[LLVMbugs] [Bug 14954] New: problem checking for invalid attribute parameters

bugzilla-daemon at llvm.org bugzilla-daemon at llvm.org
Mon Jan 14 13:29:38 PST 2013


http://llvm.org/bugs/show_bug.cgi?id=14954

             Bug #: 14954
           Summary: problem checking for invalid attribute parameters
           Product: clang
           Version: unspecified
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: Frontend
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: rkotler at mips.com
                CC: llvmbugs at cs.uiuc.edu
    Classification: Unclassified


It seems that almost uniformly, checking for the presence of unwanted attribute
parameters is done incorrectly.

For example, if you compile a program with the following for an x86 target,
there will be no warnings.

void  __attribute__((force_align_arg_pointer(foo))) p();

The reason for this is that the error condition is being checked as follows:

   if (Attr.getNumArgs() != 0) {
     S.Diag(Attr.getLoc(), diag::err_attribute_wrong_number_arguments)
<< 0;
     return;
   } 

Whereas the check should be 


if (Attr.hasParameterOrArguments()) {


An attribute with a single parameter (unless it is a number like for aligned)
is treated as a parameter and not part of a list of arguments.
So when there is a single parameter, #args is still 0.

/// AttributeList - Represents GCC's __attribute__ declaration. There are
/// 4 forms of this construct...they are:
///
/// 1: __attribute__(( const )). ParmName/Args/NumArgs will all be unused.
/// 2: __attribute__(( mode(byte) )). ParmName used, Args/NumArgs unused.
/// 3: __attribute__(( format(printf, 1, 2) )). ParmName/Args/NumArgs all used.
/// 4: __attribute__(( aligned(16) )). ParmName is unused, Args/Num used.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.



More information about the llvm-bugs mailing list