linux-kernel - Re: [PATCH] filelock: add stubs for new functions when CONFIG_FILE

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <ZcAhOW0A8U3h2OTf@tissot.1015granger.net>
Date: Sun, 4 Feb 2024 18:43:53 -0500
From: Chuck Lever <chuck.lever@...cle.com>
To: Jeff Layton <jlayton@...nel.org>
Cc: Christian Brauner <brauner@...nel.org>, Al Viro <viro@...iv.linux.org.uk>,
        linux-kernel@...r.kernel.org, linux-fsdevel@...r.kernel.org,
        kernel test robot <lkp@...el.com>
Subject: Re: [PATCH] filelock: add stubs for new functions when
 CONFIG_FILE_LOCKING=n

On Sun, Feb 04, 2024 at 07:32:55AM -0500, Jeff Layton wrote:
> We recently added several functions to the file locking API. Add stubs
> for those functions for when CONFIG_FILE_LOCKING is set to n.
> 
> Fixes: 403594111407 ("filelock: add some new helper functions")
> Reported-by: kernel test robot <lkp@...el.com>
> Closes: https://lore.kernel.org/oe-kbuild-all/202402041412.6YvtlflL-lkp@intel.com/
> Signed-off-by: Jeff Layton <jlayton@...nel.org>
> ---
> Just a small follow-on fix for CONFIG_FILE_LOCKING=n builds for the
> file_lease split. Christian, it might be best to squash this into
> the patch it Fixes.
> 
> That said, I'm starting to wonder if we ought to just hardcode
> CONFIG_FILE_LOCKING to y. Does anyone ship kernels with it disabled? I
> guess maybe people with stripped-down embedded builds might?

One thing you might try is building a kernel with both settings
and compare the resulting object sizes.

CONFIG_FILE_LOCKING was added during the git era, actually, so we
have some reasonable archaeology available:

commit bfcd17a6c5529bc37234cfa720a047cf9397bcfc
Author:     Thomas Petazzoni <thomas.petazzoni@...e-electrons.com>
AuthorDate: Wed Aug 6 15:12:22 2008 +0200
Commit:     J. Bruce Fields <bfields@...ldses.org>
CommitDate: Mon Sep 29 17:56:57 2008 -0400

    Configure out file locking features
    
    This patch adds the CONFIG_FILE_LOCKING option which allows to remove
    support for advisory locks. With this patch enabled, the flock()
    system call, the F_GETLK, F_SETLK and F_SETLKW operations of fcntl()
    and NFS support are disabled. These features are not necessarly needed
    on embedded systems. It allows to save ~11 Kb of kernel code and data:
    
       text          data     bss     dec     hex filename
    1125436        118764  212992 1457192  163c28 vmlinux.old
    1114299        118564  212992 1445855  160fdf vmlinux
     -11137    -200       0  -11337   -2C49 +/-
    
    This patch has originally been written by Matt Mackall
    <mpm@...enic.com>, and is part of the Linux Tiny project.


Embedded folks might want to keep CONFIG_FILE_LOCKING.


> Another thought too: "locks_" as a prefix is awfully generic. Might it be
> better to rename these new functions with a "filelock_" prefix instead?
> That would better distinguish to the casual reader that this is dealing
> with a file_lock object. I'm happy to respin the set if that's the
> consensus.

"posix_lock" might be even better, but no-one likes to make function
names longer.


> ---
>  include/linux/filelock.h | 21 +++++++++++++++++++++
>  1 file changed, 21 insertions(+)
> 
> diff --git a/include/linux/filelock.h b/include/linux/filelock.h
> index 4a5ad26962c1..553d65a88048 100644
> --- a/include/linux/filelock.h
> +++ b/include/linux/filelock.h
> @@ -263,6 +263,27 @@ static inline int fcntl_getlease(struct file *filp)
>  	return F_UNLCK;
>  }
>  
> +static inline bool lock_is_unlock(struct file_lock *fl)
> +{
> +	return false;
> +}
> +
> +static inline bool lock_is_read(struct file_lock *fl)
> +{
> +	return false;
> +}
> +
> +static inline bool lock_is_write(struct file_lock *fl)
> +{
> +	return false;
> +}
> +
> +static inline void locks_wake_up(struct file_lock *fl)
> +{
> +}
> +
> +#define for_each_file_lock(_fl, _head)	while(false)
> +
>  static inline void
>  locks_free_lock_context(struct inode *inode)
>  {
> 
> ---
> base-commit: 1499e59af376949b062cdc039257f811f6c1697f
> change-id: 20240204-flsplit3-da666d82b7b4
> 
> Best regards,
> -- 
> Jeff Layton <jlayton@...nel.org>
> 
> 

-- 
Chuck Lever