Hi all,

I request you to please review the source modification done to support split DWARF.

    Existing DWZ when deals with executable having split DWARF information
    in separate file (DW_AT_GNU_dwo_name/DW_AT_dwo_name), it ignores the
    separate file and only attempts to optimize main executable which has
    very less information as almost all the debug information is present
    in sepaterate file.

    Now DWZ is modified to extract the separate file name from
    atributes DW_AT_GNU_dwo_name/DW_AT_dwo_name in main executable and
    calls function DWZ on separate file as well.

    While processing DWO file, information is read from corresponding ".dwo"
    sections in place of original one. This is assisted by macros
    DEFINE_DEBUG (defining section enum) and DEFINE_SECTION (defining section
    names along with other fields) in a way that very next section to the
    original one is the ".dwo" variant of it. When we process DWO file we
    access section using macro "INDEX(DEBUG_SECTION)", which is defined as
    "(DEBUG_SECTION + is_dwo_file)". This way we avoid any duplicacy and
    use existing code.

Regards,
Alok