On 02/19/2019 11:24 AM, Jesper Dangaard Brouer wrote: > The skb->queue_mapping already have read access, via __sk_buff->queue_mapping. > > This patch allow BPF tc qdisc clsact write access to the queue_mapping via > tc_cls_act_is_valid_access. > > It is already possible to change this via TC filter action skbedit > tc-skbedit(8). Due to the lack of TC examples, lets show one: > > # tc qdisc add dev ixgbe1 handle ffff: ingress > # tc filter add dev ixgbe1 parent ffff: matchall action skbedit > queue_mapping 5 > # tc filter list dev ixgbe1 parent ffff:
Using handles was in the old days, if we add examples, then lets do something more user friendly ;) # tc qdisc add dev ixgbe1 clsact # tc filter replace dev ixgbe1 ingress matchall action skbedit queue_mapping 5 # tc filter list dev ixgbe1 ingress > The most common mistake is that XPS (Transmit Packet Steering) takes > precedence over setting skb->queue_mapping. XPS is configured per DEVICE > via /sys/class/net/DEVICE/queues/tx-*/xps_cpus via a CPU hex mask. To > disable set mask=00. > > The purpose of changing skb->queue_mapping is to influence the selection of > the net_device "txq" (struct netdev_queue), which influence selection of > the qdisc "root_lock" (via txq->qdisc->q.lock) and txq->_xmit_lock. When > using the MQ qdisc the txq->qdisc points to different qdiscs and associated > locks, and HARD_TX_LOCK (txq->_xmit_lock), allowing for CPU scalability. > > Due to lack of TC examples, lets show howto attach clsact BPF programs: > > # tc qdisc add dev ixgbe2 clsact > # tc filter replace dev ixgbe2 egress bpf da obj XXX_kern.o sec tc_qmap2cpu > # tc filter list dev ixgbe2 egress > > Signed-off-by: Jesper Dangaard Brouer <bro...@redhat.com> > --- > net/core/filter.c | 14 +++++++++++--- > 1 file changed, 11 insertions(+), 3 deletions(-) > > diff --git a/net/core/filter.c b/net/core/filter.c > index 353735575204..d05ae8d05397 100644 > --- a/net/core/filter.c > +++ b/net/core/filter.c > @@ -6238,6 +6238,7 @@ static bool tc_cls_act_is_valid_access(int off, int > size, > case bpf_ctx_range(struct __sk_buff, tc_classid): > case bpf_ctx_range_till(struct __sk_buff, cb[0], cb[4]): > case bpf_ctx_range(struct __sk_buff, tstamp): > + case bpf_ctx_range(struct __sk_buff, queue_mapping): > break; > default: > return false; > @@ -6642,9 +6643,16 @@ static u32 bpf_convert_ctx_access(enum bpf_access_type > type, > break; > > case offsetof(struct __sk_buff, queue_mapping): > - *insn++ = BPF_LDX_MEM(BPF_H, si->dst_reg, si->src_reg, > - bpf_target_off(struct sk_buff, > queue_mapping, 2, > - target_size)); > + if (type == BPF_WRITE) > + *insn++ = BPF_STX_MEM(BPF_H, si->dst_reg, si->src_reg, > + bpf_target_off(struct sk_buff, > + queue_mapping, > + 2, target_size)); > + else > + *insn++ = BPF_LDX_MEM(BPF_H, si->dst_reg, si->src_reg, > + bpf_target_off(struct sk_buff, > + queue_mapping, > + 2, target_size)); One thing we should avoid would be to allow user to write NO_QUEUE_MAPPING into skb->queue_mapping so we don't hit the warn in sk_tx_queue_set(), I'd add this into the ctx rewrite here. > break; > > case offsetof(struct __sk_buff, vlan_present): >
