Re: [PATCH bpf-next] bpf: check flags in 'bpf_ringbuf_discard()' and 'bpf_ringbuf_submit()'

[Pedro Tammela <pctammela@gmail.com>]

Em seg., 29 de mar. de 2021 às 13:10, Song Liu <songliubraving@fb.com> escreveu:
>
>
>
> > On Mar 28, 2021, at 9:10 AM, Pedro Tammela <pctammela@gmail.com> wrote:
> >
> > The current code only checks flags in 'bpf_ringbuf_output()'.
> >
> > Signed-off-by: Pedro Tammela <pctammela@mojatatu.com>
> > ---
> > include/uapi/linux/bpf.h       |  8 ++++----
> > kernel/bpf/ringbuf.c           | 13 +++++++++++--
> > tools/include/uapi/linux/bpf.h |  8 ++++----
> > 3 files changed, 19 insertions(+), 10 deletions(-)
> >
> > diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h
> > index 100cb2e4c104..232b5e5dd045 100644
> > --- a/include/uapi/linux/bpf.h
> > +++ b/include/uapi/linux/bpf.h
> > @@ -4073,7 +4073,7 @@ union bpf_attr {
> >  *            Valid pointer with *size* bytes of memory available; NULL,
> >  *            otherwise.
> >  *
> > - * void bpf_ringbuf_submit(void *data, u64 flags)
> > + * int bpf_ringbuf_submit(void *data, u64 flags)
>
> This should be "long" instead of "int".
>
> >  *    Description
> >  *            Submit reserved ring buffer sample, pointed to by *data*.
> >  *            If **BPF_RB_NO_WAKEUP** is specified in *flags*, no notification
> > @@ -4083,9 +4083,9 @@ union bpf_attr {
> >  *            If **BPF_RB_FORCE_WAKEUP** is specified in *flags*, notification
> >  *            of new data availability is sent unconditionally.
> >  *    Return
> > - *           Nothing. Always succeeds.
> > + *           0 on success, or a negative error in case of failure.
> >  *
> > - * void bpf_ringbuf_discard(void *data, u64 flags)
> > + * int bpf_ringbuf_discard(void *data, u64 flags)
>
> Ditto. And same for tools/include/uapi/linux/bpf.h
>
> >  *    Description
> >  *            Discard reserved ring buffer sample, pointed to by *data*.
> >  *            If **BPF_RB_NO_WAKEUP** is specified in *flags*, no notification
> > @@ -4095,7 +4095,7 @@ union bpf_attr {
> >  *            If **BPF_RB_FORCE_WAKEUP** is specified in *flags*, notification
> >  *            of new data availability is sent unconditionally.
> >  *    Return
> > - *           Nothing. Always succeeds.
> > + *           0 on success, or a negative error in case of failure.
> >  *
> >  * u64 bpf_ringbuf_query(void *ringbuf, u64 flags)
> >  *    Description
> > diff --git a/kernel/bpf/ringbuf.c b/kernel/bpf/ringbuf.c
> > index f25b719ac786..f76dafe2427e 100644
> > --- a/kernel/bpf/ringbuf.c
> > +++ b/kernel/bpf/ringbuf.c
> > @@ -397,26 +397,35 @@ static void bpf_ringbuf_commit(void *sample, u64 flags, bool discard)
> >
> > BPF_CALL_2(bpf_ringbuf_submit, void *, sample, u64, flags)
> > {
> > +     if (unlikely(flags & ~(BPF_RB_NO_WAKEUP | BPF_RB_FORCE_WAKEUP)))
> > +             return -EINVAL;
>
> We can move this check to bpf_ringbuf_commit().

I don't believe we can because in 'bpf_ringbuf_output()' the flag
checking in 'bpf_ringbuf_commit()' is already
too late.

>
> Thanks,
> Song
>
> [...]

Pedro