From: kernel test robot <lkp@intel.com>
To: Jakub Kicinski <kuba@kernel.org>,
davem@davemloft.net, edumazet@google.com
Cc: kbuild-all@lists.01.org, dsahern@gmail.com, pabeni@redhat.com,
herbert@gondor.apana.org.au, netdev@vger.kernel.org,
Jakub Kicinski <kuba@kernel.org>
Subject: Re: [PATCH net] ipv6: gro: flush instead of assuming different flows on hop_limit mismatch
Date: Tue, 25 Jan 2022 03:23:55 +0800 [thread overview]
Message-ID: <202201250210.roaIok2H-lkp@intel.com> (raw)
In-Reply-To: <20220121011941.1123392-1-kuba@kernel.org>
Hi Jakub,
I love your patch! Perhaps something to improve:
[auto build test WARNING on net/master]
url: https://github.com/0day-ci/linux/commits/Jakub-Kicinski/ipv6-gro-flush-instead-of-assuming-different-flows-on-hop_limit-mismatch/20220121-092033
base: https://git.kernel.org/pub/scm/linux/kernel/git/davem/net.git 57afdc0aab094b4c811b3fe030b2567812a495f3
config: x86_64-randconfig-s022 (https://download.01.org/0day-ci/archive/20220125/202201250210.roaIok2H-lkp@intel.com/config)
compiler: gcc-9 (Debian 9.3.0-22) 9.3.0
reproduce:
# apt-get install sparse
# sparse version: v0.6.4-dirty
# https://github.com/0day-ci/linux/commit/6f8f3e541288381a67df8b670068d5add231d082
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Jakub-Kicinski/ipv6-gro-flush-instead-of-assuming-different-flows-on-hop_limit-mismatch/20220121-092033
git checkout 6f8f3e541288381a67df8b670068d5add231d082
# save the config file to linux build tree
mkdir build_dir
make W=1 C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' O=build_dir ARCH=x86_64 SHELL=/bin/bash net/ipv6/
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
sparse warnings: (new ones prefixed by >>)
>> net/ipv6/ip6_offload.c:264:57: sparse: sparse: restricted __be32 degrades to integer
>> net/ipv6/ip6_offload.c:263:48: sparse: sparse: dubious: x | !y
vim +264 net/ipv6/ip6_offload.c
182
183 INDIRECT_CALLABLE_SCOPE struct sk_buff *ipv6_gro_receive(struct list_head *head,
184 struct sk_buff *skb)
185 {
186 const struct net_offload *ops;
187 struct sk_buff *pp = NULL;
188 struct sk_buff *p;
189 struct ipv6hdr *iph;
190 unsigned int nlen;
191 unsigned int hlen;
192 unsigned int off;
193 u16 flush = 1;
194 int proto;
195
196 off = skb_gro_offset(skb);
197 hlen = off + sizeof(*iph);
198 iph = skb_gro_header_fast(skb, off);
199 if (skb_gro_header_hard(skb, hlen)) {
200 iph = skb_gro_header_slow(skb, hlen, off);
201 if (unlikely(!iph))
202 goto out;
203 }
204
205 skb_set_network_header(skb, off);
206 skb_gro_pull(skb, sizeof(*iph));
207 skb_set_transport_header(skb, skb_gro_offset(skb));
208
209 flush += ntohs(iph->payload_len) != skb_gro_len(skb);
210
211 proto = iph->nexthdr;
212 ops = rcu_dereference(inet6_offloads[proto]);
213 if (!ops || !ops->callbacks.gro_receive) {
214 __pskb_pull(skb, skb_gro_offset(skb));
215 skb_gro_frag0_invalidate(skb);
216 proto = ipv6_gso_pull_exthdrs(skb, proto);
217 skb_gro_pull(skb, -skb_transport_offset(skb));
218 skb_reset_transport_header(skb);
219 __skb_push(skb, skb_gro_offset(skb));
220
221 ops = rcu_dereference(inet6_offloads[proto]);
222 if (!ops || !ops->callbacks.gro_receive)
223 goto out;
224
225 iph = ipv6_hdr(skb);
226 }
227
228 NAPI_GRO_CB(skb)->proto = proto;
229
230 flush--;
231 nlen = skb_network_header_len(skb);
232
233 list_for_each_entry(p, head, list) {
234 const struct ipv6hdr *iph2;
235 __be32 first_word; /* <Version:4><Traffic_Class:8><Flow_Label:20> */
236
237 if (!NAPI_GRO_CB(p)->same_flow)
238 continue;
239
240 iph2 = (struct ipv6hdr *)(p->data + off);
241 first_word = *(__be32 *)iph ^ *(__be32 *)iph2;
242
243 /* All fields must match except length and Traffic Class.
244 * XXX skbs on the gro_list have all been parsed and pulled
245 * already so we don't need to compare nlen
246 * (nlen != (sizeof(*iph2) + ipv6_exthdrs_len(iph2, &ops)))
247 * memcmp() alone below is sufficient, right?
248 */
249 if ((first_word & htonl(0xF00FFFFF)) ||
250 !ipv6_addr_equal(&iph->saddr, &iph2->saddr) ||
251 !ipv6_addr_equal(&iph->daddr, &iph2->daddr) ||
252 iph->nexthdr != iph2->nexthdr) {
253 not_same_flow:
254 NAPI_GRO_CB(p)->same_flow = 0;
255 continue;
256 }
257 if (unlikely(nlen > sizeof(struct ipv6hdr))) {
258 if (memcmp(iph + 1, iph2 + 1,
259 nlen - sizeof(struct ipv6hdr)))
260 goto not_same_flow;
261 }
262 /* flush if Traffic Class fields are different */
> 263 NAPI_GRO_CB(p)->flush |= flush |
> 264 !!((first_word & htonl(0x0FF00000)) |
265 (iph->hop_limit ^ iph2->hop_limit));
266
267 /* If the previous IP ID value was based on an atomic
268 * datagram we can overwrite the value and ignore it.
269 */
270 if (NAPI_GRO_CB(skb)->is_atomic)
271 NAPI_GRO_CB(p)->flush_id = 0;
272 }
273
274 NAPI_GRO_CB(skb)->is_atomic = true;
275 NAPI_GRO_CB(skb)->flush |= flush;
276
277 skb_gro_postpull_rcsum(skb, iph, nlen);
278
279 pp = indirect_call_gro_receive_l4(tcp6_gro_receive, udp6_gro_receive,
280 ops->callbacks.gro_receive, head, skb);
281
282 out:
283 skb_gro_flush_final(skb, pp, flush);
284
285 return pp;
286 }
287
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
prev parent reply other threads:[~2022-01-24 19:33 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-01-21 1:19 [PATCH net] ipv6: gro: flush instead of assuming different flows on hop_limit mismatch Jakub Kicinski
2022-01-21 8:55 ` Eric Dumazet
2022-01-21 15:15 ` Jakub Kicinski
2022-01-21 16:37 ` Eric Dumazet
2022-01-25 0:02 ` Jakub Kicinski
2022-01-24 19:23 ` kernel test robot [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=202201250210.roaIok2H-lkp@intel.com \
--to=lkp@intel.com \
--cc=davem@davemloft.net \
--cc=dsahern@gmail.com \
--cc=edumazet@google.com \
--cc=herbert@gondor.apana.org.au \
--cc=kbuild-all@lists.01.org \
--cc=kuba@kernel.org \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).