From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stephen Rothwell Subject: linux-next: build failure after merge of the vfs tree Date: Thu, 22 Dec 2011 11:15:52 +1100 Message-ID: <20111222111552.f47d2033aa6d5dea00919e47@canb.auug.org.au> Mime-Version: 1.0 Content-Type: multipart/signed; protocol="application/pgp-signature"; micalg="PGP-SHA256"; boundary="Signature=_Thu__22_Dec_2011_11_15_52_+1100_d40QUZOaK1t4jQRO" Return-path: Received: from calzone.tip.net.au ([203.10.76.15]:42674 "EHLO calzone.tip.net.au" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754464Ab1LVAQD (ORCPT ); Wed, 21 Dec 2011 19:16:03 -0500 Sender: linux-next-owner@vger.kernel.org List-ID: To: Al Viro Cc: linux-next@vger.kernel.org, linux-kernel@vger.kernel.org, James Morris --Signature=_Thu__22_Dec_2011_11_15_52_+1100_d40QUZOaK1t4jQRO Content-Type: text/plain; charset=US-ASCII Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Hi Al, After merging the vfs tree, today's linux-next build (powerpc ppc64_defconf= ig) failed like this: In file included from ipc/util.c:28:0: include/linux/security.h: In function 'security_real_capable': include/linux/security.h:1886:25: error: dereferencing pointer to incomplet= e type include/linux/security.h:1886:25: error: dereferencing pointer to incomplet= e type include/linux/security.h:1886:25: error: dereferencing pointer to incomplet= e type include/linux/security.h:1886:25: error: dereferencing pointer to incomplet= e type include/linux/security.h:1886:25: error: dereferencing pointer to incomplet= e type And so on ... These are the references to __task_cred() which dereferences a "struct task_struct *". Adding an include of linux/sched.h back to linux/security.h (which was removed by commit ae1a442f9d95 ("trim security.h")) results in this error (to be fair, the include of linux-sched.h really belongs in cred.h): arch/powerpc/platforms/cell/spu_syscalls.c:68:22: error: expected ')' befor= e 'const' Which was caused by commit 2dbc0a0f41d4 ("switch spu_create(2) to use of SYSCALL4, make it use umode_t"). That commit (along with a whole string of others around it) was supposedly authored July 27, 2011, so it would have been nice to see them earlier then this in linux-next. :-( I have used the vfs tree from next-20111216 (which is unfortunately empty) again for today. BTW, the x86_64 allmodconfig build (with out the sched.h include restored) does not get any of these errors, so sches.h must be being indirectly included, so you are not saving anything on x86_64 by removing the sched.h include from security.h. --=20 Cheers, Stephen Rothwell sfr@canb.auug.org.au http://www.canb.auug.org.au/~sfr/ --Signature=_Thu__22_Dec_2011_11_15_52_+1100_d40QUZOaK1t4jQRO Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) iQIcBAEBCAAGBQJO8na4AAoJEECxmPOUX5FEVB0QAIlK2+hKqssCKSXyJn+n0lmz TyOknMmLRpOZSKcM24FYt19iXVSlBO4BnMfMt9q3DcUSCneYaWWrQSRw477RrEHk BwiITpi7zjBNXlqoEc6xurN0GIueLn/++Y/zg5QBjjNXnuodr3dmet+4RQogVxAo zItLSNJXs2J6RLBD76Wa3f6RMO4UrgkUuH17tOrAtRFOlysjO/0dWsVmsOyTSj5T kOvL0wgZzWg8/57kUnIDl0oKBl2QlDOrEYSzCdF9FBu6nFDkXPjNNikDnzIEY8Vt +F2dVzqEbR9vh0Xc/7Z9f1flzgzNXFQRVgJ/BHAfdO76bdl0Um1dMOwYfXzf+hqR N3E9lWCUN4h/h1NlLHV5I/1OFKNlCAl16/RQ5pcHRLV72nQ3sSo5cNZ7/bXkpa0d ufq377a8ZmSwGY0eEDs0zontllNHKtMHqrDmQqQ+LMe90uaiP2AnmZkfotfTXzAy 7aTO+BGehoxHAdFunTYVbMHzuZHpOiz3BhwxX6ygXR8WuI8VHzaexyumFDgXtADt RiT4pZqnZY0EOw1XUFsPJOv97eMsIJAXotb8kr6fFHl/Bua5aWhN3j9qIzrCM+c6 FsWN2Zh0t3g6VwGesT78QFRSRGsDeQi4HVAHzA/Mk8eYGgInMJgNJ4/ZbUsBh8E4 lbbEfmpXQidogrWRFF3X =Quy3 -----END PGP SIGNATURE----- --Signature=_Thu__22_Dec_2011_11_15_52_+1100_d40QUZOaK1t4jQRO--