From: Christian Brauner <christian.brauner@canonical.com>
To: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Christian Brauner <christian.brauner@ubuntu.com>,
linux-kernel@vger.kernel.org, stgraber@ubuntu.com,
tycho@tycho.ws, serge@hallyn.com
Subject: Re: [PATCH 1/2 v3] user namespace: use union in {g,u}idmap struct
Date: Thu, 19 Oct 2017 18:15:19 +0200 [thread overview]
Message-ID: <20171019161518.b6r3ql5xk4wwr5wu@gmail.com> (raw)
In-Reply-To: <87r2u0hsj5.fsf@xmission.com>
On Wed, Oct 18, 2017 at 07:48:14PM -0500, Eric W. Biederman wrote:
> Christian Brauner <christian.brauner@canonical.com> writes:
>
> > I'm not sure why the build is complaining about how the union is initialized
> > here. This looks legitimate to me and I can't reproduce this locally with or
> > without the appended config. The struct introduced here is:
> >
> > #define UID_GID_MAP_MAX_EXTENTS 5
> >
> > struct uid_gid_extent {
> > u32 first;
> > u32 lower_first;
> > u32 count;
> > };
> >
> > struct uid_gid_map { /* 64 bytes -- 1 cache line */
> > u32 nr_extents;
> > union {
> > struct uid_gid_extent extent[UID_GID_MAP_MAX_EXTENTS];
> > struct {
> > struct uid_gid_extent *forward;
> > struct uid_gid_extent *reverse;
> > };
> > };
> > };
> >
> > And the initialization in kernel/user.c which I didn't change looks correct.
> > But maybe I'm missing the point.
>
> You may want to check your compiler version this feels like a compiler
> dependent error.
>
> It looks like gcc isn't happy about not having braces for the anonymous
> union of extent and the anonymouns structure that holds forward and
> reverse.
>
> FYI since I am commenting. I took a quick skim through your code today
> and at first glance everything looks good. The performance is nice and
> fast, and the changes look reasonable at first glance.
Thanks. Glad to hear.
>
> I think there are some nits that can be picked but nothing yet that
> indicates the code is working incorrectly.
Do you want me to wait for your feedback? If not I'd send a new version of the
patch with an additonal patch for kernel/user.c to use enclosing brackets when
initializing the union in the struct.
Christian
>
> Eric
>
> > On Thu, Oct 19, 2017 at 06:51:48AM +0800, kbuild test robot wrote:
> >> Hi Christian,
> >>
> >> [auto build test ERROR on linus/master]
> >> [also build test ERROR on v4.14-rc5 next-20171017]
> >> [if your patch is applied to the wrong git tree, please drop us a note to help improve the system]
> >>
> >> url: https://github.com/0day-ci/linux/commits/Christian-Brauner/user-namespace-use-union-in-g-u-idmap-struct/20171019-022142
> >> config: x86_64-randconfig-a0-10190527 (attached as .config)
> >> compiler: gcc-4.4 (Debian 4.4.7-8) 4.4.7
> >> reproduce:
> >> # save the attached .config to linux build tree
> >> make ARCH=x86_64
> >>
> >> All error/warnings (new ones prefixed by >>):
> >>
> >> >> kernel/user.c:29: error: unknown field 'extent' specified in initializer
> >> >> kernel/user.c:30: error: unknown field 'first' specified in initializer
> >> >> kernel/user.c:30: warning: missing braces around initializer
> >> kernel/user.c:30: warning: (near initialization for 'init_user_ns.uid_map.<anonymous>.extent')
> >> >> kernel/user.c:31: error: unknown field 'lower_first' specified in initializer
> >> >> kernel/user.c:31: warning: excess elements in union initializer
> >> kernel/user.c:31: warning: (near initialization for 'init_user_ns.uid_map.<anonymous>')
> >> >> kernel/user.c:32: error: unknown field 'count' specified in initializer
> >> kernel/user.c:32: warning: excess elements in union initializer
> >> kernel/user.c:32: warning: (near initialization for 'init_user_ns.uid_map.<anonymous>')
> >> kernel/user.c:37: error: unknown field 'extent' specified in initializer
> >> kernel/user.c:38: error: unknown field 'first' specified in initializer
> >> kernel/user.c:39: error: unknown field 'lower_first' specified in initializer
> >> kernel/user.c:39: warning: excess elements in union initializer
> >> kernel/user.c:39: warning: (near initialization for 'init_user_ns.gid_map.<anonymous>')
> >> kernel/user.c:40: error: unknown field 'count' specified in initializer
> >> kernel/user.c:40: warning: excess elements in union initializer
> >> kernel/user.c:40: warning: (near initialization for 'init_user_ns.gid_map.<anonymous>')
> >> kernel/user.c:45: error: unknown field 'extent' specified in initializer
> >> kernel/user.c:46: error: unknown field 'first' specified in initializer
> >> kernel/user.c:47: error: unknown field 'lower_first' specified in initializer
> >> kernel/user.c:47: warning: excess elements in union initializer
> >> kernel/user.c:47: warning: (near initialization for 'init_user_ns.projid_map.<anonymous>')
> >> kernel/user.c:48: error: unknown field 'count' specified in initializer
> >> kernel/user.c:48: warning: excess elements in union initializer
> >> kernel/user.c:48: warning: (near initialization for 'init_user_ns.projid_map.<anonymous>')
> >>
> >> vim +/lower_first +31 kernel/user.c
> >>
> >> ^1da177e4 Linus Torvalds 2005-04-16 21
> >> 59607db36 Serge E. Hallyn 2011-03-23 22 /*
> >> 59607db36 Serge E. Hallyn 2011-03-23 23 * userns count is 1 for root user, 1 for init_uts_ns,
> >> 59607db36 Serge E. Hallyn 2011-03-23 24 * and 1 for... ?
> >> 59607db36 Serge E. Hallyn 2011-03-23 25 */
> >> aee16ce73 Pavel Emelyanov 2008-02-08 26 struct user_namespace init_user_ns = {
> >> 22d917d80 Eric W. Biederman 2011-11-17 27 .uid_map = {
> >> 22d917d80 Eric W. Biederman 2011-11-17 28 .nr_extents = 1,
> >> 22d917d80 Eric W. Biederman 2011-11-17 @29 .extent[0] = {
> >> 22d917d80 Eric W. Biederman 2011-11-17 @30 .first = 0,
> >> 22d917d80 Eric W. Biederman 2011-11-17 @31 .lower_first = 0,
> >> 4b06a81f1 Eric W. Biederman 2012-05-19 @32 .count = 4294967295U,
> >> 22d917d80 Eric W. Biederman 2011-11-17 33 },
> >> 22d917d80 Eric W. Biederman 2011-11-17 34 },
> >> 22d917d80 Eric W. Biederman 2011-11-17 35 .gid_map = {
> >> 22d917d80 Eric W. Biederman 2011-11-17 36 .nr_extents = 1,
> >> 22d917d80 Eric W. Biederman 2011-11-17 37 .extent[0] = {
> >> 22d917d80 Eric W. Biederman 2011-11-17 38 .first = 0,
> >> 22d917d80 Eric W. Biederman 2011-11-17 39 .lower_first = 0,
> >> 4b06a81f1 Eric W. Biederman 2012-05-19 40 .count = 4294967295U,
> >> 22d917d80 Eric W. Biederman 2011-11-17 41 },
> >> 22d917d80 Eric W. Biederman 2011-11-17 42 },
> >> f76d207a6 Eric W. Biederman 2012-08-30 43 .projid_map = {
> >> f76d207a6 Eric W. Biederman 2012-08-30 44 .nr_extents = 1,
> >> f76d207a6 Eric W. Biederman 2012-08-30 45 .extent[0] = {
> >> f76d207a6 Eric W. Biederman 2012-08-30 46 .first = 0,
> >> f76d207a6 Eric W. Biederman 2012-08-30 47 .lower_first = 0,
> >> f76d207a6 Eric W. Biederman 2012-08-30 48 .count = 4294967295U,
> >> f76d207a6 Eric W. Biederman 2012-08-30 49 },
> >> f76d207a6 Eric W. Biederman 2012-08-30 50 },
> >> c61a2810a Eric W. Biederman 2012-12-28 51 .count = ATOMIC_INIT(3),
> >> 783291e69 Eric W. Biederman 2011-11-17 52 .owner = GLOBAL_ROOT_UID,
> >> 783291e69 Eric W. Biederman 2011-11-17 53 .group = GLOBAL_ROOT_GID,
> >> 435d5f4bb Al Viro 2014-10-31 54 .ns.inum = PROC_USER_INIT_INO,
> >> 33c429405 Al Viro 2014-11-01 55 #ifdef CONFIG_USER_NS
> >> 33c429405 Al Viro 2014-11-01 56 .ns.ops = &userns_operations,
> >> 33c429405 Al Viro 2014-11-01 57 #endif
> >> 9cc46516d Eric W. Biederman 2014-12-02 58 .flags = USERNS_INIT_FLAGS,
> >> 6bd364d82 Xiao Guangrong 2013-12-13 59 #ifdef CONFIG_PERSISTENT_KEYRINGS
> >> 6bd364d82 Xiao Guangrong 2013-12-13 60 .persistent_keyring_register_sem =
> >> 6bd364d82 Xiao Guangrong 2013-12-13 61 __RWSEM_INITIALIZER(init_user_ns.persistent_keyring_register_sem),
> >> f36f8c75a David Howells 2013-09-24 62 #endif
> >> aee16ce73 Pavel Emelyanov 2008-02-08 63 };
> >> aee16ce73 Pavel Emelyanov 2008-02-08 64 EXPORT_SYMBOL_GPL(init_user_ns);
> >> aee16ce73 Pavel Emelyanov 2008-02-08 65
> >>
> >> :::::: The code at line 31 was first introduced by commit
> >> :::::: 22d917d80e842829d0ca0a561967d728eb1d6303 userns: Rework the user_namespace adding uid/gid mapping support
> >>
> >> :::::: TO: Eric W. Biederman <ebiederm@xmission.com>
> >> :::::: CC: Eric W. Biederman <ebiederm@xmission.com>
> >>
> >> ---
> >> 0-DAY kernel test infrastructure Open Source Technology Center
> >> https://lists.01.org/pipermail/kbuild-all Intel Corporation
next prev parent reply other threads:[~2017-10-19 16:15 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-10-16 15:34 [PATCH 1/2 v3] user namespace: use union in {g,u}idmap struct Christian Brauner
2017-10-16 15:34 ` [PATCH 2/2 v3] user namespaces: bump idmap limits to 340 Christian Brauner
2017-10-18 22:51 ` [PATCH 1/2 v3] user namespace: use union in {g,u}idmap struct kbuild test robot
2017-10-18 23:42 ` Christian Brauner
2017-10-19 0:48 ` Eric W. Biederman
2017-10-19 16:15 ` Christian Brauner [this message]
2017-10-19 16:38 ` Eric W. Biederman
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=20171019161518.b6r3ql5xk4wwr5wu@gmail.com \
--to=christian.brauner@canonical.com \
--cc=christian.brauner@ubuntu.com \
--cc=ebiederm@xmission.com \
--cc=linux-kernel@vger.kernel.org \
--cc=serge@hallyn.com \
--cc=stgraber@ubuntu.com \
--cc=tycho@tycho.ws \
/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