Your message dated Tue, 23 Sep 2025 06:28:48 +0300 with message-id <capzxpqea1dmgkqpzccmv7afnaye-jhfi6+sk02y8atk24r2...@mail.gmail.com> and subject line Re: dhcpcd segfault on boot has caused the Debian Bug report #1114964, regarding dhcpcd segfault on boot to be marked as done.
This means that you claim that the problem has been dealt with. If this is not the case it is now your responsibility to reopen the Bug report if necessary, and/or fix the problem forthwith. (NB: If you are a system administrator and have no idea what this message is talking about, this may indicate a serious mail system misconfiguration somewhere. Please contact [email protected] immediately.) -- 1114964: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1114964 Debian Bug Tracking System Contact [email protected] with problems
--- Begin Message ---Package: dhcpcd-base Version: 1:10.1.0-11 I am seeing a segfault from dhcpcd during the boot process. It is recorded in syslog with the following: 2025-09-09T23:16:19.340385-04:00 test-debian64 kernel: dhcpcd.real[775]: segfault at 20 ip 000055e44c77ed0f sp 00007fff3392fca0 error 4 in dhcpcd.real[2cd0f,55e44c75b000+3e000] likely on CPU 0 (core 0, socket 0) 2025-09-09T23:16:19.340392-04:00 test-debian64 kernel: Code: e8 fd ff ff 31 d2 66 41 89 54 24 60 e9 8e fe ff ff e8 15 c9 fd ff 0f 1f 44 00 00 f3 0f 1e fa 41 54 48 8d 57 38 41 89 f4 55 53 <48> 8b 47 20 48 89 fb 48 8d 3d 14 bf 01 00 48 8d 70 18 31 c0 e8 48 The daemon is being invoked from NetworkManager. The crash appears to be the cause of an unusual delay during boot for the systemd unit "NetworkManager-wait-online.service", but the system does eventually join the network. I was not able to obtain a core dump until I rebuilt the package with the --disable-privsep compile-time option. Here is the backtrace from the core dump: Program terminated with signal SIGSEGV, Segmentation fault. #0 0x00005555734227da in ipv4_deladdr (addr=0x0, keeparp=0) at ./src/ipv4.c:527 527 logdebugx("%s: deleting IP address %s", (gdb) bt #0 0x00005555734227da in ipv4_deladdr (addr=0x0, keeparp=0) at ./src/ipv4.c:527 #1 0x0000555573422e9a in ipv4_applyaddr (arg=0x5555829fd9b0) at ./src/ipv4.c:763 #2 0x000055557341d3c5 in dhcp_drop (ifp=0x5555829fd9b0, reason=0x55557344258f "EXPIRE") at ./src/dhcp.c:2806 #3 0x00005555734211c2 in dhcp_handleifa (cmd=21, ia=0x5555829da860, pid=753) at ./src/dhcp.c:4215 #4 0x000055557342343b in ipv4_handleifa (ctx=0x7fffff4f2500, cmd=21, ifs=0x5555829fcc70, ifname=0x5555829fd9c8 "enp0s3", addr=0x7fffff4ee184, mask=0x7fffff4ee188, brd=0x7fffff4ee18c, addrflags=0, pid=753) at ./src/ipv4.c:953 #5 0x00005555734130ad in link_addr (ctx=0x7fffff4f2500, ifp=0x5555829fd9b0, nlm=0x7fffff4ee370) at ./src/if-linux.c:886 #6 0x000055557341347e in link_netlink (ctx=0x7fffff4f2500, arg=0x0, nlm=0x7fffff4ee370) at ./src/if-linux.c:998 #7 0x000055557341288d in if_getnetlink (ctx=0x7fffff4f2500, iov=0x7fffff4ee360, fd=8, flags=64, cb=0x555573413403 <link_netlink>, cbarg=0x0) at ./src/if-linux.c:650 #8 0x000055557341389e in if_handlelink (ctx=0x7fffff4f2500) at ./src/if-linux.c:1102 #9 0x00005555733f7d75 in dhcpcd_handlelink (arg=0x7fffff4f2500, events=1) at ./src/dhcpcd.c:1141 #10 0x00005555733fda4e in eloop_run_ppoll (eloop=0x5555829d8690, ts=0x7fffff4f2440, signals=0x7fffff4f2740) at ./src/eloop.c:1106 #11 0x00005555733fdc8e in eloop_start (eloop=0x5555829d8690, signals=0x7fffff4f2740) at ./src/eloop.c:1228 #12 0x00005555733fbd33 in main (argc=2, argv=0x7fffff4f2a48, envp=0x7fffff4f2a60) at ./src/dhcpcd.c:2651 If I put a simple workaround into ipv4_deladdr() of the form if (!addr) return 0; then dhcpcd does not segfault, but behaves somewhat erratically, setting up and then releasing the network connection two or three times before stabilizing (online). Likely something more is needed for an interim fix.
--- End Message ---
--- Begin Message ---ti 23.9.2025 klo 0.39 Daniel Richard G. ([email protected]) kirjoitti: > On Sun, 2025 Sep 21 04:23-04:00, Martin-Éric Racine wrote: > > > > Does it also segfault with the 1:10.2.4-4 that currently sits in Testing? > > I installed that version of dhcpcd-base on my trixie test system, and > confirmed that the segfault no longer occurs. Thanks. Closing this accordingly. Martin-Éric
--- End Message ---
