[llvm-bugs] [Bug 40973] New: Incorrect debug values for parameters that are smaller than stack slots

[via llvm-bugs]

https://bugs.llvm.org/show_bug.cgi?id=40973

            Bug ID: 40973
           Summary: Incorrect debug values for parameters that are smaller
                    than stack slots
           Product: libraries
           Version: trunk
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: DebugInfo
          Assignee: unassignedbugs at nondot.org
          Reporter: david.stenberg at ericsson.com
                CC: jdevlieghere at apple.com, keith.walker at arm.com,
                    llvm-bugs at lists.llvm.org,
                    paul_robinson at playstation.sony.com

Spawned out from D58831.

When compiling the following test case:

    extern void bar(short);

    void foo(int p0, int p1, int p2, int p3, short p4, short p5) {
      bar(p4);
    }

using clang from trunk (r355379):

    $ clang --target=arm -mbig-endian -O2 -g

we land up with the following debug information for p4 and p5:

0x0000006d:     DW_TAG_formal_parameter
                  DW_AT_name    ("p4")
                  DW_AT_decl_file      
("/home/user/upstream/master/stack_dbg_values.c")
                  DW_AT_decl_line       (3)
                  DW_AT_type    (0x0000008e "short")

0x00000078:     DW_TAG_formal_parameter
                  DW_AT_location        (DW_OP_breg13 SP+4)
                  DW_AT_name    ("p5")
                  DW_AT_decl_file      
("/home/user/upstream/master/stack_dbg_values.c")
                  DW_AT_decl_line       (3)
                  DW_AT_type    (0x0000008e "short")

I think there are two issues here. For p4, we do not consider truncates when
searching for the FI in SelectionDAGBuilder::EmitFuncArgumentDbgValue(),
meaning that we are left with a $noreg DBG_VALUE, ultimately leading to no
location being emitted for the parameter. For p5, we say that it is located on
the stack at offset 4, but as we are on a big-endian target the two-byte short
parameter is located in the lower two bytes of the 4-byte stack slot, meaning
that the offset should really be 6.

Compare this with the output from GCC (5.4.1):

$ arm-none-eabi-gcc -O1 -g -mbig-endian stack_dbg_values.c -c

0x00000072:     DW_TAG_formal_parameter
                  DW_AT_name    ("p4")
                  DW_AT_decl_file      
("/home/user/upstream/master/stack_dbg_values.c")
                  DW_AT_decl_line       (3)
                  DW_AT_type    (0x000000a6 "short int")
                  DW_AT_location        (DW_OP_fbreg +2)

0x0000007f:     DW_TAG_formal_parameter
                  DW_AT_name    ("p5")
                  DW_AT_decl_file      
("/home/user/upstream/master/stack_dbg_values.c")
                  DW_AT_decl_line       (3)
                  DW_AT_type    (0x000000a6 "short int")
                  DW_AT_location        (DW_OP_fbreg +6)

-- 
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/20190306/304bd51c/attachment.html>