public inbox for gcc@gcc.gnu.org
 help / color / mirror / Atom feed
From: "Ulrich Weigand" <Ulrich.Weigand@de.ibm.com>
To: gcc@gcc.gnu.org
Subject: DATA_ALIGNMENT vs. DECL_USER_ALIGNMENT
Date: Tue, 08 Apr 2003 17:47:00 -0000	[thread overview]
Message-ID: <OF725729DE.A7D59996-ONC1256D02.005653C1@de.ibm.com> (raw)

Hello,

the target macro DATA_ALIGNMENT is supposed to be able to override the
alignment used for global data variables.  We are using this on s390x
to make sure all data objects are at least 2-byte aligned.  This is not
an optimization, but essential for correctness, as we cannot load the
address of a label that is not 2-byte aligned.

Unfortunately, DATA_ALIGNMENT is simply ignored when outputting a
variable that has the DECL_USER_ALIGNMENT flag set.  While this does
not usually occur with C code unless using source code like

  char x __attribute__ ((__aligned__(1)));

this flag is apparently much more frequently set by the Ada frontend.
In fact, this is currently causing bootstrap errors for Ada on s390x
because an elaboration check flag variable is not correctly aligned.


My question is now, why is DATA_ALIGNMENT ignored in such cases?
Is this a bug or a feature?  If the latter, what other options does
a backend have to absolutely enforce a minimum alignment for global
variables?


Mit freundlichen Gruessen / Best Regards

Ulrich Weigand

--
  Dr. Ulrich Weigand
  Linux for S/390 Design & Development
  IBM Deutschland Entwicklung GmbH, Schoenaicher Str. 220, 71032 Boeblingen
  Phone: +49-7031/16-3727   ---   Email: Ulrich.Weigand@de.ibm.com

             reply	other threads:[~2003-04-08 15:50 UTC|newest]

Thread overview: 101+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2003-04-08 17:47 Ulrich Weigand [this message]
2003-04-09 23:31 Ulrich Weigand
2003-04-10  2:27 ` Richard Henderson
2003-04-10  3:10   ` Ulrich Weigand
2003-04-10 16:20     ` Ulrich Weigand
2003-04-10 16:39       ` Geoff Keating
2003-04-10 16:44         ` Ulrich Weigand
2003-04-10 18:06       ` Richard Henderson
2003-04-10  5:33 Richard Kenner
2003-04-10  7:20 ` Ulrich Weigand
2003-04-10 19:20 Richard Kenner
2003-04-10 20:13 Richard Kenner
2003-04-10 20:28 ` Ulrich Weigand
2003-04-10 20:39 Richard Kenner
2003-04-10 21:00 ` Geoff Keating
2003-04-10 21:01 ` Ulrich Weigand
2003-04-10 21:44 ` Richard Henderson
2003-04-10 20:52 Robert Dewar
2003-04-10 20:57 Robert Dewar
2003-04-10 21:07 ` Ulrich Weigand
2003-04-10 21:04 Richard Kenner
2003-04-10 21:41 ` Geoff Keating
2003-04-10 21:10 Richard Kenner
2003-04-10 21:10 Robert Dewar
2003-04-10 21:17 ` David Edelsohn
2003-04-10 21:29 Robert Dewar
2003-04-10 21:31 Richard Kenner
2003-04-10 21:32 Robert Dewar
2003-04-10 21:32 ` David Edelsohn
2003-04-10 21:44 Richard Kenner
2003-04-10 21:45 Richard Kenner
2003-04-10 21:52 ` David Edelsohn
2003-04-10 21:57 ` Richard Henderson
2003-04-10 21:56 Richard Kenner
2003-04-10 22:02 ` David Edelsohn
2003-04-10 22:02 Richard Kenner
2003-04-10 22:24 ` Richard Henderson
2003-04-11 11:20   ` Laurent Guerby
2003-04-11 11:57     ` Arnaud Charlet
2003-04-11 14:16       ` Laurent Guerby
2003-04-10 22:05 Robert Dewar
2003-04-10 22:07 Richard Kenner
2003-04-10 22:23 ` David Edelsohn
2003-04-10 22:16 Richard Kenner
2003-04-10 22:23 Richard Kenner
2003-04-10 22:43 Richard Kenner
2003-04-11  0:31 ` Richard Henderson
2003-04-11  2:19   ` David Edelsohn
2003-04-11  0:45 Ulrich Weigand
2003-04-11  4:45 Richard Kenner
2003-04-17  5:39 ` Alexandre Oliva
2003-04-11 14:52 Robert Dewar
2003-04-11 16:51 Robert Dewar
2003-04-11 16:59 Robert Dewar
2003-04-11 18:49 ` Laurent Guerby
2003-04-11 22:26 Robert Dewar
2003-04-17  8:12 Richard Kenner
2003-04-17  8:38 ` Alexandre Oliva
2003-04-17 10:44 Robert Dewar
2003-04-17 20:30 ` Alexandre Oliva
2003-04-17 11:56 Richard Kenner
2003-04-17 20:35 ` Alexandre Oliva
2003-04-17 22:41   ` Geert Bosch
2003-04-17 23:19     ` Alexandre Oliva
2003-04-17 21:40 Richard Kenner
2003-04-17 21:41 Richard Kenner
2003-04-17 23:20 ` Alexandre Oliva
2003-04-18  1:16 ` Richard Henderson
2003-04-18  8:06 Richard Kenner
2003-04-18  8:59 ` Richard Henderson
2003-04-18 11:41 Richard Kenner
2003-04-20 21:19 Robert Dewar
2003-04-20 21:28 Robert Dewar
2003-04-21 17:05 ` Alexandre Oliva
2003-04-21 17:25 Richard Kenner
2003-04-21 18:14 ` Jamie Lokier
2003-04-22  9:00 ` Alexandre Oliva
2003-04-21 18:14 Richard Kenner
2003-04-21 20:33 ` Jamie Lokier
2003-04-21 20:45 Richard Kenner
2003-04-21 20:58 ` Jamie Lokier
2003-04-21 20:59 Richard Kenner
2003-04-21 21:14 ` Jamie Lokier
2003-04-21 21:25 Richard Kenner
2003-04-22  0:11 Robert Dewar
2003-04-22  1:35 ` Jamie Lokier
2003-04-22  0:15 Robert Dewar
2003-04-22  1:58 ` Jamie Lokier
2003-04-22 15:08 Richard Kenner
2003-04-23  1:07 ` Alexandre Oliva
2003-04-23  2:20 Richard Kenner
2003-04-23  5:31 ` Alexandre Oliva
2003-04-24  6:23   ` Jason Merrill
2003-04-24  8:43     ` Geert Bosch
2003-04-24 17:48       ` Jason Merrill
2003-04-24 18:04         ` Zack Weinberg
2003-04-24 19:31           ` Jason Merrill
2003-04-24 21:32         ` Geert Bosch
2003-04-24 23:24           ` Jason Merrill
2003-04-25  2:56     ` Richard Henderson
2003-04-23  3:29 Robert Dewar

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=OF725729DE.A7D59996-ONC1256D02.005653C1@de.ibm.com \
    --to=ulrich.weigand@de.ibm.com \
    --cc=gcc@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).