Re: [libvirt] [PATCH] build: avoid compile failure on linux kernels older than 2.6.19

On Mon, May 17, 2010 at 04:59:51PM +0200, Jim Meyering wrote: > Daniel P. Berrange wrote: > > On Mon, May 17, 2010 at 04:25:43PM +0200, Jim Meyering wrote: > >> This is required for any kernel prior to 2.6.19, > >> since <linux/magic.h> didn't exist back then. > >> Now that file is provided by the kernel-headers package. > >> > >> >From d14ef1669968ffeb65076b007e318934ed99aa61 Mon Sep 17 00:00:00 2001 > >> From: Jim Meyering <meyering(a)redhat.com&gt; > >> Date: Mon, 17 May 2010 16:17:08 +0200 > >> Subject: [PATCH] build: avoid compile failure on linux kernels older than 2.6.19 > >> > >> * configure.ac: Check for <linux/magic.h>. > >> * src/util/storage_file.c: Include <linux/magic.h> only if present. > >> Linux kernels prior to 2.6.19 lacked it. > >> [__linux__] (NFS_SUPER_MAGIC): Define if not already defined. > >> --- > >> configure.ac | 2 +- > >> src/util/storage_file.c | 7 ++++++- > >> 2 files changed, 7 insertions(+), 2 deletions(-) > >> > >> diff --git a/configure.ac b/configure.ac > >> index c187420..ebd2082 100644 > >> --- a/configure.ac > >> +++ b/configure.ac > >> @@ -108,7 +108,7 @@ LIBS=$old_libs > >> > >> dnl Availability of various common headers (non-fatal if missing). > >> AC_CHECK_HEADERS([pwd.h paths.h regex.h sys/syslimits.h \ > >> - termios.h sys/poll.h syslog.h mntent.h net/ethernet.h]) > >> + termios.h sys/poll.h syslog.h mntent.h net/ethernet.h linux/magic.h]) > >> > >> dnl Where are the XDR functions? > >> dnl If portablexdr is installed, prefer that. > >> diff --git a/src/util/storage_file.c b/src/util/storage_file.c > >> index 5f15a64..a07bedc 100644 > >> --- a/src/util/storage_file.c > >> +++ b/src/util/storage_file.c > >> @@ -27,7 +27,9 @@ > >> #include <unistd.h> > >> #include <fcntl.h> > >> #ifdef __linux__ > >> -# include <linux/magic.h> > >> +# if HAVE_LINUX_MAGIC_H > >> +# include <linux/magic.h> > >> +# endif > > > > Don't bother with this - just kill this include off completely. It is a > > waste of time using it if we can't rely on it existing & have to re-define > > everything ourselves. > > I admit that those numbers will probably never change, but I see > including the recommended/modern header as the preferred approach. > Then, eventually when the old systems are no longer relevant (yeah, in > a long time) we'll be able to remove our hackish define-if-required code. If we're going to plan todo that then we need to open a bug report against the current linux/magic.h because it is still missing many filesystems that exist today.