linux-next: build failure after merge of the pidfd tree

linux-next: build failure after merge of the pidfd tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 2571 bytes --]

Hi Christian,

After merging the pidfd tree, today's linux-next build (arm
multi_v7_defconfig) failed like this:

kernel/pid.c: In function '__do_sys_pidfd_open':
kernel/pid.c:652:7: error: 'fd' redeclared as different kind of symbol
  long fd = -EINVAL;
       ^~
In file included from kernel/pid.c:40:
kernel/pid.c:649:46: note: previous definition of 'fd' was here
 SYSCALL_DEFINE4(pidfd_open, pid_t, pid, int, fd, int, pidfd,
                                         ~~~~~^~
include/linux/syscalls.h:117:27: note: in definition of macro '__SC_DECL'
 #define __SC_DECL(t, a) t a
                           ^
include/linux/syscalls.h:112:35: note: in expansion of macro '__MAP3'
 #define __MAP4(m,t,a,...) m(t,a), __MAP3(m,__VA_ARGS__)
                                   ^~~~~~
include/linux/syscalls.h:115:22: note: in expansion of macro '__MAP4'
 #define __MAP(n,...) __MAP##n(__VA_ARGS__)
                      ^~~~~
include/linux/syscalls.h:253:36: note: in expansion of macro '__MAP'
  static inline long __do_sys##name(__MAP(x,__SC_DECL,__VA_ARGS__))
                                    ^~~~~
include/linux/syscalls.h:226:2: note: in expansion of macro '__SYSCALL_DEFINEx'
  __SYSCALL_DEFINEx(x, sname, __VA_ARGS__)
  ^~~~~~~~~~~~~~~~~
include/linux/syscalls.h:218:36: note: in expansion of macro 'SYSCALL_DEFINEx'
 #define SYSCALL_DEFINE4(name, ...) SYSCALL_DEFINEx(4, _##name, __VA_ARGS__)
                                    ^~~~~~~~~~~~~~~
kernel/pid.c:649:1: note: in expansion of macro 'SYSCALL_DEFINE4'
 SYSCALL_DEFINE4(pidfd_open, pid_t, pid, int, fd, int, pidfd,
 ^~~~~~~~~~~~~~~
kernel/pid.c:663:7: error: 'procfd' undeclared (first use in this function); did you mean 'pidfd'?
   if (procfd != -1 || pidfd != -1)
       ^~~~~~
       pidfd
kernel/pid.c:663:7: note: each undeclared identifier is reported only once for each function it appears in

Caused by commit

  9170fba40db0 ("pid: add pidfd_open()")

I have used the version of the pifd tree from next-20190327 for today.

Please do *not* use linux-next as a development tree (I have seen several
different version of this code over the past few days :-().  The rules
for linux-next included code include:
"
     * posted to the relevant mailing list,
     * reviewed by you (or another maintainer of your subsystem tree),
     * successfully unit tested, and 
     * destined for the current or next Linux merge window.

Basically, this should be just what you would send to Linus (or ask him
to fetch)."
-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: linux-next: build failure after merge of the pidfd tree

[Christian Brauner]

On March 28, 2019 3:04:49 AM GMT+01:00, Stephen Rothwell <sfr@canb.auug.org.au> wrote:
>Hi Christian,
>
>After merging the pidfd tree, today's linux-next build (arm
>multi_v7_defconfig) failed like this:
>
>kernel/pid.c: In function '__do_sys_pidfd_open':
>kernel/pid.c:652:7: error: 'fd' redeclared as different kind of symbol
>  long fd = -EINVAL;
>       ^~
>In file included from kernel/pid.c:40:
>kernel/pid.c:649:46: note: previous definition of 'fd' was here
> SYSCALL_DEFINE4(pidfd_open, pid_t, pid, int, fd, int, pidfd,
>                                         ~~~~~^~
>include/linux/syscalls.h:117:27: note: in definition of macro
>'__SC_DECL'
> #define __SC_DECL(t, a) t a
>                           ^
>include/linux/syscalls.h:112:35: note: in expansion of macro '__MAP3'
> #define __MAP4(m,t,a,...) m(t,a), __MAP3(m,__VA_ARGS__)
>                                   ^~~~~~
>include/linux/syscalls.h:115:22: note: in expansion of macro '__MAP4'
> #define __MAP(n,...) __MAP##n(__VA_ARGS__)
>                      ^~~~~
>include/linux/syscalls.h:253:36: note: in expansion of macro '__MAP'
>  static inline long __do_sys##name(__MAP(x,__SC_DECL,__VA_ARGS__))
>                                    ^~~~~
>include/linux/syscalls.h:226:2: note: in expansion of macro
>'__SYSCALL_DEFINEx'
>  __SYSCALL_DEFINEx(x, sname, __VA_ARGS__)
>  ^~~~~~~~~~~~~~~~~
>include/linux/syscalls.h:218:36: note: in expansion of macro
>'SYSCALL_DEFINEx'
>#define SYSCALL_DEFINE4(name, ...) SYSCALL_DEFINEx(4, _##name,
>__VA_ARGS__)
>                                    ^~~~~~~~~~~~~~~
>kernel/pid.c:649:1: note: in expansion of macro 'SYSCALL_DEFINE4'
> SYSCALL_DEFINE4(pidfd_open, pid_t, pid, int, fd, int, pidfd,
> ^~~~~~~~~~~~~~~
>kernel/pid.c:663:7: error: 'procfd' undeclared (first use in this
>function); did you mean 'pidfd'?
>   if (procfd != -1 || pidfd != -1)
>       ^~~~~~
>       pidfd
>kernel/pid.c:663:7: note: each undeclared identifier is reported only
>once for each function it appears in
>
>Caused by commit
>
>  9170fba40db0 ("pid: add pidfd_open()")
>
>I have used the version of the pifd tree from next-20190327 for today.
>
>Please do *not* use linux-next as a development tree (I have seen
>several
>different version of this code over the past few days :-().  The rules
>for linux-next included code include:
>"
>     * posted to the relevant mailing list,
>     * reviewed by you (or another maintainer of your subsystem tree),
>     * successfully unit tested, and 
>     * destined for the current or next Linux merge window.
>
>Basically, this should be just what you would send to Linus (or ask him
>to fetch)."

Yeah, that should not have ended up there.
This was caused by a faulty regex in my push script that pushes to different servers.
One of them always builds a kernel and runs the tests that come with all patches on the for-next and work branches. 

Sorry about that. Should be fixed now.
Christian

Re: linux-next: build failure after merge of the pidfd tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 191 bytes --]

Hi Christian,

On Thu, 28 Mar 2019 03:28:25 +0100 Christian Brauner <christian@brauner.io> wrote:
>
> Sorry about that. Should be fixed now.

Thanks.

-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: linux-next: build failure after merge of the pidfd tree

[Christian Brauner]

On Tue, May 12, 2020 at 12:20:24PM +0200, Christian Brauner wrote:
> On Tue, May 12, 2020 at 08:18:11PM +1000, Stephen Rothwell wrote:
> > Hi all,
> > 
> > After merging the pidfd tree, today's linux-next build (x86_64
> > allnoconfig) failed like this:
> > 
> > fs/nsfs.c:232:6: error: redefinition of 'proc_ns_file'
> >   232 | bool proc_ns_file(const struct file *file)
> >       |      ^~~~~~~~~~~~
> > In file included from fs/nsfs.c:6:
> > include/linux/proc_fs.h:194:20: note: previous definition of 'proc_ns_file' was here
> >   194 | static inline bool proc_ns_file(const struct file *file)
> >       |                    ^~~~~~~~~~~~
> > 
> > Caused by commit
> > 
> >   1e76b8ad203a ("nsproxy: attach to namespaces via pidfds")
> > 
> > I have applied the following hack for today:
> 
> Thanks for spotting this. I'll fix this now.

diff --git a/include/linux/proc_fs.h b/include/linux/proc_fs.h
index acfd5012db4e..592a6e47b235 100644
--- a/include/linux/proc_fs.h
+++ b/include/linux/proc_fs.h
@@ -70,7 +70,7 @@ struct proc_dir_entry *proc_create_single_data(const char *name, umode_t mode,
                int (*show)(struct seq_file *, void *), void *data);
 #define proc_create_single(name, mode, parent, show) \
        proc_create_single_data(name, mode, parent, show, NULL)
-
+
 extern struct proc_dir_entry *proc_create_data(const char *, umode_t,
                                               struct proc_dir_entry *,
                                               const struct proc_ops *,
@@ -104,7 +104,6 @@ struct proc_dir_entry *proc_create_net_single_write(const char *name, umode_t mo
                                                    proc_write_t write,
                                                    void *data);
 extern struct pid *tgid_pidfd_to_pid(const struct file *file);
-extern bool proc_ns_file(const struct file *file);

 #ifdef CONFIG_PROC_PID_ARCH_STATUS
 /*
@@ -160,11 +159,6 @@ static inline struct pid *tgid_pidfd_to_pid(const struct file *file)
        return ERR_PTR(-EBADF);
 }

-static inline bool proc_ns_file(const struct file *file)
-{
-       return false;
-}
-
 #endif /* CONFIG_PROC_FS */

 struct net;
@@ -185,4 +179,6 @@ static inline struct pid_namespace *proc_pid_ns(const struct inode *inode)
        return inode->i_sb->s_fs_info;
 }

+bool proc_ns_file(const struct file *file);
+
 #endif /* _LINUX_PROC_FS_H */

should fix it cleanly.

Thanks, I've added an allnoconfig to my local tests now!

Christian

Re: linux-next: build failure after merge of the pidfd tree

[Christian Brauner]

On Tue, May 12, 2020 at 08:18:11PM +1000, Stephen Rothwell wrote:
> Hi all,
> 
> After merging the pidfd tree, today's linux-next build (x86_64
> allnoconfig) failed like this:
> 
> fs/nsfs.c:232:6: error: redefinition of 'proc_ns_file'
>   232 | bool proc_ns_file(const struct file *file)
>       |      ^~~~~~~~~~~~
> In file included from fs/nsfs.c:6:
> include/linux/proc_fs.h:194:20: note: previous definition of 'proc_ns_file' was here
>   194 | static inline bool proc_ns_file(const struct file *file)
>       |                    ^~~~~~~~~~~~
> 
> Caused by commit
> 
>   1e76b8ad203a ("nsproxy: attach to namespaces via pidfds")
> 
> I have applied the following hack for today:

Thanks for spotting this. I'll fix this now.

Christian

> 
> From 07065344c6fb39c440b0de5f75842066bc97a675 Mon Sep 17 00:00:00 2001
> From: Stephen Rothwell <sfr@canb.auug.org.au>
> Date: Tue, 12 May 2020 20:08:51 +1000
> Subject: [PATCH] nsproxy: protect proc_ns_file() when CONFIG_PROC_FS is not set
> 
> Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
> ---
>  fs/nsfs.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/fs/nsfs.c b/fs/nsfs.c
> index 800c1d0eb0d0..9215ad7597d6 100644
> --- a/fs/nsfs.c
> +++ b/fs/nsfs.c
> @@ -229,10 +229,12 @@ int ns_get_name(char *buf, size_t size, struct task_struct *task,
>  	return res;
>  }
>  
> +#ifdef CONFIG_PROC_FS
>  bool proc_ns_file(const struct file *file)
>  {
>  	return file->f_op == &ns_file_operations;
>  }
> +#endif
>  
>  struct file *proc_ns_fget(int fd)
>  {
> -- 
> 2.26.2
> 
> -- 
> Cheers,
> Stephen Rothwell

linux-next: build failure after merge of the pidfd tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1406 bytes --]

Hi all,

After merging the pidfd tree, today's linux-next build (x86_64
allnoconfig) failed like this:

fs/nsfs.c:232:6: error: redefinition of 'proc_ns_file'
  232 | bool proc_ns_file(const struct file *file)
      |      ^~~~~~~~~~~~
In file included from fs/nsfs.c:6:
include/linux/proc_fs.h:194:20: note: previous definition of 'proc_ns_file' was here
  194 | static inline bool proc_ns_file(const struct file *file)
      |                    ^~~~~~~~~~~~

Caused by commit

  1e76b8ad203a ("nsproxy: attach to namespaces via pidfds")

I have applied the following hack for today:

From 07065344c6fb39c440b0de5f75842066bc97a675 Mon Sep 17 00:00:00 2001
From: Stephen Rothwell <sfr@canb.auug.org.au>
Date: Tue, 12 May 2020 20:08:51 +1000
Subject: [PATCH] nsproxy: protect proc_ns_file() when CONFIG_PROC_FS is not set

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
---
 fs/nsfs.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/fs/nsfs.c b/fs/nsfs.c
index 800c1d0eb0d0..9215ad7597d6 100644
--- a/fs/nsfs.c
+++ b/fs/nsfs.c
@@ -229,10 +229,12 @@ int ns_get_name(char *buf, size_t size, struct task_struct *task,
 	return res;
 }
 
+#ifdef CONFIG_PROC_FS
 bool proc_ns_file(const struct file *file)
 {
 	return file->f_op == &ns_file_operations;
 }
+#endif
 
 struct file *proc_ns_fget(int fd)
 {
-- 
2.26.2

-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]