[PATCH 3.19.y-ckt 169/251] arm64: bpf: fix out-of-bounds read in bpf2a64_offset()

Kamal Mostafa kamal at canonical.com
Thu Jul 16 01:08:10 UTC 2015


3.19.8-ckt4 -stable review patch.  If anyone has any objections, please let me know.

------------------

From: Xi Wang <xi.wang at gmail.com>

commit 8eee539ddea09bccae2426f09b0ba6a18b72b691 upstream.

Problems occur when bpf_to or bpf_from has value prog->len - 1 (e.g.,
"Very long jump backwards" in test_bpf where the last instruction is a
jump): since ctx->offset has length prog->len, ctx->offset[bpf_to + 1]
or ctx->offset[bpf_from + 1] will cause an out-of-bounds read, leading
to a bogus jump offset and kernel panic.

This patch moves updating ctx->offset to after calling build_insn(),
and changes indexing to use bpf_to and bpf_from without + 1.

Fixes: e54bcde3d69d ("arm64: eBPF JIT compiler")
Cc: Zi Shen Lim <zlim.lnx at gmail.com>
Cc: Will Deacon <will.deacon at arm.com>
Acked-by: Alexei Starovoitov <ast at plumgrid.com>
Signed-off-by: Xi Wang <xi.wang at gmail.com>
Signed-off-by: Catalin Marinas <catalin.marinas at arm.com>
Signed-off-by: Kamal Mostafa <kamal at canonical.com>
---
 arch/arm64/net/bpf_jit_comp.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/arch/arm64/net/bpf_jit_comp.c b/arch/arm64/net/bpf_jit_comp.c
index dc6a484..c81ddd4 100644
--- a/arch/arm64/net/bpf_jit_comp.c
+++ b/arch/arm64/net/bpf_jit_comp.c
@@ -113,9 +113,9 @@ static inline void emit_a64_mov_i(const int is64, const int reg,
 static inline int bpf2a64_offset(int bpf_to, int bpf_from,
 				 const struct jit_ctx *ctx)
 {
-	int to = ctx->offset[bpf_to + 1];
+	int to = ctx->offset[bpf_to];
 	/* -1 to account for the Branch instruction */
-	int from = ctx->offset[bpf_from + 1] - 1;
+	int from = ctx->offset[bpf_from] - 1;
 
 	return to - from;
 }
@@ -640,10 +640,11 @@ static int build_body(struct jit_ctx *ctx)
 		const struct bpf_insn *insn = &prog->insnsi[i];
 		int ret;
 
+		ret = build_insn(insn, ctx);
+
 		if (ctx->image == NULL)
 			ctx->offset[i] = ctx->idx;
 
-		ret = build_insn(insn, ctx);
 		if (ret > 0) {
 			i++;
 			continue;
-- 
1.9.1





More information about the kernel-team mailing list