lxc-start crashed with SIGSEGV in cgfsng_payload_create()

[Lukas Jelinek]

https://errors.ubuntu.com/problem/8e0f1b9682f08bab5dedd4100e42f59cfe2cc004

Steps to reproduce:
1) Prepare creating unprivileged containers as described here: https://linuxcontainers.org/lxc/getting-started/
2) Create a new container (e.g. "lxc-create -n container -t download" and then "debian", "stretch" and "amd64").
3) Start the created container in foreground using "lxc-start -n container -F"

ProblemType: Crash
DistroRelease: Ubuntu 18.10
Package: lxc-utils 3.0.3-0ubuntu1~18.10.1
ProcVersionSignature: Ubuntu 4.18.0-17.18-generic 4.18.20
Uname: Linux 4.18.0-17-generic x86_64
ApportVersion: 2.20.10-0ubuntu13.2
Architecture: amd64
CrashCounter: 1
CurrentDesktop: KDE
Date: Wed Apr 17 12:03:37 2019
ExecutablePath: /usr/bin/lxc-start
InstallationDate: Installed on 2019-04-15 (1 days ago)
InstallationMedia: Kubuntu 18.10 "Cosmic Cuttlefish" - Release amd64 (20181017.2)
ProcCmdline: BOOT_IMAGE=/vmlinuz-4.18.0-17-generic root=UUID=0739f905-33c9-4de3-9ba9-a63b8d69e5e5 ro quiet splash vt.handoff=1
SegvAnalysis:
 Segfault happened at: 0x7f6292c5dbbd: mov (%rax),%r15
 PC (0x7f6292c5dbbd) ok
 source "(%rax)" (0x00000000) not located in a known VMA region (needed readable region)!
 destination "%r15" ok
SegvReason: reading NULL VMA
Signal: 11
SourcePackage: lxc
StacktraceTop:
 ?? () from /usr/lib/x86_64-linux-gnu/liblxc.so.1
 __lxc_start () from /usr/lib/x86_64-linux-gnu/liblxc.so.1
 lxc_start () from /usr/lib/x86_64-linux-gnu/liblxc.so.1
 ?? () from /usr/lib/x86_64-linux-gnu/liblxc.so.1
 ?? () from /usr/lib/x86_64-linux-gnu/liblxc.so.1
Title: lxc-start crashed with SIGSEGV in __lxc_start()
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups: adm cdrom dip lpadmin plugdev sambashare sudo
defaults.conf:
 lxc.net.0.type = veth
 lxc.net.0.link = lxcbr0
 lxc.net.0.flags = up
 lxc.net.0.hwaddr = 00:16:3e:xx:xx:xx
mtime.conffile..etc.apport.crashdb.conf: 2019-04-17T11:58:15.556166

[Apport retracing service (apport)]

StacktraceTop:
 cgfsng_payload_create (ops=0x5605602f54e0, handler=<optimized out>) at cgroups/cgfsng.c:1303
 lxc_spawn (handler=0x5605602f4340) at start.c:1633
 __lxc_start (name=name@entry=0x5605602dab60 "perccli1", handler=handler@entry=0x5605602f4340, ops=ops@entry=0x7f6292cf8340 <start_ops>, data=data@entry=0x7ffdb2f002f8, lxcpath=lxcpath@entry=0x5605602f1590 "/home/luk/.local/share/lxc", daemonize=daemonize@entry=false, error_num=0x5605602f16e0) at start.c:1937
 lxc_start (name=0x5605602dab60 "perccli1", argv=argv@entry=0x7ffdb2f003b0, handler=handler@entry=0x5605602f4340, lxcpath=0x5605602f1590 "/home/luk/.local/share/lxc", daemonize=<optimized out>, error_num=error_num@entry=0x5605602f16e0) at start.c:2053
 do_lxcapi_start (c=c@entry=0x5605602f16a0, useinit=useinit@entry=0, argv=0x7ffdb2f003b0, argv@entry=0x0) at lxccontainer.c:1096

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Daniel van Vugt (vanvugt)]

Also tracking in:
https://errors.ubuntu.com/problem/8e0f1b9682f08bab5dedd4100e42f59cfe2cc004

[Scott Moser (smoser)]

I bumped into this yesterday on bionic.
The commit 1e04bb71da3ed829 [1] reports to fix it.

  [1] https://github.com/lxc/lxc/commit/1e04bb71da3ed829761ae8c729c3d021a6a709df

Hopefully there will be a 3.0.x update to bionic at some point.