diff options
| author | Christian Brauner <brauner@kernel.org> | 2023-07-01 09:44:44 +0300 |
|---|---|---|
| committer | Linus Torvalds <torvalds@linux-foundation.org> | 2023-07-01 18:26:23 +0300 |
| commit | dd546618ba704be4f3724a11e5a194052c551f08 (patch) | |
| tree | ff54f98f4a99a521c1de628502955dc359d49d11 /kernel/pid_namespace.c | |
| parent | 408579cd627a15bd703fe3eeb8485fd02726e9d3 (diff) | |
| download | linux-dd546618ba704be4f3724a11e5a194052c551f08.tar.xz | |
pid: use struct_size_t() helper
Before commit d67790ddf021 ("overflow: Add struct_size_t() helper") only
struct_size() existed, which expects a valid pointer instance containing
the flexible array.
However, when we determine the default struct pid allocation size for
the associated kmem cache of a pid namespace we need to take the nesting
depth of the pid namespace into account without an variable instance
necessarily being available.
In commit b69f0aeb0689 ("pid: Replace struct pid 1-element array with
flex-array") we used to handle this the old fashioned way and cast NULL
to a struct pid pointer type. However, we do apparently have a dedicated
struct_size_t() helper for exactly this case. So switch to that.
Suggested-by: Kees Cook <keescook@chromium.org>
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Diffstat (limited to 'kernel/pid_namespace.c')
| -rw-r--r-- | kernel/pid_namespace.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/kernel/pid_namespace.c b/kernel/pid_namespace.c index 70a929784a5d..0bf44afe04dd 100644 --- a/kernel/pid_namespace.c +++ b/kernel/pid_namespace.c @@ -48,7 +48,7 @@ static struct kmem_cache *create_pid_cachep(unsigned int level) return kc; snprintf(name, sizeof(name), "pid_%u", level + 1); - len = struct_size((struct pid *)NULL, numbers, level + 1); + len = struct_size_t(struct pid, numbers, level + 1); mutex_lock(&pid_caches_mutex); /* Name collision forces to do allocation under mutex. */ if (!*pkc) |