Re: [PATCH 2/3] Check data is GC'ed

[Yao Qi <yao@codesourcery.com>]

Doug Evans <dje@google.com> writes:

>  >        if (pdi->d.locdesc
>  >  	  && addr == 0
>  > -	  && !dwarf2_per_objfile->has_section_at_zero)
>  > +	  /* When the address is 0, if the object file doesn't have
>  > +	     section at zero or the section at zero is executable,
>  > +	     we think address 0 means the corresponding variable is
>  > +	     removed by linker, instead of there is a data at address
>  > +	     0.  */
>  > +	  && (dwarf2_per_objfile->section_at_zero == NULL
>  > +	      || dwarf2_per_objfile->section_at_zero->flags & SEC_CODE))
>
> Could there be a DW_TAG_variable in SEC_CODE?
> Someone might put one there for a particular reason.

I don't figure out a reason to do so, but it is possible via:

__attribute__((section(".text"))) int foo = 1;

I can't think of other heuristics, so if GDB has a section at address
zero, GDB can't tell address zero in debug information means the
corresponding variable is GC'ed by linker or the variable address is
zero.

How about open a PR about GDB is unable to know whether a variable is
GC'ed by linker if there is a section at address zero? and kfail
dw2-var-zero-addr.exp conditionally?  Something like:

    if [is_address_zero_readable] {
	setup_kfail "gdb/PR" "*-*-*"
    }
    gdb_test "print &var" {No symbol "var" in current context\.}

What do you think?

-- 
Yao (齐尧)