On 9/11/24 6:40 AM, Yonghong Song wrote:
Zac Ecob reported a problem where a bpf program may cause kernel crash due
to the following error:
Oops: divide error: 0000 [#1] PREEMPT SMP KASAN PTI
The failure is due to the below signed divide:
LLONG_MIN/-1 where LLONG_MIN equals to -9,223,372,036,854,775,808.
LLONG_MIN/-1 is supposed to give a positive number 9,223,372,036,854,775,808,
but it is impossible since for 64-bit system, the maximum positive
number is 9,223,372,036,854,775,807. On x86_64, LLONG_MIN/-1 will
cause a kernel exception. On arm64, the result for LLONG_MIN/-1 is
LLONG_MIN.
So for 64-bit signed divide (sdiv), some additional insns are patched
to check LLONG_MIN/-1 pattern. If such a pattern does exist, the result
will be LLONG_MIN. Otherwise, it follows normal sdiv operation.
I presume this could be follow-up but it would also need an update to [0]
to describe the behavior.
[0] Documentation/bpf/standardization/instruction-set.rst
[1] https://lore.kernel.org/bpf/tPJLTEh7S_DxFEqAI2Ji5MBSoZVg7_G-Py2iaZpAaWtM961fFTWtsnlzwvTbzBzaUzwQAoNATXKUlt0LZOFgnDcIyKCswAnAGdUF3LBrhGQ=@protonmail.com/
Reported-by: Zac Ecob <zacecob@xxxxxxxxxxxxxx>
Signed-off-by: Yonghong Song <yonghong.song@xxxxxxxxx>
---
kernel/bpf/verifier.c | 29 ++++++++++++++++++++++++++---
1 file changed, 26 insertions(+), 3 deletions(-)
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index f35b80c16cda..d77f1a05a065 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -20506,6 +20506,7 @@ static int do_misc_fixups(struct bpf_verifier_env *env)
insn->code == (BPF_ALU | BPF_DIV | BPF_X)) {
bool is64 = BPF_CLASS(insn->code) == BPF_ALU64;
bool isdiv = BPF_OP(insn->code) == BPF_DIV;
+ bool is_sdiv64 = is64 && isdiv && insn->off == 1;
struct bpf_insn *patchlet;
struct bpf_insn chk_and_div[] = {
/* [R,W]x div 0 -> 0 */
@@ -20525,10 +20526,32 @@ static int do_misc_fixups(struct bpf_verifier_env *env)
BPF_JMP_IMM(BPF_JA, 0, 0, 1),
BPF_MOV32_REG(insn->dst_reg, insn->dst_reg),
};
+ struct bpf_insn chk_and_sdiv64[] = {
+ /* Rx sdiv 0 -> 0 */
+ BPF_RAW_INSN(BPF_JMP | BPF_JNE | BPF_K, insn->src_reg,
+ 0, 2, 0),
+ BPF_ALU32_REG(BPF_XOR, insn->dst_reg, insn->dst_reg),
+ BPF_JMP_IMM(BPF_JA, 0, 0, 8),
+ /* LLONG_MIN sdiv -1 -> LLONG_MIN */
+ BPF_RAW_INSN(BPF_JMP | BPF_JNE | BPF_K, insn->src_reg,
+ 0, 6, -1),
+ BPF_LD_IMM64(insn->src_reg, LLONG_MIN),
+ BPF_RAW_INSN(BPF_JMP | BPF_JNE | BPF_X, insn->dst_reg,
+ insn->src_reg, 2, 0),
+ BPF_MOV64_IMM(insn->src_reg, -1),
+ BPF_JMP_IMM(BPF_JA, 0, 0, 2),
+ BPF_MOV64_IMM(insn->src_reg, -1),
Looks good, we could probably shrink this snippet via BPF_REG_AX ?
Untested, like below:
+ /* Rx sdiv 0 -> 0 */
+ BPF_RAW_INSN(BPF_JMP | BPF_JNE | BPF_K, insn->src_reg, 0, 2, 0),
+ BPF_ALU32_REG(BPF_XOR, insn->dst_reg, insn->dst_reg),
+ BPF_JMP_IMM(BPF_JA, 0, 0, 5),
+ /* LLONG_MIN sdiv -1 -> LLONG_MIN */
+ BPF_RAW_INSN(BPF_JMP | BPF_JNE | BPF_K, insn->src_reg, 0, 2, -1),
+ BPF_LD_IMM64(BPF_REG_AX, LLONG_MIN),
+ BPF_RAW_INSN(BPF_JMP | BPF_JEQ | BPF_X, insn->dst_reg, BPF_REG_AX, 1, 0),
+ *insn,
Then we don't need to restore the src_reg in both paths.
+ *insn,
+ };
Have you also looked into rejecting this pattern upfront on load when its a known
constant as we do with div by 0 in check_alu_op()?
Otherwise lgtm if this is equivalent to arm64 as you describe.
- patchlet = isdiv ? chk_and_div : chk_and_mod;
- cnt = isdiv ? ARRAY_SIZE(chk_and_div) :
- ARRAY_SIZE(chk_and_mod) - (is64 ? 2 : 0);
+ if (is_sdiv64) {
+ patchlet = chk_and_sdiv64;
+ cnt = ARRAY_SIZE(chk_and_sdiv64);
+ } else {
+ patchlet = isdiv ? chk_and_div : chk_and_mod;
+ cnt = isdiv ? ARRAY_SIZE(chk_and_div) :
+ ARRAY_SIZE(chk_and_mod) - (is64 ? 2 : 0);
+ }
new_prog = bpf_patch_insn_data(env, i + delta, patchlet, cnt);
if (!new_prog)
