Re: [PATCH bpf-next v8 1/3] net: Rename mono_delivery_time to tstamp_type for scalabilty

[Abhishek Chauhan (ABC)]

On 5/9/2024 2:18 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 on enum defined
> in 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 tstamp_type to distinguish between mono and real time.
> 
> Also skb_set_delivery_type_by_clockid is a new function which accepts
> clockid to determine the tstamp_type.
> 
> 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>
> Reviewed-by: Willem de Bruijn <willemb@xxxxxxxxxx>
> Reviewed-by: Martin KaFai Lau <martin.lau@xxxxxxxxxx>
> ---
> Changes since v7
> - Added reviewed by tags and removed RFC 
> 
> Changes since v6
> - Moved documentation comment from patch 2 to patch 1 (Minor)
> - Instead of calling the wrapper api to set tstamp_type
>   for tcp, directly call main api to set the tstamp_type
>   as suggested by Willem
> 
> Changes since v5
> - Avoided using garble function names as mentioned by
>   Willem.
> - Implemented a conversion function stead of duplicating 
>   the same logic as mentioned by Willem.
> - Fixed indentation problems and minor documentation issues
>   which mentions tstamp_type as a whole instead of bitfield
>   notations. (Mentioned both by Willem and Martin)
>   
> Changes since v4
> - Introduce new function to directly delivery_time and
>   another to set tstamp_type based on clockid. 
> - Removed un-necessary comments in skbuff.h as 
>   enums were obvious and understood.
> 
> Changes since v3
> - Fixed inconsistent capitalization in skbuff.h
> - remove reference to MONO_DELIVERY_TIME_MASK in skbuff.h
>   and point it to skb_tstamp_type now.
> - Explicitely setting SKB_CLOCK_MONO if valid transmit_time
>   ip_send_unicast_reply 
> - Keeping skb_tstamp inline with skb_clear_tstamp. 
> - skb_set_delivery_time checks if timstamp is 0 and 
>   sets the tstamp_type to SKB_CLOCK_REAL.
> - Above comments are given by Willem 
> - Found out that skbuff.h has access to uapi/linux/time.h
>   So now instead of using  CLOCK_REAL/CLOCK_MONO 
>   i am checking actual clockid_t directly to set tstamp_type 
>   example:- CLOCK_REALTIME/CLOCK_MONOTONIC 
> - Compilation error fixed in 
>   net/ieee802154/6lowpan/reassembly.c
> 
> 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                     | 52 ++++++++++++++++------
>  include/net/inet_frag.h                    |  4 +-
>  net/bridge/netfilter/nf_conntrack_bridge.c |  6 +--
>  net/core/dev.c                             |  2 +-
>  net/core/filter.c                          | 10 ++---
>  net/ieee802154/6lowpan/reassembly.c        |  2 +-
>  net/ipv4/inet_fragment.c                   |  2 +-
>  net/ipv4/ip_fragment.c                     |  2 +-
>  net/ipv4/ip_output.c                       |  9 ++--
>  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 +-
>  17 files changed, 78 insertions(+), 51 deletions(-)
> 
> diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
> index 1c2902eaebd3..05aec712d16d 100644
> --- a/include/linux/skbuff.h
> +++ b/include/linux/skbuff.h
> @@ -706,6 +706,11 @@ typedef unsigned int sk_buff_data_t;
>  typedef unsigned char *sk_buff_data_t;
>  #endif
>  
> +enum skb_tstamp_type {
> +	SKB_CLOCK_REALTIME,
> +	SKB_CLOCK_MONOTONIC,
> +};
> +
>  /**
>   * DOC: Basic sk_buff geometry
>   *
> @@ -823,10 +828,8 @@ 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
> - *		delivery_time in mono clock base (i.e. EDT).  Otherwise, the
> - *		skb->tstamp has the (rcv) timestamp at ingress and
> - *		delivery_time at egress.
> + *	@tstamp_type: When set, skb->tstamp has the
> + *		delivery_time clock base of skb->tstamp.
>   *	@napi_id: id of the NAPI struct this skb came from
>   *	@sender_cpu: (aka @napi_id) source CPU in XPS
>   *	@alloc_cpu: CPU which did the skb allocation.
> @@ -954,7 +957,7 @@ 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_tstamp_type */
>  #ifdef CONFIG_NET_XGRESS
>  	__u8			tc_at_ingress:1;	/* See TC_AT_INGRESS_MASK */
>  	__u8			tc_skip_classify:1;
> @@ -4179,7 +4182,7 @@ 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 = SKB_CLOCK_REALTIME;
>  }
>  
>  static inline ktime_t net_timedelta(ktime_t t)
> @@ -4188,10 +4191,33 @@ 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;
> +
> +	if (kt)
> +		skb->tstamp_type = tstamp_type;
> +	else
> +		skb->tstamp_type = SKB_CLOCK_REALTIME;
> +}
> +
> +static inline void skb_set_delivery_type_by_clockid(struct sk_buff *skb,
> +						    ktime_t kt, clockid_t clockid)
> +{
> +	u8 tstamp_type = SKB_CLOCK_REALTIME;
> +
> +	switch (clockid) {
> +	case CLOCK_REALTIME:
> +		break;
> +	case CLOCK_MONOTONIC:
> +		tstamp_type = SKB_CLOCK_MONOTONIC;
> +		break;
> +	default:

Willem and Martin, I was thinking we should remove this warn_on_once from below line. Some systems also use panic on warn. 
So i think this might result in unnecessary crashes. 

Let me know what you think. 

Logs which are complaining. 
https://syzkaller.appspot.com/x/log.txt?x=118c3ae8980000

> +		WARN_ON_ONCE(1);
> +		kt = 0;
> +	}
> +
> +	skb_set_delivery_time(skb, kt, tstamp_type);
>  }
>