Merge branch 'master' into drcachesim_external_simulators

api/docs/release.dox

@@ -157,6 +157,7 @@ Further non-compatibility-affecting changes include:
  - Added several routines to the #dynamorio::drmemtrace::memtrace_stream_t interface
    for drmemtrace analysis tools: get_output_cpuid(), get_workload_id(),
    get_input_id(), get_input_interface().
+ - Added -record_syscall to drmemtrace for recording syscall parameters.
  - Added opportunity to run multiple drcachesim analysis tools simultaneously.
  - Added support of loading separately-built analysis tools to drcachesim dynamically.
 

clients/drcachesim/common/options.cpp

@@ -748,6 +748,24 @@ droption_t<bool> op_record_replace_retaddr(
     "replacement, which has lower overhead, but runs the risk of breaking an "
     "application that examines or changes its own return addresses in the recorded "
     "functions.");
+droption_t<std::string> op_record_syscall(
+    DROPTION_SCOPE_CLIENT, "record_syscall", DROPTION_FLAG_ACCUMULATE,
+    OP_RECORD_FUNC_ITEM_SEP, "", "Record parameters for the specified syscall number(s).",
+    "Record the parameters and success of the specified system call number(s)."
+    " The option value should fit this format:"
+    " sycsall_number|parameter_number"
+    " E.g., -record_syscall \"2|2\" will record SYS_open's 2 parameters and whether"
+    " successful (1 for success or 0 for failure, in a function return value record)"
+    " for x86 Linux.  SYS_futex is recorded by default on Linux and this option's value"
+    " adds to futex rather than replacing it (setting futex to 0 parameters disables)."
+    " The trace identifies which syscall owns each set of parameter and return value"
+    " records via a numeric ID equal to the syscall number + TRACE_FUNC_ID_SYSCALL_BASE."
+    " Recording multiple syscalls can be achieved by using the separator"
+    " \"" OP_RECORD_FUNC_ITEM_SEP
+    "\" (e.g., -record_syscall \"202|6" OP_RECORD_FUNC_ITEM_SEP "3|1\"), or"
+    " specifying multiple -record_syscall options."
+    " It is up to the user to ensure the values are correct; a too-large parameter"
+    " count may cause tracing to fail with an error mid-run.");
 droption_t<unsigned int> op_miss_count_threshold(
     DROPTION_SCOPE_FRONTEND, "miss_count_threshold", 50000,
     "For cache miss analysis: minimum LLC miss count for a load to be eligible for "

clients/drcachesim/common/options.h

@@ -57,6 +57,10 @@
 #define CACHE_TYPE_DATA "data"
 #define CACHE_TYPE_UNIFIED "unified"
 #define CACHE_PARENT_MEMORY "memory"
+// The expected pattern for a single_op_value is:
+//     function_name|function_id|arguments_num
+// where function_name can contain spaces (for instance, C++ namespace prefix)
+#define PATTERN_SEPARATOR "|"
 
 #ifdef HAS_ZIP
 #    define DEFAULT_TRACE_COMPRESSION_TYPE "zip"
@@ -169,6 +173,7 @@ extern dynamorio::droption::droption_t<bool> op_record_heap;
 extern dynamorio::droption::droption_t<std::string> op_record_heap_value;
 extern dynamorio::droption::droption_t<bool> op_record_dynsym_only;
 extern dynamorio::droption::droption_t<bool> op_record_replace_retaddr;
+extern dynamorio::droption::droption_t<std::string> op_record_syscall;
 extern dynamorio::droption::droption_t<unsigned int> op_miss_count_threshold;
 extern dynamorio::droption::droption_t<double> op_miss_frac_threshold;
 extern dynamorio::droption::droption_t<double> op_confidence_threshold;

clients/drcachesim/common/utils.h

@@ -39,6 +39,7 @@
 #include <iomanip>
 #include <sstream>
 #include <string>
+#include <vector>
 
 namespace dynamorio {
 namespace drmemtrace {
@@ -166,6 +167,21 @@ starts_with(const std::string &str, const std::string &with)
     return pos == 0;
 }
 
+static inline std::vector<std::string>
+split_by(std::string s, const std::string &sep)
+{
+    size_t pos;
+    std::vector<std::string> vec;
+    if (s.empty())
+        return vec;
+    do {
+        pos = s.find(sep);
+        vec.push_back(s.substr(0, pos));
+        s.erase(0, pos + sep.length());
+    } while (pos != std::string::npos);
+    return vec;
+}
+
 } // namespace drmemtrace
 } // namespace dynamorio
 

clients/drcachesim/docs/drcachesim.dox.in

@@ -1482,6 +1482,11 @@ The -record_heap parameter requests recording of a pre-determined set
 of functions related to heap allocation.  The -record_heap_value
 paramter controls the contents of this set.
 
+The tracer also supports recording system call argument and success
+values via the option -record_syscall, which functions similarly to
+-record_function with the system call number replacing the function
+name.
+
 ****************************************************************************
 \page sec_drcachesim_newtool Creating New Analysis Tools
 

clients/drcachesim/tests/offline-allasm-record-syscall.templatex

@@ -0,0 +1,24 @@
+Adios world!
+Adios world!
+Adios world!
+Adios world!
+Adios world!
+Adios world!
+Adios world!
+Adios world!
+Adios world!
+Adios world!
+.*
+          43          20:   .* ifetch       2 byte\(s\) @ 0x.* 0f 05                syscall
+          44          20:   .* <marker: timestamp .*>
+          45          20:   .* <marker: tid .* on core .*>
+          46          20:   .* <marker: system call 1>
+          47          20:   .* <marker: maybe-blocking system call>
+          48          20:   .* <marker: function==syscall #1>
+          49          20:   .* <marker: function argument 0x2>
+          50          20:   .* <marker: function argument 0x.*>
+          51          20:   .* <marker: function argument 0xd>
+          52          20:   .* <marker: function==syscall #1>
+          53          20:   .* <marker: function return value 0x1>
+          54          20:   .* <marker: timestamp .*>
+.*