[PATCH] D56244: [XRay][docs] XRay Framework Usage Guide

David Greene via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Mon Jan 7 07:58:11 PST 2019


greened added inline comments.


================
Comment at: llvm/docs/XRayFramework.rst:42
+implementations plugged in through the XRay framework as "modes" throughout
+this document. There are four modes that come with the LLVM XRay distribution:
+Basic, Flight Data Recorder (FDR), and Profiling.
----------------
Three modes?


================
Comment at: llvm/docs/XRayFramework.rst:48
+
+You can control the various modes programmatically using functions exposed in
+the ``xray_log_interface.h`` header. This header comes with the compiler-rt
----------------
Use "we" here for consistency with the following text?


================
Comment at: llvm/docs/XRayFramework.rst:84
+intercept function entry and exit events, along with other special "custom" and
+tyuped events that might be used. Because we've initialised the specific mode
+with a specific set of options, once the patched functions run, the handlers
----------------
"tyuped?"


================
Comment at: llvm/docs/XRayFramework.rst:330
+    // We specifically ignore the result of unpatching a specific function.
+    __xray_unpatch_function(function_d);
+
----------------
`function_id`?


================
Comment at: llvm/docs/XRayFramework.rst:396
+    if (finalized.exchange(1, std::memory_order_acq_rel) == 1)
+      XRayLogInitStatus::XRAY_LOG_FINALIZED;
+
----------------
What is this doing?


================
Comment at: llvm/docs/XRayFramework.rst:423
+    // pointer so that it can be treated as a global variable by the lambda we
+    // provide to the call to __xray_log_process_buffers(...).
+    static FILE* tmpf = 0;
----------------
I don't understand this comment.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D56244/new/

https://reviews.llvm.org/D56244





More information about the llvm-commits mailing list