Hi!

On 2022-09-01T12:05:23+0200, Martin Liška <mliska@suse.cz> wrote:
> Patch can bootstrap on x86_64-linux-gnu and survives regression tests.
>
> I've also built all cross compilers.

First: thanks for that: clean up plus "built all cross compilers"!

But yet, I've now tracked down an issue related to these changes,
apparently only visible via the nvptx back end -- and quite
non-obvious...  ;-)

> --- a/gcc/config/nvptx/nvptx.cc
> +++ b/gcc/config/nvptx/nvptx.cc
> @@ -52,7 +52,6 @@
>  #include "tm-preds.h"
>  #include "tm-constrs.h"
>  #include "langhooks.h"
> -#include "dbxout.h"
>  #include "cfgrtl.h"
>  #include "gimple.h"
>  #include "stor-layout.h"

> --- a/gcc/dbxout.cc
> +++ /dev/null
> @@ -1,3936 +0,0 @@
> -/* Output dbx-format symbol table information from GNU compiler.

The "dbx-format symbol table information" stuff indeed is not relevant
anymore, but:

> -/* Record an element in the table of global destructors.  SYMBOL is
> -   a SYMBOL_REF of the function to be called; PRIORITY is a number
> -   between 0 and MAX_INIT_PRIORITY.  */
> -
> -void
> -default_stabs_asm_out_destructor (rtx symbol ATTRIBUTE_UNUSED,
> -                               int priority ATTRIBUTE_UNUSED)
> -{
> -#if defined DBX_DEBUGGING_INFO || defined XCOFF_DEBUGGING_INFO
> -  /* Tell GNU LD that this is part of the static destructor set.
> -     This will work for any system that uses stabs, most usefully
> -     aout systems.  */
> -  dbxout_begin_simple_stabs ("___DTOR_LIST__", 22 /* N_SETT */);
> -  dbxout_stab_value_label (XSTR (symbol, 0));
> -#else
> -  sorry ("global destructors not supported on this target");
> -#endif
> -}
> -
> -/* Likewise for global constructors.  */
> -
> -void
> -default_stabs_asm_out_constructor (rtx symbol ATTRIBUTE_UNUSED,
> -                                int priority ATTRIBUTE_UNUSED)
> -{
> -#if defined DBX_DEBUGGING_INFO || defined XCOFF_DEBUGGING_INFO
> -  /* Tell GNU LD that this is part of the static destructor set.
> -     This will work for any system that uses stabs, most usefully
> -     aout systems.  */
> -  dbxout_begin_simple_stabs ("___CTOR_LIST__", 22 /* N_SETT */);
> -  dbxout_stab_value_label (XSTR (symbol, 0));
> -#else
> -  sorry ("global constructors not supported on this target");
> -#endif
> -}

> --- a/gcc/dbxout.h
> +++ /dev/null
> @@ -1,60 +0,0 @@
> -/* dbxout.h - Various declarations for functions found in dbxout.cc

> -extern void default_stabs_asm_out_destructor (rtx, int);
> -extern void default_stabs_asm_out_constructor (rtx, int);

... these two functions, 'default_stabs_asm_out_constructor',
'default_stabs_asm_out_destructor' (specifically, now their 'sorry'
branches only) used to serve as default 'TARGET_ASM_CONSTRUCTOR',
'TARGET_ASM_DESTRUCTOR' via...

> --- a/gcc/target-def.h
> +++ b/gcc/target-def.h

|  #if !defined(TARGET_ASM_CONSTRUCTOR) && !defined(USE_COLLECT2)
|  # ifdef CTORS_SECTION_ASM_OP
|  #  define TARGET_ASM_CONSTRUCTOR default_ctor_section_asm_out_constructor
>  # else
>  #  ifdef TARGET_ASM_NAMED_SECTION
>  #   define TARGET_ASM_CONSTRUCTOR default_named_section_asm_out_constructor
> -#  else
> -#   define TARGET_ASM_CONSTRUCTOR default_stabs_asm_out_constructor
>  #  endif
>  # endif
>  #endif
> @@ -74,8 +72,6 @@
|  #if !defined(TARGET_ASM_DESTRUCTOR) && !defined(USE_COLLECT2)
|  # ifdef DTORS_SECTION_ASM_OP
|  #  define TARGET_ASM_DESTRUCTOR default_dtor_section_asm_out_destructor
>  # else
>  #  ifdef TARGET_ASM_NAMED_SECTION
>  #   define TARGET_ASM_DESTRUCTOR default_named_section_asm_out_destructor
> -#  else
> -#   define TARGET_ASM_DESTRUCTOR default_stabs_asm_out_destructor
>  #  endif
>  # endif
>  #endif

... this setup here (manually added some more context to the 'diff').

That is, if a back end was not 'USE_COLLECT2', nor manually defined
'TARGET_ASM_CONSTRUCTOR', 'TARGET_ASM_DESTRUCTOR', or got pointed to the
respective 'default_[...]'  functions due to 'CTORS_SECTION_ASM_OP',
'DTORS_SECTION_ASM_OP', or 'TARGET_ASM_NAMED_SECTION', it got pointed to
'default_stabs_asm_out_constructor', 'default_stabs_asm_out_destructor'.
These would emit 'sorry' for any global constructor/destructor they're
run into.

This is now gone, and thus in such a back end configuration case
'TARGET_ASM_CONSTRUCTOR', 'TARGET_ASM_DESTRUCTOR' don't get defined
anymore, and thus the subsequently following:

    #if !defined(TARGET_HAVE_CTORS_DTORS)
    # if defined(TARGET_ASM_CONSTRUCTOR) && defined(TARGET_ASM_DESTRUCTOR)
    # define TARGET_HAVE_CTORS_DTORS true
    # endif
    #endif

... doesn't define 'TARGET_HAVE_CTORS_DTORS' anymore, and thus per my
understanding, 'gcc/final.cc:rest_of_handle_final':

    if (DECL_STATIC_CONSTRUCTOR (current_function_decl)
        && targetm.have_ctors_dtors)
      targetm.asm_out.constructor (XEXP (DECL_RTL (current_function_decl), 0),
                                   decl_init_priority_lookup
                                     (current_function_decl));
    if (DECL_STATIC_DESTRUCTOR (current_function_decl)
        && targetm.have_ctors_dtors)
      targetm.asm_out.destructor (XEXP (DECL_RTL (current_function_decl), 0),
                                  decl_fini_priority_lookup
                                    (current_function_decl));

... simply does nothing anymore for a 'DECL_STATIC_CONSTRUCTOR',
'DECL_STATIC_DESTRUCTOR'.

This, effectively, means that GCC/nvptx now suddenly appears to "support"
global constructors/destructors, which means that a ton of test cases now
erroneously PASS that previously used to FAIL:

    sorry, unimplemented: global constructors not supported on this target

Of course, such support didn't magically happen due to
"STABS: remove -gstabs and -gxcoff functionality", so this is bad.  And,
corresponding execution testing then regularly FAILs (due to the global
constructor/destructor functions never being invoked), for example:

    [-UNSUPPORTED:-]{+PASS:+} gcc.dg/initpri1.c {+(test for excess errors)+}
    {+FAIL: gcc.dg/initpri1.c execution test+}

    [-UNSUPPORTED:-]{+PASS:+} g++.dg/special/conpr-1.C {+(test for excess errors)+}
    {+FAIL: g++.dg/special/conpr-1.C execution test+}

To restore the previous GCC/nvptx behavior, I seek permission to first
push the attached
"Restore default 'sorry' 'TARGET_ASM_CONSTRUCTOR', 'TARGET_ASM_DESTRUCTOR'".
For traceability, this simply restores the previous code, stripped down
to the bare minimum.

As a next step, I'd then prepare a patch to re-work this thing, to make
it more obvious what's actually happening there.  (Details to be done.
For example, do we want to continue to have 'sorry' functions called via
'gcc/target-def.h', as done previously?  Or, should that be an '#error'
case there, and the nvptx back end provides its own 'sorry' functions for
'TARGET_ASM_CONSTRUCTOR', 'TARGET_ASM_DESTRUCTOR'?  Or, emit 'sorry' in
'gcc/final.cc:rest_of_handle_final'?  Or, something else?)


In addition to 'gcc/config/nvptx/nvptx.cc' (as quoted above) three more
files did '#include "dbxout.h"', and thus potentially could use these
'sorry' default functions.

Per my testing of '--target=pdp11-aout' as well as
'--target=powerpc64-darwin', re-introducing '#include "dbxout.h"' is not
relevant for 'gcc/config/pdp11/pdp11.cc' as well as for
'gcc/config/rs6000/rs6000-call.cc' and
'gcc/config/rs6000/rs6000-logue.cc' (for 'TARGET_MACHO' only).

Target pdp11-aout is not affected, because of 'gcc/config.gcc' setting:

    # On a.out targets, we need to use collect2.
    case ${target} in
    *-*-*aout*)
            use_collect2=yes
            ;;
    esac

Target powerpc64-darwin is not affected, because it defines
'TARGET_ASM_NAMED_SECTION'.

I'll thus remove those changes/TODO markers that are still present in the
attached
"Restore default 'sorry' 'TARGET_ASM_CONSTRUCTOR', 'TARGET_ASM_DESTRUCTOR'".
With that, OK to push?


Grüße
 Thomas


-----------------
Siemens Electronic Design Automation GmbH; Anschrift: Arnulfstraße 201, 80634 München; Gesellschaft mit beschränkter Haftung; Geschäftsführer: Thomas Heurung, Frank Thürauf; Sitz der Gesellschaft: München; Registergericht München, HRB 106955