From: Song Liu <song@kernel.org>
To: He Fengqing <hefengqing@huawei.com>
Cc: Alexei Starovoitov <ast@kernel.org>,
Daniel Borkmann <daniel@iogearbox.net>,
Andrii Nakryiko <andrii@kernel.org>,
Martin KaFai Lau <kafai@fb.com>, Song Liu <songliubraving@fb.com>,
Yonghong Song <yhs@fb.com>,
John Fastabend <john.fastabend@gmail.com>,
KP Singh <kpsingh@kernel.org>,
"David S . Miller" <davem@davemloft.net>,
Jakub Kicinski <kuba@kernel.org>,
Networking <netdev@vger.kernel.org>, bpf <bpf@vger.kernel.org>,
open list <linux-kernel@vger.kernel.org>
Subject: Re: [bpf-next 3/3] bpf: Fix a use after free in bpf_check()
Date: Wed, 7 Jul 2021 00:25:24 -0700 [thread overview]
Message-ID: <CAPhsuW7ssFzvS5-kdZa3tY-2EJk8QUdVpQCJYVBr+vD11JzrsQ@mail.gmail.com> (raw)
In-Reply-To: <20210707043811.5349-4-hefengqing@huawei.com>
On Tue, Jul 6, 2021 at 8:53 PM He Fengqing <hefengqing@huawei.com> wrote:
>
> In bpf_patch_insn_data, env->prog was input parameter of
> bpf_patch_insn_single function. bpf_patch_insn_single call
> bpf_prog_realloc to realloc ebpf prog. When we need to malloc new prog,
> bpf_prog_realloc will free the old prog, in this scenery is the
> env->prog.
> Then bpf_patch_insn_data function call adjust_insn_aux_data function, if
> adjust_insn_aux_data function return error, bpf_patch_insn_data will
> return NULL.
> In bpf_check->convert_ctx_accesses->bpf_patch_insn_data call chain, if
> bpf_patch_insn_data return NULL, env->prog has been freed in
> bpf_prog_realloc, then bpf_check will use the freed env->prog.
Besides "what is the bug", please also describe "how to fix it". For example,
add "Fix it by adding a free_old argument to bpf_prog_realloc(), and ...".
Also, for the subject of 0/3, it is better to say "fix potential
memory leak and ...".
>
> Signed-off-by: He Fengqing <hefengqing@huawei.com>
> ---
> include/linux/filter.h | 2 +-
> kernel/bpf/core.c | 9 ++++---
> kernel/bpf/verifier.c | 53 ++++++++++++++++++++++++++++++++----------
> net/core/filter.c | 2 +-
> 4 files changed, 49 insertions(+), 17 deletions(-)
>
> diff --git a/include/linux/filter.h b/include/linux/filter.h
> index f39e008a377d..ec11a5ae92c2 100644
> --- a/include/linux/filter.h
> +++ b/include/linux/filter.h
> @@ -881,7 +881,7 @@ void bpf_prog_jit_attempt_done(struct bpf_prog *prog);
> struct bpf_prog *bpf_prog_alloc(unsigned int size, gfp_t gfp_extra_flags);
> struct bpf_prog *bpf_prog_alloc_no_stats(unsigned int size, gfp_t gfp_extra_flags);
> struct bpf_prog *bpf_prog_realloc(struct bpf_prog *fp_old, unsigned int size,
> - gfp_t gfp_extra_flags);
> + gfp_t gfp_extra_flags, bool free_old);
> void __bpf_prog_free(struct bpf_prog *fp);
>
> static inline void bpf_prog_clone_free(struct bpf_prog *fp)
> diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
> index 49b0311f48c1..e5616bb1665b 100644
> --- a/kernel/bpf/core.c
> +++ b/kernel/bpf/core.c
> @@ -218,7 +218,7 @@ void bpf_prog_fill_jited_linfo(struct bpf_prog *prog,
> }
>
> struct bpf_prog *bpf_prog_realloc(struct bpf_prog *fp_old, unsigned int size,
> - gfp_t gfp_extra_flags)
> + gfp_t gfp_extra_flags, bool free_old)
> {
> gfp_t gfp_flags = GFP_KERNEL_ACCOUNT | __GFP_ZERO | gfp_extra_flags;
> struct bpf_prog *fp;
> @@ -238,7 +238,8 @@ struct bpf_prog *bpf_prog_realloc(struct bpf_prog *fp_old, unsigned int size,
> /* We keep fp->aux from fp_old around in the new
> * reallocated structure.
> */
> - bpf_prog_clone_free(fp_old);
> + if (free_old)
> + bpf_prog_clone_free(fp_old);
> }
>
> return fp;
> @@ -456,7 +457,7 @@ struct bpf_prog *bpf_patch_insn_single(struct bpf_prog *prog, u32 off,
> * last page could have large enough tailroom.
> */
> prog_adj = bpf_prog_realloc(prog, bpf_prog_size(insn_adj_cnt),
> - GFP_USER);
> + GFP_USER, false);
> if (!prog_adj)
> return ERR_PTR(-ENOMEM);
>
> @@ -1150,6 +1151,8 @@ struct bpf_prog *bpf_jit_blind_constants(struct bpf_prog *prog)
> return tmp;
> }
>
> + if (tmp != clone)
> + bpf_prog_clone_free(clone);
> clone = tmp;
> insn_delta = rewritten - 1;
>
> diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
> index 41109f49b724..e75b933f69e4 100644
> --- a/kernel/bpf/verifier.c
> +++ b/kernel/bpf/verifier.c
> @@ -11855,7 +11855,10 @@ static int opt_subreg_zext_lo32_rnd_hi32(struct bpf_verifier_env *env,
> new_prog = bpf_patch_insn_data(env, adj_idx, patch, patch_len);
> if (!new_prog)
> return -ENOMEM;
> - env->prog = new_prog;
> + if (new_prog != env->prog) {
> + bpf_prog_clone_free(env->prog);
> + env->prog = new_prog;
> + }
Can we move this check into bpf_patch_insn_data()?
> insns = new_prog->insnsi;
> aux = env->insn_aux_data;
> delta += patch_len - 1;
[...]
> diff --git a/net/core/filter.c b/net/core/filter.c
> index d70187ce851b..8a8d1a3ba5c2 100644
> --- a/net/core/filter.c
> +++ b/net/core/filter.c
> @@ -1268,7 +1268,7 @@ static struct bpf_prog *bpf_migrate_filter(struct bpf_prog *fp)
>
> /* Expand fp for appending the new filter representation. */
> old_fp = fp;
> - fp = bpf_prog_realloc(old_fp, bpf_prog_size(new_len), 0);
> + fp = bpf_prog_realloc(old_fp, bpf_prog_size(new_len), 0, true);
Can we add some logic here and not add free_old to bpf_prog_realloc()?
Thanks,
Song
next prev parent reply other threads:[~2021-07-07 7:25 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-07-07 4:38 [bpf-next 0/3] potential memleak and use after free in bpf verifier He Fengqing
2021-07-07 4:38 ` [bpf-next 1/3] bpf: Move bpf_prog_clone_free into filter.h file He Fengqing
2021-07-07 7:02 ` Song Liu
2021-07-07 4:38 ` [bpf-next 2/3] bpf: Fix a memory leak in an error handling path in 'bpf_patch_insn_data()' He Fengqing
2021-07-07 4:38 ` [bpf-next 3/3] bpf: Fix a use after free in bpf_check() He Fengqing
2021-07-07 7:25 ` Song Liu [this message]
2021-07-08 3:00 ` He Fengqing
2021-07-08 3:09 ` Alexei Starovoitov
2021-07-09 11:11 ` He Fengqing
2021-07-09 15:12 ` Alexei Starovoitov
2021-07-12 2:17 ` He Fengqing
2021-07-13 23:17 ` Alexei Starovoitov
2021-07-14 1:53 ` He Fengqing
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=CAPhsuW7ssFzvS5-kdZa3tY-2EJk8QUdVpQCJYVBr+vD11JzrsQ@mail.gmail.com \
--to=song@kernel.org \
--cc=andrii@kernel.org \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=davem@davemloft.net \
--cc=hefengqing@huawei.com \
--cc=john.fastabend@gmail.com \
--cc=kafai@fb.com \
--cc=kpsingh@kernel.org \
--cc=kuba@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=songliubraving@fb.com \
--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
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).