[llvm-bugs] [Bug 40642] New: After 350404, clang drops volatile load

[via llvm-bugs]

https://bugs.llvm.org/show_bug.cgi?id=40642

            Bug ID: 40642
           Summary: After 350404, clang drops volatile load
           Product: clang
           Version: trunk
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: -New Bugs
          Assignee: unassignedclangbugs at nondot.org
          Reporter: hans at chromium.org
                CC: aaron at aaronballman.com, htmldeveloper at gmail.com,
                    llvm-bugs at lists.llvm.org, neeilans at live.com,
                    richard-llvm at metafoo.co.uk

Reproducer:

$ cat /tmp/c.cc
template <class T> struct S {
  void foo() {
    if (true)
      reinterpret_cast<const volatile unsigned char*>(m_ptr)[0];
  }
  int* m_ptr;
};

void f(S<int> *x) {
  x->foo();
}


With r350403:

$ /work/llvm.monorepo/build.release/bin/clang.good -cc1 -triple
x86_64-unknown-linux-gnu -S -emit-llvm /tmp/c.cc -o - | grep -A999 'define
linkonce_odr void @_ZN1SIiE3fooEv'
define linkonce_odr void @_ZN1SIiE3fooEv(%struct.S* %this) #0 comdat align 2 {
entry:
  %this.addr = alloca %struct.S*, align 8
  store %struct.S* %this, %struct.S** %this.addr, align 8
  %this1 = load %struct.S*, %struct.S** %this.addr, align 8
  %m_ptr = getelementptr inbounds %struct.S, %struct.S* %this1, i32 0, i32 0
  %0 = load i32*, i32** %m_ptr, align 8
  %1 = bitcast i32* %0 to i8*
  %arrayidx = getelementptr inbounds i8, i8* %1, i64 0
  %2 = load volatile i8, i8* %arrayidx, align 1
  ret void
}


With r350404:

$ /work/llvm.monorepo/build.release/bin/clang.bad -cc1 -triple
x86_64-unknown-linux-gnu -S -emit-llvm /tmp/c.cc -o - | grep -A999 'define
linkonce_odr void @_ZN1SIiE3fooEv'
define linkonce_odr void @_ZN1SIiE3fooEv(%struct.S* %this) #0 comdat align 2 {
entry:
  %this.addr = alloca %struct.S*, align 8
  store %struct.S* %this, %struct.S** %this.addr, align 8
  %this1 = load %struct.S*, %struct.S** %this.addr, align 8
  %m_ptr = getelementptr inbounds %struct.S, %struct.S* %this1, i32 0, i32 0
  %0 = load i32*, i32** %m_ptr, align 8
  %1 = bitcast i32* %0 to i8*
  %arrayidx = getelementptr inbounds i8, i8* %1, i64 0
  ret void
}

Note that "%2 = load volatile i8, i8* %arrayidx, align 1" is missing.



Adding a void cast, i.e. "(void)reinterpret_cast<const volatile unsigned
char*>(m_ptr)[0];" makes the load appear also with r350404.


The regression was caught by test failures in pdfium
(https://bugs.chromium.org/p/chromium/issues/detail?id=929475)

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20190207/ea461eb2/attachment.html>