* [PATCH bpf-next v3 0/2] bpf: Improve linked register tracking
@ 2026-02-03 22:26 Puranjay Mohan
2026-02-03 22:26 ` [PATCH bpf-next v3 1/2] bpf: Support negative offsets, BPF_SUB, and alu32 for " Puranjay Mohan
2026-02-03 22:26 ` [PATCH bpf-next v3 2/2] selftests/bpf: Add tests for improved " Puranjay Mohan
0 siblings, 2 replies; 3+ messages in thread
From: Puranjay Mohan @ 2026-02-03 22:26 UTC (permalink / raw)
To: bpf
Cc: Puranjay Mohan, Puranjay Mohan, Alexei Starovoitov,
Andrii Nakryiko, Daniel Borkmann, Martin KaFai Lau,
Eduard Zingerman, Kumar Kartikeya Dwivedi, Mykyta Yatsenko,
kernel-team
V2: https://lore.kernel.org/all/20260113152529.3217648-1-puranjay@kernel.org/
Changes in v2->v3:
- Added another selftest showing a real usage pattern
- Rebased on bpf-next/master
v1: https://lore.kernel.org/bpf/20260107203941.1063754-1-puranjay@kernel.org/
Changes in v1->v2:
- Add support for alu32 operations in linked register tracking (Alexei)
- Squash the selftest fix with the first patch (Eduard)
- Add more selftests to detect edge cases
This series extends the BPF verifier's linked register tracking to handle
negative offsets, BPF_SUB operations, and alu32 operations, enabling better bounds propagation for
common arithmetic patterns.
The verifier previously only tracked positive constant deltas between linked
registers using BPF_ADD. This meant patterns using negative offsets or
subtraction couldn't benefit from bounds propagation:
void alu32_negative_offset(void)
{
volatile char path[5];
volatile int offset = bpf_get_prandom_u32();
int off = offset;
if (off >= 5 && off < 10)
path[off - 5] = '.';
}
this gets compiled to:
0000000000000478 <alu32_negative_offset>:
143: call 0x7
144: *(u32 *)(r10 - 0xc) = w0
145: w1 = *(u32 *)(r10 - 0xc)
146: w2 = w1 // w2 and w1 share the same id
147: w2 += -0x5 // verifier knows w1 = w2 + 5
148: if w2 > 0x4 goto +0x5 <L0> // in fall-through: verifier knows w2 ∈ [0,4] => w1 ∈ [5, 9]
149: r2 = r10
150: r2 += -0x5 // r2 = fp - 5
151: r2 += r1 // r2 = fp - 5 + r1 (∈ [5, 9]) => r2 ∈ [fp, fp + 4]
152: w1 = 0x2e
153: *(u8 *)(r2 - 0x5) = w1 // r2 ∈ [fp, fp + 4] => r2 - 5 ∈ [fp - 5, fp - 1]
<L0>:
154: exit
After the changes, the verifier could link 32-bit scalars and also supported -ve offsets for linking:
146: w2 = w1
147: w2 += -0x5
It allowed the verifier to correctly propagate bounds, without the
changes in this patchset, verifier would reject this program with:
invalid unbounded variable-offset write to stack R2
This program has been added as a selftest in the second patch.
Veristat comparison on programs from sched_ext, selftests, and some meta internal programs:
Scx Progs
File Program Verdict (A) Verdict (B) Verdict (DIFF) Insns (A) Insns (B) Insns (DIFF)
----------------- ---------------- ----------- ----------- -------------- --------- --------- -------------
scx_layered.bpf.o layered_runnable success success MATCH 5674 6077 +403 (+7.10%)
FB Progs
File Program Verdict (A) Verdict (B) Verdict (DIFF) Insns (A) Insns (B) Insns (DIFF)
------------ ---------------- ----------- ----------- -------------- --------- --------- -----------------
bpf232.bpf.o layered_dump success success MATCH 1151 1218 +67 (+5.82%)
bpf257.bpf.o layered_runnable success success MATCH 5743 6143 +400 (+6.97%)
bpf252.bpf.o layered_runnable success success MATCH 5677 6075 +398 (+7.01%)
bpf227.bpf.o layered_dump success success MATCH 915 982 +67 (+7.32%)
bpf239.bpf.o layered_runnable success success MATCH 5459 5861 +402 (+7.36%)
bpf246.bpf.o layered_runnable success success MATCH 5562 6008 +446 (+8.02%)
bpf229.bpf.o layered_runnable success success MATCH 2559 3011 +452 (+17.66%)
bpf231.bpf.o layered_runnable success success MATCH 2559 3011 +452 (+17.66%)
bpf234.bpf.o layered_runnable success success MATCH 2549 3001 +452 (+17.73%)
bpf019.bpf.o do_sendmsg success success MATCH 124823 153523 +28700 (+22.99%)
bpf019.bpf.o do_parse success success MATCH 124809 153509 +28700 (+23.00%)
bpf227.bpf.o layered_runnable success success MATCH 1915 2356 +441 (+23.03%)
bpf228.bpf.o layered_runnable success success MATCH 1700 2152 +452 (+26.59%)
bpf232.bpf.o layered_runnable success success MATCH 1499 1951 +452 (+30.15%)
bpf312.bpf.o mount_exit success success MATCH 19253 62883 +43630 (+226.61%)
bpf312.bpf.o umount_exit success success MATCH 19253 62883 +43630 (+226.61%)
bpf311.bpf.o mount_exit success success MATCH 19226 62863 +43637 (+226.97%)
bpf311.bpf.o umount_exit success success MATCH 19226 62863 +43637 (+226.97%)
The above four programs have specific patters that make the verifier explore a lot more states:
for (; depth < MAX_DIR_DEPTH; depth++) {
const unsigned char* name = BPF_CORE_READ(dentry, d_name.name);
if (offset >= MAX_PATH_LEN - MAX_DIR_LEN) {
return depth;
}
int len = bpf_probe_read_kernel_str(&path[offset], MAX_DIR_LEN, name);
offset += len;
if (len == MAX_DIR_LEN) {
if (offset - 2 < MAX_PATH_LEN) { // <---- (a)
path[offset - 2] = '.';
}
if (offset - 3 < MAX_PATH_LEN) { // <---- (b)
path[offset - 3] = '.';
}
if (offset - 4 < MAX_PATH_LEN) { // <---- (c)
path[offset - 4] = '.';
}
}
}
When at some depth == N false branches of conditions (a), (b) and (c) are scheduled for
verification, constraints for offset at depth == N+1 are:
1. offset >= MAX_PATH_LEN + 2
2. offset >= MAX_PATH_LEN + 3
3. offset >= MAX_PATH_LEN + 4 (visited before others)
And after offset += len it becomes:
1. offset >= MAX_PATH_LEN - 4093
2. offset >= MAX_PATH_LEN - 4092
3. offset >= MAX_PATH_LEN - 4091 (visited before others)
Because of the DFS states exploration logic, the states above are visited in order 3, 2, 1; 3 is not
a subset of 2 and 1 is not a subset of 2, so pruning logic does not kick in.
Previously this was not a problem, because range for offset was not propagated through the
statements (a), (b), (c).
As the root cause of this regression is understood, this is not a blocker for this change.
Selftest Progs
File Program Verdict (A) Verdict (B) Verdict (DIFF) Insns (A) Insns (B) Insns (DIFF)
---------------------------------- ------------------------ ----------- ----------- -------------- --------- --------- --------------
linked_list_peek.bpf.o list_peek success success MATCH 152 88 -64 (-42.11%)
verifier_iterating_callbacks.bpf.o cond_break2 success success MATCH 110 88 -22 (-20.00%)
These are the added selftests that failed earlier but are passing now:
verifier_linked_scalars.bpf.o alu32_negative_offset failure success MISMATCH 11 13 +2 (+18.18%)
verifier_linked_scalars.bpf.o scalars_alu32_big_offset failure success MISMATCH 7 10 +3 (+42.86%)
verifier_linked_scalars.bpf.o scalars_neg_alu32_add failure success MISMATCH 7 10 +3 (+42.86%)
verifier_linked_scalars.bpf.o scalars_neg_alu32_sub failure success MISMATCH 7 10 +3 (+42.86%)
verifier_linked_scalars.bpf.o scalars_neg failure success MISMATCH 7 10 +3 (+42.86%)
verifier_linked_scalars.bpf.o scalars_neg_sub failure success MISMATCH 7 10 +3 (+42.86%)
verifier_linked_scalars.bpf.o scalars_sub_neg_imm failure success MISMATCH 7 10 +3 (+42.86%)
iters.bpf.o iter_obfuscate_counter success success MATCH 83 119 +36 (+43.37%)
bpf_cubic.bpf.o bpf_cubic_acked success success MATCH 243 430 +187 (+76.95%)
Puranjay Mohan (2):
bpf: Support negative offsets, BPF_SUB, and alu32 for linked register
tracking
selftests/bpf: Add tests for improved linked register tracking
include/linux/bpf_verifier.h | 6 +-
kernel/bpf/verifier.c | 49 ++-
.../selftests/bpf/progs/verifier_bounds.c | 2 +-
.../bpf/progs/verifier_linked_scalars.c | 300 +++++++++++++++++-
4 files changed, 342 insertions(+), 15 deletions(-)
base-commit: f11f7cf90ee09dbcf76413818063ffc38ed2d9fe
--
2.47.3
^ permalink raw reply [flat|nested] 3+ messages in thread
* [PATCH bpf-next v3 1/2] bpf: Support negative offsets, BPF_SUB, and alu32 for linked register tracking
2026-02-03 22:26 [PATCH bpf-next v3 0/2] bpf: Improve linked register tracking Puranjay Mohan
@ 2026-02-03 22:26 ` Puranjay Mohan
2026-02-03 22:26 ` [PATCH bpf-next v3 2/2] selftests/bpf: Add tests for improved " Puranjay Mohan
1 sibling, 0 replies; 3+ messages in thread
From: Puranjay Mohan @ 2026-02-03 22:26 UTC (permalink / raw)
To: bpf
Cc: Puranjay Mohan, Puranjay Mohan, Alexei Starovoitov,
Andrii Nakryiko, Daniel Borkmann, Martin KaFai Lau,
Eduard Zingerman, Kumar Kartikeya Dwivedi, Mykyta Yatsenko,
kernel-team
Previously, the verifier only tracked positive constant deltas between
linked registers using BPF_ADD. This limitation meant patterns like:
r1 = r0;
r1 += -4;
if r1 s>= 0 goto l0_%=; // r1 >= 0 implies r0 >= 4
// verifier couldn't propagate bounds back to r0
if r0 != 0 goto l0_%=;
r0 /= 0; // Verifier thinks this is reachable
l0_%=:
Similar limitation exists for 32-bit registers.
With this change, the verifier can now track negative deltas in reg->off
enabling bound propagation for the above pattern.
For alu32, we make sure the destination register has the upper 32 bits
as 0s before creating the link. BPF_ADD_CONST is split into
BPF_ADD_CONST64 and BPF_ADD_CONST32, the latter is used in case of alu32
and sync_linked_regs uses this to zext the result if known_reg has this
flag.
Signed-off-by: Puranjay Mohan <puranjay@kernel.org>
---
include/linux/bpf_verifier.h | 6 ++-
kernel/bpf/verifier.c | 49 ++++++++++++++-----
.../selftests/bpf/progs/verifier_bounds.c | 2 +-
3 files changed, 44 insertions(+), 13 deletions(-)
diff --git a/include/linux/bpf_verifier.h b/include/linux/bpf_verifier.h
index 746025df82c8..ef8e45a362d9 100644
--- a/include/linux/bpf_verifier.h
+++ b/include/linux/bpf_verifier.h
@@ -147,8 +147,12 @@ struct bpf_reg_state {
* registers. Example:
* r1 = r2; both will have r1->id == r2->id == N
* r1 += 10; r1->id == N | BPF_ADD_CONST and r1->off == 10
+ * r3 = r2; both will have r3->id == r2->id == N
+ * w3 += 10; r3->id == N | BPF_ADD_CONST32 and r3->off == 10
*/
-#define BPF_ADD_CONST (1U << 31)
+#define BPF_ADD_CONST64 (1U << 31)
+#define BPF_ADD_CONST32 (1U << 30)
+#define BPF_ADD_CONST (BPF_ADD_CONST64 | BPF_ADD_CONST32)
u32 id;
/* PTR_TO_SOCKET and PTR_TO_TCP_SOCK could be a ptr returned
* from a pointer-cast helper, bpf_sk_fullsock() and
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index 6a616dc4dc54..7f73f73adfe0 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -16122,6 +16122,13 @@ static int adjust_reg_min_max_vals(struct bpf_verifier_env *env,
verbose(env, "verifier internal error: no src_reg\n");
return -EFAULT;
}
+ /*
+ * For alu32 linked register tracking, we need to check dst_reg's
+ * umax_value before the ALU operation. After adjust_scalar_min_max_vals(),
+ * alu32 ops will have zero-extended the result, making umax_value <= U32_MAX.
+ */
+ u64 dst_umax = dst_reg->umax_value;
+
err = adjust_scalar_min_max_vals(env, insn, dst_reg, *src_reg);
if (err)
return err;
@@ -16131,26 +16138,44 @@ static int adjust_reg_min_max_vals(struct bpf_verifier_env *env,
* r1 += 0x1
* if r2 < 1000 goto ...
* use r1 in memory access
- * So for 64-bit alu remember constant delta between r2 and r1 and
- * update r1 after 'if' condition.
+ * So remember constant delta between r2 and r1 and update r1 after
+ * 'if' condition.
*/
if (env->bpf_capable &&
- BPF_OP(insn->code) == BPF_ADD && !alu32 &&
- dst_reg->id && is_reg_const(src_reg, false)) {
- u64 val = reg_const_value(src_reg, false);
+ (BPF_OP(insn->code) == BPF_ADD || BPF_OP(insn->code) == BPF_SUB) &&
+ dst_reg->id && is_reg_const(src_reg, alu32)) {
+ u64 val = reg_const_value(src_reg, alu32);
+ s32 off;
+
+ if (!alu32 && ((s64)val < S32_MIN || (s64)val > S32_MAX))
+ goto clear_id;
+
+ if (alu32 && (dst_umax > U32_MAX))
+ goto clear_id;
- if ((dst_reg->id & BPF_ADD_CONST) ||
- /* prevent overflow in sync_linked_regs() later */
- val > (u32)S32_MAX) {
+ off = (s32)val;
+
+ if (BPF_OP(insn->code) == BPF_SUB) {
+ /* Negating S32_MIN would overflow */
+ if (off == S32_MIN)
+ goto clear_id;
+ off = -off;
+ }
+
+ if (dst_reg->id & BPF_ADD_CONST) {
/*
* If the register already went through rX += val
* we cannot accumulate another val into rx->off.
*/
+clear_id:
dst_reg->off = 0;
dst_reg->id = 0;
} else {
- dst_reg->id |= BPF_ADD_CONST;
- dst_reg->off = val;
+ if (alu32)
+ dst_reg->id |= BPF_ADD_CONST32;
+ else
+ dst_reg->id |= BPF_ADD_CONST64;
+ dst_reg->off = off;
}
} else {
/*
@@ -17247,7 +17272,7 @@ static void sync_linked_regs(struct bpf_verifier_env *env, struct bpf_verifier_s
u32 saved_id = reg->id;
fake_reg.type = SCALAR_VALUE;
- __mark_reg_known(&fake_reg, (s32)reg->off - (s32)known_reg->off);
+ __mark_reg_known(&fake_reg, (s64)reg->off - (s64)known_reg->off);
/* reg = known_reg; reg += delta */
copy_register_state(reg, known_reg);
@@ -17262,6 +17287,8 @@ static void sync_linked_regs(struct bpf_verifier_env *env, struct bpf_verifier_s
scalar32_min_max_add(reg, &fake_reg);
scalar_min_max_add(reg, &fake_reg);
reg->var_off = tnum_add(reg->var_off, fake_reg.var_off);
+ if (known_reg->id & BPF_ADD_CONST32)
+ zext_32_to_64(reg);
}
if (e->is_reg)
mark_reg_scratched(env, e->regno);
diff --git a/tools/testing/selftests/bpf/progs/verifier_bounds.c b/tools/testing/selftests/bpf/progs/verifier_bounds.c
index 411a18437d7e..560531404bce 100644
--- a/tools/testing/selftests/bpf/progs/verifier_bounds.c
+++ b/tools/testing/selftests/bpf/progs/verifier_bounds.c
@@ -1477,7 +1477,7 @@ __naked void sub64_full_overflow(void)
SEC("socket")
__description("64-bit subtraction, partial overflow, result in unbounded reg")
__success __log_level(2)
-__msg("3: (1f) r3 -= r2 {{.*}} R3=scalar()")
+__msg("3: (1f) r3 -= r2 {{.*}} R3=scalar(id=1-1)")
__retval(0)
__naked void sub64_partial_overflow(void)
{
--
2.47.3
^ permalink raw reply related [flat|nested] 3+ messages in thread
* [PATCH bpf-next v3 2/2] selftests/bpf: Add tests for improved linked register tracking
2026-02-03 22:26 [PATCH bpf-next v3 0/2] bpf: Improve linked register tracking Puranjay Mohan
2026-02-03 22:26 ` [PATCH bpf-next v3 1/2] bpf: Support negative offsets, BPF_SUB, and alu32 for " Puranjay Mohan
@ 2026-02-03 22:26 ` Puranjay Mohan
1 sibling, 0 replies; 3+ messages in thread
From: Puranjay Mohan @ 2026-02-03 22:26 UTC (permalink / raw)
To: bpf
Cc: Puranjay Mohan, Puranjay Mohan, Alexei Starovoitov,
Andrii Nakryiko, Daniel Borkmann, Martin KaFai Lau,
Eduard Zingerman, Kumar Kartikeya Dwivedi, Mykyta Yatsenko,
kernel-team
Add tests for linked register tracking with negative offsets, BPF_SUB,
and alu32. These test for all edge cases like overflows, etc.
Signed-off-by: Puranjay Mohan <puranjay@kernel.org>
---
.../bpf/progs/verifier_linked_scalars.c | 300 +++++++++++++++++-
1 file changed, 298 insertions(+), 2 deletions(-)
diff --git a/tools/testing/selftests/bpf/progs/verifier_linked_scalars.c b/tools/testing/selftests/bpf/progs/verifier_linked_scalars.c
index 5f41bbb730a7..5ff3983c48b7 100644
--- a/tools/testing/selftests/bpf/progs/verifier_linked_scalars.c
+++ b/tools/testing/selftests/bpf/progs/verifier_linked_scalars.c
@@ -1,6 +1,7 @@
// SPDX-License-Identifier: GPL-2.0
#include <linux/bpf.h>
+#include <limits.h>
#include <bpf/bpf_helpers.h>
#include "bpf_misc.h"
@@ -18,9 +19,9 @@ __naked void scalars(void)
r4 = r1; \
w2 += 0x7FFFFFFF; \
w4 += 0; \
- if r2 == 0 goto l1; \
+ if r2 == 0 goto l0_%=; \
exit; \
-l1: \
+l0_%=: \
r4 >>= 63; \
r3 = 1; \
r3 -= r4; \
@@ -64,4 +65,299 @@ l0_%=: \
: __clobber_all);
}
+SEC("socket")
+__success
+__naked void scalars_neg(void)
+{
+ asm volatile (" \
+ call %[bpf_get_prandom_u32]; \
+ r0 &= 0xff; \
+ r1 = r0; \
+ r1 += -4; \
+ if r1 s< 0 goto l0_%=; \
+ if r0 != 0 goto l0_%=; \
+ r0 /= 0; \
+l0_%=: \
+ r0 = 0; \
+ exit; \
+" :
+ : __imm(bpf_get_prandom_u32)
+ : __clobber_all);
+}
+
+/* Same test but using BPF_SUB instead of BPF_ADD with negative immediate */
+SEC("socket")
+__success
+__naked void scalars_neg_sub(void)
+{
+ asm volatile (" \
+ call %[bpf_get_prandom_u32]; \
+ r0 &= 0xff; \
+ r1 = r0; \
+ r1 -= 4; \
+ if r1 s< 0 goto l0_%=; \
+ if r0 != 0 goto l0_%=; \
+ r0 /= 0; \
+l0_%=: \
+ r0 = 0; \
+ exit; \
+" :
+ : __imm(bpf_get_prandom_u32)
+ : __clobber_all);
+}
+
+/* alu32 with negative offset */
+SEC("socket")
+__success
+__naked void scalars_neg_alu32_add(void)
+{
+ asm volatile (" \
+ call %[bpf_get_prandom_u32]; \
+ w0 &= 0xff; \
+ w1 = w0; \
+ w1 += -4; \
+ if w1 s< 0 goto l0_%=; \
+ if w0 != 0 goto l0_%=; \
+ r0 /= 0; \
+l0_%=: \
+ r0 = 0; \
+ exit; \
+" :
+ : __imm(bpf_get_prandom_u32)
+ : __clobber_all);
+}
+
+/* alu32 with negative offset using SUB */
+SEC("socket")
+__success
+__naked void scalars_neg_alu32_sub(void)
+{
+ asm volatile (" \
+ call %[bpf_get_prandom_u32]; \
+ w0 &= 0xff; \
+ w1 = w0; \
+ w1 -= 4; \
+ if w1 s< 0 goto l0_%=; \
+ if w0 != 0 goto l0_%=; \
+ r0 /= 0; \
+l0_%=: \
+ r0 = 0; \
+ exit; \
+" :
+ : __imm(bpf_get_prandom_u32)
+ : __clobber_all);
+}
+
+/* Positive offset: r1 = r0 + 4, then if r1 >= 6, r0 >= 2, so r0 != 0 */
+SEC("socket")
+__success
+__naked void scalars_pos(void)
+{
+ asm volatile (" \
+ call %[bpf_get_prandom_u32]; \
+ r0 &= 0xff; \
+ r1 = r0; \
+ r1 += 4; \
+ if r1 < 6 goto l0_%=; \
+ if r0 != 0 goto l0_%=; \
+ r0 /= 0; \
+l0_%=: \
+ r0 = 0; \
+ exit; \
+" :
+ : __imm(bpf_get_prandom_u32)
+ : __clobber_all);
+}
+
+/* SUB with negative immediate: r1 -= -4 is equivalent to r1 += 4 */
+SEC("socket")
+__success
+__naked void scalars_sub_neg_imm(void)
+{
+ asm volatile (" \
+ call %[bpf_get_prandom_u32]; \
+ r0 &= 0xff; \
+ r1 = r0; \
+ r1 -= -4; \
+ if r1 < 6 goto l0_%=; \
+ if r0 != 0 goto l0_%=; \
+ r0 /= 0; \
+l0_%=: \
+ r0 = 0; \
+ exit; \
+" :
+ : __imm(bpf_get_prandom_u32)
+ : __clobber_all);
+}
+
+/* Double ADD clears the ID (can't accumulate offsets) */
+SEC("socket")
+__failure
+__msg("div by zero")
+__naked void scalars_double_add(void)
+{
+ asm volatile (" \
+ call %[bpf_get_prandom_u32]; \
+ r0 &= 0xff; \
+ r1 = r0; \
+ r1 += 2; \
+ r1 += 2; \
+ if r1 < 6 goto l0_%=; \
+ if r0 != 0 goto l0_%=; \
+ r0 /= 0; \
+l0_%=: \
+ r0 = 0; \
+ exit; \
+" :
+ : __imm(bpf_get_prandom_u32)
+ : __clobber_all);
+}
+
+/*
+ * Test that sync_linked_regs() correctly handles large offset differences.
+ * r1.off = S32_MIN, r2.off = 1, delta = S32_MIN - 1 requires 64-bit math.
+ */
+SEC("socket")
+__success
+__naked void scalars_sync_delta_overflow(void)
+{
+ asm volatile (" \
+ call %[bpf_get_prandom_u32]; \
+ r0 &= 0xff; \
+ r1 = r0; \
+ r2 = r0; \
+ r1 += %[s32_min]; \
+ r2 += 1; \
+ if r2 s< 100 goto l0_%=; \
+ if r1 s< 0 goto l0_%=; \
+ r0 /= 0; \
+l0_%=: \
+ r0 = 0; \
+ exit; \
+" :
+ : __imm(bpf_get_prandom_u32),
+ [s32_min]"i"(INT_MIN)
+ : __clobber_all);
+}
+
+/*
+ * Another large delta case: r1.off = S32_MAX, r2.off = -1.
+ * delta = S32_MAX - (-1) = S32_MAX + 1 requires 64-bit math.
+ */
+SEC("socket")
+__success
+__naked void scalars_sync_delta_overflow_large_range(void)
+{
+ asm volatile (" \
+ call %[bpf_get_prandom_u32]; \
+ r0 &= 0xff; \
+ r1 = r0; \
+ r2 = r0; \
+ r1 += %[s32_max]; \
+ r2 += -1; \
+ if r2 s< 0 goto l0_%=; \
+ if r1 s>= 0 goto l0_%=; \
+ r0 /= 0; \
+l0_%=: \
+ r0 = 0; \
+ exit; \
+" :
+ : __imm(bpf_get_prandom_u32),
+ [s32_max]"i"(INT_MAX)
+ : __clobber_all);
+}
+
+/*
+ * Test linked scalar tracking with alu32 and large positive offset (0x7FFFFFFF).
+ * After w1 += 0x7FFFFFFF, w1 wraps to negative for any r0 >= 1.
+ * If w1 is signed-negative, then r0 >= 1, so r0 != 0.
+ */
+SEC("socket")
+__success
+__naked void scalars_alu32_big_offset(void)
+{
+ asm volatile (" \
+ call %[bpf_get_prandom_u32]; \
+ w0 &= 0xff; \
+ w1 = w0; \
+ w1 += 0x7FFFFFFF; \
+ if w1 s>= 0 goto l0_%=; \
+ if w0 != 0 goto l0_%=; \
+ r0 /= 0; \
+l0_%=: \
+ r0 = 0; \
+ exit; \
+" :
+ : __imm(bpf_get_prandom_u32)
+ : __clobber_all);
+}
+
+SEC("socket")
+__failure
+__msg("div by zero")
+__naked void scalars_alu32_basic(void)
+{
+ asm volatile (" \
+ call %[bpf_get_prandom_u32]; \
+ r1 = r0; \
+ w1 += 1; \
+ if r1 > 10 goto 1f; \
+ r0 >>= 32; \
+ if r0 == 0 goto 1f; \
+ r0 /= 0; \
+1: \
+ r0 = 0; \
+ exit; \
+" :
+ : __imm(bpf_get_prandom_u32)
+ : __clobber_all);
+}
+
+/*
+ * Test alu32 linked register tracking with wrapping.
+ * R0 is bounded to [0xffffff00, 0xffffffff] (high 32-bit values)
+ * w1 += 0x100 causes R1 to wrap to [0, 0xff]
+ *
+ * After sync_linked_regs, if bounds are computed correctly:
+ * R0 should be [0x00000000_ffffff00, 0x00000000_ffffff80]
+ * R0 >> 32 == 0, so div by zero is unreachable
+ *
+ * If bounds are computed incorrectly (64-bit underflow):
+ * R0 becomes [0xffffffff_ffffff00, 0xffffffff_ffffff80]
+ * R0 >> 32 == 0xffffffff != 0, so div by zero is reachable
+ */
+SEC("socket")
+__success
+__naked void scalars_alu32_wrap(void)
+{
+ asm volatile (" \
+ call %[bpf_get_prandom_u32]; \
+ w0 |= 0xffffff00; \
+ r1 = r0; \
+ w1 += 0x100; \
+ if r1 > 0x80 goto l0_%=; \
+ r2 = r0; \
+ r2 >>= 32; \
+ if r2 == 0 goto l0_%=; \
+ r0 /= 0; \
+l0_%=: \
+ r0 = 0; \
+ exit; \
+" :
+ : __imm(bpf_get_prandom_u32)
+ : __clobber_all);
+}
+
+SEC("socket")
+__success
+void alu32_negative_offset(void)
+{
+ volatile char path[5];
+ volatile int offset = bpf_get_prandom_u32();
+ int off = offset;
+
+ if (off >= 5 && off < 10)
+ path[off - 5] = '.';
+}
+
char _license[] SEC("license") = "GPL";
--
2.47.3
^ permalink raw reply related [flat|nested] 3+ messages in thread
end of thread, other threads:[~2026-02-03 22:27 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-02-03 22:26 [PATCH bpf-next v3 0/2] bpf: Improve linked register tracking Puranjay Mohan
2026-02-03 22:26 ` [PATCH bpf-next v3 1/2] bpf: Support negative offsets, BPF_SUB, and alu32 for " Puranjay Mohan
2026-02-03 22:26 ` [PATCH bpf-next v3 2/2] selftests/bpf: Add tests for improved " Puranjay Mohan
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox