[llvm-bugs] [Bug 50564] New: atomic_signal_fence should not generate a dmb on aarch64

[via llvm-bugs]

https://bugs.llvm.org/show_bug.cgi?id=50564

            Bug ID: 50564
           Summary: atomic_signal_fence should not generate a dmb on
                    aarch64
           Product: new-bugs
           Version: 11.0
          Hardware: Macintosh
                OS: MacOS X
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: new bugs
          Assignee: unassignedbugs at nondot.org
          Reporter: saagar at saagarjha.com
                CC: htmldeveloper at gmail.com, llvm-bugs at lists.llvm.org

This is related to https://bugs.llvm.org/show_bug.cgi?id=27545, which is
fixed–but apparently only for optimized code. Without optimization I'm still
seeing a dmb ish for atomic_signal_fence on aarch64. The standard (7.17.4.2)
specifically mentions that "hardware fence instructions that
atomic_thread_fence would have inserted are not emitted", so this I think this
behavior is somewhat surprising if not incorrect. Here's a Godbolt link with
sample code compiled without optimization demonstrating this bug:
https://godbolt.org/z/qh55n184G

-- 
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/20210603/6a5e1ea5/attachment-0001.html>