Re: [RFC PATCH net-next v6 08/13] net-timestamp: support hw SCM_TSTAMP_SND for bpf extension

Martin KaFai Lau <martin.lau@xxxxxxxxx> · Fri, 24 Jan 2025 17:29:53 -0800




On 1/24/25 5:18 PM, Jason Xing wrote:
@@ -5577,9 +5578,9 @@ static void skb_tstamp_tx_bpf(struct sk_buff *skb, struct sock *sk,
               op = BPF_SOCK_OPS_TS_SCHED_OPT_CB;
               break;
       case SCM_TSTAMP_SND:
+             op = sw ? BPF_SOCK_OPS_TS_SW_OPT_CB : BPF_SOCK_OPS_TS_HW_OPT_CB;
               if (!sw)
-                     return;
-             op = BPF_SOCK_OPS_TS_SW_OPT_CB;
+                     *skb_hwtstamps(skb) = *hwtstamps;
hwtstamps may still be NULL, no?
Right, it can be zero if something wrong happens.

Then it needs a NULL check, no?







