[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <3bc5824f-41c3-41c4-bcfc-9cb7d6c37937@intel.com>
Date: Mon, 22 Jul 2024 11:23:54 +0200
From: Mateusz Polchlopek <mateusz.polchlopek@...el.com>
To: Jakub Kicinski <kuba@...nel.org>
CC: <intel-wired-lan@...ts.osuosl.org>, <apw@...onical.com>,
<joe@...ches.com>, <dwaipayanray1@...il.com>, <lukas.bulwahn@...il.com>,
<akpm@...ux-foundation.org>, <willemb@...gle.com>, <edumazet@...gle.com>,
<linux-kernel@...r.kernel.org>, <netdev@...r.kernel.org>, Przemek Kitszel
<przemyslaw.kitszel@...el.com>, Wojciech Drewek <wojciech.drewek@...el.com>,
Simon Horman <horms@...nel.org>
Subject: Re: [Intel-wired-lan] [PATCH iwl-next v2 6/6] ice: devlink health:
dump also skb on Tx hang
On 7/14/2024 4:30 PM, Jakub Kicinski wrote:
> On Fri, 12 Jul 2024 05:32:51 -0400 Mateusz Polchlopek wrote:
>> + buf_pos = ice_emit_to_buf(buf, buf_size, buf_pos,
>> + "skb len=%u headroom=%u headlen=%u tailroom=%u\n"
>> + "mac=(%d,%d) net=(%d,%d) trans=%d\n"
>> + "shinfo(txflags=%u nr_frags=%u gso(size=%hu type=%u segs=%hu))\n"
>> + "csum(0x%x ip_summed=%u complete_sw=%u valid=%u level=%u)\n"
>> + "hash(0x%x sw=%u l4=%u) proto=0x%04x pkttype=%u iif=%d\n",
>> + skb->len, headroom, skb_headlen(skb), tailroom,
>> + has_mac ? skb->mac_header : -1,
>> + has_mac ? skb_mac_header_len(skb) : -1,
>> + skb->network_header,
>> + has_trans ? skb_network_header_len(skb) : -1,
>> + has_trans ? skb->transport_header : -1,
>> + sh->tx_flags, sh->nr_frags,
>> + sh->gso_size, sh->gso_type, sh->gso_segs,
>> + skb->csum, skb->ip_summed, skb->csum_complete_sw,
>> + skb->csum_valid, skb->csum_level,
>> + skb->hash, skb->sw_hash, skb->l4_hash,
>> + ntohs(skb->protocol), skb->pkt_type, skb->skb_iif);
>
> Make it a generic helper in devlink?
Makes sense for me but need to consult with an author
>
>> + if (dev)
>> + buf_pos = ice_emit_to_buf(buf, buf_size, buf_pos,
>> + "dev name=%s feat=%pNF\n", dev->name,
>> + &dev->features);
>> + if (sk)
>> + buf_pos = ice_emit_to_buf(buf, buf_size, buf_pos,
>> + "sk family=%hu type=%u proto=%u\n",
>> + sk->sk_family, sk->sk_type,
>> + sk->sk_protocol);
>> +
>> + if (headroom)
>> + buf_pos = ice_emit_hex_to_buf(buf, buf_size, buf_pos,
>> + "skb headroom: ", skb->head,
>> + headroom);
>> +
>> + seg_len = min_t(int, skb_headlen(skb), len);
>> + if (seg_len)
>> + buf_pos = ice_emit_hex_to_buf(buf, buf_size, buf_pos,
>> + "skb linear: ", skb->data,
>> + seg_len);
>> + len -= seg_len;
>> +
>> + if (tailroom)
>> + buf_pos = ice_emit_hex_to_buf(buf, buf_size, buf_pos,
>> + "skb tailroom: ",
>> + skb_tail_pointer(skb), tailroom);
>
> The printing on tailroom, headroom and frag data seems a bit much.
> I guess you're only printing the head SKB so it may be fine. But
> I don't think it's useful. The device will probably only care about
> the contents of the headers, for other parts only the metadata matters.
> No strong preference tho.
>
>> + for (i = 0; len && i < skb_shinfo(skb)->nr_frags; i++) {
>> + skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
>> + u32 p_off, p_len, copied;
>> + struct page *p;
>> + u8 *vaddr;
>> +
>> + skb_frag_foreach_page(frag, skb_frag_off(frag),
>> + skb_frag_size(frag), p, p_off, p_len,
>> + copied) {
>> + seg_len = min_t(int, p_len, len);
>> + vaddr = kmap_local_page(p);
>> + buf_pos = ice_emit_hex_to_buf(buf, buf_size, buf_pos,
>> + "skb frag: ",
>> + vaddr + p_off, seg_len);
>> + kunmap_local(vaddr);
>> + len -= seg_len;
>> +
>> + if (!len || buf_pos == buf_size)
>> + break;
>> + }
>> + }
>> +
>> + if (skb_has_frag_list(skb)) {
>> + buf_pos = ice_emit_to_buf(buf, buf_size, buf_pos,
>> + "skb fraglist:\n");
>> + skb_walk_frags(skb, list_skb) {
>> + buf_pos = ice_skb_dump_buf(buf, buf_size, buf_pos,
>> + list_skb);
>> +
>> + if (buf_pos == buf_size)
>> + break;
>> + }
>> + }
>
> You support transmitting skbs with fraglist? 🤨️
Powered by blists - more mailing lists