From: Daniel Borkmann <daniel@xxxxxxxxxxxxx>
commit b9b34ddbe2076ade359cd5ce7537d5ed019e9807 upstream.
The negation logic for the case where the off_reg is sitting in the
dst register is not correct given then we cannot just invert the add
to a sub or vice versa. As a fix, perform the final bitwise and-op
unconditionally into AX from the off_reg, then move the pointer from
the src to dst and finally use AX as the source for the original
pointer arithmetic operation such that the inversion yields a correct
result. The single non-AX mov in between is possible given constant
blinding is retaining it as it's not an immediate based operation.
Fixes: 979d63d50c0c ("bpf: prevent out of bounds speculation on pointer arithmetic")
Signed-off-by: Daniel Borkmann <daniel@xxxxxxxxxxxxx>
Tested-by: Piotr Krysiuk <piotras@xxxxxxxxx>
Reviewed-by: Piotr Krysiuk <piotras@xxxxxxxxx>
Reviewed-by: John Fastabend <john.fastabend@xxxxxxxxx>
Acked-by: Alexei Starovoitov <ast@xxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
kernel/bpf/verifier.c | 12 ++++--------
1 file changed, 4 insertions(+), 8 deletions(-)
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -4816,14 +4816,10 @@ static int fixup_bpf_calls(struct bpf_ve
*patch++ = BPF_ALU64_REG(BPF_OR, BPF_REG_AX, off_reg);
*patch++ = BPF_ALU64_IMM(BPF_NEG, BPF_REG_AX, 0);
*patch++ = BPF_ALU64_IMM(BPF_ARSH, BPF_REG_AX, 63);
- if (issrc) {
- *patch++ = BPF_ALU64_REG(BPF_AND, BPF_REG_AX,
- off_reg);
- insn->src_reg = BPF_REG_AX;
- } else {
- *patch++ = BPF_ALU64_REG(BPF_AND, off_reg,
- BPF_REG_AX);
- }
+ *patch++ = BPF_ALU64_REG(BPF_AND, BPF_REG_AX, off_reg);
+ if (!issrc)
+ *patch++ = BPF_MOV64_REG(insn->dst_reg, insn->src_reg);
+ insn->src_reg = BPF_REG_AX;
if (isneg)
insn->code = insn->code == code_add ?
code_sub : code_add;
