[llvm-bugs] [Bug 27525] New: Possible bug in ARMAsmBackend::processFixupValue()?

[via llvm-bugs]

https://llvm.org/bugs/show_bug.cgi?id=27525

            Bug ID: 27525
           Summary: Possible bug in ARMAsmBackend::processFixupValue()?
           Product: new-bugs
           Version: trunk
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
          Assignee: unassignedbugs at nondot.org
          Reporter: daniel.sanders at imgtec.com
                CC: llvm-bugs at lists.llvm.org
    Classification: Unclassified

While looking into a MIPS issue with our fixup handling, I noticed that the
caller of processFixupValue() ignores any modifications to the 'Value' result.
The caller (MCAssembler::handleFixup()) makes it look like the modified value
is used but if you look inside ELFObjectWriter::recordRelocation() you'll see
that FixedValue is never read before it is overwritten. The same is true of
WinCOFFObjectWriter::recordRelocation(), and is true for most code paths in
ARMMachObjectWriter::recordRelocation().

ARM is currently the only target to modify the Value argument and does so for
certain fixups involving Thumb functions. Deleting the relevant code doesn't
cause any test failures so it seems that there are either no tests covering it
or something else is doing the same job.

Hope this helps.

-- 
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/20160426/428f4ad0/attachment.html>