Re: [PATCH] sfc: handle NULL returned by xdp_convert_buff_to_frame()
From: Kunwu Chan <hidden>
Date: 2025-07-30 07:38:44
Also in:
bpf
On 2025/7/28 22:28, Edward Cree wrote:
On 25/07/2025 13:38, Kunwu Chan wrote:quoted
Proposed refinement:...quoted
if (net_ratelimit()) netif_err(efx, rx_err, efx->net_dev, - "XDP TX failed (%d)\n", err); + "XDP TX failed (%d)%s\n", err, + err == -ENOBUFS ? " [frame conversion]" : "");Unnecessary, since efx_xdp_tx_buffers() never returns ENOBUFS.
You're correct that efx_siena_xdp_tx_buffers() never returns -ENOBUFS,
and xdp_convert_buff_to_frame() returns NULL on failure.
Whether we need a log to distinguish where the errors come from?
+ if (unlikely(!xdpf))
+ err = -ENOBUFS;
+ else
+ err = efx_xdp_tx_buffers(efx, 1, &xdpf, true);
- err = efx_siena_xdp_tx_buffers(efx, 1, &xdpf, true);
if (unlikely(err != 1)) {
efx_siena_free_rx_buffers(rx_queue, rx_buf, 1);
if (net_ratelimit())
netif_err(efx, rx_err, efx->net_dev,
- "XDP TX failed (%d)\n", err);
+ "XDP TX failed (%d)%s\n", err,
+ err == -ENOBUFS ? " [xdp_convert_buff_to_frame]" :
"efx_siena_xdp_tx_buffers");
Just a personal thought.
quoted
channel->n_rx_xdp_bad_drops++; - trace_xdp_exception(efx->net_dev, xdp_prog, xdp_act); + if (err != -ENOBUFS) + trace_xdp_exception(efx->net_dev, xdp_prog, xdp_act);Why prevent the tracepoint in this case??
You're correct, my mistake.
--
Thanks,
TAO.
---
“Life finds a way.”