[06/14] bpf: Add bpf_kallsyms_tree tree

The bpf_tree is used both for kallsyms iterations and searching
for exception tables of bpf programs, which is needed only for
bpf programs.

Adding bpf_kallsyms_tree that will hold symbols for all bpf_prog,
bpf_trampoline and bpf_dispatcher objects and keeping bpf_tree
only for bpf_prog objects exception tables search to keep it fast.

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
---
 include/linux/bpf.h |  1 +
 kernel/bpf/core.c   | 60 ++++++++++++++++++++++++++++++++++++++++-----
 2 files changed, 55 insertions(+), 6 deletions(-)

On Sat, Feb 8, 2020 at 7:43 AM Jiri Olsa <jolsa@kernel.org> wrote:
>
> The bpf_tree is used both for kallsyms iterations and searching
> for exception tables of bpf programs, which is needed only for
> bpf programs.
>
> Adding bpf_kallsyms_tree that will hold symbols for all bpf_prog,
> bpf_trampoline and bpf_dispatcher objects and keeping bpf_tree
> only for bpf_prog objects exception tables search to keep it fast.
>
> Signed-off-by: Jiri Olsa <jolsa@kernel.org>
> ---
>  include/linux/bpf.h |  1 +
>  kernel/bpf/core.c   | 60 ++++++++++++++++++++++++++++++++++++++++-----
>  2 files changed, 55 insertions(+), 6 deletions(-)
>
> diff --git a/include/linux/bpf.h b/include/linux/bpf.h
> index da67ca3afa2f..151d7b1c8435 100644
> --- a/include/linux/bpf.h
> +++ b/include/linux/bpf.h
> @@ -468,6 +468,7 @@ struct bpf_ksym {
>         unsigned long            end;
>         char                     name[KSYM_NAME_LEN];
>         struct list_head         lnode;
> +       struct latch_tree_node   tnode;
>  };
>
>  enum bpf_tramp_prog_type {
> diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
> index b9b7077e60f3..1daa72341450 100644
> --- a/kernel/bpf/core.c
> +++ b/kernel/bpf/core.c
> @@ -606,8 +606,46 @@ static const struct latch_tree_ops bpf_tree_ops = {
>         .comp   = bpf_tree_comp,
>  };
>
> +static __always_inline unsigned long
> +bpf_get_ksym_start(struct latch_tree_node *n)

I thought static functions are never marked as inline in kernel
sources. Are there some special cases when its ok/necessary?

> +{
> +       const struct bpf_ksym *ksym;
> +
> +       ksym = container_of(n, struct bpf_ksym, tnode);
> +       return ksym->start;
> +}
> +
> +static __always_inline bool
> +bpf_ksym_tree_less(struct latch_tree_node *a,
> +                  struct latch_tree_node *b)
> +{
> +       return bpf_get_ksym_start(a) < bpf_get_ksym_start(b);
> +}
> +
> +static __always_inline int
> +bpf_ksym_tree_comp(void *key, struct latch_tree_node *n)
> +{
> +       unsigned long val = (unsigned long)key;
> +       const struct bpf_ksym *ksym;
> +
> +       ksym = container_of(n, struct bpf_ksym, tnode);
> +
> +       if (val < ksym->start)
> +               return -1;
> +       if (val >= ksym->end)
> +               return  1;
> +
> +       return 0;
> +}
> +
> +static const struct latch_tree_ops bpf_kallsyms_tree_ops = {

Given all the helper functions use bpf_ksym_tree and bpf_ksym
(bpf_ksym_find) prefixes, call this bpf_ksym_tree_ops?

> +       .less   = bpf_ksym_tree_less,
> +       .comp   = bpf_ksym_tree_comp,
> +};
> +
>  static DEFINE_SPINLOCK(bpf_lock);
>  static LIST_HEAD(bpf_kallsyms);
> +static struct latch_tree_root bpf_kallsyms_tree __cacheline_aligned;

same as above, bpf_ksym_tree for consistency?

>  static struct latch_tree_root bpf_tree __cacheline_aligned;
>
>  static void bpf_prog_ksym_node_add(struct bpf_prog_aux *aux)
> @@ -615,6 +653,7 @@ static void bpf_prog_ksym_node_add(struct bpf_prog_aux *aux)
>         WARN_ON_ONCE(!list_empty(&aux->ksym.lnode));
>         list_add_tail_rcu(&aux->ksym.lnode, &bpf_kallsyms);
>         latch_tree_insert(&aux->ksym_tnode, &bpf_tree, &bpf_tree_ops);
> +       latch_tree_insert(&aux->ksym.tnode, &bpf_kallsyms_tree, &bpf_kallsyms_tree_ops);
>  }
>

[...]

On Tue, Feb 11, 2020 at 10:21:10AM -0800, Andrii Nakryiko wrote:
> On Sat, Feb 8, 2020 at 7:43 AM Jiri Olsa <jolsa@kernel.org> wrote:
> >
> > The bpf_tree is used both for kallsyms iterations and searching
> > for exception tables of bpf programs, which is needed only for
> > bpf programs.
> >
> > Adding bpf_kallsyms_tree that will hold symbols for all bpf_prog,
> > bpf_trampoline and bpf_dispatcher objects and keeping bpf_tree
> > only for bpf_prog objects exception tables search to keep it fast.
> >
> > Signed-off-by: Jiri Olsa <jolsa@kernel.org>
> > ---
> >  include/linux/bpf.h |  1 +
> >  kernel/bpf/core.c   | 60 ++++++++++++++++++++++++++++++++++++++++-----
> >  2 files changed, 55 insertions(+), 6 deletions(-)
> >
> > diff --git a/include/linux/bpf.h b/include/linux/bpf.h
> > index da67ca3afa2f..151d7b1c8435 100644
> > --- a/include/linux/bpf.h
> > +++ b/include/linux/bpf.h
> > @@ -468,6 +468,7 @@ struct bpf_ksym {
> >         unsigned long            end;
> >         char                     name[KSYM_NAME_LEN];
> >         struct list_head         lnode;
> > +       struct latch_tree_node   tnode;
> >  };
> >
> >  enum bpf_tramp_prog_type {
> > diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
> > index b9b7077e60f3..1daa72341450 100644
> > --- a/kernel/bpf/core.c
> > +++ b/kernel/bpf/core.c
> > @@ -606,8 +606,46 @@ static const struct latch_tree_ops bpf_tree_ops = {
> >         .comp   = bpf_tree_comp,
> >  };
> >
> > +static __always_inline unsigned long
> > +bpf_get_ksym_start(struct latch_tree_node *n)
> 
> I thought static functions are never marked as inline in kernel
> sources. Are there some special cases when its ok/necessary?

I followed the other latch tree ops functions and did not think
much about that.. will check

> 
> > +{
> > +       const struct bpf_ksym *ksym;
> > +
> > +       ksym = container_of(n, struct bpf_ksym, tnode);
> > +       return ksym->start;
> > +}
> > +
> > +static __always_inline bool
> > +bpf_ksym_tree_less(struct latch_tree_node *a,
> > +                  struct latch_tree_node *b)
> > +{
> > +       return bpf_get_ksym_start(a) < bpf_get_ksym_start(b);
> > +}
> > +
> > +static __always_inline int
> > +bpf_ksym_tree_comp(void *key, struct latch_tree_node *n)
> > +{
> > +       unsigned long val = (unsigned long)key;
> > +       const struct bpf_ksym *ksym;
> > +
> > +       ksym = container_of(n, struct bpf_ksym, tnode);
> > +
> > +       if (val < ksym->start)
> > +               return -1;
> > +       if (val >= ksym->end)
> > +               return  1;
> > +
> > +       return 0;
> > +}
> > +
> > +static const struct latch_tree_ops bpf_kallsyms_tree_ops = {
> 
> Given all the helper functions use bpf_ksym_tree and bpf_ksym
> (bpf_ksym_find) prefixes, call this bpf_ksym_tree_ops?

right, should be bpf_ksym_tree_ops as you said

> 
> > +       .less   = bpf_ksym_tree_less,
> > +       .comp   = bpf_ksym_tree_comp,
> > +};
> > +
> >  static DEFINE_SPINLOCK(bpf_lock);
> >  static LIST_HEAD(bpf_kallsyms);
> > +static struct latch_tree_root bpf_kallsyms_tree __cacheline_aligned;
> 
> same as above, bpf_ksym_tree for consistency?

right, thanks

jirka