hi-glenn

https://networkop.co.uk/post/2021-03-ebpf-intro/ After read this link, you will get a better understanding.

> It looks like ci hangs on `tcp_buffer_select`, and it looks like the kernel that ci is currently using doesn't support this feature. we may need to mark it as...

> I can confirm this. > > What's happening is that we're not connected yet so [we attach the connect timeout](https://github.com/redis/hiredis/blob/be32bcdc8e84ae7dc091ceeffca2c5d4126f415c/async_private.h#L65) but since there is only one "timeout" struct in...

过期了，重来吧。 搞个 qq 群，多方便，管理功能也更强。

Of course, it support. Attention: 1. nic queue id 2. bpf_redirect_map(&xsks_map, af_xdp_socket_index, 0), af_xdp_socket_index should be corresponding to the correct socket.

@kewinrausch advanced03-AF_XDP is ok in my environment. My environment is Fedora29 with bpf-next kernel v5.3-rc1 ( https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next.git ) My xdp-tutorial git from https://github.com/chaudron/xdp-tutorial You also can refer to： https://github.com/xdp-project/xdp-tutorial/issues/60#issue-482401608

Hi, @tohojo @netoptimizer a strange problem： running on SKB mode xdp receive udp and reply. ctx->data_end - ctx->data == 92 and call ``` long ret = bpf_xdp_adjust_tail(ctx , grow_val); ```...

> In skb mode the size is limited by whatever is available in the skb. Which I guess is 356 bytes in your case :) I'm confused. When I have...

> Not too surprising that different drivers and setups results in different allocation sizes. I don't think there's any way to influence that either. You may be able to use...

> skbs are allocated in the drivers, though. The allocation size can depend on the packet size as well, so I suspect you'd get different results if you test with...