Skip to contentSkip to conversation

bpf: Fix an incorrect verification success with movsx insn

PATCH Applied Yonghong Song 3 participants 4 messages Aug 7, 2023

Subsystems (4)
lore
Share & tools
lore

Yonghong Song

author reviewer
#0 lore raw

PATCH 1/2 bpf: Fix an incorrect verification success with movsx insn

syzbot reports a verifier bug which triggers a runtime panic. The test bpf program is:
0: (62) *(u32 *)(r10 -8) = 553656332
1: (bf) r1 = (s16)r10
2: (07) r1 += -8
3: (b7) r2 = 3
4: (bd) if r2 <= r1 goto pc+0
5: (85) call bpf_trace_printk#-138320
6: (b7) r0 = 0
7: (95) exit

At insn 1, the current implementation keeps 'r1' as a frame pointer, which caused later bpf_trace_printk helper call crash since frame pointer address is not valid any more. Note that at insn 4, the 'pointer vs. scalar' comparison is allowed for privileged prog run.

To fix the problem with above insn 1, the fix in the patch adopts similar pattern to existing 'R1 = (u32) R2' handling. For unprivileged prog run, verification will fail with 'R<num> sign-extension part of pointer'. For privileged prog run, the dst_reg 'r1' will be marked as
an unknown scalar, so later 'bpf_trace_pointk' helper will complain since it expected certain pointers.

Reported-by [email protected]
Fixes 8100928c8814 ("bpf: Support new sign-extension mov insns")
---
 kernel/bpf/verifier.c | 31 ++++++++++++++++++++-----------
 1 file changed, 20 insertions(+), 11 deletions(-)
kernel/bpf/verifier.c +20 -11
index 132f25dab931..4ccca1f6c998 100644
@@ -13165,17 +13165,26 @@ static int check_alu_op(struct bpf_verifier_env *env, struct bpf_insn *insn)
 					dst_reg->subreg_def = DEF_NOT_SUBREG;
 				} else {
 					/* case: R1 = (s8, s16 s32)R2 */
-					bool no_sext;
-
-					no_sext = src_reg->umax_value < (1ULL << (insn->off - 1));
-					if (no_sext && need_id)
-						src_reg->id = ++env->id_gen;
-					copy_register_state(dst_reg, src_reg);
-					if (!no_sext)
-						dst_reg->id = 0;
-					coerce_reg_to_size_sx(dst_reg, insn->off >> 3);
-					dst_reg->live |= REG_LIVE_WRITTEN;
-					dst_reg->subreg_def = DEF_NOT_SUBREG;
+					if (is_pointer_value(env, insn->src_reg)) {
+						verbose(env,
+							"R%d sign-extension part of pointer\n",
+							insn->src_reg);
+						return -EACCES;
+					} else if (src_reg->type == SCALAR_VALUE) {
+						bool no_sext;
+
+						no_sext = src_reg->umax_value < (1ULL << (insn->off - 1));
+						if (no_sext && need_id)
+							src_reg->id = ++env->id_gen;
+						copy_register_state(dst_reg, src_reg);
+						if (!no_sext)
+							dst_reg->id = 0;
+						coerce_reg_to_size_sx(dst_reg, insn->off >> 3);
+						dst_reg->live |= REG_LIVE_WRITTEN;
+						dst_reg->subreg_def = DEF_NOT_SUBREG;
+					} else {
+						mark_reg_unknown(env, regs, insn->dst_reg);
+					}
 				}
 			} else {
 				/* R1 = (u32) R2 */
Signature · 1 line
2.34.1
author reviewer
#1 lore raw

PATCH 2/2 selftests/bpf: Add a movsx selftest for sign-extension of R10

A movsx selftest is added for sign-extension of frame pointer R10. The verification fails for both privileged and unprivileged prog runs.

---
 .../selftests/bpf/progs/verifier_movsx.c      | 22 +++++++++++++++++++
 1 file changed, 22 insertions(+)
tools/testing/selftests/bpf/progs/verifier_movsx.c +22 -0
index 9568089932d7..be6f69a6b659 100644
@@ -198,6 +198,28 @@ l0_%=:							\
 	: __clobber_all);
 }
 
+SEC("socket")
+__description("MOV64SX, S16, R10 Sign Extension")
+__failure __msg("R1 type=scalar expected=fp, pkt, pkt_meta, map_key, map_value, mem, ringbuf_mem, buf, trusted_ptr_")
+__failure_unpriv __msg_unpriv("R10 sign-extension part of pointer")
+__naked void mov64sx_s16_r10(void)
+{
+	asm volatile ("					\
+	r1 = 553656332;					\
+	*(u32 *)(r10 - 8) = r1; 			\
+	r1 = (s16)r10;					\
+	r1 += -8;					\
+	r2 = 3;						\
+	if r2 <= r1 goto l0_%=;				\
+l0_%=:							\
+	call %[bpf_trace_printk];			\
+	r0 = 0;						\
+	exit;						\
+"	:
+	: __imm(bpf_trace_printk)
+	: __clobber_all);
+}
+
 #else
 
 SEC("socket")
Signature · 1 line
2.34.1
maintainer Acked-by on v1 1/2
#2 lore raw
In reply to Yonghong Song
Quoted 24 lines · Yonghong Song

syzbot reports a verifier bug which triggers a runtime panic. The test bpf program is:

   0: (62) *(u32 *)(r10 -8) = 553656332
   1: (bf) r1 = (s16)r10
   2: (07) r1 += -8
   3: (b7) r2 = 3
   4: (bd) if r2 <= r1 goto pc+0
   5: (85) call bpf_trace_printk#-138320
   6: (b7) r0 = 0
   7: (95) exit

At insn 1, the current implementation keeps 'r1' as a frame pointer, which caused later bpf_trace_printk helper call crash since frame pointer address is not valid any more. Note that at insn 4, the 'pointer vs. scalar' comparison is allowed for privileged prog run.

To fix the problem with above insn 1, the fix in the patch adopts similar pattern to existing 'R1 = (u32) R2' handling. For unprivileged prog run, verification will fail with 'R<num> sign-extension part of pointer'. For privileged prog run, the dst_reg 'r1' will be marked as
an unknown scalar, so later 'bpf_trace_pointk' helper will complain since it expected certain pointers.

Quoted 3 lines
Reported-by [email protected]
Fixes 8100928c8814 ("bpf: Support new sign-extension mov insns")

All works on my side.
Nitpick: the test case could be simplified.

Quoted 46 lines
---
 kernel/bpf/verifier.c | 31 ++++++++++++++++++++-----------
 1 file changed, 20 insertions(+), 11 deletions(-)
kernel/bpf/verifier.c +20 -11
index 132f25dab931..4ccca1f6c998 100644
@@ -13165,17 +13165,26 @@ static int check_alu_op(struct bpf_verifier_env *env, struct bpf_insn *insn)
 					dst_reg->subreg_def = DEF_NOT_SUBREG;
 				} else {
 					/* case: R1 = (s8, s16 s32)R2 */
-					bool no_sext;
-
-					no_sext = src_reg->umax_value < (1ULL << (insn->off - 1));
-					if (no_sext && need_id)
-						src_reg->id = ++env->id_gen;
-					copy_register_state(dst_reg, src_reg);
-					if (!no_sext)
-						dst_reg->id = 0;
-					coerce_reg_to_size_sx(dst_reg, insn->off >> 3);
-					dst_reg->live |= REG_LIVE_WRITTEN;
-					dst_reg->subreg_def = DEF_NOT_SUBREG;
+					if (is_pointer_value(env, insn->src_reg)) {
+						verbose(env,
+							"R%d sign-extension part of pointer\n",
+							insn->src_reg);
+						return -EACCES;
+					} else if (src_reg->type == SCALAR_VALUE) {
+						bool no_sext;
+
+						no_sext = src_reg->umax_value < (1ULL << (insn->off - 1));
+						if (no_sext && need_id)
+							src_reg->id = ++env->id_gen;
+						copy_register_state(dst_reg, src_reg);
+						if (!no_sext)
+							dst_reg->id = 0;
+						coerce_reg_to_size_sx(dst_reg, insn->off >> 3);
+						dst_reg->live |= REG_LIVE_WRITTEN;
+						dst_reg->subreg_def = DEF_NOT_SUBREG;
+					} else {
+						mark_reg_unknown(env, regs, insn->dst_reg);
+					}
 				}
 			} else {
 				/* R1 = (u32) R2 */
#3 lore raw
In reply to Yonghong Song
This series was applied to bpf/bpf-next.git (master) by Martin KaFai Lau <[email protected]>:

Hello:

This series was applied to bpf/bpf-next.git (master) by Martin KaFai Lau <[email protected]>:

Quoted 11 lines · On Mon, 7 Aug 2023 10:57:21 -0700 you wrote:

syzbot reports a verifier bug which triggers a runtime panic. The test bpf program is:

   0: (62) *(u32 *)(r10 -8) = 553656332
   1: (bf) r1 = (s16)r10
   2: (07) r1 += -8
   3: (b7) r2 = 3
   4: (bd) if r2 <= r1 goto pc+0
   5: (85) call bpf_trace_printk#-138320
   6: (b7) r0 = 0
   7: (95) exit
Quoted 1 line

[...]

Here is the summary with links:
- [bpf-next,1/2] bpf: Fix an incorrect verification success with movsx insn

    https://git.kernel.org/bpf/bpf-next/c/db2baf82b098
  - [bpf-next,2/2] selftests/bpf: Add a movsx selftest for sign-extension of R10
    https://git.kernel.org/bpf/bpf-next/c/a5c0a42bd374

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html