[PATCH] fs: predict not reaching the limit in alloc_empty_file()

[PATCH] fs: predict not reaching the limit in alloc_empty_file()

[Mateusz Guzik]

Eliminates a jump over a call to capable() in the common case.

By default the limit is not even set, in which case the check can't even
fail to begin with. It remains unset at least on Debian and Ubuntu.
For this cases this can probably become a static branch instead.

In the meantime tidy it up.

I note the check separate from the bump makes the entire thing racy.

Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
---
 fs/file_table.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/fs/file_table.c b/fs/file_table.c
index 5b4cc9da1344..c04ed94cdc4b 100644
--- a/fs/file_table.c
+++ b/fs/file_table.c
@@ -221,7 +221,8 @@ struct file *alloc_empty_file(int flags, const struct cred *cred)
 	/*
 	 * Privileged users can go above max_files
 	 */
-	if (get_nr_files() >= files_stat.max_files && !capable(CAP_SYS_ADMIN)) {
+	if (unlikely(get_nr_files() >= files_stat.max_files) &&
+	    !capable(CAP_SYS_ADMIN)) {
 		/*
 		 * percpu_counters are inaccurate.  Do an expensive check before
 		 * we go and fail.
-- 
2.43.0

Re: [PATCH] fs: predict not reaching the limit in alloc_empty_file()

[Christian Brauner]

On Wed, 19 Mar 2025 13:49:23 +0100, Mateusz Guzik wrote:
> Eliminates a jump over a call to capable() in the common case.
> 
> By default the limit is not even set, in which case the check can't even
> fail to begin with. It remains unset at least on Debian and Ubuntu.
> For this cases this can probably become a static branch instead.
> 
> In the meantime tidy it up.
> 
> [...]

Applied to the vfs-6.15.misc branch of the vfs/vfs.git tree.
Patches in the vfs-6.15.misc branch should appear in linux-next soon.

Please report any outstanding bugs that were missed during review in a
new review to the original patch series allowing us to drop it.

It's encouraged to provide Acked-bys and Reviewed-bys even though the
patch has now been applied. If possible patch trailers will be updated.

Note that commit hashes shown below are subject to change due to rebase,
trailer updates or similar. If in doubt, please check the listed branch.

tree:   https://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs.git
branch: vfs-6.15.misc

[1/1] fs: predict not reaching the limit in alloc_empty_file()
      https://git.kernel.org/vfs/vfs/c/9ae7e5a1cd17

Re: [PATCH] fs: predict not reaching the limit in alloc_empty_file()

[Jan Kara]

On Wed 19-03-25 13:49:23, Mateusz Guzik wrote:
> Eliminates a jump over a call to capable() in the common case.
> 
> By default the limit is not even set, in which case the check can't even
> fail to begin with. It remains unset at least on Debian and Ubuntu.
> For this cases this can probably become a static branch instead.
> 
> In the meantime tidy it up.
> 
> I note the check separate from the bump makes the entire thing racy.
> 
> Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>

Looks good. Feel free to add:

Reviewed-by: Jan Kara <jack@suse.cz>

								Honza

> ---
>  fs/file_table.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/file_table.c b/fs/file_table.c
> index 5b4cc9da1344..c04ed94cdc4b 100644
> --- a/fs/file_table.c
> +++ b/fs/file_table.c
> @@ -221,7 +221,8 @@ struct file *alloc_empty_file(int flags, const struct cred *cred)
>  	/*
>  	 * Privileged users can go above max_files
>  	 */
> -	if (get_nr_files() >= files_stat.max_files && !capable(CAP_SYS_ADMIN)) {
> +	if (unlikely(get_nr_files() >= files_stat.max_files) &&
> +	    !capable(CAP_SYS_ADMIN)) {
>  		/*
>  		 * percpu_counters are inaccurate.  Do an expensive check before
>  		 * we go and fail.
> -- 
> 2.43.0
> 
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR