Re: [RFC PATCH bpf-next v3 1/2] net: Rename mono_delivery_time to tstamp_type for scalabilty

From: Willem de Bruijn
Date: Sat Apr 13 2024 - 14:54:58 EST

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/bc037db4-58bb-4861-ac31-a361a93841d3@xxxxxxxxx/
> Signed-off-by: Abhishek Chauhan <quic_abchauha@xxxxxxxxxxx>
> ---
> 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);