[Devel] [PATCH VZ10 1/2] anon_inode: rework assertions

[Vladimir Riabchun]

On 7/27/26 14:22, Pavel Tikhomirov wrote:
> From: Christian Brauner <brauner at kernel.org>
> 
> Making anonymous inodes regular files comes with a lot of risk and
> regression potential as evidenced by a recent hickup in io_uring. We're
> better of continuing to not have them be regular files. Since we have
> S_ANON_INODE we can port all of our assertions easily.
> 
> Link: https://lore.kernel.org/20250702-work-fixes-v1-1-ff76ea589e33@kernel.org
> Fixes: cfd86ef7e8e7 ("anon_inode: use a proper mode internally")
> Acked-by: Jens Axboe <axboe at kernel.dk>
> Cc: stable at kernel.org
> Reported-by: Jens Axboe <axboe at kernel.dk>
> Signed-off-by: Christian Brauner <brauner at kernel.org>
> 
> (cherry picked from commit 1e7ab6f67824343ee3e96f100f0937c393749a8a)
> 
> This fixes CRIU dump of processes that use AIO, since CRIU does not
> expect S_IFREG on AIO inodes.
> 
> vz10 backport notes:
> - drop debug with missing VFS_WARN_ON_ONCE/VFS_BUG_ON_INODE

Any particular reason for this drop? These checks seem harmless and
dropping them may add some work during the rebase.

Other than that looks good.
Reviewed-by: Vladimir Riabchun <vladimir.riabchun at virtuozzo.com>

> - add fput(file) on the introduced error paths in do_open_execat()
> 
> Feature: ve: CRIU support
> https://virtuozzo.atlassian.net/browse/VSTOR-139264
> Signed-off-by: Pavel Tikhomirov <ptikhomirov at virtuozzo.com>
> ---
>   fs/exec.c  | 8 ++++++--
>   fs/libfs.c | 8 +++-----
>   2 files changed, 9 insertions(+), 7 deletions(-)
> 
> diff --git a/fs/exec.c b/fs/exec.c
> index dd5de1a4a9d6a..e861a54b491f8 100644
> --- a/fs/exec.c
> +++ b/fs/exec.c
> @@ -921,13 +921,17 @@ static struct file *do_open_execat(int fd, struct filename *name, int flags)
>   	if (IS_ERR(file))
>   		return file;
>   
> +	if (path_noexec(&file->f_path)) {
> +		fput(file);
> +		return ERR_PTR(-EACCES);
> +	}
> +
>   	/*
>   	 * In the past the regular type check was here. It moved to may_open() in
>   	 * 633fb6ac3980 ("exec: move S_ISREG() check earlier"). Since then it is
>   	 * an invariant that all non-regular files error out before we get here.
>   	 */
> -	if (WARN_ON_ONCE(!S_ISREG(file_inode(file)->i_mode)) ||
> -	    path_noexec(&file->f_path)) {
> +	if (WARN_ON_ONCE(!S_ISREG(file_inode(file)->i_mode))) {
>   		fput(file);
>   		return ERR_PTR(-EACCES);
>   	}
> diff --git a/fs/libfs.c b/fs/libfs.c
> index c4a305967619f..af191a388303c 100644
> --- a/fs/libfs.c
> +++ b/fs/libfs.c
> @@ -1652,12 +1652,10 @@ struct inode *alloc_anon_inode(struct super_block *s)
>   	 */
>   	inode->i_state = I_DIRTY;
>   	/*
> -	 * Historically anonymous inodes didn't have a type at all and
> -	 * userspace has come to rely on this. Internally they're just
> -	 * regular files but S_IFREG is masked off when reporting
> -	 * information to userspace.
> +	 * Historically anonymous inodes don't have a type at all and
> +	 * userspace has come to rely on this.
>   	 */
> -	inode->i_mode = S_IFREG | S_IRUSR | S_IWUSR;
> +	inode->i_mode = S_IRUSR | S_IWUSR;
>   	inode->i_uid = current_fsuid();
>   	inode->i_gid = current_fsgid();
>   	inode->i_flags |= S_PRIVATE | S_ANON_INODE;

-- 
Best regards, Riabchun Vladimir
Linux Kernel Developer, Virtuozzo