Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 27 Mar 2014 14:54:10 +0100
From:      Ralf Wenk <iz-rpi03@hs-karlsruhe.de>
To:        arm@freebsd.org
Subject:   screen(1) crashes plus weird output for screen -ls
Message-ID:  <E1WTAl0-004Ooa-Au@smtp.hs-karlsruhe.de>

next in thread | raw e-mail | index | archive | help
Hi,

while working with screen(1) it dumps core very often on arm.
Using gdb on the core shows
  (gdb) bt
  =230  0x2029bc18 in strlen () from /lib/libc.so.7
  =231  0x0001e9a4 in ?? ()
  (gdb)
as the last function called every time.

Interestingly the output of =22screen -ls=22 is a little bit weird an loo=
ks

  There is a screen on:
          814.ttyu0.rpi  (Attached)
  -1073746664 Socket
                    =D0=A0=E1 in =B0=F5=FF=BF8=E8.

instead of the expected=20

  There is a screen on:
          814.ttyu0.rpi  (Attached)
  1 Socket in /tmp/screens/S-root


/tmp is a tmpfs. kernel and world are from a fresh checked-out r263766.
screen(1) is newly compiled with clang from that version an pk info shows=

  =5B...=5D
  Name           : screen
  Version        : 4.0.3_14
  Installed on   : Wed Mar 26 16:07:57 CET 2014
  Origin         : sysutils/screen
  Architecture   : freebsd:11:armv6:32:el:eabi:softfp
  =5B...=5D

ktrace/kdump verified that the weird =22Socket=22 line(s) are already in =
the
write(2) buffer.

As screen(1) on i386 does not dump core and =22screen -ls=22 shows the ex=
pected
output I think this is a arm related problem.

Ralf




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?E1WTAl0-004Ooa-Au>