pxz crashed with SIGSEGV in _IO_vsnprintf()

[Dimitri John Ledkov]

used pxz....

ProblemType: Crash
DistroRelease: Ubuntu 13.10
Package: pxz 4.999.99~beta3+git659fc9b-1
ProcVersionSignature: Ubuntu 3.10.0-2.10-generic 3.10.0
Uname: Linux 3.10.0-2-generic x86_64
ApportVersion: 2.10.2-0ubuntu4
Architecture: amd64
Date: Wed Jul 10 17:47:18 2013
ExecutablePath: /usr/bin/pxz
InstallationDate: Installed on 2012-01-12 (544 days ago)
InstallationMedia: Ubuntu 13.04 "Raring Ringtail" - Alpha amd64 (20130318)
MarkForUpload: True
ProcCmdline: pxz ../android_0.20130709.orig.tar
SegvAnalysis:
 Segfault happened at: 0x2aaaab39163e <_IO_vsnprintf+126>: movb $0x0,0x0(%rbp)
 PC (0x2aaaab39163e) ok
 source "$0x0" ok
 destination "0x0(%rbp)" (0x00000000) not located in a known VMA region (needed writable region)!
SegvReason: writing NULL VMA
Signal: 11
SourcePackage: pxz
StacktraceTop:
 _IO_vsnprintf (string=0x0, maxlen=<optimised out>, format=0x402745 "xz", args=args@entry=0x7fffa34ae378) at vsnprintf.c:117
 __snprintf (s=<optimised out>, maxlen=<optimised out>, format=<optimised out>) at snprintf.c:34
 ?? ()
 __libc_start_main (main=0x401270, argc=2, ubp_av=0x7fffa34ae858, init=<optimised out>, fini=<optimised out>, rtld_fini=<optimised out>, stack_end=0x7fffa34ae848) at libc-start.c:260
 ?? ()
Title: pxz crashed with SIGSEGV in _IO_vsnprintf()
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups: adm autopilot cdrom dip libvirtd lpadmin plugdev sambashare sbuild sudo

[Apport retracing service (apport)]

StacktraceTop:
 _IO_vsnprintf (string=0x0, maxlen=<optimized out>, format=0x402745 "xz", args=args@entry=0x7fffa34ae378) at vsnprintf.c:117
 __snprintf (s=<optimized out>, maxlen=<optimized out>, format=format@entry=0x402745 "xz") at snprintf.c:34
 snprintf (__fmt=0x402745 "xz", __n=<optimized out>, __s=<optimized out>) at /usr/include/x86_64-linux-gnu/bits/stdio2.h:64
 main (argc=2, argv=0x7fffa34ae858) at pxz.c:245

[Curt Brune (curt-3)]

This was happening to me also. I was running pxz from within a Makefile.

The same command would work fine if I typed it on a BASH command line.

On a whim I compared the output of "ulimit -a" for the two cases.

The only difference was in the Makefile context the stack size was "unlimited",
whereas from BASH my stack size was 8192.

The problem went away for me when I ran "ulimit -s 8192" before running
pxz in the Makefile context.

See the attached strace output from the failing case which gave me some clues
about the stack size.

[Launchpad Janitor (janitor)]

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