qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Peter Maydell <peter.maydell@linaro.org>
To: qemu-devel@nongnu.org
Cc: "Andreas Färber" <andreas.faerber@web.de>,
	"Markus Armbruster" <armbru@redhat.com>,
	patches@linaro.org
Subject: [Qemu-devel] [PATCH] configure: Always build with -fno-common
Date: Wed, 26 Feb 2014 21:53:30 +0000	[thread overview]
Message-ID: <1393451610-24617-1-git-send-email-peter.maydell@linaro.org> (raw)

MacOSX doesn't pull .o files from .a archives if the symbol that it
requires is one which the .o file defines as a common symbol.
(Common symbols are those declared without "extern"; the linker
will merge together common symbols with the same name, so
redeclaring the same variable in two compilation units results in
them referring to the same symbol rather than a compilation error).

This MacOSX difference from traditional linker behaviour means that
"make check" produces link errors:

Undefined symbols for architecture x86_64:
  "_cur_mon", referenced from:
      _error_vprintf in libqemuutil.a(qemu-error.o)
      _error_printf in libqemuutil.a(qemu-error.o)
      _error_printf_unless_qmp in libqemuutil.a(qemu-error.o)
      _error_print_loc in libqemuutil.a(qemu-error.o)
      _error_report in libqemuutil.a(qemu-error.o)
ld: symbol(s) not found for architecture x86_64

in this case because "cur_mon" is a common symbol in
libqemustub.a(mon-set-error.o).

In QEMU we don't make any use at all of the common symbol
functionality, so we can avoid this problem entirely simply
by compiling with -fno-common. Enable this option for all
builds, not just MacOSX, so that if we ever inadvertently
introduce multiple definitions of some variable that will
be immediately spotted as a build error rather than only
breaking the MacOSX build.

Suggested-by:  Markus Armbruster <armbru@redhat.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
---
Thanks to Markus for pointing out that -fno-common was the
simple solution to this problem. make check now works on MacOSX,
although due to a race condition in the version of glib I'm stuck
on (2.22.4) I have to periodically use "killall -CHLD gtester"
to unwedge it...

 configure | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/configure b/configure
index 423f435..ff438e4 100755
--- a/configure
+++ b/configure
@@ -369,7 +369,7 @@ sdl_config="${SDL_CONFIG-${cross_prefix}sdl-config}"
 ARFLAGS="${ARFLAGS-rv}"
 
 # default flags for all hosts
-QEMU_CFLAGS="-fno-strict-aliasing $QEMU_CFLAGS"
+QEMU_CFLAGS="-fno-strict-aliasing -fno-common $QEMU_CFLAGS"
 QEMU_CFLAGS="-Wall -Wundef -Wwrite-strings -Wmissing-prototypes $QEMU_CFLAGS"
 QEMU_CFLAGS="-Wstrict-prototypes -Wredundant-decls $QEMU_CFLAGS"
 QEMU_CFLAGS="-D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE $QEMU_CFLAGS"
-- 
1.8.5.4

             reply	other threads:[~2014-02-26 21:53 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-02-26 21:53 Peter Maydell [this message]
2014-02-28 17:19 ` [Qemu-devel] [PATCH] configure: Always build with -fno-common Richard Henderson
2014-03-06 21:42   ` Peter Maydell

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=1393451610-24617-1-git-send-email-peter.maydell@linaro.org \
    --to=peter.maydell@linaro.org \
    --cc=andreas.faerber@web.de \
    --cc=armbru@redhat.com \
    --cc=patches@linaro.org \
    --cc=qemu-devel@nongnu.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 NNTP newsgroup(s).