mono_delivery_time was added to check if skb->tstamp has delivery
time in mono clock base (i.e. EDT) otherwise skb->tstamp has
timestamp in ingress and delivery_time at egress.
Renaming the bitfield from mono_delivery_time to tstamp_type is for
extensibilty for other timestamps such as userspace timestamp
(i.e. SO_TXTIME) set via sock opts.
As we are renaming the mono_delivery_time to tstamp_type, it makes
sense to start assigning tstamp_type based out if enum defined as
part of this commit
Earlier we used bool arg flag to check if the tstamp is mono in
function skb_set_delivery_time, Now the signature of the functions
accepts enum to distinguish between mono and real time
Bridge driver today has no support to forward the userspace timestamp
packets and ends up resetting the timestamp. ETF qdisc checks the
packet coming from userspace and encounters to be 0 thereby dropping
time sensitive packets. These changes will allow userspace timestamps
packets to be forwarded from the bridge to NIC drivers.
In future tstamp_type:1 can be extended to support userspace timestamp
by increasing the bitfield.
Link: https://lore.kernel.org/netdev/[email protected]/
Signed-off-by: Abhishek Chauhan <[email protected]>
---
Changes since v2
- Minor changes to commit subject
Changes since v1
- Squashed the two commits into one as mentioned by Willem.
- Introduced switch in skb_set_delivery_time.
- Renamed and removed directionality aspects w.r.t tstamp_type
as mentioned by Willem.
include/linux/skbuff.h | 33 +++++++++++++++-------
include/net/inet_frag.h | 4 +--
net/bridge/netfilter/nf_conntrack_bridge.c | 6 ++--
net/core/dev.c | 2 +-
net/core/filter.c | 8 +++---
net/ipv4/inet_fragment.c | 2 +-
net/ipv4/ip_fragment.c | 2 +-
net/ipv4/ip_output.c | 8 +++---
net/ipv4/tcp_output.c | 14 ++++-----
net/ipv6/ip6_output.c | 6 ++--
net/ipv6/netfilter.c | 6 ++--
net/ipv6/netfilter/nf_conntrack_reasm.c | 2 +-
net/ipv6/reassembly.c | 2 +-
net/ipv6/tcp_ipv6.c | 2 +-
net/sched/act_bpf.c | 4 +--
net/sched/cls_bpf.c | 4 +--
16 files changed, 59 insertions(+), 46 deletions(-)
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 7135a3e94afd..a83a2120b57f 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -702,6 +702,11 @@ typedef unsigned int sk_buff_data_t;
typedef unsigned char *sk_buff_data_t;
#endif
+enum skb_tstamp_type {
+ CLOCK_REAL = 0, /* Time base is realtime */
+ CLOCK_MONO = 1, /* Time base is Monotonic */
+};
+
/**
* DOC: Basic sk_buff geometry
*
@@ -819,7 +824,7 @@ typedef unsigned char *sk_buff_data_t;
* @dst_pending_confirm: need to confirm neighbour
* @decrypted: Decrypted SKB
* @slow_gro: state present at GRO time, slower prepare step required
- * @mono_delivery_time: When set, skb->tstamp has the
+ * @tstamp_type: When set, skb->tstamp has the
* delivery_time in mono clock base (i.e. EDT). Otherwise, the
* skb->tstamp has the (rcv) timestamp at ingress and
* delivery_time at egress.
@@ -950,7 +955,7 @@ struct sk_buff {
/* private: */
__u8 __mono_tc_offset[0];
/* public: */
- __u8 mono_delivery_time:1; /* See SKB_MONO_DELIVERY_TIME_MASK */
+ __u8 tstamp_type:1; /* See SKB_MONO_DELIVERY_TIME_MASK */
#ifdef CONFIG_NET_XGRESS
__u8 tc_at_ingress:1; /* See TC_AT_INGRESS_MASK */
__u8 tc_skip_classify:1;
@@ -4237,7 +4242,7 @@ static inline void skb_get_new_timestampns(const struct sk_buff *skb,
static inline void __net_timestamp(struct sk_buff *skb)
{
skb->tstamp = ktime_get_real();
- skb->mono_delivery_time = 0;
+ skb->tstamp_type = CLOCK_REAL;
}
static inline ktime_t net_timedelta(ktime_t t)
@@ -4246,10 +4251,18 @@ static inline ktime_t net_timedelta(ktime_t t)
}
static inline void skb_set_delivery_time(struct sk_buff *skb, ktime_t kt,
- bool mono)
+ u8 tstamp_type)
{
skb->tstamp = kt;
- skb->mono_delivery_time = kt && mono;
+
+ switch (tstamp_type) {
+ case CLOCK_REAL:
+ skb->tstamp_type = CLOCK_REAL;
+ break;
+ case CLOCK_MONO:
+ skb->tstamp_type = kt && tstamp_type;
+ break;
+ }
}
DECLARE_STATIC_KEY_FALSE(netstamp_needed_key);
@@ -4259,8 +4272,8 @@ DECLARE_STATIC_KEY_FALSE(netstamp_needed_key);
*/
static inline void skb_clear_delivery_time(struct sk_buff *skb)
{
- if (skb->mono_delivery_time) {
- skb->mono_delivery_time = 0;
+ if (skb->tstamp_type) {
+ skb->tstamp_type = CLOCK_REAL;
if (static_branch_unlikely(&netstamp_needed_key))
skb->tstamp = ktime_get_real();
else
@@ -4270,7 +4283,7 @@ static inline void skb_clear_delivery_time(struct sk_buff *skb)
static inline void skb_clear_tstamp(struct sk_buff *skb)
{
- if (skb->mono_delivery_time)
+ if (skb->tstamp_type)
return;
skb->tstamp = 0;
@@ -4278,7 +4291,7 @@ static inline void skb_clear_tstamp(struct sk_buff *skb)
static inline ktime_t skb_tstamp(const struct sk_buff *skb)
{
- if (skb->mono_delivery_time)
+ if (skb->tstamp_type == CLOCK_MONO)
return 0;
return skb->tstamp;
@@ -4286,7 +4299,7 @@ static inline ktime_t skb_tstamp(const struct sk_buff *skb)
static inline ktime_t skb_tstamp_cond(const struct sk_buff *skb, bool cond)
{
- if (!skb->mono_delivery_time && skb->tstamp)
+ if (skb->tstamp_type != CLOCK_MONO && skb->tstamp)
return skb->tstamp;
if (static_branch_unlikely(&netstamp_needed_key) || cond)
diff --git a/include/net/inet_frag.h b/include/net/inet_frag.h
index 153960663ce4..5af6eb14c5db 100644
--- a/include/net/inet_frag.h
+++ b/include/net/inet_frag.h
@@ -76,7 +76,7 @@ struct frag_v6_compare_key {
* @stamp: timestamp of the last received fragment
* @len: total length of the original datagram
* @meat: length of received fragments so far
- * @mono_delivery_time: stamp has a mono delivery time (EDT)
+ * @tstamp_type: stamp has a mono delivery time (EDT)
* @flags: fragment queue flags
* @max_size: maximum received fragment size
* @fqdir: pointer to struct fqdir
@@ -97,7 +97,7 @@ struct inet_frag_queue {
ktime_t stamp;
int len;
int meat;
- u8 mono_delivery_time;
+ u8 tstamp_type;
__u8 flags;
u16 max_size;
struct fqdir *fqdir;
diff --git a/net/bridge/netfilter/nf_conntrack_bridge.c b/net/bridge/netfilter/nf_conntrack_bridge.c
index c3c51b9a6826..816bb0fde718 100644
--- a/net/bridge/netfilter/nf_conntrack_bridge.c
+++ b/net/bridge/netfilter/nf_conntrack_bridge.c
@@ -32,7 +32,7 @@ static int nf_br_ip_fragment(struct net *net, struct sock *sk,
struct sk_buff *))
{
int frag_max_size = BR_INPUT_SKB_CB(skb)->frag_max_size;
- bool mono_delivery_time = skb->mono_delivery_time;
+ u8 tstamp_type = skb->tstamp_type;
unsigned int hlen, ll_rs, mtu;
ktime_t tstamp = skb->tstamp;
struct ip_frag_state state;
@@ -82,7 +82,7 @@ static int nf_br_ip_fragment(struct net *net, struct sock *sk,
if (iter.frag)
ip_fraglist_prepare(skb, &iter);
- skb_set_delivery_time(skb, tstamp, mono_delivery_time);
+ skb_set_delivery_time(skb, tstamp, tstamp_type);
err = output(net, sk, data, skb);
if (err || !iter.frag)
break;
@@ -113,7 +113,7 @@ static int nf_br_ip_fragment(struct net *net, struct sock *sk,
goto blackhole;
}
- skb_set_delivery_time(skb2, tstamp, mono_delivery_time);
+ skb_set_delivery_time(skb2, tstamp, tstamp_type);
err = output(net, sk, data, skb2);
if (err)
goto blackhole;
diff --git a/net/core/dev.c b/net/core/dev.c
index 854a3a28a8d8..77a43c05dfe3 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -2146,7 +2146,7 @@ EXPORT_SYMBOL(net_disable_timestamp);
static inline void net_timestamp_set(struct sk_buff *skb)
{
skb->tstamp = 0;
- skb->mono_delivery_time = 0;
+ skb->tstamp_type = CLOCK_REAL;
if (static_branch_unlikely(&netstamp_needed_key))
skb->tstamp = ktime_get_real();
}
diff --git a/net/core/filter.c b/net/core/filter.c
index 8d185d99a643..8bb45423df52 100644
--- a/net/core/filter.c
+++ b/net/core/filter.c
@@ -7709,13 +7709,13 @@ BPF_CALL_3(bpf_skb_set_tstamp, struct sk_buff *, skb,
if (!tstamp)
return -EINVAL;
skb->tstamp = tstamp;
- skb->mono_delivery_time = 1;
+ skb->tstamp_type = CLOCK_MONO;
break;
case BPF_SKB_TSTAMP_UNSPEC:
if (tstamp)
return -EINVAL;
skb->tstamp = 0;
- skb->mono_delivery_time = 0;
+ skb->tstamp_type = CLOCK_REAL;
break;
default:
return -EINVAL;
@@ -9422,7 +9422,7 @@ static struct bpf_insn *bpf_convert_tstamp_read(const struct bpf_prog *prog,
TC_AT_INGRESS_MASK | SKB_MONO_DELIVERY_TIME_MASK);
*insn++ = BPF_JMP32_IMM(BPF_JNE, tmp_reg,
TC_AT_INGRESS_MASK | SKB_MONO_DELIVERY_TIME_MASK, 2);
- /* skb->tc_at_ingress && skb->mono_delivery_time,
+ /* skb->tc_at_ingress && skb->tstamp_type:1,
* read 0 as the (rcv) timestamp.
*/
*insn++ = BPF_MOV64_IMM(value_reg, 0);
@@ -9447,7 +9447,7 @@ static struct bpf_insn *bpf_convert_tstamp_write(const struct bpf_prog *prog,
* the bpf prog is aware the tstamp could have delivery time.
* Thus, write skb->tstamp as is if tstamp_type_access is true.
* Otherwise, writing at ingress will have to clear the
- * mono_delivery_time bit also.
+ * mono_delivery_time (skb->tstamp_type:1)bit also.
*/
if (!prog->tstamp_type_access) {
__u8 tmp_reg = BPF_REG_AX;
diff --git a/net/ipv4/inet_fragment.c b/net/ipv4/inet_fragment.c
index faaec92a46ac..d179a2c84222 100644
--- a/net/ipv4/inet_fragment.c
+++ b/net/ipv4/inet_fragment.c
@@ -619,7 +619,7 @@ void inet_frag_reasm_finish(struct inet_frag_queue *q, struct sk_buff *head,
skb_mark_not_on_list(head);
head->prev = NULL;
head->tstamp = q->stamp;
- head->mono_delivery_time = q->mono_delivery_time;
+ head->tstamp_type = q->tstamp_type;
if (sk)
refcount_add(sum_truesize - head_truesize, &sk->sk_wmem_alloc);
diff --git a/net/ipv4/ip_fragment.c b/net/ipv4/ip_fragment.c
index fb947d1613fe..787aa86800f5 100644
--- a/net/ipv4/ip_fragment.c
+++ b/net/ipv4/ip_fragment.c
@@ -355,7 +355,7 @@ static int ip_frag_queue(struct ipq *qp, struct sk_buff *skb)
qp->iif = dev->ifindex;
qp->q.stamp = skb->tstamp;
- qp->q.mono_delivery_time = skb->mono_delivery_time;
+ qp->q.tstamp_type = skb->tstamp_type;
qp->q.meat += skb->len;
qp->ecn |= ecn;
add_frag_mem_limit(qp->q.fqdir, skb->truesize);
diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c
index 1fe794967211..62e457f7c02c 100644
--- a/net/ipv4/ip_output.c
+++ b/net/ipv4/ip_output.c
@@ -764,7 +764,7 @@ int ip_do_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
{
struct iphdr *iph;
struct sk_buff *skb2;
- bool mono_delivery_time = skb->mono_delivery_time;
+ u8 tstamp_type = skb->tstamp_type;
struct rtable *rt = skb_rtable(skb);
unsigned int mtu, hlen, ll_rs;
struct ip_fraglist_iter iter;
@@ -856,7 +856,7 @@ int ip_do_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
}
}
- skb_set_delivery_time(skb, tstamp, mono_delivery_time);
+ skb_set_delivery_time(skb, tstamp, tstamp_type);
err = output(net, sk, skb);
if (!err)
@@ -912,7 +912,7 @@ int ip_do_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
/*
* Put this fragment into the sending queue.
*/
- skb_set_delivery_time(skb2, tstamp, mono_delivery_time);
+ skb_set_delivery_time(skb2, tstamp, tstamp_type);
err = output(net, sk, skb2);
if (err)
goto fail;
@@ -1649,7 +1649,7 @@ void ip_send_unicast_reply(struct sock *sk, struct sk_buff *skb,
arg->csumoffset) = csum_fold(csum_add(nskb->csum,
arg->csum));
nskb->ip_summed = CHECKSUM_NONE;
- nskb->mono_delivery_time = !!transmit_time;
+ nskb->tstamp_type = !!transmit_time;
if (txhash)
skb_set_hash(nskb, txhash, PKT_HASH_TYPE_L4);
ip_push_pending_frames(sk, &fl4);
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 9282fafc0e61..42e6ed1decf4 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -1299,7 +1299,7 @@ static int __tcp_transmit_skb(struct sock *sk, struct sk_buff *skb,
tp = tcp_sk(sk);
prior_wstamp = tp->tcp_wstamp_ns;
tp->tcp_wstamp_ns = max(tp->tcp_wstamp_ns, tp->tcp_clock_cache);
- skb_set_delivery_time(skb, tp->tcp_wstamp_ns, true);
+ skb_set_delivery_time(skb, tp->tcp_wstamp_ns, CLOCK_MONO);
if (clone_it) {
oskb = skb;
@@ -1649,7 +1649,7 @@ int tcp_fragment(struct sock *sk, enum tcp_queue tcp_queue,
skb_split(skb, buff, len);
- skb_set_delivery_time(buff, skb->tstamp, true);
+ skb_set_delivery_time(buff, skb->tstamp, CLOCK_MONO);
tcp_fragment_tstamp(skb, buff);
old_factor = tcp_skb_pcount(skb);
@@ -2730,7 +2730,7 @@ static bool tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle,
if (unlikely(tp->repair) && tp->repair_queue == TCP_SEND_QUEUE) {
/* "skb_mstamp_ns" is used as a start point for the retransmit timer */
tp->tcp_wstamp_ns = tp->tcp_clock_cache;
- skb_set_delivery_time(skb, tp->tcp_wstamp_ns, true);
+ skb_set_delivery_time(skb, tp->tcp_wstamp_ns, CLOCK_MONO);
list_move_tail(&skb->tcp_tsorted_anchor, &tp->tsorted_sent_queue);
tcp_init_tso_segs(skb, mss_now);
goto repair; /* Skip network transmission */
@@ -3713,11 +3713,11 @@ struct sk_buff *tcp_make_synack(const struct sock *sk, struct dst_entry *dst,
#ifdef CONFIG_SYN_COOKIES
if (unlikely(synack_type == TCP_SYNACK_COOKIE && ireq->tstamp_ok))
skb_set_delivery_time(skb, cookie_init_timestamp(req, now),
- true);
+ CLOCK_MONO);
else
#endif
{
- skb_set_delivery_time(skb, now, true);
+ skb_set_delivery_time(skb, now, CLOCK_MONO);
if (!tcp_rsk(req)->snt_synack) /* Timestamp first SYNACK */
tcp_rsk(req)->snt_synack = tcp_skb_timestamp_us(skb);
}
@@ -3804,7 +3804,7 @@ struct sk_buff *tcp_make_synack(const struct sock *sk, struct dst_entry *dst,
bpf_skops_write_hdr_opt((struct sock *)sk, skb, req, syn_skb,
synack_type, &opts);
- skb_set_delivery_time(skb, now, true);
+ skb_set_delivery_time(skb, now, CLOCK_MONO);
tcp_add_tx_delay(skb, tp);
return skb;
@@ -3988,7 +3988,7 @@ static int tcp_send_syn_data(struct sock *sk, struct sk_buff *syn)
err = tcp_transmit_skb(sk, syn_data, 1, sk->sk_allocation);
- skb_set_delivery_time(syn, syn_data->skb_mstamp_ns, true);
+ skb_set_delivery_time(syn, syn_data->skb_mstamp_ns, CLOCK_MONO);
/* Now full SYN+DATA was cloned and sent (or not),
* remove the SYN from the original skb (syn_data)
diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c
index b9dd3a66e423..a9e819115622 100644
--- a/net/ipv6/ip6_output.c
+++ b/net/ipv6/ip6_output.c
@@ -859,7 +859,7 @@ int ip6_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
struct rt6_info *rt = (struct rt6_info *)skb_dst(skb);
struct ipv6_pinfo *np = skb->sk && !dev_recursion_level() ?
inet6_sk(skb->sk) : NULL;
- bool mono_delivery_time = skb->mono_delivery_time;
+ u8 tstamp_type = skb->tstamp_type;
struct ip6_frag_state state;
unsigned int mtu, hlen, nexthdr_offset;
ktime_t tstamp = skb->tstamp;
@@ -955,7 +955,7 @@ int ip6_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
if (iter.frag)
ip6_fraglist_prepare(skb, &iter);
- skb_set_delivery_time(skb, tstamp, mono_delivery_time);
+ skb_set_delivery_time(skb, tstamp, tstamp_type);
err = output(net, sk, skb);
if (!err)
IP6_INC_STATS(net, ip6_dst_idev(&rt->dst),
@@ -1016,7 +1016,7 @@ int ip6_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
/*
* Put this fragment into the sending queue.
*/
- skb_set_delivery_time(frag, tstamp, mono_delivery_time);
+ skb_set_delivery_time(frag, tstamp, tstamp_type);
err = output(net, sk, frag);
if (err)
goto fail;
diff --git a/net/ipv6/netfilter.c b/net/ipv6/netfilter.c
index 53d255838e6a..e0c2347b4dc6 100644
--- a/net/ipv6/netfilter.c
+++ b/net/ipv6/netfilter.c
@@ -126,7 +126,7 @@ int br_ip6_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
struct sk_buff *))
{
int frag_max_size = BR_INPUT_SKB_CB(skb)->frag_max_size;
- bool mono_delivery_time = skb->mono_delivery_time;
+ u8 tstamp_type = skb->tstamp_type;
ktime_t tstamp = skb->tstamp;
struct ip6_frag_state state;
u8 *prevhdr, nexthdr = 0;
@@ -192,7 +192,7 @@ int br_ip6_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
if (iter.frag)
ip6_fraglist_prepare(skb, &iter);
- skb_set_delivery_time(skb, tstamp, mono_delivery_time);
+ skb_set_delivery_time(skb, tstamp, tstamp_type);
err = output(net, sk, data, skb);
if (err || !iter.frag)
break;
@@ -225,7 +225,7 @@ int br_ip6_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
goto blackhole;
}
- skb_set_delivery_time(skb2, tstamp, mono_delivery_time);
+ skb_set_delivery_time(skb2, tstamp, tstamp_type);
err = output(net, sk, data, skb2);
if (err)
goto blackhole;
diff --git a/net/ipv6/netfilter/nf_conntrack_reasm.c b/net/ipv6/netfilter/nf_conntrack_reasm.c
index d0dcbaca1994..5cc5d823d33f 100644
--- a/net/ipv6/netfilter/nf_conntrack_reasm.c
+++ b/net/ipv6/netfilter/nf_conntrack_reasm.c
@@ -264,7 +264,7 @@ static int nf_ct_frag6_queue(struct frag_queue *fq, struct sk_buff *skb,
fq->iif = dev->ifindex;
fq->q.stamp = skb->tstamp;
- fq->q.mono_delivery_time = skb->mono_delivery_time;
+ fq->q.tstamp_type = skb->tstamp_type;
fq->q.meat += skb->len;
fq->ecn |= ecn;
if (payload_len > fq->q.max_size)
diff --git a/net/ipv6/reassembly.c b/net/ipv6/reassembly.c
index acb4f119e11f..ea724ff558b4 100644
--- a/net/ipv6/reassembly.c
+++ b/net/ipv6/reassembly.c
@@ -198,7 +198,7 @@ static int ip6_frag_queue(struct frag_queue *fq, struct sk_buff *skb,
fq->iif = dev->ifindex;
fq->q.stamp = skb->tstamp;
- fq->q.mono_delivery_time = skb->mono_delivery_time;
+ fq->q.tstamp_type = skb->tstamp_type;
fq->q.meat += skb->len;
fq->ecn |= ecn;
add_frag_mem_limit(fq->q.fqdir, skb->truesize);
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 3aa9da5c9a66..b60196061489 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -975,7 +975,7 @@ static void tcp_v6_send_response(const struct sock *sk, struct sk_buff *skb, u32
mark = inet_twsk(sk)->tw_mark;
else
mark = READ_ONCE(sk->sk_mark);
- skb_set_delivery_time(buff, tcp_transmit_time(sk), true);
+ skb_set_delivery_time(buff, tcp_transmit_time(sk), CLOCK_MONO);
}
if (txhash) {
/* autoflowlabel/skb_get_hash_flowi6 rely on buff->hash */
diff --git a/net/sched/act_bpf.c b/net/sched/act_bpf.c
index 0e3cf11ae5fc..1f8b5a3f065e 100644
--- a/net/sched/act_bpf.c
+++ b/net/sched/act_bpf.c
@@ -54,8 +54,8 @@ TC_INDIRECT_SCOPE int tcf_bpf_act(struct sk_buff *skb,
bpf_compute_data_pointers(skb);
filter_res = bpf_prog_run(filter, skb);
}
- if (unlikely(!skb->tstamp && skb->mono_delivery_time))
- skb->mono_delivery_time = 0;
+ if (unlikely(!skb->tstamp && skb->tstamp_type))
+ skb->tstamp_type = CLOCK_REAL;
if (skb_sk_is_prefetched(skb) && filter_res != TC_ACT_OK)
skb_orphan(skb);
diff --git a/net/sched/cls_bpf.c b/net/sched/cls_bpf.c
index 5e83e890f6a4..3f843e0eea3c 100644
--- a/net/sched/cls_bpf.c
+++ b/net/sched/cls_bpf.c
@@ -104,8 +104,8 @@ TC_INDIRECT_SCOPE int cls_bpf_classify(struct sk_buff *skb,
bpf_compute_data_pointers(skb);
filter_res = bpf_prog_run(prog->filter, skb);
}
- if (unlikely(!skb->tstamp && skb->mono_delivery_time))
- skb->mono_delivery_time = 0;
+ if (unlikely(!skb->tstamp && skb->tstamp_type))
+ skb->tstamp_type = CLOCK_REAL;
if (prog->exts_integrated) {
res->class = 0;
--
2.25.1
On 4/12/2024 2:01 PM, Abhishek Chauhan wrote:
> mono_delivery_time was added to check if skb->tstamp has delivery
> time in mono clock base (i.e. EDT) otherwise skb->tstamp has
> timestamp in ingress and delivery_time at egress.
>
> Renaming the bitfield from mono_delivery_time to tstamp_type is for
> extensibilty for other timestamps such as userspace timestamp
> (i.e. SO_TXTIME) set via sock opts.
>
> As we are renaming the mono_delivery_time to tstamp_type, it makes
> sense to start assigning tstamp_type based out if enum defined as
> part of this commit
>
> Earlier we used bool arg flag to check if the tstamp is mono in
> function skb_set_delivery_time, Now the signature of the functions
> accepts enum to distinguish between mono and real time
>
> Bridge driver today has no support to forward the userspace timestamp
> packets and ends up resetting the timestamp. ETF qdisc checks the
> packet coming from userspace and encounters to be 0 thereby dropping
> time sensitive packets. These changes will allow userspace timestamps
> packets to be forwarded from the bridge to NIC drivers.
>
> In future tstamp_type:1 can be extended to support userspace timestamp
> by increasing the bitfield.
>
> Link: https://lore.kernel.org/netdev/[email protected]/
> Signed-off-by: Abhishek Chauhan <[email protected]>
> ---
> Changes since v2
> - Minor changes to commit subject
>
> Changes since v1
> - Squashed the two commits into one as mentioned by Willem.
> - Introduced switch in skb_set_delivery_time.
> - Renamed and removed directionality aspects w.r.t tstamp_type
> as mentioned by Willem.
>
> include/linux/skbuff.h | 33 +++++++++++++++-------
> include/net/inet_frag.h | 4 +--
> net/bridge/netfilter/nf_conntrack_bridge.c | 6 ++--
> net/core/dev.c | 2 +-
> net/core/filter.c | 8 +++---
> net/ipv4/inet_fragment.c | 2 +-
> net/ipv4/ip_fragment.c | 2 +-
> net/ipv4/ip_output.c | 8 +++---
> net/ipv4/tcp_output.c | 14 ++++-----
> net/ipv6/ip6_output.c | 6 ++--
> net/ipv6/netfilter.c | 6 ++--
> net/ipv6/netfilter/nf_conntrack_reasm.c | 2 +-
> net/ipv6/reassembly.c | 2 +-
> net/ipv6/tcp_ipv6.c | 2 +-
> net/sched/act_bpf.c | 4 +--
> net/sched/cls_bpf.c | 4 +--
> 16 files changed, 59 insertions(+), 46 deletions(-)
>
self review :-
One more file needs to be updated here net/ieee802154/6lowpan/reassembly.c :(
Unfortunately when i compile kdev with defconfig this file never gets compiled on my internal workspace.
I will find all other instances and update accordingly.
> diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
> index 7135a3e94afd..a83a2120b57f 100644
> --- a/include/linux/skbuff.h
> +++ b/include/linux/skbuff.h
> @@ -702,6 +702,11 @@ typedef unsigned int sk_buff_data_t;
> typedef unsigned char *sk_buff_data_t;
> #endif
>
> +enum skb_tstamp_type {
> + CLOCK_REAL = 0, /* Time base is realtime */
> + CLOCK_MONO = 1, /* Time base is Monotonic */
> +};
> +
> /**
> * DOC: Basic sk_buff geometry
> *
> @@ -819,7 +824,7 @@ typedef unsigned char *sk_buff_data_t;
> * @dst_pending_confirm: need to confirm neighbour
> * @decrypted: Decrypted SKB
> * @slow_gro: state present at GRO time, slower prepare step required
> - * @mono_delivery_time: When set, skb->tstamp has the
> + * @tstamp_type: When set, skb->tstamp has the
> * delivery_time in mono clock base (i.e. EDT). Otherwise, the
> * skb->tstamp has the (rcv) timestamp at ingress and
> * delivery_time at egress.
> @@ -950,7 +955,7 @@ struct sk_buff {
> /* private: */
> __u8 __mono_tc_offset[0];
> /* public: */
> - __u8 mono_delivery_time:1; /* See SKB_MONO_DELIVERY_TIME_MASK */
> + __u8 tstamp_type:1; /* See SKB_MONO_DELIVERY_TIME_MASK */
> #ifdef CONFIG_NET_XGRESS
> __u8 tc_at_ingress:1; /* See TC_AT_INGRESS_MASK */
> __u8 tc_skip_classify:1;
> @@ -4237,7 +4242,7 @@ static inline void skb_get_new_timestampns(const struct sk_buff *skb,
> static inline void __net_timestamp(struct sk_buff *skb)
> {
> skb->tstamp = ktime_get_real();
> - skb->mono_delivery_time = 0;
> + skb->tstamp_type = CLOCK_REAL;
> }
>
> static inline ktime_t net_timedelta(ktime_t t)
> @@ -4246,10 +4251,18 @@ static inline ktime_t net_timedelta(ktime_t t)
> }
>
> static inline void skb_set_delivery_time(struct sk_buff *skb, ktime_t kt,
> - bool mono)
> + u8 tstamp_type)
> {
> skb->tstamp = kt;
> - skb->mono_delivery_time = kt && mono;
> +
> + switch (tstamp_type) {
> + case CLOCK_REAL:
> + skb->tstamp_type = CLOCK_REAL;
> + break;
> + case CLOCK_MONO:
> + skb->tstamp_type = kt && tstamp_type;
> + break;
> + }
> }
>
> DECLARE_STATIC_KEY_FALSE(netstamp_needed_key);
> @@ -4259,8 +4272,8 @@ DECLARE_STATIC_KEY_FALSE(netstamp_needed_key);
> */
> static inline void skb_clear_delivery_time(struct sk_buff *skb)
> {
> - if (skb->mono_delivery_time) {
> - skb->mono_delivery_time = 0;
> + if (skb->tstamp_type) {
> + skb->tstamp_type = CLOCK_REAL;
> if (static_branch_unlikely(&netstamp_needed_key))
> skb->tstamp = ktime_get_real();
> else
> @@ -4270,7 +4283,7 @@ static inline void skb_clear_delivery_time(struct sk_buff *skb)
>
> static inline void skb_clear_tstamp(struct sk_buff *skb)
> {
> - if (skb->mono_delivery_time)
> + if (skb->tstamp_type)
> return;
>
> skb->tstamp = 0;
> @@ -4278,7 +4291,7 @@ static inline void skb_clear_tstamp(struct sk_buff *skb)
>
> static inline ktime_t skb_tstamp(const struct sk_buff *skb)
> {
> - if (skb->mono_delivery_time)
> + if (skb->tstamp_type == CLOCK_MONO)
> return 0;
>
> return skb->tstamp;
> @@ -4286,7 +4299,7 @@ static inline ktime_t skb_tstamp(const struct sk_buff *skb)
>
> static inline ktime_t skb_tstamp_cond(const struct sk_buff *skb, bool cond)
> {
> - if (!skb->mono_delivery_time && skb->tstamp)
> + if (skb->tstamp_type != CLOCK_MONO && skb->tstamp)
> return skb->tstamp;
>
> if (static_branch_unlikely(&netstamp_needed_key) || cond)
> diff --git a/include/net/inet_frag.h b/include/net/inet_frag.h
> index 153960663ce4..5af6eb14c5db 100644
> --- a/include/net/inet_frag.h
> +++ b/include/net/inet_frag.h
> @@ -76,7 +76,7 @@ struct frag_v6_compare_key {
> * @stamp: timestamp of the last received fragment
> * @len: total length of the original datagram
> * @meat: length of received fragments so far
> - * @mono_delivery_time: stamp has a mono delivery time (EDT)
> + * @tstamp_type: stamp has a mono delivery time (EDT)
> * @flags: fragment queue flags
> * @max_size: maximum received fragment size
> * @fqdir: pointer to struct fqdir
> @@ -97,7 +97,7 @@ struct inet_frag_queue {
> ktime_t stamp;
> int len;
> int meat;
> - u8 mono_delivery_time;
> + u8 tstamp_type;
> __u8 flags;
> u16 max_size;
> struct fqdir *fqdir;
> diff --git a/net/bridge/netfilter/nf_conntrack_bridge.c b/net/bridge/netfilter/nf_conntrack_bridge.c
> index c3c51b9a6826..816bb0fde718 100644
> --- a/net/bridge/netfilter/nf_conntrack_bridge.c
> +++ b/net/bridge/netfilter/nf_conntrack_bridge.c
> @@ -32,7 +32,7 @@ static int nf_br_ip_fragment(struct net *net, struct sock *sk,
> struct sk_buff *))
> {
> int frag_max_size = BR_INPUT_SKB_CB(skb)->frag_max_size;
> - bool mono_delivery_time = skb->mono_delivery_time;
> + u8 tstamp_type = skb->tstamp_type;
> unsigned int hlen, ll_rs, mtu;
> ktime_t tstamp = skb->tstamp;
> struct ip_frag_state state;
> @@ -82,7 +82,7 @@ static int nf_br_ip_fragment(struct net *net, struct sock *sk,
> if (iter.frag)
> ip_fraglist_prepare(skb, &iter);
>
> - skb_set_delivery_time(skb, tstamp, mono_delivery_time);
> + skb_set_delivery_time(skb, tstamp, tstamp_type);
> err = output(net, sk, data, skb);
> if (err || !iter.frag)
> break;
> @@ -113,7 +113,7 @@ static int nf_br_ip_fragment(struct net *net, struct sock *sk,
> goto blackhole;
> }
>
> - skb_set_delivery_time(skb2, tstamp, mono_delivery_time);
> + skb_set_delivery_time(skb2, tstamp, tstamp_type);
> err = output(net, sk, data, skb2);
> if (err)
> goto blackhole;
> diff --git a/net/core/dev.c b/net/core/dev.c
> index 854a3a28a8d8..77a43c05dfe3 100644
> --- a/net/core/dev.c
> +++ b/net/core/dev.c
> @@ -2146,7 +2146,7 @@ EXPORT_SYMBOL(net_disable_timestamp);
> static inline void net_timestamp_set(struct sk_buff *skb)
> {
> skb->tstamp = 0;
> - skb->mono_delivery_time = 0;
> + skb->tstamp_type = CLOCK_REAL;
> if (static_branch_unlikely(&netstamp_needed_key))
> skb->tstamp = ktime_get_real();
> }
> diff --git a/net/core/filter.c b/net/core/filter.c
> index 8d185d99a643..8bb45423df52 100644
> --- a/net/core/filter.c
> +++ b/net/core/filter.c
> @@ -7709,13 +7709,13 @@ BPF_CALL_3(bpf_skb_set_tstamp, struct sk_buff *, skb,
> if (!tstamp)
> return -EINVAL;
> skb->tstamp = tstamp;
> - skb->mono_delivery_time = 1;
> + skb->tstamp_type = CLOCK_MONO;
> break;
> case BPF_SKB_TSTAMP_UNSPEC:
> if (tstamp)
> return -EINVAL;
> skb->tstamp = 0;
> - skb->mono_delivery_time = 0;
> + skb->tstamp_type = CLOCK_REAL;
> break;
> default:
> return -EINVAL;
> @@ -9422,7 +9422,7 @@ static struct bpf_insn *bpf_convert_tstamp_read(const struct bpf_prog *prog,
> TC_AT_INGRESS_MASK | SKB_MONO_DELIVERY_TIME_MASK);
> *insn++ = BPF_JMP32_IMM(BPF_JNE, tmp_reg,
> TC_AT_INGRESS_MASK | SKB_MONO_DELIVERY_TIME_MASK, 2);
> - /* skb->tc_at_ingress && skb->mono_delivery_time,
> + /* skb->tc_at_ingress && skb->tstamp_type:1,
> * read 0 as the (rcv) timestamp.
> */
> *insn++ = BPF_MOV64_IMM(value_reg, 0);
> @@ -9447,7 +9447,7 @@ static struct bpf_insn *bpf_convert_tstamp_write(const struct bpf_prog *prog,
> * the bpf prog is aware the tstamp could have delivery time.
> * Thus, write skb->tstamp as is if tstamp_type_access is true.
> * Otherwise, writing at ingress will have to clear the
> - * mono_delivery_time bit also.
> + * mono_delivery_time (skb->tstamp_type:1)bit also.
> */
> if (!prog->tstamp_type_access) {
> __u8 tmp_reg = BPF_REG_AX;
> diff --git a/net/ipv4/inet_fragment.c b/net/ipv4/inet_fragment.c
> index faaec92a46ac..d179a2c84222 100644
> --- a/net/ipv4/inet_fragment.c
> +++ b/net/ipv4/inet_fragment.c
> @@ -619,7 +619,7 @@ void inet_frag_reasm_finish(struct inet_frag_queue *q, struct sk_buff *head,
> skb_mark_not_on_list(head);
> head->prev = NULL;
> head->tstamp = q->stamp;
> - head->mono_delivery_time = q->mono_delivery_time;
> + head->tstamp_type = q->tstamp_type;
>
> if (sk)
> refcount_add(sum_truesize - head_truesize, &sk->sk_wmem_alloc);
> diff --git a/net/ipv4/ip_fragment.c b/net/ipv4/ip_fragment.c
> index fb947d1613fe..787aa86800f5 100644
> --- a/net/ipv4/ip_fragment.c
> +++ b/net/ipv4/ip_fragment.c
> @@ -355,7 +355,7 @@ static int ip_frag_queue(struct ipq *qp, struct sk_buff *skb)
> qp->iif = dev->ifindex;
>
> qp->q.stamp = skb->tstamp;
> - qp->q.mono_delivery_time = skb->mono_delivery_time;
> + qp->q.tstamp_type = skb->tstamp_type;
> qp->q.meat += skb->len;
> qp->ecn |= ecn;
> add_frag_mem_limit(qp->q.fqdir, skb->truesize);
> diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c
> index 1fe794967211..62e457f7c02c 100644
> --- a/net/ipv4/ip_output.c
> +++ b/net/ipv4/ip_output.c
> @@ -764,7 +764,7 @@ int ip_do_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
> {
> struct iphdr *iph;
> struct sk_buff *skb2;
> - bool mono_delivery_time = skb->mono_delivery_time;
> + u8 tstamp_type = skb->tstamp_type;
> struct rtable *rt = skb_rtable(skb);
> unsigned int mtu, hlen, ll_rs;
> struct ip_fraglist_iter iter;
> @@ -856,7 +856,7 @@ int ip_do_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
> }
> }
>
> - skb_set_delivery_time(skb, tstamp, mono_delivery_time);
> + skb_set_delivery_time(skb, tstamp, tstamp_type);
> err = output(net, sk, skb);
>
> if (!err)
> @@ -912,7 +912,7 @@ int ip_do_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
> /*
> * Put this fragment into the sending queue.
> */
> - skb_set_delivery_time(skb2, tstamp, mono_delivery_time);
> + skb_set_delivery_time(skb2, tstamp, tstamp_type);
> err = output(net, sk, skb2);
> if (err)
> goto fail;
> @@ -1649,7 +1649,7 @@ void ip_send_unicast_reply(struct sock *sk, struct sk_buff *skb,
> arg->csumoffset) = csum_fold(csum_add(nskb->csum,
> arg->csum));
> nskb->ip_summed = CHECKSUM_NONE;
> - nskb->mono_delivery_time = !!transmit_time;
> + nskb->tstamp_type = !!transmit_time;
> if (txhash)
> skb_set_hash(nskb, txhash, PKT_HASH_TYPE_L4);
> ip_push_pending_frames(sk, &fl4);
> diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
> index 9282fafc0e61..42e6ed1decf4 100644
> --- a/net/ipv4/tcp_output.c
> +++ b/net/ipv4/tcp_output.c
> @@ -1299,7 +1299,7 @@ static int __tcp_transmit_skb(struct sock *sk, struct sk_buff *skb,
> tp = tcp_sk(sk);
> prior_wstamp = tp->tcp_wstamp_ns;
> tp->tcp_wstamp_ns = max(tp->tcp_wstamp_ns, tp->tcp_clock_cache);
> - skb_set_delivery_time(skb, tp->tcp_wstamp_ns, true);
> + skb_set_delivery_time(skb, tp->tcp_wstamp_ns, CLOCK_MONO);
> if (clone_it) {
> oskb = skb;
>
> @@ -1649,7 +1649,7 @@ int tcp_fragment(struct sock *sk, enum tcp_queue tcp_queue,
>
> skb_split(skb, buff, len);
>
> - skb_set_delivery_time(buff, skb->tstamp, true);
> + skb_set_delivery_time(buff, skb->tstamp, CLOCK_MONO);
> tcp_fragment_tstamp(skb, buff);
>
> old_factor = tcp_skb_pcount(skb);
> @@ -2730,7 +2730,7 @@ static bool tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle,
> if (unlikely(tp->repair) && tp->repair_queue == TCP_SEND_QUEUE) {
> /* "skb_mstamp_ns" is used as a start point for the retransmit timer */
> tp->tcp_wstamp_ns = tp->tcp_clock_cache;
> - skb_set_delivery_time(skb, tp->tcp_wstamp_ns, true);
> + skb_set_delivery_time(skb, tp->tcp_wstamp_ns, CLOCK_MONO);
> list_move_tail(&skb->tcp_tsorted_anchor, &tp->tsorted_sent_queue);
> tcp_init_tso_segs(skb, mss_now);
> goto repair; /* Skip network transmission */
> @@ -3713,11 +3713,11 @@ struct sk_buff *tcp_make_synack(const struct sock *sk, struct dst_entry *dst,
> #ifdef CONFIG_SYN_COOKIES
> if (unlikely(synack_type == TCP_SYNACK_COOKIE && ireq->tstamp_ok))
> skb_set_delivery_time(skb, cookie_init_timestamp(req, now),
> - true);
> + CLOCK_MONO);
> else
> #endif
> {
> - skb_set_delivery_time(skb, now, true);
> + skb_set_delivery_time(skb, now, CLOCK_MONO);
> if (!tcp_rsk(req)->snt_synack) /* Timestamp first SYNACK */
> tcp_rsk(req)->snt_synack = tcp_skb_timestamp_us(skb);
> }
> @@ -3804,7 +3804,7 @@ struct sk_buff *tcp_make_synack(const struct sock *sk, struct dst_entry *dst,
> bpf_skops_write_hdr_opt((struct sock *)sk, skb, req, syn_skb,
> synack_type, &opts);
>
> - skb_set_delivery_time(skb, now, true);
> + skb_set_delivery_time(skb, now, CLOCK_MONO);
> tcp_add_tx_delay(skb, tp);
>
> return skb;
> @@ -3988,7 +3988,7 @@ static int tcp_send_syn_data(struct sock *sk, struct sk_buff *syn)
>
> err = tcp_transmit_skb(sk, syn_data, 1, sk->sk_allocation);
>
> - skb_set_delivery_time(syn, syn_data->skb_mstamp_ns, true);
> + skb_set_delivery_time(syn, syn_data->skb_mstamp_ns, CLOCK_MONO);
>
> /* Now full SYN+DATA was cloned and sent (or not),
> * remove the SYN from the original skb (syn_data)
> diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c
> index b9dd3a66e423..a9e819115622 100644
> --- a/net/ipv6/ip6_output.c
> +++ b/net/ipv6/ip6_output.c
> @@ -859,7 +859,7 @@ int ip6_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
> struct rt6_info *rt = (struct rt6_info *)skb_dst(skb);
> struct ipv6_pinfo *np = skb->sk && !dev_recursion_level() ?
> inet6_sk(skb->sk) : NULL;
> - bool mono_delivery_time = skb->mono_delivery_time;
> + u8 tstamp_type = skb->tstamp_type;
> struct ip6_frag_state state;
> unsigned int mtu, hlen, nexthdr_offset;
> ktime_t tstamp = skb->tstamp;
> @@ -955,7 +955,7 @@ int ip6_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
> if (iter.frag)
> ip6_fraglist_prepare(skb, &iter);
>
> - skb_set_delivery_time(skb, tstamp, mono_delivery_time);
> + skb_set_delivery_time(skb, tstamp, tstamp_type);
> err = output(net, sk, skb);
> if (!err)
> IP6_INC_STATS(net, ip6_dst_idev(&rt->dst),
> @@ -1016,7 +1016,7 @@ int ip6_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
> /*
> * Put this fragment into the sending queue.
> */
> - skb_set_delivery_time(frag, tstamp, mono_delivery_time);
> + skb_set_delivery_time(frag, tstamp, tstamp_type);
> err = output(net, sk, frag);
> if (err)
> goto fail;
> diff --git a/net/ipv6/netfilter.c b/net/ipv6/netfilter.c
> index 53d255838e6a..e0c2347b4dc6 100644
> --- a/net/ipv6/netfilter.c
> +++ b/net/ipv6/netfilter.c
> @@ -126,7 +126,7 @@ int br_ip6_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
> struct sk_buff *))
> {
> int frag_max_size = BR_INPUT_SKB_CB(skb)->frag_max_size;
> - bool mono_delivery_time = skb->mono_delivery_time;
> + u8 tstamp_type = skb->tstamp_type;
> ktime_t tstamp = skb->tstamp;
> struct ip6_frag_state state;
> u8 *prevhdr, nexthdr = 0;
> @@ -192,7 +192,7 @@ int br_ip6_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
> if (iter.frag)
> ip6_fraglist_prepare(skb, &iter);
>
> - skb_set_delivery_time(skb, tstamp, mono_delivery_time);
> + skb_set_delivery_time(skb, tstamp, tstamp_type);
> err = output(net, sk, data, skb);
> if (err || !iter.frag)
> break;
> @@ -225,7 +225,7 @@ int br_ip6_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
> goto blackhole;
> }
>
> - skb_set_delivery_time(skb2, tstamp, mono_delivery_time);
> + skb_set_delivery_time(skb2, tstamp, tstamp_type);
> err = output(net, sk, data, skb2);
> if (err)
> goto blackhole;
> diff --git a/net/ipv6/netfilter/nf_conntrack_reasm.c b/net/ipv6/netfilter/nf_conntrack_reasm.c
> index d0dcbaca1994..5cc5d823d33f 100644
> --- a/net/ipv6/netfilter/nf_conntrack_reasm.c
> +++ b/net/ipv6/netfilter/nf_conntrack_reasm.c
> @@ -264,7 +264,7 @@ static int nf_ct_frag6_queue(struct frag_queue *fq, struct sk_buff *skb,
> fq->iif = dev->ifindex;
>
> fq->q.stamp = skb->tstamp;
> - fq->q.mono_delivery_time = skb->mono_delivery_time;
> + fq->q.tstamp_type = skb->tstamp_type;
> fq->q.meat += skb->len;
> fq->ecn |= ecn;
> if (payload_len > fq->q.max_size)
> diff --git a/net/ipv6/reassembly.c b/net/ipv6/reassembly.c
> index acb4f119e11f..ea724ff558b4 100644
> --- a/net/ipv6/reassembly.c
> +++ b/net/ipv6/reassembly.c
> @@ -198,7 +198,7 @@ static int ip6_frag_queue(struct frag_queue *fq, struct sk_buff *skb,
> fq->iif = dev->ifindex;
>
> fq->q.stamp = skb->tstamp;
> - fq->q.mono_delivery_time = skb->mono_delivery_time;
> + fq->q.tstamp_type = skb->tstamp_type;
> fq->q.meat += skb->len;
> fq->ecn |= ecn;
> add_frag_mem_limit(fq->q.fqdir, skb->truesize);
> diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
> index 3aa9da5c9a66..b60196061489 100644
> --- a/net/ipv6/tcp_ipv6.c
> +++ b/net/ipv6/tcp_ipv6.c
> @@ -975,7 +975,7 @@ static void tcp_v6_send_response(const struct sock *sk, struct sk_buff *skb, u32
> mark = inet_twsk(sk)->tw_mark;
> else
> mark = READ_ONCE(sk->sk_mark);
> - skb_set_delivery_time(buff, tcp_transmit_time(sk), true);
> + skb_set_delivery_time(buff, tcp_transmit_time(sk), CLOCK_MONO);
> }
> if (txhash) {
> /* autoflowlabel/skb_get_hash_flowi6 rely on buff->hash */
> diff --git a/net/sched/act_bpf.c b/net/sched/act_bpf.c
> index 0e3cf11ae5fc..1f8b5a3f065e 100644
> --- a/net/sched/act_bpf.c
> +++ b/net/sched/act_bpf.c
> @@ -54,8 +54,8 @@ TC_INDIRECT_SCOPE int tcf_bpf_act(struct sk_buff *skb,
> bpf_compute_data_pointers(skb);
> filter_res = bpf_prog_run(filter, skb);
> }
> - if (unlikely(!skb->tstamp && skb->mono_delivery_time))
> - skb->mono_delivery_time = 0;
> + if (unlikely(!skb->tstamp && skb->tstamp_type))
> + skb->tstamp_type = CLOCK_REAL;
> if (skb_sk_is_prefetched(skb) && filter_res != TC_ACT_OK)
> skb_orphan(skb);
>
> diff --git a/net/sched/cls_bpf.c b/net/sched/cls_bpf.c
> index 5e83e890f6a4..3f843e0eea3c 100644
> --- a/net/sched/cls_bpf.c
> +++ b/net/sched/cls_bpf.c
> @@ -104,8 +104,8 @@ TC_INDIRECT_SCOPE int cls_bpf_classify(struct sk_buff *skb,
> bpf_compute_data_pointers(skb);
> filter_res = bpf_prog_run(prog->filter, skb);
> }
> - if (unlikely(!skb->tstamp && skb->mono_delivery_time))
> - skb->mono_delivery_time = 0;
> + if (unlikely(!skb->tstamp && skb->tstamp_type))
> + skb->tstamp_type = CLOCK_REAL;
>
> if (prog->exts_integrated) {
> res->class = 0;
Abhishek Chauhan wrote:
> mono_delivery_time was added to check if skb->tstamp has delivery
> time in mono clock base (i.e. EDT) otherwise skb->tstamp has
> timestamp in ingress and delivery_time at egress.
>
> Renaming the bitfield from mono_delivery_time to tstamp_type is for
> extensibilty for other timestamps such as userspace timestamp
> (i.e. SO_TXTIME) set via sock opts.
>
> As we are renaming the mono_delivery_time to tstamp_type, it makes
> sense to start assigning tstamp_type based out if enum defined as
> part of this commit
>
> Earlier we used bool arg flag to check if the tstamp is mono in
> function skb_set_delivery_time, Now the signature of the functions
> accepts enum to distinguish between mono and real time
>
> Bridge driver today has no support to forward the userspace timestamp
> packets and ends up resetting the timestamp. ETF qdisc checks the
> packet coming from userspace and encounters to be 0 thereby dropping
> time sensitive packets. These changes will allow userspace timestamps
> packets to be forwarded from the bridge to NIC drivers.
>
> In future tstamp_type:1 can be extended to support userspace timestamp
> by increasing the bitfield.
>
> Link: https://lore.kernel.org/netdev/[email protected]/
> Signed-off-by: Abhishek Chauhan <[email protected]>
> ---
> Changes since v2
> - Minor changes to commit subject
>
> Changes since v1
> - Squashed the two commits into one as mentioned by Willem.
> - Introduced switch in skb_set_delivery_time.
> - Renamed and removed directionality aspects w.r.t tstamp_type
> as mentioned by Willem.
>
> include/linux/skbuff.h | 33 +++++++++++++++-------
> include/net/inet_frag.h | 4 +--
> net/bridge/netfilter/nf_conntrack_bridge.c | 6 ++--
> net/core/dev.c | 2 +-
> net/core/filter.c | 8 +++---
> net/ipv4/inet_fragment.c | 2 +-
> net/ipv4/ip_fragment.c | 2 +-
> net/ipv4/ip_output.c | 8 +++---
> net/ipv4/tcp_output.c | 14 ++++-----
> net/ipv6/ip6_output.c | 6 ++--
> net/ipv6/netfilter.c | 6 ++--
> net/ipv6/netfilter/nf_conntrack_reasm.c | 2 +-
> net/ipv6/reassembly.c | 2 +-
> net/ipv6/tcp_ipv6.c | 2 +-
> net/sched/act_bpf.c | 4 +--
> net/sched/cls_bpf.c | 4 +--
> 16 files changed, 59 insertions(+), 46 deletions(-)
>
> diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
> index 7135a3e94afd..a83a2120b57f 100644
> --- a/include/linux/skbuff.h
> +++ b/include/linux/skbuff.h
> @@ -702,6 +702,11 @@ typedef unsigned int sk_buff_data_t;
> typedef unsigned char *sk_buff_data_t;
> #endif
>
> +enum skb_tstamp_type {
> + CLOCK_REAL = 0, /* Time base is realtime */
> + CLOCK_MONO = 1, /* Time base is Monotonic */
> +};
Minor: inconsistent capitalization
> +
> /**
> * DOC: Basic sk_buff geometry
> *
> @@ -819,7 +824,7 @@ typedef unsigned char *sk_buff_data_t;
> * @dst_pending_confirm: need to confirm neighbour
> * @decrypted: Decrypted SKB
> * @slow_gro: state present at GRO time, slower prepare step required
> - * @mono_delivery_time: When set, skb->tstamp has the
> + * @tstamp_type: When set, skb->tstamp has the
> * delivery_time in mono clock base (i.e. EDT). Otherwise, the
> * skb->tstamp has the (rcv) timestamp at ingress and
> * delivery_time at egress.
> @@ -950,7 +955,7 @@ struct sk_buff {
> /* private: */
> __u8 __mono_tc_offset[0];
> /* public: */
> - __u8 mono_delivery_time:1; /* See SKB_MONO_DELIVERY_TIME_MASK */
> + __u8 tstamp_type:1; /* See SKB_MONO_DELIVERY_TIME_MASK */
Also remove reference to MONO_DELIVERY_TIME_MASK, or instead refer to
skb_tstamp_type.
> #ifdef CONFIG_NET_XGRESS
> __u8 tc_at_ingress:1; /* See TC_AT_INGRESS_MASK */
> __u8 tc_skip_classify:1;
> @@ -4237,7 +4242,7 @@ static inline void skb_get_new_timestampns(const struct sk_buff *skb,
> static inline void __net_timestamp(struct sk_buff *skb)
> {
> skb->tstamp = ktime_get_real();
> - skb->mono_delivery_time = 0;
> + skb->tstamp_type = CLOCK_REAL;
> }
>
> static inline ktime_t net_timedelta(ktime_t t)
> @@ -4246,10 +4251,18 @@ static inline ktime_t net_timedelta(ktime_t t)
> }
>
> static inline void skb_set_delivery_time(struct sk_buff *skb, ktime_t kt,
> - bool mono)
> + u8 tstamp_type)
> {
> skb->tstamp = kt;
> - skb->mono_delivery_time = kt && mono;
> +
> + switch (tstamp_type) {
> + case CLOCK_REAL:
> + skb->tstamp_type = CLOCK_REAL;
> + break;
> + case CLOCK_MONO:
> + skb->tstamp_type = kt && tstamp_type;
> + break;
> + }
Technically this leaves the tstamp_type undefined if (skb, 0, CLOCK_REAL)
> }
>
> DECLARE_STATIC_KEY_FALSE(netstamp_needed_key);
> @@ -4259,8 +4272,8 @@ DECLARE_STATIC_KEY_FALSE(netstamp_needed_key);
> */
> static inline void skb_clear_delivery_time(struct sk_buff *skb)
> {
> - if (skb->mono_delivery_time) {
> - skb->mono_delivery_time = 0;
> + if (skb->tstamp_type) {
> + skb->tstamp_type = CLOCK_REAL;
> if (static_branch_unlikely(&netstamp_needed_key))
> skb->tstamp = ktime_get_real();
> else
> @@ -4270,7 +4283,7 @@ static inline void skb_clear_delivery_time(struct sk_buff *skb)
>
> static inline void skb_clear_tstamp(struct sk_buff *skb)
> {
> - if (skb->mono_delivery_time)
> + if (skb->tstamp_type)
> return;
>
> skb->tstamp = 0;
> @@ -4278,7 +4291,7 @@ static inline void skb_clear_tstamp(struct sk_buff *skb)
>
> static inline ktime_t skb_tstamp(const struct sk_buff *skb)
> {
> - if (skb->mono_delivery_time)
> + if (skb->tstamp_type == CLOCK_MONO)
> return 0;
Should this be if (skb->tstamp_type), in line with skb_clear_tstamp,
right above?
> @@ -1649,7 +1649,7 @@ void ip_send_unicast_reply(struct sock *sk, struct sk_buff *skb,
> arg->csumoffset) = csum_fold(csum_add(nskb->csum,
> arg->csum));
> nskb->ip_summed = CHECKSUM_NONE;
> - nskb->mono_delivery_time = !!transmit_time;
> + nskb->tstamp_type = !!transmit_time;
In anticipation of more tstamp_types, explicitly set to CLOCK_MONO.
> if (txhash)
> skb_set_hash(nskb, txhash, PKT_HASH_TYPE_L4);
> ip_push_pending_frames(sk, &fl4);
On 4/13/2024 11:54 AM, Willem de Bruijn wrote:
> Abhishek Chauhan wrote:
>> mono_delivery_time was added to check if skb->tstamp has delivery
>> time in mono clock base (i.e. EDT) otherwise skb->tstamp has
>> timestamp in ingress and delivery_time at egress.
>>
>> Renaming the bitfield from mono_delivery_time to tstamp_type is for
>> extensibilty for other timestamps such as userspace timestamp
>> (i.e. SO_TXTIME) set via sock opts.
>>
>> As we are renaming the mono_delivery_time to tstamp_type, it makes
>> sense to start assigning tstamp_type based out if enum defined as
>> part of this commit
>>
>> Earlier we used bool arg flag to check if the tstamp is mono in
>> function skb_set_delivery_time, Now the signature of the functions
>> accepts enum to distinguish between mono and real time
>>
>> Bridge driver today has no support to forward the userspace timestamp
>> packets and ends up resetting the timestamp. ETF qdisc checks the
>> packet coming from userspace and encounters to be 0 thereby dropping
>> time sensitive packets. These changes will allow userspace timestamps
>> packets to be forwarded from the bridge to NIC drivers.
>>
>> In future tstamp_type:1 can be extended to support userspace timestamp
>> by increasing the bitfield.
>>
>> Link: https://lore.kernel.org/netdev/[email protected]/
>> Signed-off-by: Abhishek Chauhan <[email protected]>
>> ---
>> Changes since v2
>> - Minor changes to commit subject
>>
>> Changes since v1
>> - Squashed the two commits into one as mentioned by Willem.
>> - Introduced switch in skb_set_delivery_time.
>> - Renamed and removed directionality aspects w.r.t tstamp_type
>> as mentioned by Willem.
>>
>> include/linux/skbuff.h | 33 +++++++++++++++-------
>> include/net/inet_frag.h | 4 +--
>> net/bridge/netfilter/nf_conntrack_bridge.c | 6 ++--
>> net/core/dev.c | 2 +-
>> net/core/filter.c | 8 +++---
>> net/ipv4/inet_fragment.c | 2 +-
>> net/ipv4/ip_fragment.c | 2 +-
>> net/ipv4/ip_output.c | 8 +++---
>> net/ipv4/tcp_output.c | 14 ++++-----
>> net/ipv6/ip6_output.c | 6 ++--
>> net/ipv6/netfilter.c | 6 ++--
>> net/ipv6/netfilter/nf_conntrack_reasm.c | 2 +-
>> net/ipv6/reassembly.c | 2 +-
>> net/ipv6/tcp_ipv6.c | 2 +-
>> net/sched/act_bpf.c | 4 +--
>> net/sched/cls_bpf.c | 4 +--
>> 16 files changed, 59 insertions(+), 46 deletions(-)
>>
>> diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
>> index 7135a3e94afd..a83a2120b57f 100644
>> --- a/include/linux/skbuff.h
>> +++ b/include/linux/skbuff.h
>> @@ -702,6 +702,11 @@ typedef unsigned int sk_buff_data_t;
>> typedef unsigned char *sk_buff_data_t;
>> #endif
>>
>> +enum skb_tstamp_type {
>> + CLOCK_REAL = 0, /* Time base is realtime */
>> + CLOCK_MONO = 1, /* Time base is Monotonic */
>> +};
>
> Minor: inconsistent capitalization
>
I will fix this.
>> +
>> /**
>> * DOC: Basic sk_buff geometry
>> *
>> @@ -819,7 +824,7 @@ typedef unsigned char *sk_buff_data_t;
>> * @dst_pending_confirm: need to confirm neighbour
>> * @decrypted: Decrypted SKB
>> * @slow_gro: state present at GRO time, slower prepare step required
>> - * @mono_delivery_time: When set, skb->tstamp has the
>> + * @tstamp_type: When set, skb->tstamp has the
>> * delivery_time in mono clock base (i.e. EDT). Otherwise, the
>> * skb->tstamp has the (rcv) timestamp at ingress and
>> * delivery_time at egress.
>> @@ -950,7 +955,7 @@ struct sk_buff {
>> /* private: */
>> __u8 __mono_tc_offset[0];
>> /* public: */
>> - __u8 mono_delivery_time:1; /* See SKB_MONO_DELIVERY_TIME_MASK */
>> + __u8 tstamp_type:1; /* See SKB_MONO_DELIVERY_TIME_MASK */
>
> Also remove reference to MONO_DELIVERY_TIME_MASK, or instead refer to
> skb_tstamp_type.
>
>> #ifdef CONFIG_NET_XGRESS
>> __u8 tc_at_ingress:1; /* See TC_AT_INGRESS_MASK */
>> __u8 tc_skip_classify:1;
>> @@ -4237,7 +4242,7 @@ static inline void skb_get_new_timestampns(const struct sk_buff *skb,
>> static inline void __net_timestamp(struct sk_buff *skb)
>> {
>> skb->tstamp = ktime_get_real();
>> - skb->mono_delivery_time = 0;
>> + skb->tstamp_type = CLOCK_REAL;
>> }
>>
>> static inline ktime_t net_timedelta(ktime_t t)
>> @@ -4246,10 +4251,18 @@ static inline ktime_t net_timedelta(ktime_t t)
>> }
>>
>> static inline void skb_set_delivery_time(struct sk_buff *skb, ktime_t kt,
>> - bool mono)
>> + u8 tstamp_type)
>> {
>> skb->tstamp = kt;
>> - skb->mono_delivery_time = kt && mono;
>> +
>> + switch (tstamp_type) {
>> + case CLOCK_REAL:
>> + skb->tstamp_type = CLOCK_REAL;
>> + break;
>> + case CLOCK_MONO:
>> + skb->tstamp_type = kt && tstamp_type;
>> + break;
>> + }
>
> Technically this leaves the tstamp_type undefined if (skb, 0, CLOCK_REAL)
Do you think i should be checking for valid value of tstamp before setting the tstamp_type ? Only then set it.
>> }
>>
>> DECLARE_STATIC_KEY_FALSE(netstamp_needed_key);
>> @@ -4259,8 +4272,8 @@ DECLARE_STATIC_KEY_FALSE(netstamp_needed_key);
>> */
>> static inline void skb_clear_delivery_time(struct sk_buff *skb)
>> {
>> - if (skb->mono_delivery_time) {
>> - skb->mono_delivery_time = 0;
>> + if (skb->tstamp_type) {
>> + skb->tstamp_type = CLOCK_REAL;
>> if (static_branch_unlikely(&netstamp_needed_key))
>> skb->tstamp = ktime_get_real();
>> else
>> @@ -4270,7 +4283,7 @@ static inline void skb_clear_delivery_time(struct sk_buff *skb)
>>
>> static inline void skb_clear_tstamp(struct sk_buff *skb)
>> {
>> - if (skb->mono_delivery_time)
>> + if (skb->tstamp_type)
>> return;
>>
>> skb->tstamp = 0;
>> @@ -4278,7 +4291,7 @@ static inline void skb_clear_tstamp(struct sk_buff *skb)
>>
>> static inline ktime_t skb_tstamp(const struct sk_buff *skb)
>> {
>> - if (skb->mono_delivery_time)
>> + if (skb->tstamp_type == CLOCK_MONO)
>> return 0;
>
> Should this be if (skb->tstamp_type), in line with skb_clear_tstamp,
> right above?
I think so too. I thought i will not alter the existing functionality and keep it as close to the previous working changes.
>
>> @@ -1649,7 +1649,7 @@ void ip_send_unicast_reply(struct sock *sk, struct sk_buff *skb,
>> arg->csumoffset) = csum_fold(csum_add(nskb->csum,
>> arg->csum));
>> nskb->ip_summed = CHECKSUM_NONE;
>> - nskb->mono_delivery_time = !!transmit_time;
>> + nskb->tstamp_type = !!transmit_time;
>
> In anticipation of more tstamp_types, explicitly set to CLOCK_MONO.
I will set it to SKB_CLOCK_MONO
>
>> if (txhash)
>> skb_set_hash(nskb, txhash, PKT_HASH_TYPE_L4);
>> ip_push_pending_frames(sk, &fl4);
Abhishek Chauhan (ABC) wrote:
>
>
> On 4/13/2024 11:54 AM, Willem de Bruijn wrote:
> > Abhishek Chauhan wrote:
> >> mono_delivery_time was added to check if skb->tstamp has delivery
> >> time in mono clock base (i.e. EDT) otherwise skb->tstamp has
> >> timestamp in ingress and delivery_time at egress.
> >>
> >> Renaming the bitfield from mono_delivery_time to tstamp_type is for
> >> extensibilty for other timestamps such as userspace timestamp
> >> (i.e. SO_TXTIME) set via sock opts.
> >>
> >> As we are renaming the mono_delivery_time to tstamp_type, it makes
> >> sense to start assigning tstamp_type based out if enum defined as
> >> part of this commit
> >>
> >> Earlier we used bool arg flag to check if the tstamp is mono in
> >> function skb_set_delivery_time, Now the signature of the functions
> >> accepts enum to distinguish between mono and real time
> >>
> >> Bridge driver today has no support to forward the userspace timestamp
> >> packets and ends up resetting the timestamp. ETF qdisc checks the
> >> packet coming from userspace and encounters to be 0 thereby dropping
> >> time sensitive packets. These changes will allow userspace timestamps
> >> packets to be forwarded from the bridge to NIC drivers.
> >>
> >> In future tstamp_type:1 can be extended to support userspace timestamp
> >> by increasing the bitfield.
> >>
> >> Link: https://lore.kernel.org/netdev/[email protected]/
> >> Signed-off-by: Abhishek Chauhan <[email protected]>
> >> ---
> >> Changes since v2
> >> - Minor changes to commit subject
> >>
> >> Changes since v1
> >> - Squashed the two commits into one as mentioned by Willem.
> >> - Introduced switch in skb_set_delivery_time.
> >> - Renamed and removed directionality aspects w.r.t tstamp_type
> >> as mentioned by Willem.
> >>
> >> include/linux/skbuff.h | 33 +++++++++++++++-------
> >> include/net/inet_frag.h | 4 +--
> >> net/bridge/netfilter/nf_conntrack_bridge.c | 6 ++--
> >> net/core/dev.c | 2 +-
> >> net/core/filter.c | 8 +++---
> >> net/ipv4/inet_fragment.c | 2 +-
> >> net/ipv4/ip_fragment.c | 2 +-
> >> net/ipv4/ip_output.c | 8 +++---
> >> net/ipv4/tcp_output.c | 14 ++++-----
> >> net/ipv6/ip6_output.c | 6 ++--
> >> net/ipv6/netfilter.c | 6 ++--
> >> net/ipv6/netfilter/nf_conntrack_reasm.c | 2 +-
> >> net/ipv6/reassembly.c | 2 +-
> >> net/ipv6/tcp_ipv6.c | 2 +-
> >> net/sched/act_bpf.c | 4 +--
> >> net/sched/cls_bpf.c | 4 +--
> >> 16 files changed, 59 insertions(+), 46 deletions(-)
> >>
> >> diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
> >> index 7135a3e94afd..a83a2120b57f 100644
> >> --- a/include/linux/skbuff.h
> >> +++ b/include/linux/skbuff.h
> >> @@ -702,6 +702,11 @@ typedef unsigned int sk_buff_data_t;
> >> typedef unsigned char *sk_buff_data_t;
> >> #endif
> >>
> >> +enum skb_tstamp_type {
> >> + CLOCK_REAL = 0, /* Time base is realtime */
> >> + CLOCK_MONO = 1, /* Time base is Monotonic */
> >> +};
> >
> > Minor: inconsistent capitalization
> >
> I will fix this.
>
> >> +
> >> /**
> >> * DOC: Basic sk_buff geometry
> >> *
> >> @@ -819,7 +824,7 @@ typedef unsigned char *sk_buff_data_t;
> >> * @dst_pending_confirm: need to confirm neighbour
> >> * @decrypted: Decrypted SKB
> >> * @slow_gro: state present at GRO time, slower prepare step required
> >> - * @mono_delivery_time: When set, skb->tstamp has the
> >> + * @tstamp_type: When set, skb->tstamp has the
> >> * delivery_time in mono clock base (i.e. EDT). Otherwise, the
> >> * skb->tstamp has the (rcv) timestamp at ingress and
> >> * delivery_time at egress.
> >> @@ -950,7 +955,7 @@ struct sk_buff {
> >> /* private: */
> >> __u8 __mono_tc_offset[0];
> >> /* public: */
> >> - __u8 mono_delivery_time:1; /* See SKB_MONO_DELIVERY_TIME_MASK */
> >> + __u8 tstamp_type:1; /* See SKB_MONO_DELIVERY_TIME_MASK */
> >
> > Also remove reference to MONO_DELIVERY_TIME_MASK, or instead refer to
> > skb_tstamp_type.
> >
> >> #ifdef CONFIG_NET_XGRESS
> >> __u8 tc_at_ingress:1; /* See TC_AT_INGRESS_MASK */
> >> __u8 tc_skip_classify:1;
> >> @@ -4237,7 +4242,7 @@ static inline void skb_get_new_timestampns(const struct sk_buff *skb,
> >> static inline void __net_timestamp(struct sk_buff *skb)
> >> {
> >> skb->tstamp = ktime_get_real();
> >> - skb->mono_delivery_time = 0;
> >> + skb->tstamp_type = CLOCK_REAL;
> >> }
> >>
> >> static inline ktime_t net_timedelta(ktime_t t)
> >> @@ -4246,10 +4251,18 @@ static inline ktime_t net_timedelta(ktime_t t)
> >> }
> >>
> >> static inline void skb_set_delivery_time(struct sk_buff *skb, ktime_t kt,
> >> - bool mono)
> >> + u8 tstamp_type)
> >> {
> >> skb->tstamp = kt;
> >> - skb->mono_delivery_time = kt && mono;
> >> +
> >> + switch (tstamp_type) {
> >> + case CLOCK_REAL:
> >> + skb->tstamp_type = CLOCK_REAL;
> >> + break;
> >> + case CLOCK_MONO:
> >> + skb->tstamp_type = kt && tstamp_type;
> >> + break;
> >> + }
> >
> > Technically this leaves the tstamp_type undefined if (skb, 0, CLOCK_REAL)
> Do you think i should be checking for valid value of tstamp before setting the tstamp_type ? Only then set it.
A kt of 0 is interpreted as resetting the type. That should probably
be maintained.
For SO_TIMESTAMPING, a mono delivery time of 0 does have some meaning.
In __sock_recv_timestamp:
/* Race occurred between timestamp enabling and packet
receiving. Fill in the current time for now. */
if (need_software_tstamp && skb->tstamp == 0) {
__net_timestamp(skb);
false_tstamp = 1;
}
On 4/15/2024 1:46 PM, Willem de Bruijn wrote:
> Abhishek Chauhan (ABC) wrote:
>>
>>
>> On 4/13/2024 11:54 AM, Willem de Bruijn wrote:
>>> Abhishek Chauhan wrote:
>>>> mono_delivery_time was added to check if skb->tstamp has delivery
>>>> time in mono clock base (i.e. EDT) otherwise skb->tstamp has
>>>> timestamp in ingress and delivery_time at egress.
>>>>
>>>> Renaming the bitfield from mono_delivery_time to tstamp_type is for
>>>> extensibilty for other timestamps such as userspace timestamp
>>>> (i.e. SO_TXTIME) set via sock opts.
>>>>
>>>> As we are renaming the mono_delivery_time to tstamp_type, it makes
>>>> sense to start assigning tstamp_type based out if enum defined as
>>>> part of this commit
>>>>
>>>> Earlier we used bool arg flag to check if the tstamp is mono in
>>>> function skb_set_delivery_time, Now the signature of the functions
>>>> accepts enum to distinguish between mono and real time
>>>>
>>>> Bridge driver today has no support to forward the userspace timestamp
>>>> packets and ends up resetting the timestamp. ETF qdisc checks the
>>>> packet coming from userspace and encounters to be 0 thereby dropping
>>>> time sensitive packets. These changes will allow userspace timestamps
>>>> packets to be forwarded from the bridge to NIC drivers.
>>>>
>>>> In future tstamp_type:1 can be extended to support userspace timestamp
>>>> by increasing the bitfield.
>>>>
>>>> Link: https://lore.kernel.org/netdev/[email protected]/
>>>> Signed-off-by: Abhishek Chauhan <[email protected]>
>>>> ---
>>>> Changes since v2
>>>> - Minor changes to commit subject
>>>>
>>>> Changes since v1
>>>> - Squashed the two commits into one as mentioned by Willem.
>>>> - Introduced switch in skb_set_delivery_time.
>>>> - Renamed and removed directionality aspects w.r.t tstamp_type
>>>> as mentioned by Willem.
>>>>
>>>> include/linux/skbuff.h | 33 +++++++++++++++-------
>>>> include/net/inet_frag.h | 4 +--
>>>> net/bridge/netfilter/nf_conntrack_bridge.c | 6 ++--
>>>> net/core/dev.c | 2 +-
>>>> net/core/filter.c | 8 +++---
>>>> net/ipv4/inet_fragment.c | 2 +-
>>>> net/ipv4/ip_fragment.c | 2 +-
>>>> net/ipv4/ip_output.c | 8 +++---
>>>> net/ipv4/tcp_output.c | 14 ++++-----
>>>> net/ipv6/ip6_output.c | 6 ++--
>>>> net/ipv6/netfilter.c | 6 ++--
>>>> net/ipv6/netfilter/nf_conntrack_reasm.c | 2 +-
>>>> net/ipv6/reassembly.c | 2 +-
>>>> net/ipv6/tcp_ipv6.c | 2 +-
>>>> net/sched/act_bpf.c | 4 +--
>>>> net/sched/cls_bpf.c | 4 +--
>>>> 16 files changed, 59 insertions(+), 46 deletions(-)
>>>>
>>>> diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
>>>> index 7135a3e94afd..a83a2120b57f 100644
>>>> --- a/include/linux/skbuff.h
>>>> +++ b/include/linux/skbuff.h
>>>> @@ -702,6 +702,11 @@ typedef unsigned int sk_buff_data_t;
>>>> typedef unsigned char *sk_buff_data_t;
>>>> #endif
>>>>
>>>> +enum skb_tstamp_type {
>>>> + CLOCK_REAL = 0, /* Time base is realtime */
>>>> + CLOCK_MONO = 1, /* Time base is Monotonic */
>>>> +};
>>>
>>> Minor: inconsistent capitalization
>>>
>> I will fix this.
>>
>>>> +
>>>> /**
>>>> * DOC: Basic sk_buff geometry
>>>> *
>>>> @@ -819,7 +824,7 @@ typedef unsigned char *sk_buff_data_t;
>>>> * @dst_pending_confirm: need to confirm neighbour
>>>> * @decrypted: Decrypted SKB
>>>> * @slow_gro: state present at GRO time, slower prepare step required
>>>> - * @mono_delivery_time: When set, skb->tstamp has the
>>>> + * @tstamp_type: When set, skb->tstamp has the
>>>> * delivery_time in mono clock base (i.e. EDT). Otherwise, the
>>>> * skb->tstamp has the (rcv) timestamp at ingress and
>>>> * delivery_time at egress.
>>>> @@ -950,7 +955,7 @@ struct sk_buff {
>>>> /* private: */
>>>> __u8 __mono_tc_offset[0];
>>>> /* public: */
>>>> - __u8 mono_delivery_time:1; /* See SKB_MONO_DELIVERY_TIME_MASK */
>>>> + __u8 tstamp_type:1; /* See SKB_MONO_DELIVERY_TIME_MASK */
>>>
>>> Also remove reference to MONO_DELIVERY_TIME_MASK, or instead refer to
>>> skb_tstamp_type.
>>>
>>>> #ifdef CONFIG_NET_XGRESS
>>>> __u8 tc_at_ingress:1; /* See TC_AT_INGRESS_MASK */
>>>> __u8 tc_skip_classify:1;
>>>> @@ -4237,7 +4242,7 @@ static inline void skb_get_new_timestampns(const struct sk_buff *skb,
>>>> static inline void __net_timestamp(struct sk_buff *skb)
>>>> {
>>>> skb->tstamp = ktime_get_real();
>>>> - skb->mono_delivery_time = 0;
>>>> + skb->tstamp_type = CLOCK_REAL;
>>>> }
>>>>
>>>> static inline ktime_t net_timedelta(ktime_t t)
>>>> @@ -4246,10 +4251,18 @@ static inline ktime_t net_timedelta(ktime_t t)
>>>> }
>>>>
>>>> static inline void skb_set_delivery_time(struct sk_buff *skb, ktime_t kt,
>>>> - bool mono)
>>>> + u8 tstamp_type)
>>>> {
>>>> skb->tstamp = kt;
>>>> - skb->mono_delivery_time = kt && mono;
>>>> +
>>>> + switch (tstamp_type) {
>>>> + case CLOCK_REAL:
>>>> + skb->tstamp_type = CLOCK_REAL;
>>>> + break;
>>>> + case CLOCK_MONO:
>>>> + skb->tstamp_type = kt && tstamp_type;
>>>> + break;
>>>> + }
>>>
>>> Technically this leaves the tstamp_type undefined if (skb, 0, CLOCK_REAL)
>> Do you think i should be checking for valid value of tstamp before setting the tstamp_type ? Only then set it.
>
> A kt of 0 is interpreted as resetting the type. That should probably
> be maintained.
>
> For SO_TIMESTAMPING, a mono delivery time of 0 does have some meaning.
> In __sock_recv_timestamp:
>
> /* Race occurred between timestamp enabling and packet
> receiving. Fill in the current time for now. */
> if (need_software_tstamp && skb->tstamp == 0) {
> __net_timestamp(skb);
> false_tstamp = 1;
> }
Well in that case the above logic still resets the tstamp and sets the tstamp_type to CLOCK_REAL(value 0).
Anyway the tstamp_type will be 0 to begin with.
The logic is still inline with previous implementation, because previously if kt was 0 then kt && mono sets the tstamp_type (previously called as mono_delivery_time) to 0 (i.e SKB_CLOCK_REAL).
> >>>> static inline void skb_set_delivery_time(struct sk_buff *skb, ktime_t kt,
> >>>> - bool mono)
> >>>> + u8 tstamp_type)
> >>>> {
> >>>> skb->tstamp = kt;
> >>>> - skb->mono_delivery_time = kt && mono;
> >>>> +
> >>>> + switch (tstamp_type) {
> >>>> + case CLOCK_REAL:
> >>>> + skb->tstamp_type = CLOCK_REAL;
> >>>> + break;
> >>>> + case CLOCK_MONO:
> >>>> + skb->tstamp_type = kt && tstamp_type;
> >>>> + break;
> >>>> + }
> >>>
> >>> Technically this leaves the tstamp_type undefined if (skb, 0, CLOCK_REAL)
> >> Do you think i should be checking for valid value of tstamp before setting the tstamp_type ? Only then set it.
> >
> > A kt of 0 is interpreted as resetting the type. That should probably
> > be maintained.
> >
> > For SO_TIMESTAMPING, a mono delivery time of 0 does have some meaning.
> > In __sock_recv_timestamp:
> >
> > /* Race occurred between timestamp enabling and packet
> > receiving. Fill in the current time for now. */
> > if (need_software_tstamp && skb->tstamp == 0) {
> > __net_timestamp(skb);
> > false_tstamp = 1;
> > }
>
> Well in that case the above logic still resets the tstamp and sets the tstamp_type to CLOCK_REAL(value 0).
> Anyway the tstamp_type will be 0 to begin with.
> The logic is still inline with previous implementation, because previously if kt was 0 then kt && mono sets the tstamp_type (previously called as mono_delivery_time) to 0 (i.e SKB_CLOCK_REAL).
Sorry, I got my defaults confused. If we maintain that a zero tstamp
resets the type, then here should be no case with skb->tstamp 0 and
skb->tstamp_type SKB_CLOCK_REAL (or SKB_CLOCK_TAI or whatever). I
think it's preferable to make that obvious in the
skb_set_delivery_time implementation, rather than depend on knowledge
of its callers.
On 4/15/2024 2:22 PM, Willem de Bruijn wrote:
>>>>>> static inline void skb_set_delivery_time(struct sk_buff *skb, ktime_t kt,
>>>>>> - bool mono)
>>>>>> + u8 tstamp_type)
>>>>>> {
>>>>>> skb->tstamp = kt;
>>>>>> - skb->mono_delivery_time = kt && mono;
>>>>>> +
>>>>>> + switch (tstamp_type) {
>>>>>> + case CLOCK_REAL:
>>>>>> + skb->tstamp_type = CLOCK_REAL;
>>>>>> + break;
>>>>>> + case CLOCK_MONO:
>>>>>> + skb->tstamp_type = kt && tstamp_type;
>>>>>> + break;
>>>>>> + }
>>>>>
>>>>> Technically this leaves the tstamp_type undefined if (skb, 0, CLOCK_REAL)
>>>> Do you think i should be checking for valid value of tstamp before setting the tstamp_type ? Only then set it.
>>>
>>> A kt of 0 is interpreted as resetting the type. That should probably
>>> be maintained.
>>>
>>> For SO_TIMESTAMPING, a mono delivery time of 0 does have some meaning.
>>> In __sock_recv_timestamp:
>>>
>>> /* Race occurred between timestamp enabling and packet
>>> receiving. Fill in the current time for now. */
>>> if (need_software_tstamp && skb->tstamp == 0) {
>>> __net_timestamp(skb);
>>> false_tstamp = 1;
>>> }
>>
>> Well in that case the above logic still resets the tstamp and sets the tstamp_type to CLOCK_REAL(value 0).
>> Anyway the tstamp_type will be 0 to begin with.
>> The logic is still inline with previous implementation, because previously if kt was 0 then kt && mono sets the tstamp_type (previously called as mono_delivery_time) to 0 (i.e SKB_CLOCK_REAL).
>
> Sorry, I got my defaults confused. If we maintain that a zero tstamp
> resets the type, then here should be no case with skb->tstamp 0 and
> skb->tstamp_type SKB_CLOCK_REAL (or SKB_CLOCK_TAI or whatever). I
> think it's preferable to make that obvious in the
> skb_set_delivery_time implementation, rather than depend on knowledge
> of its callers.
Noted!. I will do the same as part of the next patchset.