Description of problem: With a kernel config that does not include CONFIG_SWAP, /proc/swaps does not exist. systemd-208-14.fc20 segfaults while enumerating swap devices in this case. systemd-208-9.fc20 worked. Version-Release number of selected component (if applicable): 208-14.fc20 How reproducible: Reliably. Steps to Reproduce: 1. Build kernel without CONFIG_SWAP. 2. Install systemd-208-14.fc20 3. Reboot the system or systemctl daemon-reexec Actual results: Feb 24 13:10:47 fedora systemd[1]: Reexecuting. Feb 24 13:10:47 fedora systemd[1]: systemd 208 running in system mode. (+PAM +LIBWRAP +AUDIT +SELINUX +IMA +SYSVINIT +LIBCRYPTSETUP +GCRYPT +ACL +XZ) Feb 24 13:10:47 fedora systemd[1]: Detected virtualization 'kvm'. Feb 24 13:10:48 fedora kernel: systemd[1]: segfault at 0 ip 00007f047f4f0b61 sp 00007fff60cd4b50 error 4 in libc-2.18.so[7f047f47c000+1b4000] Feb 24 13:10:48 fedora systemd[1]: Caught <SEGV>, dumped core as pid 7646. Feb 24 13:10:48 fedora systemd[1]: Freezing execution. Expected results: Feb 24 14:17:32 fedora systemd[1]: Reexecuting. Feb 24 14:17:32 fedora systemd[1]: systemd 208 running in system mode. (+PAM +LIBWRAP +AUDIT +SELINUX +IMA +SYSVINIT +LIBCRYPTSETUP +GCRYPT +ACL +XZ) Feb 24 14:17:32 fedora systemd[1]: Detected virtualization 'kvm'. Additional info: Backtrace is: (gdb) bt #0 0x00007f047f84a62b in raise (sig=sig@entry=11) at ../nptl/sysdeps/unix/sysv/linux/pt-raise.c:37 #1 0x00007f04813959ae in crash (sig=11) at ../src/core/main.c:152 #2 <signal handler called> #3 __GI_rewind (fp=0x0) at rewind.c:35 #4 0x00007f04813b6fee in swap_load_proc_swaps (m=m@entry=0x7f04822f9560, set_flags=set_flags@entry=false) at ../src/core/swap.c:1033 #5 0x00007f04813b7361 in swap_enumerate (m=0x7f04822f9560) at ../src/core/swap.c:1255 #6 0x00007f0481398554 in manager_enumerate (m=0x7f04822f9560) at ../src/core/manager.c:789 #7 0x00007f048139afdc in manager_startup (m=0x7f04822f9560, serialization=0x7f04822ef2c0, fds=0x7f04822f7550) at ../src/core/manager.c:905 #8 0x00007f0481393827 in main (argc=4, argv=0x7fff60cd5578) at ../src/core/main.c:1583 It looks like the refactoring in SOURCES/0147-core-do-not-segfault-if-swap-activity-happens-when-p.patch has caused swap_enumerate to continue if fopen() sets errno=ENOENT. Previously, the error would cause swap_enumerate to terminate successfully, but now it is open_proc_swaps that completes "successfully". Both swap_dispatch_reload and swap_enumerate continue even though m->proc_swaps is NULL.
Created attachment 867215 [details] patch to check for ENOENT and return from swap_dispatch_reload and swap_enumerate Attaching a patch that works against a machine without /proc/swaps. I also reproduced #969795 on 208-9.fc20 by creating a bogus swap unit and attempting to start it (without CONFIG_SWAP enabled). I verified that with this patch, systemd successfully boots the machine without swap, and trying to enable the bogus swap unit also does not segfault.
Thank you for the patch. I'll go out in the next update.
systemd-208-15.fc20 has been submitted as an update for Fedora 20. https://admin.fedoraproject.org/updates/systemd-208-15.fc20
Package systemd-208-15.fc20: * should fix your issue, * was pushed to the Fedora 20 testing repository, * should be available at your local mirror within two days. Update it with: # su -c 'yum update --enablerepo=updates-testing systemd-208-15.fc20' as soon as you are able to. Please go to the following url: https://admin.fedoraproject.org/updates/FEDORA-2014-3522/systemd-208-15.fc20 then log in and leave karma (feedback).
systemd-208-15.fc20 has been pushed to the Fedora 20 stable repository. If problems still persist, please make note of it in this bug report.