public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
From: "redi at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug libstdc++/103626] _GLIBCXX_HOSTED should respect -ffreestanding
Date: Fri, 16 Sep 2022 20:06:26 +0000	[thread overview]
Message-ID: <bug-103626-4-96hubo2Zvo@http.gcc.gnu.org/bugzilla/> (raw)
In-Reply-To: <bug-103626-4@http.gcc.gnu.org/bugzilla/>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103626

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2022-09-16

--- Comment #1 from Jonathan Wakely <redi at gcc dot gnu.org> ---
(In reply to Jonathan Wakely from comment #0)
> Libstdc++ should not define _GLIBCXX_HOSTED to 1 when __STD_HOSTED__ == 0
> i.e. when -ffreestanding is used.

That should be as simple as:

> diff --git a/libstdc++-v3/acinclude.m4 b/libstdc++-v3/acinclude.m4
> index 9aa81e1ac13..1b404d7df22 100644
> --- a/libstdc++-v3/acinclude.m4
> +++ b/libstdc++-v3/acinclude.m4
> @@ -2992,7 +2992,7 @@ AC_DEFUN([GLIBCXX_ENABLE_HOSTED], [
>      fi
>    else
>      is_hosted=yes
> -    hosted_define=1
> +    hosted_define=__STDC_HOSTED__
>    fi
>    GLIBCXX_CONDITIONAL(GLIBCXX_HOSTED, test $is_hosted = yes)
>    AC_DEFINE_UNQUOTED(_GLIBCXX_HOSTED, $hosted_define,
> diff --git a/libstdc++-v3/include/Makefile.am b/libstdc++-v3/include/Makefile.am
> index 3eeb407a57f..3472e84f75c 100644
> --- a/libstdc++-v3/include/Makefile.am
> +++ b/libstdc++-v3/include/Makefile.am
> @@ -1326,7 +1326,7 @@ ${host_builddir}/c++config.h: ${CONFIG_HEADER} \
>       ${CONFIG_HEADER} > /dev/null 2>&1 \
>       && ldbl_alt128_compat='s,^#undef _GLIBCXX_LONG_DOUBLE_ALT128_COMPAT$$,#define _GLIBCXX_LONG_DOUBLE_ALT128_COMPAT 1,' ;\
>       verbose_assert='s,g,g,' ; \
> -     grep "^[         ]*#[    ]*define[       ][      ]*_GLIBCXX_HOSTED[      ][      ]*1[    ]*$$" \
> +     grep "^[         ]*#[    ]*define[       ][      ]*_GLIBCXX_HOSTED[      ][      ]*__STDC_HOSTED__[      ]*$$" \
>       ${CONFIG_HEADER} > /dev/null 2>&1 \
>       && grep "^[      ]*#[    ]*define[       ][      ]*_GLIBCXX_VERBOSE[     ][      ]*1[    ]*$$" \
>       ${CONFIG_HEADER} > /dev/null 2>&1 \


> The --disable-hosted-libstdcxx configure option will still allow
> installation of a library that can never be used as hosted, but it will also
> be possible to use an existing hosted implementation in freestanding mode.

This would mean that a full set of library headers would be available for
inclusion in programs compiled with -ffreestanding, so to prevent the
non-freestanding ones being included we should add to them:

#if !_GLIBCXX_HOSTED
# error "This header is not supported in freestanding mode"
#endif

Then we can teach the testsuite to recognise that error and skip a test as
UNSUPPORTED. That will allow the full testsuite to be run without changing
thousands of individual tests with something like "{
dg-require-effective-target hosted }". The tests that aren't valid for
freestanding will just be UNSUPPORTED when -ffreestanding is in the
runtestflags.

We might need to also strip -ffreestanding from the flags when building the
libtestc++.a support lib used by the testsuite, because that lib requires
hosted features.

This all seems doable though.

  reply	other threads:[~2022-09-16 20:06 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-12-09 11:05 [Bug libstdc++/103626] New: " redi at gcc dot gnu.org
2022-09-16 20:06 ` redi at gcc dot gnu.org [this message]
2022-09-17 19:37 ` [Bug libstdc++/103626] " arsen at aarsen dot me
2022-09-17 22:19 ` redi at gcc dot gnu.org
2022-09-18  1:40 ` arsen at aarsen dot me
2022-09-18  6:41 ` redi at gcc dot gnu.org
2022-09-18 12:36 ` arsen at aarsen dot me
2022-10-03 14:45 ` cvs-commit at gcc dot gnu.org
2022-10-03 14:45 ` cvs-commit at gcc dot gnu.org
2022-10-03 14:45 ` cvs-commit at gcc dot gnu.org
2022-10-03 14:46 ` cvs-commit at gcc dot gnu.org
2022-10-03 14:46 ` cvs-commit at gcc dot gnu.org
2022-10-03 14:46 ` cvs-commit at gcc dot gnu.org
2022-10-03 14:50 ` redi at gcc dot gnu.org

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=bug-103626-4-96hubo2Zvo@http.gcc.gnu.org/bugzilla/ \
    --to=gcc-bugzilla@gcc.gnu.org \
    --cc=gcc-bugs@gcc.gnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).