BPF Archive on lore.kernel.org
 help / color / Atom feed
From: Andrii Nakryiko <andrii.nakryiko@gmail.com>
To: Jiri Olsa <jolsa@kernel.org>
Cc: Alexei Starovoitov <ast@kernel.org>,
	Daniel Borkmann <daniel@iogearbox.net>,
	Andrii Nakryiko <andriin@fb.com>,
	Networking <netdev@vger.kernel.org>, bpf <bpf@vger.kernel.org>,
	Song Liu <songliubraving@fb.com>, Yonghong Song <yhs@fb.com>,
	Martin KaFai Lau <kafai@fb.com>, David Miller <davem@redhat.com>,
	John Fastabend <john.fastabend@gmail.com>,
	Wenbo Zhang <ethercflow@gmail.com>,
	KP Singh <kpsingh@chromium.org>,
	Brendan Gregg <bgregg@netflix.com>,
	Florent Revest <revest@chromium.org>,
	Al Viro <viro@zeniv.linux.org.uk>
Subject: Re: [PATCH v8 bpf-next 06/13] bpf: Factor btf_struct_access function
Date: Tue, 28 Jul 2020 16:27:21 -0700
Message-ID: <CAEf4BzbS_JFW70Z_68hDtN4VTkLfohkR0PV0d9jCJRjZEhc01Q@mail.gmail.com> (raw)
In-Reply-To: <20200722211223.1055107-7-jolsa@kernel.org>

On Wed, Jul 22, 2020 at 2:13 PM Jiri Olsa <jolsa@kernel.org> wrote:
>
> Adding btf_struct_walk function that walks through the
> struct type + given offset and returns following values:
>
>   enum walk_return {
>        /* < 0 error */
>        walk_scalar = 0,
>        walk_ptr,
>        walk_struct,
>   };
>
> walk_scalar - when SCALAR_VALUE is found
> walk_ptr    - when pointer value is found, its ID is stored
>               in 'rid' output param
> walk_struct - when nested struct object is found, its ID is stored
>               in 'rid' output param
>
> It will be used in following patches to get all nested
> struct objects for given type and offset.
>
> The btf_struct_access now calls btf_struct_walk function,
> as long as it gets nested structs as return value.
>
> Signed-off-by: Jiri Olsa <jolsa@kernel.org>
> ---

It actually worked out to a pretty minimal changes to
btf_struct_access, I'm pleasantly surprised :))

Logic seems correct, just have few nits about naming and a bit more
safe handling in btf_struct_access, see below.


>  kernel/bpf/btf.c | 73 +++++++++++++++++++++++++++++++++++++++---------
>  1 file changed, 60 insertions(+), 13 deletions(-)
>
> diff --git a/kernel/bpf/btf.c b/kernel/bpf/btf.c
> index 841be6c49f11..1ab5fd5bf992 100644
> --- a/kernel/bpf/btf.c
> +++ b/kernel/bpf/btf.c
> @@ -3873,16 +3873,22 @@ bool btf_ctx_access(int off, int size, enum bpf_access_type type,
>         return true;
>  }
>
> -int btf_struct_access(struct bpf_verifier_log *log,
> -                     const struct btf_type *t, int off, int size,
> -                     enum bpf_access_type atype,
> -                     u32 *next_btf_id)
> +enum walk_return {
> +       /* < 0 error */
> +       walk_scalar = 0,
> +       walk_ptr,
> +       walk_struct,
> +};

let's keep enum values in ALL_CAPS? walk_return is also a bit generic,
maybe something like bpf_struct_walk_result?

> +
> +static int btf_struct_walk(struct bpf_verifier_log *log,
> +                          const struct btf_type *t, int off, int size,
> +                          u32 *rid)
>  {
>         u32 i, moff, mtrue_end, msize = 0, total_nelems = 0;
>         const struct btf_type *mtype, *elem_type = NULL;
>         const struct btf_member *member;
>         const char *tname, *mname;
> -       u32 vlen;
> +       u32 vlen, elem_id, mid;
>
>  again:
>         tname = __btf_name_by_offset(btf_vmlinux, t->name_off);
> @@ -3924,8 +3930,7 @@ int btf_struct_access(struct bpf_verifier_log *log,
>                         goto error;
>
>                 off = (off - moff) % elem_type->size;
> -               return btf_struct_access(log, elem_type, off, size, atype,
> -                                        next_btf_id);
> +               return btf_struct_walk(log, elem_type, off, size, rid);

oh, btw, this is a recursion in the kernel, let's fix that? I think it
could easily be just `goto again` here?
>
>  error:
>                 bpf_log(log, "access beyond struct %s at off %u size %u\n",
> @@ -3954,7 +3959,7 @@ int btf_struct_access(struct bpf_verifier_log *log,
>                          */
>                         if (off <= moff &&
>                             BITS_ROUNDUP_BYTES(end_bit) <= off + size)
> -                               return SCALAR_VALUE;
> +                               return walk_scalar;
>
>                         /* off may be accessing a following member
>                          *

[...]

> @@ -4066,11 +4080,10 @@ int btf_struct_access(struct bpf_verifier_log *log,
>                                         mname, moff, tname, off, size);
>                                 return -EACCES;
>                         }
> -
>                         stype = btf_type_skip_modifiers(btf_vmlinux, mtype->type, &id);
>                         if (btf_type_is_struct(stype)) {
> -                               *next_btf_id = id;
> -                               return PTR_TO_BTF_ID;
> +                               *rid = id;

nit: rid is a very opaque name, I find next_btf_id more appropriate
(even if it's meaning changes depending on walk_ptr vs walk_struct.

> +                               return walk_ptr;
>                         }
>                 }
>
> @@ -4087,12 +4100,46 @@ int btf_struct_access(struct bpf_verifier_log *log,
>                         return -EACCES;
>                 }
>
> -               return SCALAR_VALUE;
> +               return walk_scalar;
>         }
>         bpf_log(log, "struct %s doesn't have field at offset %d\n", tname, off);
>         return -EINVAL;
>  }
>
> +int btf_struct_access(struct bpf_verifier_log *log,
> +                     const struct btf_type *t, int off, int size,
> +                     enum bpf_access_type atype __maybe_unused,
> +                     u32 *next_btf_id)
> +{
> +       int err;
> +       u32 id;
> +
> +       do {
> +               err = btf_struct_walk(log, t, off, size, &id);
> +               if (err < 0)
> +                       return err;
> +
> +               /* We found the pointer or scalar on t+off,
> +                * we're done.
> +                */
> +               if (err == walk_ptr) {
> +                       *next_btf_id = id;
> +                       return PTR_TO_BTF_ID;
> +               }
> +               if (err == walk_scalar)
> +                       return SCALAR_VALUE;
> +
> +               /* We found nested struct, so continue the search
> +                * by diving in it. At this point the offset is
> +                * aligned with the new type, so set it to 0.
> +                */
> +               t = btf_type_by_id(btf_vmlinux, id);
> +               off = 0;

It's very easy to miss that this case corresponds to walk_struct here.
If someone in the future adds a 4th special value, it will be too easy
to forget to update this piece of logic. So when dealing with enums, I
generally prefer this approach:

switch (err) {
case walk_ptr:
    ...
case walk_scalar:
    ...
case walk_struct:
    ...
default: /* complain loudly here */
}

WDYT?

> +       } while (t);
> +
> +       return -EINVAL;
> +}
> +
>  int btf_resolve_helper_id(struct bpf_verifier_log *log,
>                           const struct bpf_func_proto *fn, int arg)
>  {
> --
> 2.25.4
>

  reply index

Thread overview: 40+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-07-22 21:12 [PATCH v8 bpf-next 00/13] bpf: Add d_path helper Jiri Olsa
2020-07-22 21:12 ` [PATCH v8 bpf-next 01/13] selftests/bpf: Fix resolve_btfids test Jiri Olsa
2020-07-28 20:27   ` Andrii Nakryiko
2020-07-29 20:06     ` Jiri Olsa
2020-07-29 21:38       ` Andrii Nakryiko
2020-07-22 21:12 ` [PATCH v8 bpf-next 02/13] tools resolve_btfids: Add support for set symbols Jiri Olsa
2020-07-28  0:53   ` Andrii Nakryiko
2020-07-28  9:35     ` Jiri Olsa
2020-07-22 21:12 ` [PATCH v8 bpf-next 03/13] bpf: Move btf_resolve_size into __btf_resolve_size Jiri Olsa
2020-07-28 20:29   ` Andrii Nakryiko
2020-07-22 21:12 ` [PATCH v8 bpf-next 04/13] bpf: Add elem_id pointer as argument to __btf_resolve_size Jiri Olsa
2020-07-28 20:30   ` Andrii Nakryiko
2020-07-22 21:12 ` [PATCH v8 bpf-next 05/13] bpf: Add type_id " Jiri Olsa
2020-07-22 21:12 ` [PATCH v8 bpf-next 06/13] bpf: Factor btf_struct_access function Jiri Olsa
2020-07-28 23:27   ` Andrii Nakryiko [this message]
2020-07-29 15:59     ` Jiri Olsa
2020-07-22 21:12 ` [PATCH v8 bpf-next 07/13] bpf: Add btf_struct_ids_match function Jiri Olsa
2020-07-28 23:35   ` Andrii Nakryiko
2020-07-29 16:04     ` Jiri Olsa
2020-07-29 17:51       ` Andrii Nakryiko
2020-07-29 18:55         ` Jiri Olsa
2020-07-22 21:12 ` [PATCH v8 bpf-next 08/13] bpf: Add BTF_SET_START/END macros Jiri Olsa
2020-07-28 19:39   ` Andrii Nakryiko
2020-07-29 11:54     ` Jiri Olsa
2020-07-22 21:12 ` [PATCH v8 bpf-next 09/13] bpf: Add d_path helper Jiri Olsa
2020-07-28 19:47   ` Andrii Nakryiko
2020-07-29 15:54     ` Jiri Olsa
2020-07-29 20:11   ` Al Viro
2020-07-30 10:22     ` Jiri Olsa
2020-07-29 20:17   ` Al Viro
2020-07-30 10:09     ` Jiri Olsa
2020-07-22 21:12 ` [PATCH v8 bpf-next 10/13] bpf: Update .BTF_ids section in btf.rst with sets info Jiri Olsa
2020-07-22 21:12 ` [PATCH v8 bpf-next 11/13] selftests/bpf: Add verifier test for d_path helper Jiri Olsa
2020-07-28 19:49   ` Andrii Nakryiko
2020-07-22 21:12 ` [PATCH v8 bpf-next 12/13] selftests/bpf: Add " Jiri Olsa
2020-07-28 19:53   ` Andrii Nakryiko
2020-07-29 11:25     ` Jiri Olsa
2020-07-22 21:12 ` [PATCH v8 bpf-next 13/13] selftests/bpf: Add set test to resolve_btfids Jiri Olsa
2020-07-28 19:56   ` Andrii Nakryiko
2020-07-29 15:34     ` Jiri Olsa

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=CAEf4BzbS_JFW70Z_68hDtN4VTkLfohkR0PV0d9jCJRjZEhc01Q@mail.gmail.com \
    --to=andrii.nakryiko@gmail.com \
    --cc=andriin@fb.com \
    --cc=ast@kernel.org \
    --cc=bgregg@netflix.com \
    --cc=bpf@vger.kernel.org \
    --cc=daniel@iogearbox.net \
    --cc=davem@redhat.com \
    --cc=ethercflow@gmail.com \
    --cc=john.fastabend@gmail.com \
    --cc=jolsa@kernel.org \
    --cc=kafai@fb.com \
    --cc=kpsingh@chromium.org \
    --cc=netdev@vger.kernel.org \
    --cc=revest@chromium.org \
    --cc=songliubraving@fb.com \
    --cc=viro@zeniv.linux.org.uk \
    --cc=yhs@fb.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

BPF Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/bpf/0 bpf/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 bpf bpf/ https://lore.kernel.org/bpf \
		bpf@vger.kernel.org
	public-inbox-index bpf

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.bpf


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git