Skip to contentSkip to conversation

bpf: improve verifier packet range checks

PATCH In review Alexei Starovoitov 2 participants 2 messages Mar 24, 2017 – Mar 25, 2017

Subsystems (6)
Networking Network drivers Kselftest BPF +2 more subsystems
lore
Share & tools
lore

Alexei Starovoitov

author
#0 lore raw

PATCH 1/1 bpf: improve verifier packet range checks

llvm can optimize the 'if (ptr > data_end)' checks to be in the order slightly different than the original C code which will confuse verifier. Like:
if (ptr + 16 > data_end)
return TC_ACT_SHOT;
// may be followed by
if (ptr + 14 > data_end)
return TC_ACT_SHOT;
while llvm can see that 'ptr' is valid for all 16 bytes,
the verifier could not.
Fix verifier logic to account for such case and add a test.

Reported-by Huapeng Zhou <[email protected]>
Fixes 969bf05eb3ce ("bpf: direct packet access")
Signed-off-by Alexei Starovoitov <[email protected]>
Acked-by Daniel Borkmann <[email protected]>
Acked-by Martin KaFai Lau <[email protected]>

---
new verifier test is added in the middle of older tests to avoid conflicts with new tests in net-next.

 kernel/bpf/verifier.c                       |  5 +++--
 tools/testing/selftests/bpf/test_verifier.c | 20 ++++++++++++++++++++
 2 files changed, 23 insertions(+), 2 deletions(-)
kernel/bpf/verifier.c +3 -2
index 796b68d00119..5e6202e62265 100644
@@ -1973,14 +1973,15 @@ static void find_good_pkt_pointers(struct bpf_verifier_state *state,
 
 	for (i = 0; i < MAX_BPF_REG; i++)
 		if (regs[i].type == PTR_TO_PACKET && regs[i].id == dst_reg->id)
-			regs[i].range = dst_reg->off;
+			/* keep the maximum range already checked */
+			regs[i].range = max(regs[i].range, dst_reg->off);
 
 	for (i = 0; i < MAX_BPF_STACK; i += BPF_REG_SIZE) {
 		if (state->stack_slot_type[i] != STACK_SPILL)
 			continue;
 		reg = &state->spilled_regs[i / BPF_REG_SIZE];
 		if (reg->type == PTR_TO_PACKET && reg->id == dst_reg->id)
-			reg->range = dst_reg->off;
+			reg->range = max(reg->range, dst_reg->off);
 	}
 }
tools/testing/selftests/bpf/test_verifier.c +20 -0
index d1555e4240c0..7d761d4cc759 100644
@@ -3418,6 +3418,26 @@ static struct bpf_test tests[] = {
 		.prog_type = BPF_PROG_TYPE_LWT_XMIT,
 	},
 	{
+		"overlapping checks for direct packet access",
+		.insns = {
+			BPF_LDX_MEM(BPF_W, BPF_REG_2, BPF_REG_1,
+				    offsetof(struct __sk_buff, data)),
+			BPF_LDX_MEM(BPF_W, BPF_REG_3, BPF_REG_1,
+				    offsetof(struct __sk_buff, data_end)),
+			BPF_MOV64_REG(BPF_REG_0, BPF_REG_2),
+			BPF_ALU64_IMM(BPF_ADD, BPF_REG_0, 8),
+			BPF_JMP_REG(BPF_JGT, BPF_REG_0, BPF_REG_3, 4),
+			BPF_MOV64_REG(BPF_REG_1, BPF_REG_2),
+			BPF_ALU64_IMM(BPF_ADD, BPF_REG_1, 6),
+			BPF_JMP_REG(BPF_JGT, BPF_REG_1, BPF_REG_3, 1),
+			BPF_LDX_MEM(BPF_H, BPF_REG_0, BPF_REG_2, 6),
+			BPF_MOV64_IMM(BPF_REG_0, 0),
+			BPF_EXIT_INSN(),
+		},
+		.result = ACCEPT,
+		.prog_type = BPF_PROG_TYPE_LWT_XMIT,
+	},
+	{
 		"invalid access of tc_classid for LWT_IN",
 		.insns = {
 			BPF_LDX_MEM(BPF_W, BPF_REG_0, BPF_REG_1,
Signature · 1 line
2.9.3
maintainer
#1 lore raw

From: Alexei Starovoitov <[email protected]>
Date: Fri, 24 Mar 2017 15:57:33 -0700

Quoted 12 lines

llvm can optimize the 'if (ptr > data_end)' checks to be in the order slightly different than the original C code which will confuse verifier. Like:
if (ptr + 16 > data_end)
return TC_ACT_SHOT;
// may be followed by
if (ptr + 14 > data_end)
return TC_ACT_SHOT;
while llvm can see that 'ptr' is valid for all 16 bytes,
the verifier could not.
Fix verifier logic to account for such case and add a test.

Quoted 5 lines
Reported-by Huapeng Zhou <[email protected]>
Fixes 969bf05eb3ce ("bpf: direct packet access")
Signed-off-by Alexei Starovoitov <[email protected]>
Acked-by Daniel Borkmann <[email protected]>
Acked-by Martin KaFai Lau <[email protected]>

Applied and queued up for -stable, thanks.