From: Yonghong Song <yonghong.song@linux.dev>
To: bpf@vger.kernel.org
Cc: Alexei Starovoitov <ast@kernel.org>,
Andrii Nakryiko <andrii@kernel.org>,
Daniel Borkmann <daniel@iogearbox.net>,
Eduard Zingerman <eddyz87@gmail.com>,
kernel-team@fb.com
Subject: [PATCH bpf-next 05/12] bpf: Account R2 of register-pair returns in live register analysis
Date: Wed, 8 Jul 2026 13:10:05 -0700 [thread overview]
Message-ID: <20260708201005.2158063-1-yonghong.song@linux.dev> (raw)
In-Reply-To: <20260708200939.2153664-1-yonghong.song@linux.dev>
A BPF_EXIT of a subprogram returning a value larger than 8 bytes (a
struct/union or an __int128) reads R2 as well as R0, since the second half
of the return value is passed back in R2. compute_insn_live_regs() only
marked R0 used at exit, so the callee's R2 could be considered dead and
cleaned from checkpointed states, which would allow unsound state pruning.
Mark R2 used at a BPF_EXIT when the enclosing subprogram returns via the
R0:R2 register pair. The call site already treats the caller-saved
registers R0-R5 as clobbered, so R2 as a result register is covered there.
For example, for a subprogram returning an __int128, the live-registers
dump (log level 2) shows both R0 and R2 live before the callee's exit
(insn 5):
Live regs before insn:
0: .12345.... (85) call pc+2
1: ..2....... (bf) r0 = r2
2: 0......... (95) exit
3: .......... (b7) r0 = 1
4: 0......... (b7) r2 = 2
5: 0.2....... (95) exit
Signed-off-by: Yonghong Song <yonghong.song@linux.dev>
---
kernel/bpf/liveness.c | 13 ++++++++++---
1 file changed, 10 insertions(+), 3 deletions(-)
diff --git a/kernel/bpf/liveness.c b/kernel/bpf/liveness.c
index 0aadfbae0acc..0977d32fcf01 100644
--- a/kernel/bpf/liveness.c
+++ b/kernel/bpf/liveness.c
@@ -2062,7 +2062,8 @@ struct insn_live_regs {
/* Compute info->{use,def} fields for the instruction */
static void compute_insn_live_regs(struct bpf_verifier_env *env,
struct bpf_insn *insn,
- struct insn_live_regs *info)
+ struct insn_live_regs *info,
+ int subprog)
{
struct bpf_call_summary cs;
u8 class = BPF_CLASS(insn->code);
@@ -2175,6 +2176,8 @@ static void compute_insn_live_regs(struct bpf_verifier_env *env,
case BPF_EXIT:
def = 0;
use = r0;
+ if (bpf_ret_reg_pair(env, subprog))
+ use |= BIT(BPF_REG_2);
break;
case BPF_CALL:
def = ALL_CALLER_SAVED_REGS;
@@ -2237,8 +2240,12 @@ int bpf_compute_live_registers(struct bpf_verifier_env *env)
goto out;
}
- for (i = 0; i < insn_cnt; ++i)
- compute_insn_live_regs(env, &insns[i], &state[i]);
+ for (i = 0, j = 0; i < insn_cnt; ++i) {
+ /* Advance to the subprog that contains instruction i. */
+ while (j + 1 < env->subprog_cnt && env->subprog_info[j + 1].start <= i)
+ j++;
+ compute_insn_live_regs(env, &insns[i], &state[i], j);
+ }
/* Forward pass: resolve stack access through FP-derived pointers */
err = bpf_compute_subprog_arg_access(env);
--
2.53.0-Meta
next prev parent reply other threads:[~2026-07-08 20:10 UTC|newest]
Thread overview: 56+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-07-08 20:09 [PATCH bpf-next 00/12] bpf: Support 16-byte return values in the R0:R2 register pair Yonghong Song
2026-07-08 20:09 ` [PATCH bpf-next 01/12] bpf: Factor check_global_ret_scalar_reg() out of the global return check Yonghong Song
2026-07-08 20:09 ` [PATCH bpf-next 02/12] bpf: Add verifier support for 16-byte returns in R0:R2 Yonghong Song
2026-07-08 21:17 ` bot+bpf-ci
2026-07-09 16:40 ` Yonghong Song
2026-07-09 22:21 ` Eduard Zingerman
2026-07-10 4:57 ` Yonghong Song
2026-07-08 20:09 ` [PATCH bpf-next 03/12] bpf: Wire up JIT support for 16-byte kfunc returns Yonghong Song
2026-07-10 0:06 ` Eduard Zingerman
2026-07-08 20:10 ` [PATCH bpf-next 04/12] bpf: Track R2 of register-pair returns in precision backtracking Yonghong Song
2026-07-08 21:10 ` bot+bpf-ci
2026-07-09 16:41 ` Yonghong Song
2026-07-10 0:31 ` Eduard Zingerman
2026-07-08 20:10 ` Yonghong Song [this message]
2026-07-08 21:10 ` [PATCH bpf-next 05/12] bpf: Account R2 of register-pair returns in live register analysis bot+bpf-ci
2026-07-09 21:07 ` Eduard Zingerman
2026-07-10 5:06 ` Yonghong Song
2026-07-08 20:10 ` [PATCH bpf-next 06/12] bpf: Force JIT for programs using the R0:R2 register pair Yonghong Song
2026-07-08 21:10 ` bot+bpf-ci
2026-07-09 3:15 ` Leon Hwang
2026-07-09 16:44 ` Yonghong Song
2026-07-09 20:30 ` Eduard Zingerman
2026-07-10 5:07 ` Yonghong Song
2026-07-09 22:27 ` Eduard Zingerman
2026-07-10 5:09 ` Yonghong Song
2026-07-08 20:10 ` [PATCH bpf-next 07/12] bpf: Reject >8 byte return values on return-reading trampoline paths Yonghong Song
2026-07-09 3:16 ` Leon Hwang
2026-07-09 16:52 ` Yonghong Song
2026-07-10 2:01 ` Leon Hwang
2026-07-09 23:08 ` Eduard Zingerman
2026-07-10 2:02 ` Leon Hwang
2026-07-10 5:27 ` Yonghong Song
2026-07-10 5:12 ` Yonghong Song
2026-07-08 20:10 ` [PATCH bpf-next 08/12] bpf: Enable 16-byte aggregate return types Yonghong Song
2026-07-08 20:28 ` sashiko-bot
2026-07-09 19:40 ` Yonghong Song
2026-07-10 0:45 ` Eduard Zingerman
2026-07-10 5:29 ` Yonghong Song
2026-07-08 20:10 ` [PATCH bpf-next 09/12] selftests/bpf: Add C tests for 16-byte returns in R0:R2 Yonghong Song
2026-07-08 20:24 ` sashiko-bot
2026-07-09 19:51 ` Yonghong Song
2026-07-08 21:10 ` bot+bpf-ci
2026-07-09 19:54 ` Yonghong Song
2026-07-10 1:19 ` Eduard Zingerman
2026-07-10 5:32 ` Yonghong Song
2026-07-08 20:10 ` [PATCH bpf-next 10/12] selftests/bpf: Add inline-asm and subprog tests for R0:R2 returns Yonghong Song
2026-07-08 20:28 ` sashiko-bot
2026-07-09 19:57 ` Yonghong Song
2026-07-10 1:38 ` Eduard Zingerman
2026-07-10 5:35 ` Yonghong Song
2026-07-08 20:10 ` [PATCH bpf-next 11/12] selftests/bpf: Cover tracing on >8 and <=16 byte return targets Yonghong Song
2026-07-08 20:10 ` [PATCH bpf-next 12/12] Documentation/bpf: Document 16-byte kfunc return values in R0:R2 Yonghong Song
2026-07-10 0:56 ` [PATCH bpf-next 00/12] bpf: Support 16-byte return values in the R0:R2 register pair Eduard Zingerman
2026-07-10 6:00 ` Yonghong Song
2026-07-10 6:13 ` Eduard Zingerman
2026-07-10 15:18 ` Yonghong Song
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=20260708201005.2158063-1-yonghong.song@linux.dev \
--to=yonghong.song@linux.dev \
--cc=andrii@kernel.org \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=eddyz87@gmail.com \
--cc=kernel-team@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