All of lore.kernel.org
 help / color / mirror / Atom feed
* xfstests configure error with xfsprogs v4.5-rc1
@ 2016-02-22  3:03 ` Eryu Guan
  0 siblings, 0 replies; 10+ messages in thread
From: Eryu Guan @ 2016-02-22  3:03 UTC (permalink / raw)
  To: xfs; +Cc: fstests

Hi,

After upgrading to xfsprogs v4.5-rc1 version, xfstests failed to built
because of configure error:

    checking xfs/xfs.h usability... no
    checking xfs/xfs.h presence... yes
    configure: WARNING: xfs/xfs.h: present but cannot be compiled
    configure: WARNING: xfs/xfs.h:     check for missing prerequisite headers?
    configure: WARNING: xfs/xfs.h: see the Autoconf documentation
    configure: WARNING: xfs/xfs.h:     section "Present But Cannot Be Compiled"
    configure: WARNING: xfs/xfs.h: proceeding with the compiler's result
    checking for xfs/xfs.h... no
    
    FATAL ERROR: cannot find a valid <xfs/xfs.h> header file.
    Run "make install-qa" from the xfsprogs source.

and config.log showed:

    configure:4543: checking xfs/xfs.h usability
    configure:4543: ccache gcc -c -g -O2  conftest.c >&5
    In file included from /usr/include/xfs/xfs.h:37:0,
    		 from conftest.c:73:
    /usr/include/xfs/linux.h:145:1: error: unknown type name 'off64_t'
     typedef off64_t  xfs_off_t;
     ^
    configure:4543: $? = 1

It's the change made by commit cb898f157f84 ("linux.h: Use off64_t
instead of loff_t") introduced this error. If I do the following
modifications xfstests compiles without errors:

diff --git a/include/linux.h b/include/linux.h
index cc0f70c..d66436c 100644
--- a/include/linux.h
+++ b/include/linux.h
@@ -142,7 +142,7 @@ platform_discard_blocks(int fd, uint64_t start, uint64_t len)
 #define EFSCORRUPTED   EUCLEAN /* Filesystem is corrupted */
 #define EFSBADCRC      EBADMSG /* Bad CRC detected */
 
-typedef off64_t                xfs_off_t;
+typedef __off64_t              xfs_off_t;
 typedef __uint64_t     xfs_ino_t;
 typedef __uint32_t     xfs_dev_t;
 typedef __int64_t      xfs_daddr_t;

But I'm not sure if it's a proper fix.

Thanks,
Eryu

^ permalink raw reply related	[flat|nested] 10+ messages in thread

* xfstests configure error with xfsprogs v4.5-rc1
@ 2016-02-22  3:03 ` Eryu Guan
  0 siblings, 0 replies; 10+ messages in thread
From: Eryu Guan @ 2016-02-22  3:03 UTC (permalink / raw)
  To: xfs; +Cc: fstests

Hi,

After upgrading to xfsprogs v4.5-rc1 version, xfstests failed to built
because of configure error:

    checking xfs/xfs.h usability... no
    checking xfs/xfs.h presence... yes
    configure: WARNING: xfs/xfs.h: present but cannot be compiled
    configure: WARNING: xfs/xfs.h:     check for missing prerequisite headers?
    configure: WARNING: xfs/xfs.h: see the Autoconf documentation
    configure: WARNING: xfs/xfs.h:     section "Present But Cannot Be Compiled"
    configure: WARNING: xfs/xfs.h: proceeding with the compiler's result
    checking for xfs/xfs.h... no
    
    FATAL ERROR: cannot find a valid <xfs/xfs.h> header file.
    Run "make install-qa" from the xfsprogs source.

and config.log showed:

    configure:4543: checking xfs/xfs.h usability
    configure:4543: ccache gcc -c -g -O2  conftest.c >&5
    In file included from /usr/include/xfs/xfs.h:37:0,
    		 from conftest.c:73:
    /usr/include/xfs/linux.h:145:1: error: unknown type name 'off64_t'
     typedef off64_t  xfs_off_t;
     ^
    configure:4543: $? = 1

It's the change made by commit cb898f157f84 ("linux.h: Use off64_t
instead of loff_t") introduced this error. If I do the following
modifications xfstests compiles without errors:

diff --git a/include/linux.h b/include/linux.h
index cc0f70c..d66436c 100644
--- a/include/linux.h
+++ b/include/linux.h
@@ -142,7 +142,7 @@ platform_discard_blocks(int fd, uint64_t start, uint64_t len)
 #define EFSCORRUPTED   EUCLEAN /* Filesystem is corrupted */
 #define EFSBADCRC      EBADMSG /* Bad CRC detected */
 
-typedef off64_t                xfs_off_t;
+typedef __off64_t              xfs_off_t;
 typedef __uint64_t     xfs_ino_t;
 typedef __uint32_t     xfs_dev_t;
 typedef __int64_t      xfs_daddr_t;

But I'm not sure if it's a proper fix.

Thanks,
Eryu

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

^ permalink raw reply related	[flat|nested] 10+ messages in thread

* Re: xfstests configure error with xfsprogs v4.5-rc1
  2016-02-22  3:03 ` Eryu Guan
@ 2016-02-22  3:24   ` Dave Chinner
  -1 siblings, 0 replies; 10+ messages in thread
From: Dave Chinner @ 2016-02-22  3:24 UTC (permalink / raw)
  To: Eryu Guan; +Cc: xfs, fstests

On Mon, Feb 22, 2016 at 11:03:04AM +0800, Eryu Guan wrote:
> Hi,
> 
> After upgrading to xfsprogs v4.5-rc1 version, xfstests failed to built
> because of configure error:
> 
>     checking xfs/xfs.h usability... no
>     checking xfs/xfs.h presence... yes
>     configure: WARNING: xfs/xfs.h: present but cannot be compiled
>     configure: WARNING: xfs/xfs.h:     check for missing prerequisite headers?
>     configure: WARNING: xfs/xfs.h: see the Autoconf documentation
>     configure: WARNING: xfs/xfs.h:     section "Present But Cannot Be Compiled"
>     configure: WARNING: xfs/xfs.h: proceeding with the compiler's result
>     checking for xfs/xfs.h... no
>     
>     FATAL ERROR: cannot find a valid <xfs/xfs.h> header file.
>     Run "make install-qa" from the xfsprogs source.
> 
> and config.log showed:
> 
>     configure:4543: checking xfs/xfs.h usability
>     configure:4543: ccache gcc -c -g -O2  conftest.c >&5
>     In file included from /usr/include/xfs/xfs.h:37:0,
>     		 from conftest.c:73:
>     /usr/include/xfs/linux.h:145:1: error: unknown type name 'off64_t'
>      typedef off64_t  xfs_off_t;
>      ^
>     configure:4543: $? = 1

That would seem like a system library problem in that it doesn't
define off64_t through stdio.h. What distro are you compiling on?
Also,  if you set _GNU_SOURCE, does it then compile? e.g. see
m4/package_libcdev.m4 for how to test whether code compiles under
certain defines.

If it does compile with _GNU_SOURCE being set, then all that needs
to be is to update the autoconf rule.

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: xfstests configure error with xfsprogs v4.5-rc1
@ 2016-02-22  3:24   ` Dave Chinner
  0 siblings, 0 replies; 10+ messages in thread
From: Dave Chinner @ 2016-02-22  3:24 UTC (permalink / raw)
  To: Eryu Guan; +Cc: fstests, xfs

On Mon, Feb 22, 2016 at 11:03:04AM +0800, Eryu Guan wrote:
> Hi,
> 
> After upgrading to xfsprogs v4.5-rc1 version, xfstests failed to built
> because of configure error:
> 
>     checking xfs/xfs.h usability... no
>     checking xfs/xfs.h presence... yes
>     configure: WARNING: xfs/xfs.h: present but cannot be compiled
>     configure: WARNING: xfs/xfs.h:     check for missing prerequisite headers?
>     configure: WARNING: xfs/xfs.h: see the Autoconf documentation
>     configure: WARNING: xfs/xfs.h:     section "Present But Cannot Be Compiled"
>     configure: WARNING: xfs/xfs.h: proceeding with the compiler's result
>     checking for xfs/xfs.h... no
>     
>     FATAL ERROR: cannot find a valid <xfs/xfs.h> header file.
>     Run "make install-qa" from the xfsprogs source.
> 
> and config.log showed:
> 
>     configure:4543: checking xfs/xfs.h usability
>     configure:4543: ccache gcc -c -g -O2  conftest.c >&5
>     In file included from /usr/include/xfs/xfs.h:37:0,
>     		 from conftest.c:73:
>     /usr/include/xfs/linux.h:145:1: error: unknown type name 'off64_t'
>      typedef off64_t  xfs_off_t;
>      ^
>     configure:4543: $? = 1

That would seem like a system library problem in that it doesn't
define off64_t through stdio.h. What distro are you compiling on?
Also,  if you set _GNU_SOURCE, does it then compile? e.g. see
m4/package_libcdev.m4 for how to test whether code compiles under
certain defines.

If it does compile with _GNU_SOURCE being set, then all that needs
to be is to update the autoconf rule.

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: xfstests configure error with xfsprogs v4.5-rc1
  2016-02-22  3:24   ` Dave Chinner
@ 2016-02-22  8:00     ` Eryu Guan
  -1 siblings, 0 replies; 10+ messages in thread
From: Eryu Guan @ 2016-02-22  8:00 UTC (permalink / raw)
  To: Dave Chinner; +Cc: xfs, fstests

On Mon, Feb 22, 2016 at 02:24:50PM +1100, Dave Chinner wrote:
> On Mon, Feb 22, 2016 at 11:03:04AM +0800, Eryu Guan wrote:
> > Hi,
> > 
> > After upgrading to xfsprogs v4.5-rc1 version, xfstests failed to built
> > because of configure error:
> > 
> >     checking xfs/xfs.h usability... no
> >     checking xfs/xfs.h presence... yes
> >     configure: WARNING: xfs/xfs.h: present but cannot be compiled
> >     configure: WARNING: xfs/xfs.h:     check for missing prerequisite headers?
> >     configure: WARNING: xfs/xfs.h: see the Autoconf documentation
> >     configure: WARNING: xfs/xfs.h:     section "Present But Cannot Be Compiled"
> >     configure: WARNING: xfs/xfs.h: proceeding with the compiler's result
> >     checking for xfs/xfs.h... no
> >     
> >     FATAL ERROR: cannot find a valid <xfs/xfs.h> header file.
> >     Run "make install-qa" from the xfsprogs source.
> > 
> > and config.log showed:
> > 
> >     configure:4543: checking xfs/xfs.h usability
> >     configure:4543: ccache gcc -c -g -O2  conftest.c >&5
> >     In file included from /usr/include/xfs/xfs.h:37:0,
> >     		 from conftest.c:73:
> >     /usr/include/xfs/linux.h:145:1: error: unknown type name 'off64_t'
> >      typedef off64_t  xfs_off_t;
> >      ^
> >     configure:4543: $? = 1
> 
> That would seem like a system library problem in that it doesn't
> define off64_t through stdio.h. What distro are you compiling on?

It's a RHEL7.2 box.

> Also,  if you set _GNU_SOURCE, does it then compile? e.g. see
> m4/package_libcdev.m4 for how to test whether code compiles under
> certain defines.

Okay, I'll go through the m4 files and test around.

Thanks,
Eryu

> 
> If it does compile with _GNU_SOURCE being set, then all that needs
> to be is to update the autoconf rule.
> 
> Cheers,
> 
> Dave.
> -- 
> Dave Chinner
> david@fromorbit.com

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: xfstests configure error with xfsprogs v4.5-rc1
@ 2016-02-22  8:00     ` Eryu Guan
  0 siblings, 0 replies; 10+ messages in thread
From: Eryu Guan @ 2016-02-22  8:00 UTC (permalink / raw)
  To: Dave Chinner; +Cc: fstests, xfs

On Mon, Feb 22, 2016 at 02:24:50PM +1100, Dave Chinner wrote:
> On Mon, Feb 22, 2016 at 11:03:04AM +0800, Eryu Guan wrote:
> > Hi,
> > 
> > After upgrading to xfsprogs v4.5-rc1 version, xfstests failed to built
> > because of configure error:
> > 
> >     checking xfs/xfs.h usability... no
> >     checking xfs/xfs.h presence... yes
> >     configure: WARNING: xfs/xfs.h: present but cannot be compiled
> >     configure: WARNING: xfs/xfs.h:     check for missing prerequisite headers?
> >     configure: WARNING: xfs/xfs.h: see the Autoconf documentation
> >     configure: WARNING: xfs/xfs.h:     section "Present But Cannot Be Compiled"
> >     configure: WARNING: xfs/xfs.h: proceeding with the compiler's result
> >     checking for xfs/xfs.h... no
> >     
> >     FATAL ERROR: cannot find a valid <xfs/xfs.h> header file.
> >     Run "make install-qa" from the xfsprogs source.
> > 
> > and config.log showed:
> > 
> >     configure:4543: checking xfs/xfs.h usability
> >     configure:4543: ccache gcc -c -g -O2  conftest.c >&5
> >     In file included from /usr/include/xfs/xfs.h:37:0,
> >     		 from conftest.c:73:
> >     /usr/include/xfs/linux.h:145:1: error: unknown type name 'off64_t'
> >      typedef off64_t  xfs_off_t;
> >      ^
> >     configure:4543: $? = 1
> 
> That would seem like a system library problem in that it doesn't
> define off64_t through stdio.h. What distro are you compiling on?

It's a RHEL7.2 box.

> Also,  if you set _GNU_SOURCE, does it then compile? e.g. see
> m4/package_libcdev.m4 for how to test whether code compiles under
> certain defines.

Okay, I'll go through the m4 files and test around.

Thanks,
Eryu

> 
> If it does compile with _GNU_SOURCE being set, then all that needs
> to be is to update the autoconf rule.
> 
> Cheers,
> 
> Dave.
> -- 
> Dave Chinner
> david@fromorbit.com

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: xfstests configure error with xfsprogs v4.5-rc1
  2016-02-22  3:24   ` Dave Chinner
@ 2016-02-22 10:16     ` Eryu Guan
  -1 siblings, 0 replies; 10+ messages in thread
From: Eryu Guan @ 2016-02-22 10:16 UTC (permalink / raw)
  To: Dave Chinner; +Cc: fstests, xfs

[-- Attachment #1: Type: text/plain, Size: 2102 bytes --]

On Mon, Feb 22, 2016 at 02:24:50PM +1100, Dave Chinner wrote:
> On Mon, Feb 22, 2016 at 11:03:04AM +0800, Eryu Guan wrote:
> > Hi,
> > 
> > After upgrading to xfsprogs v4.5-rc1 version, xfstests failed to built
> > because of configure error:
> > 
> >     checking xfs/xfs.h usability... no
> >     checking xfs/xfs.h presence... yes
> >     configure: WARNING: xfs/xfs.h: present but cannot be compiled
> >     configure: WARNING: xfs/xfs.h:     check for missing prerequisite headers?
> >     configure: WARNING: xfs/xfs.h: see the Autoconf documentation
> >     configure: WARNING: xfs/xfs.h:     section "Present But Cannot Be Compiled"
> >     configure: WARNING: xfs/xfs.h: proceeding with the compiler's result
> >     checking for xfs/xfs.h... no
> >     
> >     FATAL ERROR: cannot find a valid <xfs/xfs.h> header file.
> >     Run "make install-qa" from the xfsprogs source.
> > 
> > and config.log showed:
> > 
> >     configure:4543: checking xfs/xfs.h usability
> >     configure:4543: ccache gcc -c -g -O2  conftest.c >&5
> >     In file included from /usr/include/xfs/xfs.h:37:0,
> >     		 from conftest.c:73:
> >     /usr/include/xfs/linux.h:145:1: error: unknown type name 'off64_t'
> >      typedef off64_t  xfs_off_t;
> >      ^
> >     configure:4543: $? = 1
> 
> That would seem like a system library problem in that it doesn't
> define off64_t through stdio.h. What distro are you compiling on?
> Also,  if you set _GNU_SOURCE, does it then compile? e.g. see
> m4/package_libcdev.m4 for how to test whether code compiles under
> certain defines.
> 
> If it does compile with _GNU_SOURCE being set, then all that needs
> to be is to update the autoconf rule.

Yes, it compiles with _GNU_SOURCE being set. And it turns out that not
only the check for xfs.h needs _GNU_SOURCE, many other checks need it
too, like libxfs.h, xfs_log_format.h, xlog_assign_lsn, xqm.h and xfsctl.

Attached file is updates I made to autoconf rules, as I'm not familiar
with autotools, I'm not sure if it's a proper fix, but it does work for
me, with both old and new version of xfsprogs.

Thanks,
Eryu

[-- Attachment #2: off64_t.patch --]
[-- Type: text/plain, Size: 2206 bytes --]

diff --git a/configure.ac b/configure.ac
index d038f95..fa48d2f 100644
--- a/configure.ac
+++ b/configure.ac
@@ -34,7 +34,9 @@ AC_HEADER_STDC
 			cifs/ioctl.h		\
     ])
 
-AC_CHECK_HEADERS([xfs/xfs_log_format.h],,,[#include <xfs/libxfs.h>])
+AC_CHECK_HEADERS([xfs/xfs_log_format.h],,,[
+#define _GNU_SOURCE
+#include <xfs/libxfs.h>])
 
 AC_PACKAGE_NEED_UUIDCOMPARE
 
diff --git a/m4/package_xfslibs.m4 b/m4/package_xfslibs.m4
index 78a3dc4..aeb3d7c 100644
--- a/m4/package_xfslibs.m4
+++ b/m4/package_xfslibs.m4
@@ -1,5 +1,5 @@
 AC_DEFUN([AC_PACKAGE_NEED_XFS_XFS_H],
-  [ AC_CHECK_HEADERS([xfs/xfs.h])
+  [ AC_CHECK_HEADERS([xfs/xfs.h],,,[#define _GNU_SOURCE])
     if test "$ac_cv_header_xfs_xfs_h" != "yes"; then
         echo
         echo 'FATAL ERROR: cannot find a valid <xfs/xfs.h> header file.'
@@ -9,19 +9,21 @@ AC_DEFUN([AC_PACKAGE_NEED_XFS_XFS_H],
   ])
 
 AC_DEFUN([AC_PACKAGE_WANT_LIBXFS_H],
-  [ AC_CHECK_HEADERS([xfs/libxfs.h], [ have_libxfs=true ], [ have_libxfs=false ])
+  [ AC_CHECK_HEADERS([xfs/libxfs.h], [ have_libxfs=true ],
+    [ have_libxfs=false ], [#define _GNU_SOURCE])
     AC_SUBST(have_libxfs)
   ])
 
 AC_DEFUN([AC_PACKAGE_WANT_XLOG_ASSIGN_LSN],
   [ AC_CHECK_DECL(xlog_assign_lsn,
-      [ have_xlog_assign_lsn=true ], [ have_xlog_assign_lsn=false ],
-      [[#include <xfs/libxfs.h>]])
+      [ have_xlog_assign_lsn=true ], [ have_xlog_assign_lsn=false ], [[
+#define _GNU_SOURCE
+#include <xfs/libxfs.h>]])
     AC_SUBST(have_xlog_assign_lsn)
   ])
 
 AC_DEFUN([AC_PACKAGE_NEED_XFS_XQM_H],
-  [ AC_CHECK_HEADERS([xfs/xqm.h])
+  [ AC_CHECK_HEADERS([xfs/xqm.h],,,[#define _GNU_SOURCE])
     if test "$ac_cv_header_xfs_xqm_h" != "yes"; then
         echo
         echo 'FATAL ERROR: cannot find a valid <xfs/xqm.h> header file.'
@@ -99,7 +101,10 @@ AC_DEFUN([AC_PACKAGE_NEED_IRIX_LIBHANDLE],
 
 AC_DEFUN([AC_PACKAGE_NEED_XFSCTL_MACRO],
   [ AC_MSG_CHECKING([xfsctl from xfs/xfs.h])
-    AC_TRY_LINK([#include <xfs/xfs.h>], [ int x = xfsctl(0, 0, 0, 0); ],
+    AC_TRY_LINK([
+#define _GNU_SOURCE
+#include <xfs/xfs.h> ],
+      [ int x = xfsctl(0, 0, 0, 0); ],
       [ echo ok ],
       [ echo
         echo 'FATAL ERROR: cannot find required macros in the XFS headers.'

^ permalink raw reply related	[flat|nested] 10+ messages in thread

* Re: xfstests configure error with xfsprogs v4.5-rc1
@ 2016-02-22 10:16     ` Eryu Guan
  0 siblings, 0 replies; 10+ messages in thread
From: Eryu Guan @ 2016-02-22 10:16 UTC (permalink / raw)
  To: Dave Chinner; +Cc: fstests, xfs

[-- Attachment #1: Type: text/plain, Size: 2102 bytes --]

On Mon, Feb 22, 2016 at 02:24:50PM +1100, Dave Chinner wrote:
> On Mon, Feb 22, 2016 at 11:03:04AM +0800, Eryu Guan wrote:
> > Hi,
> > 
> > After upgrading to xfsprogs v4.5-rc1 version, xfstests failed to built
> > because of configure error:
> > 
> >     checking xfs/xfs.h usability... no
> >     checking xfs/xfs.h presence... yes
> >     configure: WARNING: xfs/xfs.h: present but cannot be compiled
> >     configure: WARNING: xfs/xfs.h:     check for missing prerequisite headers?
> >     configure: WARNING: xfs/xfs.h: see the Autoconf documentation
> >     configure: WARNING: xfs/xfs.h:     section "Present But Cannot Be Compiled"
> >     configure: WARNING: xfs/xfs.h: proceeding with the compiler's result
> >     checking for xfs/xfs.h... no
> >     
> >     FATAL ERROR: cannot find a valid <xfs/xfs.h> header file.
> >     Run "make install-qa" from the xfsprogs source.
> > 
> > and config.log showed:
> > 
> >     configure:4543: checking xfs/xfs.h usability
> >     configure:4543: ccache gcc -c -g -O2  conftest.c >&5
> >     In file included from /usr/include/xfs/xfs.h:37:0,
> >     		 from conftest.c:73:
> >     /usr/include/xfs/linux.h:145:1: error: unknown type name 'off64_t'
> >      typedef off64_t  xfs_off_t;
> >      ^
> >     configure:4543: $? = 1
> 
> That would seem like a system library problem in that it doesn't
> define off64_t through stdio.h. What distro are you compiling on?
> Also,  if you set _GNU_SOURCE, does it then compile? e.g. see
> m4/package_libcdev.m4 for how to test whether code compiles under
> certain defines.
> 
> If it does compile with _GNU_SOURCE being set, then all that needs
> to be is to update the autoconf rule.

Yes, it compiles with _GNU_SOURCE being set. And it turns out that not
only the check for xfs.h needs _GNU_SOURCE, many other checks need it
too, like libxfs.h, xfs_log_format.h, xlog_assign_lsn, xqm.h and xfsctl.

Attached file is updates I made to autoconf rules, as I'm not familiar
with autotools, I'm not sure if it's a proper fix, but it does work for
me, with both old and new version of xfsprogs.

Thanks,
Eryu

[-- Attachment #2: off64_t.patch --]
[-- Type: text/plain, Size: 2206 bytes --]

diff --git a/configure.ac b/configure.ac
index d038f95..fa48d2f 100644
--- a/configure.ac
+++ b/configure.ac
@@ -34,7 +34,9 @@ AC_HEADER_STDC
 			cifs/ioctl.h		\
     ])
 
-AC_CHECK_HEADERS([xfs/xfs_log_format.h],,,[#include <xfs/libxfs.h>])
+AC_CHECK_HEADERS([xfs/xfs_log_format.h],,,[
+#define _GNU_SOURCE
+#include <xfs/libxfs.h>])
 
 AC_PACKAGE_NEED_UUIDCOMPARE
 
diff --git a/m4/package_xfslibs.m4 b/m4/package_xfslibs.m4
index 78a3dc4..aeb3d7c 100644
--- a/m4/package_xfslibs.m4
+++ b/m4/package_xfslibs.m4
@@ -1,5 +1,5 @@
 AC_DEFUN([AC_PACKAGE_NEED_XFS_XFS_H],
-  [ AC_CHECK_HEADERS([xfs/xfs.h])
+  [ AC_CHECK_HEADERS([xfs/xfs.h],,,[#define _GNU_SOURCE])
     if test "$ac_cv_header_xfs_xfs_h" != "yes"; then
         echo
         echo 'FATAL ERROR: cannot find a valid <xfs/xfs.h> header file.'
@@ -9,19 +9,21 @@ AC_DEFUN([AC_PACKAGE_NEED_XFS_XFS_H],
   ])
 
 AC_DEFUN([AC_PACKAGE_WANT_LIBXFS_H],
-  [ AC_CHECK_HEADERS([xfs/libxfs.h], [ have_libxfs=true ], [ have_libxfs=false ])
+  [ AC_CHECK_HEADERS([xfs/libxfs.h], [ have_libxfs=true ],
+    [ have_libxfs=false ], [#define _GNU_SOURCE])
     AC_SUBST(have_libxfs)
   ])
 
 AC_DEFUN([AC_PACKAGE_WANT_XLOG_ASSIGN_LSN],
   [ AC_CHECK_DECL(xlog_assign_lsn,
-      [ have_xlog_assign_lsn=true ], [ have_xlog_assign_lsn=false ],
-      [[#include <xfs/libxfs.h>]])
+      [ have_xlog_assign_lsn=true ], [ have_xlog_assign_lsn=false ], [[
+#define _GNU_SOURCE
+#include <xfs/libxfs.h>]])
     AC_SUBST(have_xlog_assign_lsn)
   ])
 
 AC_DEFUN([AC_PACKAGE_NEED_XFS_XQM_H],
-  [ AC_CHECK_HEADERS([xfs/xqm.h])
+  [ AC_CHECK_HEADERS([xfs/xqm.h],,,[#define _GNU_SOURCE])
     if test "$ac_cv_header_xfs_xqm_h" != "yes"; then
         echo
         echo 'FATAL ERROR: cannot find a valid <xfs/xqm.h> header file.'
@@ -99,7 +101,10 @@ AC_DEFUN([AC_PACKAGE_NEED_IRIX_LIBHANDLE],
 
 AC_DEFUN([AC_PACKAGE_NEED_XFSCTL_MACRO],
   [ AC_MSG_CHECKING([xfsctl from xfs/xfs.h])
-    AC_TRY_LINK([#include <xfs/xfs.h>], [ int x = xfsctl(0, 0, 0, 0); ],
+    AC_TRY_LINK([
+#define _GNU_SOURCE
+#include <xfs/xfs.h> ],
+      [ int x = xfsctl(0, 0, 0, 0); ],
       [ echo ok ],
       [ echo
         echo 'FATAL ERROR: cannot find required macros in the XFS headers.'

[-- Attachment #3: Type: text/plain, Size: 121 bytes --]

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

^ permalink raw reply related	[flat|nested] 10+ messages in thread

* Re: xfstests configure error with xfsprogs v4.5-rc1
  2016-02-22 10:16     ` Eryu Guan
@ 2016-02-22 21:58       ` Dave Chinner
  -1 siblings, 0 replies; 10+ messages in thread
From: Dave Chinner @ 2016-02-22 21:58 UTC (permalink / raw)
  To: Eryu Guan; +Cc: fstests, xfs

On Mon, Feb 22, 2016 at 06:16:48PM +0800, Eryu Guan wrote:
> On Mon, Feb 22, 2016 at 02:24:50PM +1100, Dave Chinner wrote:
> > On Mon, Feb 22, 2016 at 11:03:04AM +0800, Eryu Guan wrote:
> > > Hi,
> > > 
> > > After upgrading to xfsprogs v4.5-rc1 version, xfstests failed to built
> > > because of configure error:
> > > 
> > >     checking xfs/xfs.h usability... no
> > >     checking xfs/xfs.h presence... yes
> > >     configure: WARNING: xfs/xfs.h: present but cannot be compiled
> > >     configure: WARNING: xfs/xfs.h:     check for missing prerequisite headers?
> > >     configure: WARNING: xfs/xfs.h: see the Autoconf documentation
> > >     configure: WARNING: xfs/xfs.h:     section "Present But Cannot Be Compiled"
> > >     configure: WARNING: xfs/xfs.h: proceeding with the compiler's result
> > >     checking for xfs/xfs.h... no
> > >     
> > >     FATAL ERROR: cannot find a valid <xfs/xfs.h> header file.
> > >     Run "make install-qa" from the xfsprogs source.
> > > 
> > > and config.log showed:
> > > 
> > >     configure:4543: checking xfs/xfs.h usability
> > >     configure:4543: ccache gcc -c -g -O2  conftest.c >&5
> > >     In file included from /usr/include/xfs/xfs.h:37:0,
> > >     		 from conftest.c:73:
> > >     /usr/include/xfs/linux.h:145:1: error: unknown type name 'off64_t'
> > >      typedef off64_t  xfs_off_t;
> > >      ^
> > >     configure:4543: $? = 1
> > 
> > That would seem like a system library problem in that it doesn't
> > define off64_t through stdio.h. What distro are you compiling on?
> > Also,  if you set _GNU_SOURCE, does it then compile? e.g. see
> > m4/package_libcdev.m4 for how to test whether code compiles under
> > certain defines.
> > 
> > If it does compile with _GNU_SOURCE being set, then all that needs
> > to be is to update the autoconf rule.
> 
> Yes, it compiles with _GNU_SOURCE being set. And it turns out that not
> only the check for xfs.h needs _GNU_SOURCE, many other checks need it
> too, like libxfs.h, xfs_log_format.h, xlog_assign_lsn, xqm.h and xfsctl.
> 
> Attached file is updates I made to autoconf rules, as I'm not familiar
> with autotools, I'm not sure if it's a proper fix, but it does work for
> me, with both old and new version of xfsprogs.

Yup, works here too, and is cleaner than the hack I did late
yesterday afternoon. Thanks, Eryu!

Cheers,

Dave.

-- 
Dave Chinner
david@fromorbit.com

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: xfstests configure error with xfsprogs v4.5-rc1
@ 2016-02-22 21:58       ` Dave Chinner
  0 siblings, 0 replies; 10+ messages in thread
From: Dave Chinner @ 2016-02-22 21:58 UTC (permalink / raw)
  To: Eryu Guan; +Cc: fstests, xfs

On Mon, Feb 22, 2016 at 06:16:48PM +0800, Eryu Guan wrote:
> On Mon, Feb 22, 2016 at 02:24:50PM +1100, Dave Chinner wrote:
> > On Mon, Feb 22, 2016 at 11:03:04AM +0800, Eryu Guan wrote:
> > > Hi,
> > > 
> > > After upgrading to xfsprogs v4.5-rc1 version, xfstests failed to built
> > > because of configure error:
> > > 
> > >     checking xfs/xfs.h usability... no
> > >     checking xfs/xfs.h presence... yes
> > >     configure: WARNING: xfs/xfs.h: present but cannot be compiled
> > >     configure: WARNING: xfs/xfs.h:     check for missing prerequisite headers?
> > >     configure: WARNING: xfs/xfs.h: see the Autoconf documentation
> > >     configure: WARNING: xfs/xfs.h:     section "Present But Cannot Be Compiled"
> > >     configure: WARNING: xfs/xfs.h: proceeding with the compiler's result
> > >     checking for xfs/xfs.h... no
> > >     
> > >     FATAL ERROR: cannot find a valid <xfs/xfs.h> header file.
> > >     Run "make install-qa" from the xfsprogs source.
> > > 
> > > and config.log showed:
> > > 
> > >     configure:4543: checking xfs/xfs.h usability
> > >     configure:4543: ccache gcc -c -g -O2  conftest.c >&5
> > >     In file included from /usr/include/xfs/xfs.h:37:0,
> > >     		 from conftest.c:73:
> > >     /usr/include/xfs/linux.h:145:1: error: unknown type name 'off64_t'
> > >      typedef off64_t  xfs_off_t;
> > >      ^
> > >     configure:4543: $? = 1
> > 
> > That would seem like a system library problem in that it doesn't
> > define off64_t through stdio.h. What distro are you compiling on?
> > Also,  if you set _GNU_SOURCE, does it then compile? e.g. see
> > m4/package_libcdev.m4 for how to test whether code compiles under
> > certain defines.
> > 
> > If it does compile with _GNU_SOURCE being set, then all that needs
> > to be is to update the autoconf rule.
> 
> Yes, it compiles with _GNU_SOURCE being set. And it turns out that not
> only the check for xfs.h needs _GNU_SOURCE, many other checks need it
> too, like libxfs.h, xfs_log_format.h, xlog_assign_lsn, xqm.h and xfsctl.
> 
> Attached file is updates I made to autoconf rules, as I'm not familiar
> with autotools, I'm not sure if it's a proper fix, but it does work for
> me, with both old and new version of xfsprogs.

Yup, works here too, and is cleaner than the hack I did late
yesterday afternoon. Thanks, Eryu!

Cheers,

Dave.

-- 
Dave Chinner
david@fromorbit.com

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

^ permalink raw reply	[flat|nested] 10+ messages in thread

end of thread, other threads:[~2016-02-22 21:58 UTC | newest]

Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-02-22  3:03 xfstests configure error with xfsprogs v4.5-rc1 Eryu Guan
2016-02-22  3:03 ` Eryu Guan
2016-02-22  3:24 ` Dave Chinner
2016-02-22  3:24   ` Dave Chinner
2016-02-22  8:00   ` Eryu Guan
2016-02-22  8:00     ` Eryu Guan
2016-02-22 10:16   ` Eryu Guan
2016-02-22 10:16     ` Eryu Guan
2016-02-22 21:58     ` Dave Chinner
2016-02-22 21:58       ` Dave Chinner

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.