i#7180: Add drmemtrace support for stdin via "-infile -"

Adds support for reading a single drmemtrace file from stdin when "-infile -" is specified. The file can be uncompressed or in gzip format. Adds a UNIX-only test which was easiest to create directly using bash. Fixes #7180
DynamoRIO · Jan 10, 2025 · 6e3a93f · 6e3a93f
1 parent e034f37
commit 6e3a93f
Show file tree

Hide file tree

Showing 5 changed files with 52 additions and 4 deletions.
diff --git a/api/docs/release.dox b/api/docs/release.dox
@@ -133,6 +133,8 @@ Further non-compatibility-affecting changes include:
    a pure-static application via a new configure_DynamoRIO_static_client() CMake
    function and new "_drstatic" static library CMake targets for the provided
    extension libraries.
+ - Added support for reading a single drmemtrace trace file from stdin
+   via "-infile -".
 
 **************************************************
 <hr>

diff --git a/clients/drcachesim/common/options.cpp b/clients/drcachesim/common/options.cpp
@@ -1,5 +1,5 @@
 /* **********************************************************
- * Copyright (c) 2015-2024 Google, Inc.  All rights reserved.
+ * Copyright (c) 2015-2025 Google, Inc.  All rights reserved.
  * **********************************************************/
 
 /*
@@ -97,7 +97,9 @@ droption_t<std::string> op_indir(
 droption_t<std::string> op_infile(
     DROPTION_SCOPE_ALL, "infile", "", "Offline legacy file for input to the simulator",
     "Directs the simulator to use a single all-threads-interleaved-into-one trace file. "
-    "This is a legacy file format that is no longer produced.");
+    "This is a legacy file format that is no longer produced as it does not support "
+    "mixing multiple inputs nor using auxiliary metadata files.  Passing '-' will read "
+    "from stdin as plain or gzip-compressed data.");
 
 droption_t<int> op_jobs(
     DROPTION_SCOPE_ALL, "jobs", -1, "Number of parallel jobs",

diff --git a/clients/drcachesim/reader/compressed_file_reader.cpp b/clients/drcachesim/reader/compressed_file_reader.cpp
@@ -1,5 +1,5 @@
 /* **********************************************************
- * Copyright (c) 2017-2023 Google, Inc.  All rights reserved.
+ * Copyright (c) 2017-2025 Google, Inc.  All rights reserved.
  * **********************************************************/
 
 /*
@@ -52,7 +52,13 @@ namespace drmemtrace {
 bool
 open_single_file_common(const std::string &path, gzFile &out)
 {
-    out = gzopen(path.c_str(), "rb");
+    if (path == "-") {
+        // We assume stdin is 0.
+        // We do not use "stdin->_fileno" as that is not portable.
+        constexpr int STDIN_FD = 0;
+        out = gzdopen(STDIN_FD, "rb");
+    } else
+        out = gzopen(path.c_str(), "rb");
     return out != nullptr;
 }
 

diff --git a/clients/drcachesim/tests/offline-stdin.expect b/clients/drcachesim/tests/offline-stdin.expect
@@ -0,0 +1,16 @@
+Output format:
+<--record#-> <--instr#->: <---tid---> <record details>
+------------------------------------------------------------
+           1           0:     1257596 <marker: version 4>
+           2           0:     1257596 <marker: filetype 0x10e40>
+           3           0:     1257596 <marker: cache line size 64>
+           4           0:     1257596 <marker: chunk instruction count 10000000>
+           5           0:     1257596 <marker: page size 4096>
+           6           0:     1257596 <marker: timestamp 13331862027607247>
+           7           0:     1257596 <marker: tid 1257596 on core 8>
+           8           1:     1257596 ifetch       4 byte\(s\) @ 0x00007f3e429fce35 48 8b 45 f8          mov    -0x08\(%rbp\), %rax
+           9           1:     1257596 read         8 byte\(s\) @ 0x00007ffefb03e128 by PC 0x00007f3e429fce35
+          10           2:     1257596 ifetch       8 byte\(s\) @ 0x00007f3e429fce39 48 8d 14 c5 00 00 00 lea    0x00\(,%rax,8\), %rdx
+          10           2:     1257596                                             00
+View tool results:
+              2 : total instructions
diff --git a/suite/tests/CMakeLists.txt b/suite/tests/CMakeLists.txt
@@ -4006,6 +4006,28 @@ if (BUILD_CLIENTS)
           # Avoid the default core-sharded from re-scheduling the trace.
           "-indir ${core_sharded_dir} -tool schedule_stats -only_shards 2,3 -no_core_sharded" "")
         set(tool.core_on_disk_rawtemp ON) # no preprocessor
+
+        # Test reading a file over stdin.
+        # We rely on there being just one subfile in this zip file: gunzip only
+        # extracts the first one (bsdtar can extract all but is not likely installed).
+        # This is UNIX only so we rely on cat and gunzip being available.
+        find_program(CAT_EXE cat DOC "path to cat")
+        find_program(GUNZIP_EXE gunzip DOC "path to gunzip")
+        find_program(BASH_EXE bash DOC "path to bash")
+        if (CAT_EXE AND GUNZIP_EXE AND BASH_EXE)
+          # Easiest to make our pipeline in bash with a direct add_test()
+          # (although this bypasses the torun() features used by all other tests).
+          set(infile "${core_sharded_dir}/drmemtrace.core.0.trace.zip")
+          add_test(tool.drcacheoff.stdin
+            ${BASH_EXE} -c
+            "${CAT_EXE} ${infile} | ${GUNZIP_EXE} | ${drrun_path} -t drmemtrace -tool view -sim_refs 10 -infile -")
+          file(READ "${PROJECT_SOURCE_DIR}/clients/drcachesim/tests/offline-stdin.expect"
+            stdin_expect)
+          set_tests_properties(tool.drcacheoff.stdin PROPERTIES
+            PASS_REGULAR_EXPRESSION "${stdin_expect}")
+        else ()
+          message("Failed to find cat, gunzip, and bash: not running stdin test")
+        endif ()
       endif ()
     endif ()