Re: [PATCHv5 bpf-next 01/28] ftrace: Add ftrace_hash_count function
Jiri Olsa <[email protected]>
| Newsgroups | dev.linux.lists.sashiko,org.kernel.vger.bpf |
|---|---|
| Message-ID | <aec7Vqu2wVDbx8QU@krava> |
On Sat, Apr 18, 2026 at 12:58:20AM +0000, [email protected] wrote: > Sashiko AI review found 1 potential issue(s): > - [Low] ftrace_hash_count() is declared under CONFIG_DYNAMIC_FTRACE but defined under CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS, leading to potential linker errors. > -- > > commit 62c46ea6c0c10abc0df16b7c684809496d439e2f > Author: Jiri Olsa <[email protected]> > > ftrace: Add ftrace_hash_count function > > This commit adds an external ftrace_hash_count function so the hash > count can be obtained outside of the ftrace object. > > > diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h > > index 28b30c6f10314..9d9ae1bc692ae 100644 > > --- a/include/linux/ftrace.h > > +++ b/include/linux/ftrace.h > > @@ -415,6 +415,7 @@ struct ftrace_hash *alloc_ftrace_hash(int size_bits); > > void free_ftrace_hash(struct ftrace_hash *hash); > > struct ftrace_func_entry *add_ftrace_hash_entry_direct(struct ftrace_hash *hash, > > unsigned long ip, unsigned long direct); > > +unsigned long ftrace_hash_count(struct ftrace_hash *hash); > > > > /* The hash used to know what functions callbacks trace */ > > struct ftrace_ops_hash { > > [ ... ] > > > diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c > > index b2611de3f5943..57ab01fd00bda 100644 > > --- a/kernel/trace/ftrace.c > > +++ b/kernel/trace/ftrace.c > > @@ -6288,11 +6288,16 @@ int modify_ftrace_direct(struct ftrace_ops *ops, unsigned long addr) > > } > > EXPORT_SYMBOL_GPL(modify_ftrace_direct); > > > > -static unsigned long hash_count(struct ftrace_hash *hash) > > +static inline unsigned long hash_count(struct ftrace_hash *hash) > > { > > return hash ? hash->count : 0; > > } > > > > +unsigned long ftrace_hash_count(struct ftrace_hash *hash) > > +{ > > + return hash_count(hash); > > +} > > + > > If a kernel is configured with CONFIG_DYNAMIC_FTRACE=y but > CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS=n, will this result in a linker error? > > The declaration in include/linux/ftrace.h is located under > CONFIG_DYNAMIC_FTRACE, but this definition in kernel/trace/ftrace.c is placed > inside a CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS block. > > Code built-in to the kernel (like BPF or tracing extensions) that attempts > to use this new API could compile successfully but fail to link on > architectures that do not support direct ftrace calls. right, will check the compilation for such config jirka