranges.size(), expectation.size());
auto iter = expectation.begin();
for (auto it = ranges.begin(), e = ranges.end(); it != e; ++it, ++iter) {
CHECK_EQ(it->first, A(iter->first));
CHECK_EQ(it->second, A(iter->second));
}
}
DisjointAllocationPool DisjointAllocationPoolTest::Make(
std::vector<std::pair<size_t, size_t>> model) {
DisjointAllocationPool ret;
for (auto& pair : model) {
ret.Merge(DisjointAllocationPool(A(pair.first), A(pair.second)));
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
}
return ret;
}
TEST_F(DisjointAllocationPoolTest, Construct) {
DisjointAllocationPool a;
CHECK(a.IsEmpty());
CHECK_EQ(a.ranges().size(), 0);
DisjointAllocationPool b = Make({{1, 5}});
CHECK(!b.IsEmpty());
CHECK_EQ(b.ranges().size(), 1);
a.Merge(std::move(b));
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
CheckLooksLike(a, {{1, 5}});
DisjointAllocationPool c;
a.Merge(std::move(c));
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
CheckLooksLike(a, {{1, 5}});
DisjointAllocationPool e, f;
e.Merge(std::move(f));
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
CHECK(e.IsEmpty());
}
TEST_F(DisjointAllocationPoolTest, SimpleExtract) {
DisjointAllocationPool a = Make({{1, 5}});
DisjointAllocationPool b = a.AllocatePool(2);
CheckLooksLike(a, {{3, 5}});
CheckLooksLike(b, {{1, 3}});
a.Merge(std::move(b));
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
CheckLooksLike(a, {{1, 5}});
CHECK_EQ(a.ranges().size(), 1);
CHECK_EQ(a.ranges().front().first, A(1));
CHECK_EQ(a.ranges().front().second, A(5));
}
TEST_F(DisjointAllocationPoolTest, ExtractAll) {
DisjointAllocationPool a(A(1), A(5));
DisjointAllocationPool b = a.AllocatePool(4);
CheckLooksLike(b, {{1, 5}});
CHECK(a.IsEmpty());
a.Merge(std::move(b));
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
CheckLooksLike(a, {{1, 5}});
}
TEST_F(DisjointAllocationPoolTest, ExtractAccross) {
DisjointAllocationPool a = Make({{1, 5}, {10, 20}});
DisjointAllocationPool b = a.AllocatePool(5);
CheckLooksLike(a, {{11, 20}});
CheckLooksLike(b, {{1, 5}, {10, 11}});
a.Merge(std::move(b));
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
CheckLooksLike(a, {{1, 5}, {10, 20}});
}
TEST_F(DisjointAllocationPoolTest, ReassembleOutOfOrder) {
DisjointAllocationPool a = Make({{1, 5}, {10, 15}});
DisjointAllocationPool b = Make({{7, 8}, {20, 22}});
a.Merge(std::move(b));
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
CheckLooksLike(a, {{1, 5}, {7, 8}, {10, 15}, {20, 22}});
DisjointAllocationPool c = Make({{1, 5}, {10, 15}});
DisjointAllocationPool d = Make({{7, 8}, {20, 22}});
d.Merge(std::move(c));
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
CheckLooksLike(d, {{1, 5}, {7, 8}, {10, 15}, {20, 22}});
}
TEST_F(DisjointAllocationPoolTest, FailToExtract) {
DisjointAllocationPool a = Make({{1, 5}});
DisjointAllocationPool b = a.AllocatePool(5);
CheckLooksLike(a, {{1, 5}});
CHECK(b.IsEmpty());
}
TEST_F(DisjointAllocationPoolTest, FailToExtractExact) {
DisjointAllocationPool a = Make({{1, 5}, {10, 14}});
DisjointAllocationPool b = a.Allocate(5);
CheckLooksLike(a, {{1, 5}, {10, 14}});
CHECK(b.IsEmpty());
}
TEST_F(DisjointAllocationPoolTest, ExtractExact) {
DisjointAllocationPool a = Make({{1, 5}, {10, 15}});
DisjointAllocationPool b = a.Allocate(5);
CheckLooksLike(a, {{1, 5}});
CheckLooksLike(b, {{10, 15}});
}
TEST_F(DisjointAllocationPoolTest, Merging) {
DisjointAllocationPool a = Make({{10, 15}, {20, 25}});
a.Merge(Make({{15, 20}}));
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
CheckLooksLike(a, {{10, 25}});
}
TEST_F(DisjointAllocationPoolTest, MergingMore) {
DisjointAllocationPool a = Make({{10, 15}, {20, 25}, {30, 35}});
a.Merge(Make({{15, 20}, {25, 30}}));
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
CheckLooksLike(a, {{10, 35}});
}
TEST_F(DisjointAllocationPoolTest, MergingSkip) {
DisjointAllocationPool a = Make({{10, 15}, {20, 25}, {30, 35}});
a.Merge(Make({{25, 30}}));
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
CheckLooksLike(a, {{10, 15}, {20, 35}});
}
TEST_F(DisjointAllocationPoolTest, MergingSkipLargerSrc) {
DisjointAllocationPool a = Make({{10, 15}, {20, 25}, {30, 35}});
a.Merge(Make({{25, 30}, {35, 40}}));
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
CheckLooksLike(a, {{10, 15}, {20, 40}});
}
TEST_F(DisjointAllocationPoolTest, MergingSkipLargerSrcWithGap) {
DisjointAllocationPool a = Make({{10, 15}, {20, 25}, {30, 35}});
a.Merge(Make({{25, 30}, {36, 40}}));
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
CheckLooksLike(a, {{10, 15}, {20, 35}, {36, 40}});
}
class WasmCodeManagerTest : public TestWithContext {
2017-11-20 21:34:04 +00:00
public:
using NativeModulePtr = std::unique_ptr<NativeModule>;
enum ModuleStyle : int { Fixed = 0, Growable = 1 };
const std::vector<ModuleStyle> styles() const {
return std::vector<ModuleStyle>({Fixed, Growable});
}
// We pretend all our modules have 10 functions and no imports, just so
// we can size up the code_table.
NativeModulePtr AllocFixedModule(WasmCodeManager* manager, size_t size) {
return manager->NewNativeModule(size, 10, 0, false);
}
NativeModulePtr AllocGrowableModule(WasmCodeManager* manager, size_t size) {
return manager->NewNativeModule(size, 10, 0, true);
}
NativeModulePtr AllocModule(WasmCodeManager* manager, size_t size,
ModuleStyle style) {
switch (style) {
case Fixed:
return AllocFixedModule(manager, size);
case Growable:
return AllocGrowableModule(manager, size);
default:
UNREACHABLE();
}
}
WasmCode* AddCode(NativeModule* native_module, uint32_t index, size_t size) {
CodeDesc desc;
memset(reinterpret_cast<void*>(&desc), 0, sizeof(CodeDesc));
std::unique_ptr<byte[]> exec_buff(new byte[size]);
desc.buffer = exec_buff.get();
desc.instr_size = static_cast<int>(size);
return native_module->AddCode(desc, 0, index, 0, 0, {}, WasmCode::kOther);
2017-11-20 21:34:04 +00:00
}
size_t page() const { return AllocatePageSize(); }
2017-11-20 21:34:04 +00:00
v8::Isolate* v8_isolate() const {
return reinterpret_cast<v8::Isolate*>(isolate());
}
};
TEST_F(WasmCodeManagerTest, EmptyCase) {
for (auto style : styles()) {
WasmCodeManager manager(v8_isolate(), 0 * page());
CHECK_EQ(0, manager.remaining_uncommitted());
NativeModulePtr native_module = AllocModule(&manager, 1 * page(), style);
CHECK(native_module);
WasmCode* code = AddCode(native_module.get(), 0, 10);
CHECK_NULL(code);
CHECK_EQ(0, manager.remaining_uncommitted());
native_module.reset();
CHECK_EQ(0, manager.remaining_uncommitted());
}
}
TEST_F(WasmCodeManagerTest, AllocateAndGoOverLimit) {
for (auto style : styles()) {
WasmCodeManager manager(v8_isolate(), 1 * page());
CHECK_EQ(1 * page(), manager.remaining_uncommitted());
NativeModulePtr native_module = AllocModule(&manager, 1 * page(), style);
CHECK(native_module);
CHECK_EQ(1 * page(), manager.remaining_uncommitted());
uint32_t index = 0;
WasmCode* code = AddCode(native_module.get(), index++, 1 * kCodeAlignment);
CHECK_NOT_NULL(code);
CHECK_EQ(0, manager.remaining_uncommitted());
code = AddCode(native_module.get(), index++, 3 * kCodeAlignment);
CHECK_NOT_NULL(code);
CHECK_EQ(0, manager.remaining_uncommitted());
code = AddCode(native_module.get(), index++, page() - 4 * kCodeAlignment);
CHECK_NOT_NULL(code);
CHECK_EQ(0, manager.remaining_uncommitted());
code = AddCode(native_module.get(), index++, 1 * kCodeAlignment);
CHECK_NULL(code);
CHECK_EQ(0, manager.remaining_uncommitted());
native_module.reset();
CHECK_EQ(1 * page(), manager.remaining_uncommitted());
}
}
TEST_F(WasmCodeManagerTest, TotalLimitIrrespectiveOfModuleCount) {
for (auto style : styles()) {
WasmCodeManager manager(v8_isolate(), 1 * page());
NativeModulePtr nm1 = AllocModule(&manager, 1 * page(), style);
NativeModulePtr nm2 = AllocModule(&manager, 1 * page(), style);
CHECK(nm1);
CHECK(nm2);
WasmCode* code = AddCode(nm1.get(), 0, 1 * page());
CHECK_NOT_NULL(code);
code = AddCode(nm2.get(), 0, 1 * page());
CHECK_NULL(code);
}
}
TEST_F(WasmCodeManagerTest, DifferentHeapsApplyLimitsIndependently) {
for (auto style : styles()) {
WasmCodeManager manager1(v8_isolate(), 1 * page());
WasmCodeManager manager2(v8_isolate(), 2 * page());
NativeModulePtr nm1 = AllocModule(&manager1, 1 * page(), style);
NativeModulePtr nm2 = AllocModule(&manager2, 1 * page(), style);
CHECK(nm1);
CHECK(nm2);
WasmCode* code = AddCode(nm1.get(), 0, 1 * page());
CHECK_NOT_NULL(code);
CHECK_EQ(0, manager1.remaining_uncommitted());
code = AddCode(nm2.get(), 0, 1 * page());
CHECK_NOT_NULL(code);
}
}
TEST_F(WasmCodeManagerTest, GrowingVsFixedModule) {
for (auto style : styles()) {
WasmCodeManager manager(v8_isolate(), 3 * page());
NativeModulePtr nm = AllocModule(&manager, 1 * page(), style);
WasmCode* code = AddCode(nm.get(), 0, 1 * page() + kCodeAlignment);
if (style == Fixed) {
CHECK_NULL(code);
CHECK_EQ(manager.remaining_uncommitted(), 3 * page());
} else {
CHECK_NOT_NULL(code);
CHECK_EQ(manager.remaining_uncommitted(), 1 * page());
}
}
}
TEST_F(WasmCodeManagerTest, CommitIncrements) {
for (auto style : styles()) {
WasmCodeManager manager(v8_isolate(), 10 * page());
NativeModulePtr nm = AllocModule(&manager, 3 * page(), style);
WasmCode* code = AddCode(nm.get(), 0, kCodeAlignment);
CHECK_NOT_NULL(code);
CHECK_EQ(manager.remaining_uncommitted(), 9 * page());
code = AddCode(nm.get(), 1, 2 * page());
CHECK_NOT_NULL(code);
CHECK_EQ(manager.remaining_uncommitted(), 7 * page());
code = AddCode(nm.get(), 2, page() - kCodeAlignment);
CHECK_NOT_NULL(code);
CHECK_EQ(manager.remaining_uncommitted(), 7 * page());
}
}
TEST_F(WasmCodeManagerTest, Lookup) {
for (auto style : styles()) {
WasmCodeManager manager(v8_isolate(), 2 * page());
NativeModulePtr nm1 = AllocModule(&manager, 1 * page(), style);
NativeModulePtr nm2 = AllocModule(&manager, 1 * page(), style);
WasmCode* code1_0 = AddCode(nm1.get(), 0, kCodeAlignment);
CHECK_EQ(nm1.get(), code1_0->native_module());
2017-11-20 21:34:04 +00:00
WasmCode* code1_1 = AddCode(nm1.get(), 1, kCodeAlignment);
WasmCode* code2_0 = AddCode(nm2.get(), 0, kCodeAlignment);
WasmCode* code2_1 = AddCode(nm2.get(), 1, kCodeAlignment);
CHECK_EQ(nm2.get(), code2_1->native_module());
2017-11-20 21:34:04 +00:00
CHECK_EQ(0, code1_0->index());
CHECK_EQ(1, code1_1->index());
CHECK_EQ(0, code2_0->index());
CHECK_EQ(1, code2_1->index());
// we know the manager object is allocated here, so we shouldn't
// find any WasmCode* associated with that ptr.
WasmCode* not_found =
manager.LookupCode(reinterpret_cast<Address>(&manager));
CHECK_NULL(not_found);
WasmCode* found = manager.LookupCode(code1_0->instructions().start());
CHECK_EQ(found, code1_0);
found = manager.LookupCode(code2_1->instructions().start() +
(code2_1->instructions().size() / 2));
CHECK_EQ(found, code2_1);
found = manager.LookupCode(code2_1->instructions().start() +
code2_1->instructions().size() - 1);
CHECK_EQ(found, code2_1);
found = manager.LookupCode(code2_1->instructions().start() +
code2_1->instructions().size());
CHECK_NULL(found);
Address mid_code1_1 =
code1_1->instructions().start() + (code1_1->instructions().size() / 2);
CHECK_EQ(code1_1, manager.LookupCode(mid_code1_1));
nm1.reset();
CHECK_NULL(manager.LookupCode(mid_code1_1));
}
}
TEST_F(WasmCodeManagerTest, MultiManagerLookup) {
for (auto style : styles()) {
WasmCodeManager manager1(v8_isolate(), 2 * page());
WasmCodeManager manager2(v8_isolate(), 2 * page());
NativeModulePtr nm1 = AllocModule(&manager1, 1 * page(), style);
NativeModulePtr nm2 = AllocModule(&manager2, 1 * page(), style);
WasmCode* code1_0 = AddCode(nm1.get(), 0, kCodeAlignment);
CHECK_EQ(nm1.get(), code1_0->native_module());
2017-11-20 21:34:04 +00:00
WasmCode* code1_1 = AddCode(nm1.get(), 1, kCodeAlignment);
WasmCode* code2_0 = AddCode(nm2.get(), 0, kCodeAlignment);
WasmCode* code2_1 = AddCode(nm2.get(), 1, kCodeAlignment);
CHECK_EQ(nm2.get(), code2_1->native_module());
2017-11-20 21:34:04 +00:00
CHECK_EQ(0, code1_0->index());
CHECK_EQ(1, code1_1->index());
CHECK_EQ(0, code2_0->index());
CHECK_EQ(1, code2_1->index());
CHECK_EQ(code1_0, manager1.LookupCode(code1_0->instructions().start()));
CHECK_NULL(manager2.LookupCode(code1_0->instructions().start()));
}
}
TEST_F(WasmCodeManagerTest, LookupWorksAfterRewrite) {
for (auto style : styles()) {
WasmCodeManager manager(v8_isolate(), 2 * page());
NativeModulePtr nm1 = AllocModule(&manager, 1 * page(), style);
WasmCode* code0 = AddCode(nm1.get(), 0, kCodeAlignment);
WasmCode* code1 = AddCode(nm1.get(), 1, kCodeAlignment);
CHECK_EQ(0, code0->index());
CHECK_EQ(1, code1->index());
CHECK_EQ(code1, manager.LookupCode(code1->instructions().start()));
WasmCode* code1_1 = AddCode(nm1.get(), 1, kCodeAlignment);
CHECK_EQ(1, code1_1->index());
CHECK_EQ(code1, manager.LookupCode(code1->instructions().start()));
CHECK_EQ(code1_1, manager.LookupCode(code1_1->instructions().start()));
}
}
} // namespace wasm_heap_unittest
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