All of lore.kernel.org
 help / color / mirror / Atom feed
From: Andrii Nakryiko <andrii@kernel.org>
To: <bpf@vger.kernel.org>, <ast@kernel.org>, <daniel@iogearbox.net>,
	<martin.lau@kernel.org>
Cc: <andrii@kernel.org>, <kernel-team@meta.com>
Subject: [PATCH v5 bpf-next 08/23] bpf: try harder to deduce register bounds from different numeric domains
Date: Fri, 27 Oct 2023 11:13:31 -0700	[thread overview]
Message-ID: <20231027181346.4019398-9-andrii@kernel.org> (raw)
In-Reply-To: <20231027181346.4019398-1-andrii@kernel.org>

There are cases (caught by subsequent reg_bounds tests in selftests/bpf)
where performing one round of __reg_deduce_bounds() doesn't propagate
all the information from, say, s32 to u32 bounds and than from newly
learned u32 bounds back to u64 and s64. So perform __reg_deduce_bounds()
twice to make sure such derivations are propagated fully after
reg_bounds_sync().

One such example is test `(s64)[0xffffffff00000001; 0] (u64)<
0xffffffff00000000` from selftest patch from this patch set. It demonstrates an
intricate dance of u64 -> s64 -> u64 -> u32 bounds adjustments, which requires
two rounds of __reg_deduce_bounds(). Here are corresponding refinement log from
selftest, showing evolution of knowledge.

REFINING (FALSE R1) (u64)SRC=[0xffffffff00000000; U64_MAX] (u64)DST_OLD=[0; U64_MAX] (u64)DST_NEW=[0xffffffff00000000; U64_MAX]
REFINING (FALSE R1) (u64)SRC=[0xffffffff00000000; U64_MAX] (s64)DST_OLD=[0xffffffff00000001; 0] (s64)DST_NEW=[0xffffffff00000001; -1]
REFINING (FALSE R1) (s64)SRC=[0xffffffff00000001; -1] (u64)DST_OLD=[0xffffffff00000000; U64_MAX] (u64)DST_NEW=[0xffffffff00000001; U64_MAX]
REFINING (FALSE R1) (u64)SRC=[0xffffffff00000001; U64_MAX] (u32)DST_OLD=[0; U32_MAX] (u32)DST_NEW=[1; U32_MAX]

R1 initially has smin/smax set to [0xffffffff00000001; -1], while umin/umax is
unknown. After (u64)< comparison, in FALSE branch we gain knowledge that
umin/umax is [0xffffffff00000000; U64_MAX]. That causes smin/smax to learn that
zero can't happen and upper bound is -1. Then smin/smax is adjusted from
umin/umax improving lower bound from 0xffffffff00000000 to 0xffffffff00000001.
And then eventually umin32/umax32 bounds are drived from umin/umax and become
[1; U32_MAX].

Selftest in the last patch is actually implementing a multi-round fixed-point
convergence logic, but so far all the tests are handled by two rounds of
reg_bounds_sync() on the verifier state, so we keep it simple for now.

Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
---
 kernel/bpf/verifier.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index 768247e3d667..6b0736c04ebe 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -2607,6 +2607,7 @@ static void reg_bounds_sync(struct bpf_reg_state *reg)
 	__update_reg_bounds(reg);
 	/* We might have learned something about the sign bit. */
 	__reg_deduce_bounds(reg);
+	__reg_deduce_bounds(reg);
 	/* We might have learned some bits from the bounds. */
 	__reg_bound_offset(reg);
 	/* Intersecting with the old var_off might have improved our bounds
-- 
2.34.1


  parent reply	other threads:[~2023-10-27 18:16 UTC|newest]

Thread overview: 77+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-10-27 18:13 [PATCH v5 bpf-next 00/23] BPF register bounds logic and testing improvements Andrii Nakryiko
2023-10-27 18:13 ` [PATCH v5 bpf-next 01/23] selftests/bpf: fix RELEASE=1 build for tc_opts Andrii Nakryiko
2023-10-27 18:13 ` [PATCH v5 bpf-next 02/23] selftests/bpf: satisfy compiler by having explicit return in btf test Andrii Nakryiko
2023-10-27 18:13 ` [PATCH v5 bpf-next 03/23] bpf: derive smin/smax from umin/max bounds Andrii Nakryiko
2023-10-31 15:37   ` Eduard Zingerman
2023-10-31 17:30     ` Andrii Nakryiko
2023-10-27 18:13 ` [PATCH v5 bpf-next 04/23] bpf: derive smin32/smax32 from umin32/umax32 bounds Andrii Nakryiko
2023-10-31 15:37   ` Eduard Zingerman
2023-10-27 18:13 ` [PATCH v5 bpf-next 05/23] bpf: derive subreg bounds from full bounds when upper 32 bits are constant Andrii Nakryiko
2023-10-31 15:37   ` Eduard Zingerman
2023-10-27 18:13 ` [PATCH v5 bpf-next 06/23] bpf: add special smin32/smax32 derivation from 64-bit bounds Andrii Nakryiko
2023-10-31 15:37   ` Eduard Zingerman
2023-10-31 17:39     ` Andrii Nakryiko
2023-10-31 18:41       ` Alexei Starovoitov
2023-10-31 18:49         ` Andrii Nakryiko
2023-10-27 18:13 ` [PATCH v5 bpf-next 07/23] bpf: improve deduction of 64-bit bounds from 32-bit bounds Andrii Nakryiko
2023-10-31 15:37   ` Eduard Zingerman
2023-10-31 20:26   ` Alexei Starovoitov
2023-10-31 20:33     ` Andrii Nakryiko
2023-10-27 18:13 ` Andrii Nakryiko [this message]
2023-10-27 18:13 ` [PATCH v5 bpf-next 09/23] bpf: drop knowledge-losing __reg_combine_{32,64}_into_{64,32} logic Andrii Nakryiko
2023-10-31 15:38   ` Eduard Zingerman
2023-10-27 18:13 ` [PATCH v5 bpf-next 10/23] selftests/bpf: BPF register range bounds tester Andrii Nakryiko
2023-11-08 22:08   ` Eduard Zingerman
2023-11-08 23:23     ` Andrii Nakryiko
2023-11-09  0:30       ` Eduard Zingerman
2023-10-27 18:13 ` [PATCH v5 bpf-next 11/23] bpf: rename is_branch_taken reg arguments to prepare for the second one Andrii Nakryiko
2023-10-30 19:39   ` Alexei Starovoitov
2023-10-31  5:19     ` Andrii Nakryiko
2023-10-27 18:13 ` [PATCH v5 bpf-next 12/23] bpf: generalize is_branch_taken() to work with two registers Andrii Nakryiko
2023-10-31 15:38   ` Eduard Zingerman
2023-10-31 17:41     ` Andrii Nakryiko
2023-10-27 18:13 ` [PATCH v5 bpf-next 13/23] bpf: move is_branch_taken() down Andrii Nakryiko
2023-10-27 18:13 ` [PATCH v5 bpf-next 14/23] bpf: generalize is_branch_taken to handle all conditional jumps in one place Andrii Nakryiko
2023-10-31 15:38   ` Eduard Zingerman
2023-10-27 18:13 ` [PATCH v5 bpf-next 15/23] bpf: unify 32-bit and 64-bit is_branch_taken logic Andrii Nakryiko
2023-10-30 19:52   ` Alexei Starovoitov
2023-10-31  5:28     ` Andrii Nakryiko
2023-10-31 17:35   ` Eduard Zingerman
2023-10-27 18:13 ` [PATCH v5 bpf-next 16/23] bpf: prepare reg_set_min_max for second set of registers Andrii Nakryiko
2023-10-27 18:13 ` [PATCH v5 bpf-next 17/23] bpf: generalize reg_set_min_max() to handle two sets of two registers Andrii Nakryiko
2023-10-31  2:02   ` Alexei Starovoitov
2023-10-31  6:03     ` Andrii Nakryiko
2023-10-31 16:23       ` Alexei Starovoitov
2023-10-31 17:50         ` Andrii Nakryiko
2023-10-31 17:56           ` Andrii Nakryiko
2023-10-31 18:04             ` Alexei Starovoitov
2023-10-31 18:06               ` Andrii Nakryiko
2023-10-31 18:14   ` Eduard Zingerman
2023-10-27 18:13 ` [PATCH v5 bpf-next 18/23] bpf: generalize reg_set_min_max() to handle non-const register comparisons Andrii Nakryiko
2023-10-31 23:25   ` Eduard Zingerman
2023-11-01 16:35     ` Andrii Nakryiko
2023-11-01 17:12       ` Eduard Zingerman
2023-10-27 18:13 ` [PATCH v5 bpf-next 19/23] bpf: generalize is_scalar_branch_taken() logic Andrii Nakryiko
2023-10-31  2:12   ` Alexei Starovoitov
2023-10-31  6:12     ` Andrii Nakryiko
2023-10-31 16:34       ` Alexei Starovoitov
2023-10-31 18:01         ` Andrii Nakryiko
2023-10-31 20:53           ` Andrii Nakryiko
2023-10-31 20:55             ` Andrii Nakryiko
2023-10-27 18:13 ` [PATCH v5 bpf-next 20/23] bpf: enhance BPF_JEQ/BPF_JNE is_branch_taken logic Andrii Nakryiko
2023-10-31  2:20   ` Alexei Starovoitov
2023-10-31  6:16     ` Andrii Nakryiko
2023-10-31 16:36       ` Alexei Starovoitov
2023-10-31 18:04         ` Andrii Nakryiko
2023-10-31 18:06           ` Alexei Starovoitov
2023-10-27 18:13 ` [PATCH v5 bpf-next 21/23] selftests/bpf: adjust OP_EQ/OP_NE handling to use subranges for branch taken Andrii Nakryiko
2023-11-08 18:22   ` Eduard Zingerman
2023-11-08 19:59     ` Andrii Nakryiko
2023-10-27 18:13 ` [PATCH v5 bpf-next 22/23] selftests/bpf: add range x range test to reg_bounds Andrii Nakryiko
2023-10-27 18:13 ` [PATCH v5 bpf-next 23/23] selftests/bpf: add iter test requiring range x range logic Andrii Nakryiko
2023-10-30 17:55 ` [PATCH v5 bpf-next 00/23] BPF register bounds logic and testing improvements Alexei Starovoitov
2023-10-31  5:19   ` Andrii Nakryiko
2023-11-01 12:37     ` Paul Chaignon
2023-11-01 17:13       ` Andrii Nakryiko
2023-11-07  6:37         ` Harishankar Vishwanathan
2023-11-07 16:38           ` 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=20231027181346.4019398-9-andrii@kernel.org \
    --to=andrii@kernel.org \
    --cc=ast@kernel.org \
    --cc=bpf@vger.kernel.org \
    --cc=daniel@iogearbox.net \
    --cc=kernel-team@meta.com \
    --cc=martin.lau@kernel.org \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.