blob: d908d17fe971b3c0c03dbf4145d0468ffafe90c7 [file] [log] [blame]
/*
* Copyright (C) 2018 The Android Open Source Project
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "src/trace_processor/importers/systrace/systrace_trace_parser.h"
#include "perfetto/base/logging.h"
#include "perfetto/ext/base/string_splitter.h"
#include "perfetto/ext/base/string_utils.h"
#include "src/trace_processor/forwarding_trace_parser.h"
#include "src/trace_processor/importers/common/process_tracker.h"
#include "src/trace_processor/sorter/trace_sorter.h"
#include <cctype>
#include <cinttypes>
#include <string>
#include <unordered_map>
namespace perfetto {
namespace trace_processor {
namespace {
std::vector<base::StringView> SplitOnSpaces(base::StringView str) {
std::vector<base::StringView> result;
for (size_t i = 0; i < str.size(); ++i) {
// Consume all spaces.
for (; i < str.size() && str.data()[i] == ' '; ++i)
;
// If we haven't reached the end consume all non-spaces and add result.
if (i != str.size()) {
size_t start = i;
for (; i < str.size() && str.data()[i] != ' '; ++i)
;
result.push_back(base::StringView(str.data() + start, i - start));
}
}
return result;
}
bool IsProcessDumpShortHeader(const std::vector<base::StringView>& tokens) {
return tokens.size() == 4 && tokens[0] == "USER" && tokens[1] == "PID" &&
tokens[2] == "TID" && tokens[3] == "CMD";
}
bool IsProcessDumpLongHeader(const std::vector<base::StringView>& tokens) {
return tokens.size() > 4 && tokens[0] == "USER" && tokens[1] == "PID" &&
tokens[2] == "PPID" && tokens[3] == "VSZ";
}
} // namespace
SystraceTraceParser::SystraceTraceParser(TraceProcessorContext* ctx)
: line_parser_(ctx), ctx_(ctx) {}
SystraceTraceParser::~SystraceTraceParser() = default;
util::Status SystraceTraceParser::Parse(TraceBlobView blob) {
if (state_ == ParseState::kEndOfSystrace)
return util::OkStatus();
partial_buf_.insert(partial_buf_.end(), blob.data(),
blob.data() + blob.size());
if (state_ == ParseState::kBeforeParse) {
// Remove anything before the TRACE:\n marker, which is emitted when
// obtaining traces via `adb shell "atrace -t 1 sched" > out.txt`.
std::array<uint8_t, 7> kAtraceMarker = {'T', 'R', 'A', 'C', 'E', ':', '\n'};
auto search_end = partial_buf_.begin() +
static_cast<int>(std::min(partial_buf_.size(),
kGuessTraceMaxLookahead));
auto it = std::search(partial_buf_.begin(), search_end,
kAtraceMarker.begin(), kAtraceMarker.end());
if (it != search_end)
partial_buf_.erase(partial_buf_.begin(), it + kAtraceMarker.size());
// Deal with HTML traces.
state_ = partial_buf_[0] == '<' ? ParseState::kHtmlBeforeSystrace
: ParseState::kSystrace;
}
// There can be multiple trace data sections in an HTML trace, we want to
// ignore any that don't contain systrace data. In the future it would be
// good to also parse the process dump section.
const char kTraceDataSection[] =
R"(<script class="trace-data" type="application/text">)";
auto start_it = partial_buf_.begin();
for (;;) {
auto line_it = std::find(start_it, partial_buf_.end(), '\n');
if (line_it == partial_buf_.end())
break;
std::string buffer(start_it, line_it);
if (state_ == ParseState::kHtmlBeforeSystrace) {
if (base::Contains(buffer, kTraceDataSection)) {
state_ = ParseState::kTraceDataSection;
}
} else if (state_ == ParseState::kTraceDataSection) {
if (base::StartsWith(buffer, "#") && base::Contains(buffer, "TASK-PID")) {
state_ = ParseState::kSystrace;
} else if (base::StartsWith(buffer, "PROCESS DUMP")) {
state_ = ParseState::kProcessDumpLong;
} else if (base::StartsWith(buffer, "CGROUP DUMP")) {
state_ = ParseState::kCgroupDump;
} else if (base::Contains(buffer, R"(</script>)")) {
state_ = ParseState::kHtmlBeforeSystrace;
}
} else if (state_ == ParseState::kSystrace) {
if (base::Contains(buffer, R"(</script>)")) {
state_ = ParseState::kEndOfSystrace;
break;
} else if (!base::StartsWith(buffer, "#") && !buffer.empty()) {
SystraceLine line;
util::Status status = line_tokenizer_.Tokenize(buffer, &line);
if (status.ok()) {
line_parser_.ParseLine(std::move(line));
} else {
ctx_->storage->IncrementStats(stats::systrace_parse_failure);
}
}
} else if (state_ == ParseState::kProcessDumpLong ||
state_ == ParseState::kProcessDumpShort) {
if (base::Contains(buffer, R"(</script>)")) {
state_ = ParseState::kHtmlBeforeSystrace;
} else {
std::vector<base::StringView> tokens =
SplitOnSpaces(base::StringView(buffer));
if (IsProcessDumpShortHeader(tokens)) {
state_ = ParseState::kProcessDumpShort;
} else if (IsProcessDumpLongHeader(tokens)) {
state_ = ParseState::kProcessDumpLong;
} else if (state_ == ParseState::kProcessDumpLong &&
tokens.size() >= 10) {
// Format is:
// user pid ppid vsz rss wchan pc s name my cmd line
const std::optional<uint32_t> pid =
base::StringToUInt32(tokens[1].ToStdString());
const std::optional<uint32_t> ppid =
base::StringToUInt32(tokens[2].ToStdString());
base::StringView name = tokens[8];
// Command line may contain spaces, merge all remaining tokens:
const char* cmd_start = tokens[9].data();
base::StringView cmd(
cmd_start,
static_cast<size_t>((buffer.data() + buffer.size()) - cmd_start));
if (!pid || !ppid) {
PERFETTO_ELOG("Could not parse line '%s'", buffer.c_str());
return util::ErrStatus("Could not parse PROCESS DUMP line");
}
ctx_->process_tracker->SetProcessMetadata(pid.value(), ppid, name,
base::StringView());
} else if (state_ == ParseState::kProcessDumpShort &&
tokens.size() >= 4) {
// Format is:
// username pid tid my cmd line
const std::optional<uint32_t> tgid =
base::StringToUInt32(tokens[1].ToStdString());
const std::optional<uint32_t> tid =
base::StringToUInt32(tokens[2].ToStdString());
// Command line may contain spaces, merge all remaining tokens:
const char* cmd_start = tokens[3].data();
base::StringView cmd(
cmd_start,
static_cast<size_t>((buffer.data() + buffer.size()) - cmd_start));
StringId cmd_id =
ctx_->storage->mutable_string_pool()->InternString(cmd);
if (!tid || !tgid) {
PERFETTO_ELOG("Could not parse line '%s'", buffer.c_str());
return util::ErrStatus("Could not parse PROCESS DUMP line");
}
UniqueTid utid =
ctx_->process_tracker->UpdateThread(tid.value(), tgid.value());
ctx_->process_tracker->UpdateThreadNameByUtid(
utid, cmd_id, ThreadNamePriority::kOther);
}
}
} else if (state_ == ParseState::kCgroupDump) {
if (base::Contains(buffer, R"(</script>)")) {
state_ = ParseState::kHtmlBeforeSystrace;
}
// TODO(lalitm): see if it is important to parse this.
}
start_it = line_it + 1;
}
if (state_ == ParseState::kEndOfSystrace) {
partial_buf_.clear();
} else {
partial_buf_.erase(partial_buf_.begin(), start_it);
}
return util::OkStatus();
}
void SystraceTraceParser::NotifyEndOfFile() {}
} // namespace trace_processor
} // namespace perfetto