src/storage/blobfs/test/unit/blob-loader-test.cc

// Copyright 2020 The Fuchsia 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 "src/storage/blobfs/blob-loader.h"

#include <lib/fzl/owned-vmo-mapper.h>
#include <lib/sync/completion.h>
#include <lib/zx/status.h>
#include <lib/zx/vmo.h>
#include <zircon/assert.h>
#include <zircon/errors.h>

#include <set>

#include <block-client/cpp/fake-device.h>
#include <digest/digest.h>
#include <digest/merkle-tree.h>
#include <digest/node-digest.h>
#include <fbl/auto_call.h>
#include <gtest/gtest.h>

#include "src/storage/blobfs/blob-layout.h"
#include "src/storage/blobfs/blob.h"
#include "src/storage/blobfs/blobfs.h"
#include "src/storage/blobfs/common.h"
#include "src/storage/blobfs/compression-settings.h"
#include "src/storage/blobfs/format.h"
#include "src/storage/blobfs/mkfs.h"
#include "src/storage/blobfs/test/blob_utils.h"
#include "src/storage/blobfs/test/unit/utils.h"

namespace blobfs {
namespace {

using ::testing::Combine;
using ::testing::TestParamInfo;
using ::testing::TestWithParam;
using ::testing::Values;
using ::testing::ValuesIn;

using block_client::FakeBlockDevice;

constexpr uint32_t kBlockSize = 512;
constexpr uint32_t kNumBlocks = 400 * kBlobfsBlockSize / kBlockSize;

using TestParamType = std::tuple<CompressionAlgorithm, BlobLayoutFormat>;

class BlobLoaderTest : public TestWithParam<TestParamType> {
 public:
  void SetUp() override {
    CompressionAlgorithm compression_algorithm;
    std::tie(compression_algorithm, blob_layout_format_) = GetParam();
    srand(testing::UnitTest::GetInstance()->random_seed());

    auto device = std::make_unique<FakeBlockDevice>(kNumBlocks, kBlockSize);
    ASSERT_TRUE(device);
    FilesystemOptions options{
        .blob_layout_format = blob_layout_format_,
    };
    switch (compression_algorithm) {
      case CompressionAlgorithm::UNCOMPRESSED:
      case CompressionAlgorithm::CHUNKED:
        break;
      case CompressionAlgorithm::ZSTD:
      case CompressionAlgorithm::ZSTD_SEEKABLE:
      case CompressionAlgorithm::LZ4:
        options.oldest_revision = kBlobfsRevisionBackupSuperblock;
        break;
    }
    ASSERT_EQ(FormatFilesystem(device.get(), options), ZX_OK);
    loop_.StartThread();

    options_ = {.compression_settings = {
                    .compression_algorithm = compression_algorithm,
                }};
    ASSERT_EQ(Blobfs::Create(loop_.dispatcher(), std::move(device), options_, zx::resource(), &fs_),
              ZX_OK);

    // Pre-seed with some random blobs.
    for (unsigned i = 0; i < 3; i++) {
      AddBlob(1024);
    }
    ASSERT_EQ(Remount(), ZX_OK);
  }

  // Remounts the filesystem, which ensures writes are flushed and caches are wiped.
  zx_status_t Remount() {
    return Blobfs::Create(loop_.dispatcher(), Blobfs::Destroy(std::move(fs_)), options_,
                          zx::resource(), &fs_);
  }

  // AddBlob creates and writes a blob of a specified size to the file system.
  // The contents of the blob are compressible at a realistic level for a typical ELF binary.
  // The returned BlobInfo describes the created blob, but its lifetime is unrelated to the lifetime
  // of the on-disk blob.
  [[maybe_unused]] std::unique_ptr<BlobInfo> AddBlob(size_t sz) {
    fbl::RefPtr<fs::Vnode> root;
    EXPECT_EQ(fs_->OpenRootNode(&root), ZX_OK);
    fs::Vnode* root_node = root.get();

    std::unique_ptr<BlobInfo> info = GenerateRealisticBlob("", sz);
    memmove(info->path, info->path + 1, strlen(info->path));  // Remove leading slash.

    fbl::RefPtr<fs::Vnode> file;
    EXPECT_EQ(root_node->Create(info->path, 0, &file), ZX_OK);

    size_t actual;
    EXPECT_EQ(file->Truncate(info->size_data), ZX_OK);
    EXPECT_EQ(file->Write(info->data.get(), info->size_data, 0, &actual), ZX_OK);
    EXPECT_EQ(actual, info->size_data);
    EXPECT_EQ(file->Close(), ZX_OK);

    return info;
  }

  BlobLoader& loader() { return fs_->loader(); }

  Blobfs* Fs() const { return fs_.get(); }

  CompressionAlgorithm ExpectedAlgorithm() const {
    return options_.compression_settings.compression_algorithm;
  }

  uint32_t LookupInode(const BlobInfo& info) {
    Digest digest;
    fbl::RefPtr<CacheNode> node;
    EXPECT_EQ(digest.Parse(info.path), ZX_OK);
    EXPECT_EQ(fs_->Cache().Lookup(digest, &node), ZX_OK);
    auto vnode = fbl::RefPtr<Blob>::Downcast(std::move(node));
    return vnode->Ino();
  }

  CompressionAlgorithm LookupCompression(const BlobInfo& info) {
    Digest digest;
    fbl::RefPtr<CacheNode> node;
    EXPECT_EQ(digest.Parse(info.path), ZX_OK);
    EXPECT_EQ(fs_->Cache().Lookup(digest, &node), ZX_OK);
    auto vnode = fbl::RefPtr<Blob>::Downcast(std::move(node));
    auto algorithm_or = AlgorithmForInode(vnode->GetNode());
    EXPECT_TRUE(algorithm_or.is_ok());
    return algorithm_or.value();
  }

  void CheckMerkleTreeContents(const fzl::OwnedVmoMapper& merkle, const BlobInfo& info) {
    std::unique_ptr<MerkleTreeInfo> merkle_tree = CreateMerkleTree(
        info.data.get(), info.size_data, ShouldUseCompactMerkleTreeFormat(blob_layout_format_));
    ASSERT_TRUE(merkle.vmo().is_valid());
    ASSERT_GE(merkle.size(), merkle_tree->merkle_tree_size);
    switch (blob_layout_format_) {
      case BlobLayoutFormat::kPaddedMerkleTreeAtStart:
        // In the padded layout the Merkle starts at the start of the vmo.
        EXPECT_EQ(
            memcmp(merkle.start(), merkle_tree->merkle_tree.get(), merkle_tree->merkle_tree_size),
            0);
        break;
      case BlobLayoutFormat::kCompactMerkleTreeAtEnd:
        // In the compact layout the Merkle tree is aligned to end at the end of the vmo.
        EXPECT_EQ(memcmp(static_cast<const uint8_t*>(merkle.start()) +
                             (merkle.size() - merkle_tree->merkle_tree_size),
                         merkle_tree->merkle_tree.get(), merkle_tree->merkle_tree_size),
                  0);
        break;
    }
  }

 protected:
  std::unique_ptr<Blobfs> fs_;
  async::Loop loop_{&kAsyncLoopConfigNoAttachToCurrentThread};
  MountOptions options_;
  BlobLayoutFormat blob_layout_format_;
};

// A separate parameterized test fixture that will only be run with compression algorithms that
// support paging.
using BlobLoaderPagedTest = BlobLoaderTest;

TEST_P(BlobLoaderTest, NullBlob) {
  size_t blob_len = 0;
  std::unique_ptr<BlobInfo> info = AddBlob(blob_len);
  ASSERT_EQ(Remount(), ZX_OK);

  fzl::OwnedVmoMapper data, merkle;
  ASSERT_EQ(loader().LoadBlob(LookupInode(*info), nullptr, &data, &merkle), ZX_OK);

  EXPECT_FALSE(data.vmo().is_valid());
  EXPECT_EQ(data.size(), 0ul);

  EXPECT_FALSE(merkle.vmo().is_valid());
  EXPECT_EQ(merkle.size(), 0ul);
}

TEST_P(BlobLoaderTest, SmallBlob) {
  size_t blob_len = 1024;
  std::unique_ptr<BlobInfo> info = AddBlob(blob_len);
  ASSERT_EQ(Remount(), ZX_OK);
  // We explicitly don't check the compression algorithm was respected here, since files this small
  // don't need to be compressed.

  fzl::OwnedVmoMapper data, merkle;
  ASSERT_EQ(loader().LoadBlob(LookupInode(*info), nullptr, &data, &merkle), ZX_OK);

  ASSERT_TRUE(data.vmo().is_valid());
  ASSERT_GE(data.size(), info->size_data);
  EXPECT_EQ(memcmp(data.start(), info->data.get(), info->size_data), 0);

  EXPECT_FALSE(merkle.vmo().is_valid());
  EXPECT_EQ(merkle.size(), 0ul);
}

TEST_P(BlobLoaderPagedTest, SmallBlob) {
  size_t blob_len = 1024;
  std::unique_ptr<BlobInfo> info = AddBlob(blob_len);
  ASSERT_EQ(Remount(), ZX_OK);
  // We explicitly don't check the compression algorithm was respected here, since files this small
  // don't need to be compressed.

  fzl::OwnedVmoMapper data, merkle;
  std::unique_ptr<pager::PageWatcher> page_watcher;
  ASSERT_EQ(loader().LoadBlobPaged(LookupInode(*info), nullptr, &page_watcher, &data, &merkle),
            ZX_OK);

  ASSERT_TRUE(data.vmo().is_valid());
  ASSERT_GE(data.size(), info->size_data);
  // Use vmo::read instead of direct read so that we can synchronously fail if the pager fails.
  fbl::Array<uint8_t> buf(new uint8_t[blob_len], blob_len);
  ASSERT_EQ(data.vmo().read(buf.get(), 0, blob_len), ZX_OK);
  EXPECT_EQ(memcmp(buf.get(), info->data.get(), info->size_data), 0);

  EXPECT_FALSE(merkle.vmo().is_valid());
  EXPECT_EQ(merkle.size(), 0ul);
}

TEST_P(BlobLoaderTest, LargeBlob) {
  size_t blob_len = 1 << 18;
  std::unique_ptr<BlobInfo> info = AddBlob(blob_len);
  ASSERT_EQ(Remount(), ZX_OK);
  ASSERT_EQ(LookupCompression(*info), ExpectedAlgorithm());

  fzl::OwnedVmoMapper data, merkle;
  ASSERT_EQ(loader().LoadBlob(LookupInode(*info), nullptr, &data, &merkle), ZX_OK);

  ASSERT_TRUE(data.vmo().is_valid());
  ASSERT_GE(data.size(), info->size_data);
  EXPECT_EQ(memcmp(data.start(), info->data.get(), info->size_data), 0);

  CheckMerkleTreeContents(merkle, *info);
}

TEST_P(BlobLoaderTest, LargeBlobWithNonAlignedLength) {
  size_t blob_len = (1 << 18) - 1;
  std::unique_ptr<BlobInfo> info = AddBlob(blob_len);
  ASSERT_EQ(Remount(), ZX_OK);
  ASSERT_EQ(LookupCompression(*info), ExpectedAlgorithm());

  fzl::OwnedVmoMapper data, merkle;
  ASSERT_EQ(loader().LoadBlob(LookupInode(*info), nullptr, &data, &merkle), ZX_OK);

  ASSERT_TRUE(data.vmo().is_valid());
  ASSERT_GE(data.size(), info->size_data);
  EXPECT_EQ(memcmp(data.start(), info->data.get(), info->size_data), 0);

  CheckMerkleTreeContents(merkle, *info);
}

TEST_P(BlobLoaderPagedTest, LargeBlob) {
  size_t blob_len = 1 << 18;
  std::unique_ptr<BlobInfo> info = AddBlob(blob_len);
  ASSERT_EQ(Remount(), ZX_OK);
  ASSERT_EQ(LookupCompression(*info), ExpectedAlgorithm());

  fzl::OwnedVmoMapper data, merkle;
  std::unique_ptr<pager::PageWatcher> page_watcher;
  ASSERT_EQ(loader().LoadBlobPaged(LookupInode(*info), nullptr, &page_watcher, &data, &merkle),
            ZX_OK);

  ASSERT_TRUE(data.vmo().is_valid());
  ASSERT_GE(data.size(), info->size_data);
  // Use vmo::read instead of direct read so that we can synchronously fail if the pager fails.
  fbl::Array<uint8_t> buf(new uint8_t[blob_len], blob_len);
  ASSERT_EQ(data.vmo().read(buf.get(), 0, blob_len), ZX_OK);
  EXPECT_EQ(memcmp(buf.get(), info->data.get(), info->size_data), 0);

  CheckMerkleTreeContents(merkle, *info);
}

TEST_P(BlobLoaderPagedTest, LargeBlobWithNonAlignedLength) {
  size_t blob_len = (1 << 18) - 1;
  std::unique_ptr<BlobInfo> info = AddBlob(blob_len);
  ASSERT_EQ(Remount(), ZX_OK);
  ASSERT_EQ(LookupCompression(*info), ExpectedAlgorithm());

  fzl::OwnedVmoMapper data, merkle;
  std::unique_ptr<pager::PageWatcher> page_watcher;
  ASSERT_EQ(loader().LoadBlobPaged(LookupInode(*info), nullptr, &page_watcher, &data, &merkle),
            ZX_OK);

  ASSERT_TRUE(data.vmo().is_valid());
  ASSERT_GE(data.size(), info->size_data);
  // Use vmo::read instead of direct read so that we can synchronously fail if the pager fails.
  fbl::Array<uint8_t> buf(new uint8_t[blob_len], blob_len);
  ASSERT_EQ(data.vmo().read(buf.get(), 0, blob_len), ZX_OK);
  EXPECT_EQ(memcmp(buf.get(), info->data.get(), info->size_data), 0);

  CheckMerkleTreeContents(merkle, *info);
}

TEST_P(BlobLoaderTest, MediumBlobWithRoomForMerkleTree) {
  // In the compact layout the Merkle tree can fit perfectly into the room leftover at the end of
  // the data.
  ASSERT_EQ(fs_->Info().block_size, digest::kDefaultNodeSize);
  size_t blob_len = (digest::kDefaultNodeSize - digest::kSha256Length) * 3;
  std::unique_ptr<BlobInfo> info = AddBlob(blob_len);
  ASSERT_EQ(Remount(), ZX_OK);

  fzl::OwnedVmoMapper data, merkle;
  ASSERT_EQ(loader().LoadBlob(LookupInode(*info), nullptr, &data, &merkle), ZX_OK);

  ASSERT_TRUE(data.vmo().is_valid());
  ASSERT_GE(data.size(), info->size_data);
  EXPECT_EQ(memcmp(data.start(), info->data.get(), info->size_data), 0);

  CheckMerkleTreeContents(merkle, *info);
}

TEST_P(BlobLoaderTest, NullBlobWithCorruptedMerkleRootFailsToLoad) {
  size_t blob_len = 0;
  std::unique_ptr<BlobInfo> info = AddBlob(blob_len);
  uint32_t inode_index = LookupInode(*info);

  // Verify the null blob can be read back.
  fzl::OwnedVmoMapper data, merkle;
  ASSERT_EQ(loader().LoadBlob(inode_index, nullptr, &data, &merkle), ZX_OK);

  uint8_t corrupt_merkle_root[digest::kSha256Length] = "-corrupt-null-blob-merkle-root-";
  {
    // Corrupt the null blob's merkle root.
    // |inode| holds a pointer into |fs_| and needs to be destroyed before remounting.
    auto inode = fs_->GetNode(inode_index);
    memcpy(inode->merkle_root_hash, corrupt_merkle_root, sizeof(corrupt_merkle_root));
    BlobTransaction transaction;
    uint64_t block = (inode_index * kBlobfsInodeSize) / kBlobfsBlockSize;
    transaction.AddOperation({.vmo = zx::unowned_vmo(fs_->GetAllocator()->GetNodeMapVmo().get()),
                              .op = {
                                  .type = storage::OperationType::kWrite,
                                  .vmo_offset = block,
                                  .dev_offset = NodeMapStartBlock(fs_->Info()) + block,
                                  .length = 1,
                              }});
    transaction.Commit(*fs_->journal());
  }

  // Remount the filesystem so the node cache will pickup the new name for the blob.
  ASSERT_EQ(Remount(), ZX_OK);

  // Verify the empty blob can be found by the corrupt name.
  BlobInfo corrupt_info;
  Digest corrupt_digest(corrupt_merkle_root);
  snprintf(corrupt_info.path, sizeof(info->path), "%s", corrupt_digest.ToString().c_str());
  EXPECT_EQ(LookupInode(corrupt_info), inode_index);

  // Verify the null blob with a corrupted Merkle root fails to load.
  ASSERT_EQ(loader().LoadBlob(inode_index, nullptr, &data, &merkle), ZX_ERR_IO_DATA_INTEGRITY);
}

TEST_P(BlobLoaderTest, LoadBlobWithAnInvalidNodeIndexIsAnError) {
  uint32_t invalid_node_index = kMaxNodeId - 1;
  fzl::OwnedVmoMapper data, merkle;
  EXPECT_EQ(loader().LoadBlob(invalid_node_index, nullptr, &data, &merkle), ZX_ERR_INVALID_ARGS);
}

TEST_P(BlobLoaderPagedTest, LoadBlobPagedWithAnInvalidNodeIndexIsAnError) {
  uint32_t invalid_node_index = kMaxNodeId - 1;
  fzl::OwnedVmoMapper data, merkle;
  std::unique_ptr<pager::PageWatcher> page_watcher;
  EXPECT_EQ(loader().LoadBlobPaged(invalid_node_index, nullptr, &page_watcher, &data, &merkle),
            ZX_ERR_INVALID_ARGS);
}

TEST_P(BlobLoaderTest, LoadBlobWithACorruptNextNodeIndexIsAnError) {
  std::unique_ptr<BlobInfo> info = AddBlob(1 << 14);
  ASSERT_EQ(Remount(), ZX_OK);

  // Corrupt the next node index of the inode.
  uint32_t invalid_node_index = kMaxNodeId - 1;
  uint32_t node_index = LookupInode(*info);
  auto inode = fs_->GetAllocator()->GetNode(node_index);
  ASSERT_TRUE(inode.is_ok());
  inode->header.next_node = invalid_node_index;
  inode->extent_count = 2;

  fzl::OwnedVmoMapper data, merkle;
  std::unique_ptr<pager::PageWatcher> page_watcher;
  EXPECT_EQ(loader().LoadBlobPaged(node_index, nullptr, &page_watcher, &data, &merkle),
            ZX_ERR_IO_DATA_INTEGRITY);
}

std::string GetTestParamName(const TestParamInfo<TestParamType>& param) {
  auto [compression_algorithm, blob_layout_format] = param.param;
  return GetBlobLayoutFormatNameForTests(blob_layout_format) +
         GetCompressionAlgorithmName(compression_algorithm);
}

constexpr std::array<CompressionAlgorithm, 4> kCompressionAlgorithms = {
    CompressionAlgorithm::UNCOMPRESSED,
    CompressionAlgorithm::ZSTD,
    CompressionAlgorithm::ZSTD_SEEKABLE,
    CompressionAlgorithm::CHUNKED,
};

constexpr std::array<CompressionAlgorithm, 2> kPagingCompressionAlgorithms = {
    CompressionAlgorithm::UNCOMPRESSED,
    CompressionAlgorithm::CHUNKED,
};

constexpr std::array<BlobLayoutFormat, 2> kBlobLayoutFormats = {
    BlobLayoutFormat::kPaddedMerkleTreeAtStart,
    BlobLayoutFormat::kCompactMerkleTreeAtEnd,
};

INSTANTIATE_TEST_SUITE_P(OldFormat, BlobLoaderTest,
                         Combine(ValuesIn(kCompressionAlgorithms),
                                 Values(BlobLayoutFormat::kPaddedMerkleTreeAtStart)),
                         GetTestParamName);

INSTANTIATE_TEST_SUITE_P(/*no prefix*/, BlobLoaderTest,
                         Combine(ValuesIn(kPagingCompressionAlgorithms),
                                 Values(BlobLayoutFormat::kCompactMerkleTreeAtEnd)),
                         GetTestParamName);

INSTANTIATE_TEST_SUITE_P(/*no prefix*/, BlobLoaderPagedTest,
                         Combine(ValuesIn(kPagingCompressionAlgorithms),
                                 ValuesIn(kBlobLayoutFormats)),
                         GetTestParamName);

}  // namespace
}  // namespace blobfs