On 12/22/21 10:46, Eelco Chaudron wrote: > > > On 21 Dec 2021, at 16:31, Aaron Conole wrote: > >> Dumitru Ceara <[email protected]> writes: >> >>> UB Sanitizer reports: >>> lib/bfd.c:748:16: runtime error: member access within misaligned address >>> 0x000001f0d6ea for type 'struct msg', which requires 4 byte alignment >>> 0x000001f0d6ea: note: pointer points here >>> 00 20 00 00 20 40 03 18 93 f9 0a 6e 00 00 00 00 00 0f 42 40 00 0f 42 >>> 40 00 00 00 00 cc 00 00 00 >>> ^ >>> #0 0x59008e in bfd_process_packet lib/bfd.c:748 >>> #1 0x52a240 in process_special ofproto/ofproto-dpif-xlate.c:3370 >>> #2 0x553452 in xlate_actions ofproto/ofproto-dpif-xlate.c:7766 >>> #3 0x4fc9e6 in upcall_xlate ofproto/ofproto-dpif-upcall.c:1237 >>> #4 0x4fdecc in process_upcall ofproto/ofproto-dpif-upcall.c:1456 >>> #5 0x4fd936 in upcall_cb ofproto/ofproto-dpif-upcall.c:1358 >>> [...] >>> lib/stp.c:754:15: runtime error: member access within misaligned address >>> 0x000002c4ea61 for type 'const struct stp_bpdu_header', which requires 2 >>> byte alignment >>> 0x000002c4ea61: note: pointer points here >>> 26 42 42 03 00 00 00 00 00 80 00 aa 66 aa 66 00 01 00 00 00 00 80 00 >>> aa 66 aa 66 00 01 80 02 00 >>> ^ >>> #0 0x8a2bce in stp_received_bpdu lib/stp.c:754 >>> #1 0x51e603 in stp_process_packet ofproto/ofproto-dpif-xlate.c:1788 >>> #2 0x52a96d in process_special ofproto/ofproto-dpif-xlate.c:3394 >>> #3 0x5534df in xlate_actions ofproto/ofproto-dpif-xlate.c:7766 >>> #4 0x4fcb49 in upcall_xlate ofproto/ofproto-dpif-upcall.c:1237 >>> [...] >>> lib/lldp/lldp.c:149:10: runtime error: load of misaligned address >>> 0x7ffcc0ae72bd for type 'ovs_be16', which requires 2 byte alignment >>> 0x7ffcc0ae72bd: note: pointer points here >>> 8e e7 84 ad 04 00 05 46 61 73 74 45 74 68 65 72 6e 65 74 20 31 2f 35 >>> e0 91 07 c9 3e 7f 00 00 80 >>> ^ >>> #0 0x718d63 in lldp_tlv_end lib/lldp/lldp.c:149 >>> #1 0x7191de in lldp_send lib/lldp/lldp.c:184 >>> #2 0x484d6c in test_aa_send tests/test-aa.c:238 >>> [...] >>> >>> Signed-off-by: Dumitru Ceara <[email protected]> >>> --- >> >> Acked-by: Aaron Conole <[email protected]> >> >> I really doubt that the stp change would impact performance, but it's >> possible that it could. Maybe Eelco has a perf script to check it? > > Sorry I have no userspace protocols performance tests/scripts. >
I thought this to be "control plane path" and imagined that memcpying 4 bytes will not make a difference. _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
