[llvm-bugs] [Bug 26590] New: Tailcall forwards arguments with mismatching zero/signext parameter attributes

via llvm-bugs llvm-bugs at lists.llvm.org
Thu Feb 11 18:18:20 PST 2016


https://llvm.org/bugs/show_bug.cgi?id=26590

            Bug ID: 26590
           Summary: Tailcall forwards arguments with mismatching
                    zero/signext parameter attributes
           Product: libraries
           Version: trunk
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Backend: X86
          Assignee: unassignedbugs at nondot.org
          Reporter: hans at chromium.org
                CC: llvm-bugs at lists.llvm.org, rnk at google.com
    Classification: Unclassified

For example:

target datalayout = "e-m:e-p:32:32-f64:32:64-f80:32-n8:16:32-S128"
target triple = "i386-unknown-linux-gnu"

declare void @g(i16 signext)
define void @f(i16 zeroext %x) {
entry:
  tail call void @g(i16 signext %x)
  ret void
}


llc will generate:

f:                                      # @f
    .cfi_startproc
# BB#0:                                 # %entry
    jmp    g                       # TAILCALL


Which means g()'s parameter doesn't get sign-extended to 32 bits as expected.


The problem is that MatchingStackOffset in X86ISelLowering.cpp only checks that
the arguments are at the same stack offset and have the same size. It should
also check the parameter attributes, but that is tricky since it has no
straight-forward way of getting at the caller's arguments.

I'm looking into this, but thought I'd file it in the meantime.

-- 
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/20160212/eb7853b8/attachment-0001.html>


More information about the llvm-bugs mailing list