From: Harishankar Vishwanathan <harishankar.vishwanathan@gmail.com>
To: bpf@vger.kernel.org
Cc: Alexei Starovoitov <ast@kernel.org>,
Daniel Borkmann <daniel@iogearbox.net>,
Andrii Nakryiko <andrii@kernel.org>,
Eduard Zingerman <eddyz87@gmail.com>,
Paul Chaignon <paul.chaignon@gmail.com>,
Shung-Hsi Yu <shung-hsi.yu@suse.com>,
Harishankar Vishwanathan <harishankar.vishwanathan@gmail.com>,
Srinivas Narayana <srinivas.narayana@rutgers.edu>,
Santosh Nagarakatte <santosh.nagarakatte@rutgers.edu>
Subject: [PATCH bpf-next 2/2] selftests/bpf: Test for empty intersection of tnum and u64
Date: Wed, 15 Apr 2026 12:07:28 -0400 [thread overview]
Message-ID: <20260415160728.657270-3-harishankar.vishwanathan@gmail.com> (raw)
In-Reply-To: <20260415160728.657270-1-harishankar.vishwanathan@gmail.com>
From: Paul Chaignon <paul.chaignon@gmail.com>
Thanks to the previous patch, if the tnum and u64 don't intersect on a
branch, we can deduce the branch is dead code. This patch adds a
verifier selftest for that case, with an invalid r10 write on the dead
branch.
Without the previous patch, this selftest fails as follows:
0: call bpf_get_prandom_u32#7 ; R0=scalar()
1: r6 = r0 ; R0=scalar(id=1) R6=scalar(id=1)
2: r1 = 0xe00000002 ; R1=0xe00000002
4: r2 = 0xf00000000 ; R2=0xf00000000
6: if r6 s< r1 goto pc+2 ; R1=0xe00000002 R6=scalar(id=1,smin=umin=0xe00000002,umax=0x7fffffffffffffff,var_off=(0x0; 0x7fffffffffffffff))
7: if r6 s> r2 goto pc+1 ; R2=0xf00000000 R6=scalar(id=1,smin=umin=0xe00000002,smax=umax=0xf00000000,var_off=(0xe00000000; 0x1ffffffff))
8: if w6 == 0x1 goto pc+1 ; R6=scalar(id=1,smin=umin=0xe00000002,smax=umax=0xf00000000,var_off=(0xe00000000; 0x1ffffffff))
9: exit
from 8 to 10: R0=scalar(id=1,smin=umin=0xe00000002,smax=umax=0xf00000000,smin32=umin32=1,smax32=umax32=1,var_off=(0xe00000001; 0x100000000)) R1=0xe00000002 R2=0xf00000000 R6=scalar(id=1,smin=umin=0xe00000002,smax=umax=0xf00000000,smin32=umin32=1,smax32=umax32=1,var_off=(0xe00000001; 0x100000000)) R10=fp0
10: R0=scalar(id=1,smin=umin=0xe00000002,smax=umax=0xf00000000,smin32=umin32=1,smax32=umax32=1,var_off=(0xe00000001; 0x100000000)) R1=0xe00000002 R2=0xf00000000 R6=scalar(id=1,smin=umin=0xe00000002,smax=umax=0xf00000000,smin32=umin32=1,smax32=umax32=1,var_off=(0xe00000001; 0x100000000)) R10=fp0
10: (b7) r10 = 0
frame pointer is read only
Signed-off-by: Paul Chaignon <paul.chaignon@gmail.com>
---
.../selftests/bpf/progs/verifier_bounds.c | 25 +++++++++++++++++++
1 file changed, 25 insertions(+)
diff --git a/tools/testing/selftests/bpf/progs/verifier_bounds.c b/tools/testing/selftests/bpf/progs/verifier_bounds.c
index c1ae013dee29..151c7a479a6b 100644
--- a/tools/testing/selftests/bpf/progs/verifier_bounds.c
+++ b/tools/testing/selftests/bpf/progs/verifier_bounds.c
@@ -2184,4 +2184,29 @@ __naked void tnums_equal_impossible_constant(void *ctx)
: __clobber_all);
}
+SEC("socket")
+__description("dead branch: tnum and u64 don't intersect on true branch")
+__success
+__naked void empty_tnum_u64_intersection(void *ctx)
+{
+ asm volatile(" \
+ call %[bpf_get_prandom_u32]; \
+ r6 = r0; \
+ r1 = 0xe00000002 ll; \
+ r2 = 0xf00000000 ll; \
+ if r6 s< r1 goto +2; \
+ /* On fallthrough, r6's u64=[0xe00000002;0xf00000000] */ \
+ /* tnum=(0xe00000000; 0x1ffffffff) */ \
+ if r6 s> r2 goto +1; \
+ /* On fallthrough, r6's tnum=(0xe00000001; 0x100000000) */ \
+ /* It doesn't intersect with the u64 so the condition is always false */ \
+ if w6 == 1 goto +1; \
+ exit; \
+ r10 = 0; \
+ exit; \
+" :
+ : __imm(bpf_get_prandom_u32)
+ : __clobber_all);
+}
+
char _license[] SEC("license") = "GPL";
--
2.51.0
next prev parent reply other threads:[~2026-04-15 16:07 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-04-15 16:07 [PATCH bpf-next 0/2] Detect and prune dead branches using intersection checks Harishankar Vishwanathan
2026-04-15 16:07 ` [PATCH bpf-next 1/2] bpf/verifier: Use intersection checks when simulating to detect dead branches Harishankar Vishwanathan
2026-04-15 17:04 ` bot+bpf-ci
2026-04-15 19:02 ` Paul Chaignon
2026-04-16 1:10 ` Eduard Zingerman
2026-04-16 17:23 ` Eduard Zingerman
2026-04-16 23:13 ` Harishankar Vishwanathan
2026-04-16 23:33 ` Eduard Zingerman
2026-04-17 5:24 ` Eduard Zingerman
2026-04-17 5:34 ` Harishankar Vishwanathan
2026-04-17 21:17 ` Eduard Zingerman
2026-04-17 23:19 ` Eduard Zingerman
2026-04-18 0:38 ` Harishankar Vishwanathan
2026-04-18 0:45 ` Eduard Zingerman
2026-04-18 0:22 ` Harishankar Vishwanathan
2026-04-15 16:07 ` Harishankar Vishwanathan [this message]
2026-04-15 18:29 ` [PATCH bpf-next 2/2] selftests/bpf: Test for empty intersection of tnum and u64 sashiko-bot
2026-04-16 17:51 ` Paul Chaignon
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=20260415160728.657270-3-harishankar.vishwanathan@gmail.com \
--to=harishankar.vishwanathan@gmail.com \
--cc=andrii@kernel.org \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=eddyz87@gmail.com \
--cc=paul.chaignon@gmail.com \
--cc=santosh.nagarakatte@rutgers.edu \
--cc=shung-hsi.yu@suse.com \
--cc=srinivas.narayana@rutgers.edu \
/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