[PATCH nft 5/9] evaluate: get length from statement instead of lhs expression

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Get length from statement, instead infering it from the expression that
is used to set the value. In the particular case of {ct|meta} mark, this
is 32 bits.

Otherwise, bytecode generation is not correct:

 # nft -c --debug=netlink 'add rule ip6 x y ct mark set ip6 dscp << 2 | 0x10'
  [ payload load 2b @ network header + 0 => reg 1 ]
  [ bitwise reg 1 = ( reg 1 & 0x0000c00f ) ^ 0x00000000 ]
  [ bitwise reg 1 = ( reg 1 >> 0x00000006 ) ]
  [ byteorder reg 1 = ntoh(reg 1, 2, 1) ]
  [ bitwise reg 1 = ( reg 1 << 0x00000002 ) ]
  [ bitwise reg 1 = ( reg 1 & 0x00000fef ) ^ 0x00000010 ]    <--- incorrect!
  [ ct set mark with reg 1 ]

the previous bitwise shift already upgraded to 32-bits (not visible from
the netlink debug output above).

After this patch, the last | 0x10 uses 32-bits:

 [ bitwise reg 1 = ( reg 1 & 0xffffffef ) ^ 0x00000010 ]

note that mask 0xffffffef is used instead of 0x00000fef.

Signed-off-by: Pablo Neira Ayuso <pablo@xxxxxxxxxxxxx>
---
 src/evaluate.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/evaluate.c b/src/evaluate.c
index 6d61cdb25f3d..2a679c90a3ac 100644
--- a/src/evaluate.c
+++ b/src/evaluate.c
@@ -1324,7 +1324,7 @@ static int expr_evaluate_bitwise(struct eval_ctx *ctx, struct expr **expr)
 
 	op->dtype     = left->dtype;
 	op->byteorder = left->byteorder;
-	op->len	      = left->len;
+	op->len	      = ctx->ectx.len;
 
 	if (expr_is_constant(left))
 		return constant_binop_simplify(ctx, expr);
-- 
2.30.2




[Index of Archives]     [Netfitler Users]     [Berkeley Packet Filter]     [LARTC]     [Bugtraq]     [Yosemite Forum]

  Powered by Linux