Re: [PATCH v2 bpf-next 1/6] bpf: add a __btf_get_by_fd helper

[Anton Protopopov <aspsk@isovalent.com>]

On 24/11/25 05:31PM, Alexei Starovoitov wrote:
> On Tue, Nov 19, 2024 at 2:17 AM Anton Protopopov <aspsk@isovalent.com> wrote:
> >
> > Add a new helper to get a pointer to a struct btf from a file
> > descriptor which doesn't increase a refcount.
> >
> > Signed-off-by: Anton Protopopov <aspsk@isovalent.com>
> > ---
> >  include/linux/btf.h | 13 +++++++++++++
> >  kernel/bpf/btf.c    | 13 ++++---------
> >  2 files changed, 17 insertions(+), 9 deletions(-)
> >
> > diff --git a/include/linux/btf.h b/include/linux/btf.h
> > index 4214e76c9168..050051a578a8 100644
> > --- a/include/linux/btf.h
> > +++ b/include/linux/btf.h
> > @@ -4,6 +4,7 @@
> >  #ifndef _LINUX_BTF_H
> >  #define _LINUX_BTF_H 1
> >
> > +#include <linux/file.h>
> >  #include <linux/types.h>
> >  #include <linux/bpfptr.h>
> >  #include <linux/bsearch.h>
> > @@ -143,6 +144,18 @@ void btf_get(struct btf *btf);
> >  void btf_put(struct btf *btf);
> >  const struct btf_header *btf_header(const struct btf *btf);
> >  int btf_new_fd(const union bpf_attr *attr, bpfptr_t uattr, u32 uattr_sz);
> > +
> > +static inline struct btf *__btf_get_by_fd(struct fd f)
> > +{
> > +       if (fd_empty(f))
> > +               return ERR_PTR(-EBADF);
> > +
> > +       if (unlikely(fd_file(f)->f_op != &btf_fops))
> > +               return ERR_PTR(-EINVAL);
> > +
> > +       return fd_file(f)->private_data;
> > +}
> 
> Maybe let's call it __btf_get() and place it next to __bpf_map_get() ?
> So names and function bodies are directly comparable?

I named it so because the corresponding helper which is taking a ref is named
btf_get_by_fd(). And btf_get() is actually increasing a refcnt. In the bpf_map
case naming is a bit different (and also not super-consistent,
bpf_map_inc/bpf_map_put to +- refcnt). Do you want me to make names more
consistent, globally?