similar to: XRay: Demo on x86_64/Linux almost done; some questions.

Thanks Nirav, I can confirm that this works when I do the compile with llc, but then when linking to an executable with clang (patched with http://reviews.llvm.org/D20352 and compiler-rt patched with http://reviews.llvm.org/D21612) on Linux, I'm getting something different. Here's a sample of the transcript, and what I'm seeing: --->8 clang invocation 8<--- [16-06-23 3:33:42]

Peter suggested just writing out '.byte 0xeb, 0x09' and that allowed the jump instruction to bypass the relaxation, so that fixes my immediate problem. The question still stands though whether it should be possible to do through the instruction builder interface. Cheers On Wed, Jun 22, 2016 at 10:40 AM Dean Michael Berris <dberris at google.com> wrote: > Thanks Nirav, > >

Hello, I am initializing FDR mode and finalizing/flushing the buffers manually. XRay does not log calls from the main thread unless there is a function call after __xray_log_finalize(). This behavior is abnormal since one would expect the trace file to contain all function calls made up to the point when __xray_log_finalize() is called. To demonstrate this behavior, I have taken the test case

Hi there, I have a problem using the function call tracing tools that is designed in llvm tools set. My aim is to record every function call that a program makes when it run. However, for whatever reason, a simple matrix multiply c program that I wrote cannot record all the function calls that happened when the program run. Here is the program: matrix.c #include <stdio.h> void

Hi all, I was interested in trying out the Xray profiling component in LLVM and was trying to compile a C program with Clang with Xray flags. Unfortunately, I keep getting a linker error everytime I try to compile the C program with Clang. However, the error goes away when I try to compile the same C program with Clang++ with -xc++ flag. Searching around for this issue, looks like somebody has

Hi cfe-dev@ and llvm-dev@, I've attached a Chrome trace-viewer readable trace of a clang built with XRay instrumentation (additional build/link flags="-fxray-instrument -fxray-instruction-threshold=75") with the latest "top-of-trunk" version of clang with the recent updates to the XRay runtime supporting record filtering. To load the attached file, go to

Hi Dean, Do you have any estimates on when https://reviews.llvm.org/D21982 will reach mainline? (As I understood it's not yet there, looking at http://llvm.org/svn/llvm-project/compiler-rt/trunk ). I would like to test ARM port of XRay, so ready logging would be handful. Thanks, Serge On 8 August 2016 at 17:41, Dean Michael Berris <dean.berris at gmail.com> wrote: > > > On 8

> On 30 Jul 2016, at 05:07, Serge Rogatch via llvm-dev <llvm-dev at lists.llvm.org> wrote: > > Thanks for pointing this out, Tim. Then maybe this approach is not the best choice for x86, though ideally measuring is needed, it is just that on ARM the current x86 approach is not applicable because ARM doesn't have a single return instruction (such as RETQ on x86_64), furthermore,

> On 4 Aug 2016, at 06:27, Serge Rogatch <serge.rogatch at gmail.com> wrote: > > Hi Dean, > > I have a question about the following piece of code in compiler-rt/trunk/lib/xray/xray_trampoline_x86.S : > movq _ZN6__xray19XRayPatchedFunctionE(%rip), %rax > testq %rax, %rax > je .Ltmp0 > > // assume that %r10d has the function id. > movl %r10d,

Sent from my iPhone > On 6 Dec 2017, at 4:17 am, Brian Cain <brian.cain at gmail.com> wrote: > > > >> On Tue, Dec 5, 2017 at 9:49 AM, Dean Michael Berris via cfe-dev <cfe-dev at lists.llvm.org> wrote: >> Hi cfe-dev@ and llvm-dev@, >> >> I've attached a Chrome trace-viewer readable trace of a clang built with XRay instrumentation (additional

On 29 July 2016 at 10:43, Dean Michael Berris <dean.berris at gmail.com> wrote: > > > On 29 Jul 2016, at 09:14, Serge Rogatch via llvm-dev < > llvm-dev at lists.llvm.org> wrote: > > > > Hello, > > > > Can I ask you why you chose to patch both function entrances and exits, > rather than just patching the entrances and (in the patches) pushing on

Hi llvm-dev, I'm trying to reproduce an msan failure in one of the bots, but I can't seem to get the right incantation of building LLVM with msan. Here's what I've been doing: 1) Build the toolchain in one build directory, including `compiler-rt`. 2) Build the toolchain again with the just built toolchain in step 1, but this time with `-DLLVM_USE_SANITIZER=MemoryWithOrigins`. I

Another option is just to run corresponding script from *https://llvm.org/svn/llvm-project/zorg/trunk/zorg/buildbot/builders/sanitizers/ <https://llvm.org/svn/llvm-project/zorg/trunk/zorg/buildbot/builders/sanitizers/>* in empty directory. On Thu, Aug 30, 2018 at 5:00 AM Peter Smith via llvm-dev < llvm-dev at lists.llvm.org> wrote: > Hello Dean, > > I've not done this

Hi everyone, TL;DR: With current pending patches applied in compiler-rt and llvm, and trunk clang, you can build your application with XRay tracing enabled on Linux with tracing enabled before main starts, and logging stops when the main thread exits. Just a quick update, I have some patches under review that when applied cleanly to LLVM and compiler-rt allows for building applications with XRay

Aside: would it be useful to execute a build of the libc++/libc++abi with msan normally during release, and change the driver to look for these msan-built C++ libs when "-fsanitize=memory"? That would drastically cut down on the complexity of using msan. On Fri, Aug 31, 2018 at 5:43 AM Dean Michael Berris via llvm-dev < llvm-dev at lists.llvm.org> wrote: > Thanks Vitaly and

Hi Dean, I have a question for 32-bit platforms. I see in the code that you used the following in compiler-rt/trunk/lib/xray/xray_interface_internal.h : struct XRaySledEntry { uint64_t Address; uint64_t Function; unsigned char Kind; unsigned char AlwaysInstrument; unsigned char Padding[14]; // Need 32 bytes }; And the peer code in llvm/trunk/lib/Target/X86/X86MCInstLower.cpp : void

> On 29 Jul 2016, at 09:14, Serge Rogatch via llvm-dev <llvm-dev at lists.llvm.org> wrote: > > Hello, > > Can I ask you why you chose to patch both function entrances and exits, rather than just patching the entrances and (in the patches) pushing on the stack the address of __xray_FunctionExit , so that the user function returns normally (with RETQ or POP RIP or whatever

Thanks for pointing this out, Tim. Then maybe this approach is not the best choice for x86, though ideally measuring is needed, it is just that on ARM the current x86 approach is not applicable because ARM doesn't have a single return instruction (such as RETQ on x86_64), furthermore, the return instructions on ARM can be conditional. I have another question: what happens if the instrumented

Hmm. Odd. I just rebuilt from scratch and it seems to work with the test/CodeGen/X86/xray-attribute-instrumentation.ll test case outputing straight to obj: llc -filetype=obj -o ~/a.o -mtriple=x86_64-apple-macosx < test/CodeGen/X86/xray-attribute-instrumentation.ll What test case are you using? In any case, the issue appears to be that llvm doesn't realize that the target address is

TL;DR: At Google we use a call tracing system called XRay which inserts patchable instrumentation points into function entries and exits. If the community is interested, we'd like to contribute this system to the LLVM project. Many more details are contained in the whitepaper at: https://storage.googleapis.com/xray-downloads/whitepaper/XRayAFunctionCallTracingSystem.pdf Who's