Skip to contentSkip to conversation

bpf: prevent r10 register from being marked as precise

PATCH Applied Andrii Nakryiko 3 participants 5 messages Apr 4, 2024 – Apr 5, 2024

Subsystems (4)
lore
Share & tools
lore

Andrii Nakryiko

author maintainer
#0 lore raw

PATCH v2 1/2 bpf: prevent r10 register from being marked as precise

r10 is a special register that is not under BPF program's control and is always effectively precise. The rest of precision logic assumes that only r0-r9 SCALAR registers are marked as precise, so prevent r10 from being marked precise.

This can happen due to signed cast instruction allowing to do something like r0 = (s8)r10;, which later, if r0 needs to be precise, would lead to an attempt to mark r10 as precise.

Prevent this with an extra check during instruction backtracking.

Fixes 8100928c8814 ("bpf: Support new sign-extension mov insns")
Reported-by [email protected]
---
 kernel/bpf/verifier.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
kernel/bpf/verifier.c +4 -2
index ffaa9f7f153c..41446bea8fab 100644
@@ -3615,7 +3615,8 @@ static int backtrack_insn(struct bpf_verifier_env *env, int idx, int subseq_idx,
 				 * sreg needs precision before this insn
 				 */
 				bt_clear_reg(bt, dreg);
-				bt_set_reg(bt, sreg);
+				if (sreg != BPF_REG_FP)
+					bt_set_reg(bt, sreg);
 			} else {
 				/* dreg = K
 				 * dreg needs precision after this insn.
@@ -3631,7 +3632,8 @@ static int backtrack_insn(struct bpf_verifier_env *env, int idx, int subseq_idx,
 				 * both dreg and sreg need precision
 				 * before this insn
 				 */
-				bt_set_reg(bt, sreg);
+				if (sreg != BPF_REG_FP)
+					bt_set_reg(bt, sreg);
 			} /* else dreg += K
 			   * dreg still needs precision before this insn
 			   */
Signature · 1 line
2.43.0
author maintainer
#1 lore raw

PATCH v2 2/2 selftests/bpf: add fp-leaking precise subprog result tests

Add selftests validating that BPF verifier handles precision marking for SCALAR registers derived from r10 (fp) register correctly.

Given r0 = (s8)r10; syntax is not supported by older Clang compilers, use the raw BPF instruction syntax to maximize compatibility.

---
 tools/include/linux/filter.h                  | 18 ++++
 .../bpf/progs/verifier_subprog_precision.c    | 89 +++++++++++++++++++
 2 files changed, 107 insertions(+)
tools/include/linux/filter.h +18 -0
index 736bdeccdfe4..65aa8ce142e5 100644
@@ -111,6 +111,24 @@
 		.off   = 0,					\
 		.imm   = IMM })
 
+/* Short form of movsx, dst_reg = (s8,s16,s32)src_reg */
+
+#define BPF_MOVSX64_REG(DST, SRC, OFF)				\
+	((struct bpf_insn) {					\
+		.code  = BPF_ALU64 | BPF_MOV | BPF_X,		\
+		.dst_reg = DST,					\
+		.src_reg = SRC,					\
+		.off   = OFF,					\
+		.imm   = 0 })
+
+#define BPF_MOVSX32_REG(DST, SRC, OFF)				\
+	((struct bpf_insn) {					\
+		.code  = BPF_ALU | BPF_MOV | BPF_X,		\
+		.dst_reg = DST,					\
+		.src_reg = SRC,					\
+		.off   = OFF,					\
+		.imm   = 0 })
+
 /* Short form of mov based on type,  BPF_X: dst_reg = src_reg, BPF_K: dst_reg = imm32 */
 
 #define BPF_MOV64_RAW(TYPE, DST, SRC, IMM)			\
tools/testing/selftests/bpf/progs/verifier_subprog_precision.c +89 -0

101 lines not on this page · raw message

Signature · 1 line
2.43.0
reviewer Acked-by on v2 2/2
#2 lore raw

Re: [PATCH v2 bpf-next 2/2] selftests/bpf: add fp-leaking precise subprog result tests

Quoted 7 lines · Andrii Nakryiko

Add selftests validating that BPF verifier handles precision marking for SCALAR registers derived from r10 (fp) register correctly.

Given r0 = (s8)r10; syntax is not supported by older Clang compilers, use the raw BPF instruction syntax to maximize compatibility.

reviewer Acked-by on v2 1/2
#3 lore raw
In reply to Andrii Nakryiko
Quoted 9 lines · Andrii Nakryiko

r10 is a special register that is not under BPF program's control and is always effectively precise. The rest of precision logic assumes that only r0-r9 SCALAR registers are marked as precise, so prevent r10 from being marked precise.

This can happen due to signed cast instruction allowing to do something like r0 = (s8)r10;, which later, if r0 needs to be precise, would lead to an attempt to mark r10 as precise.

Quoted 5 lines

Prevent this with an extra check during instruction backtracking.

Fixes 8100928c8814 ("bpf: Support new sign-extension mov insns")
Reported-by [email protected]
#4 lore raw
In reply to Andrii Nakryiko
This series was applied to bpf/bpf-next.git (master) by Alexei Starovoitov <[email protected]>:

Hello:

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

Quoted 10 lines · On Thu, 4 Apr 2024 14:45:35 -0700 you wrote:

r10 is a special register that is not under BPF program's control and is always effectively precise. The rest of precision logic assumes that only r0-r9 SCALAR registers are marked as precise, so prevent r10 from being marked precise.

This can happen due to signed cast instruction allowing to do something like r0 = (s8)r10;, which later, if r0 needs to be precise, would lead to an attempt to mark r10 as precise.

[...]

Here is the summary with links:
- [v2,bpf-next,1/2] bpf: prevent r10 register from being marked as precise

    https://git.kernel.org/bpf/bpf-next/c/1f2a74b41ea8
  - [v2,bpf-next,2/2] selftests/bpf: add fp-leaking precise subprog result tests
    https://git.kernel.org/bpf/bpf-next/c/343ca8131c35

You are awesome, thank you!

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