pulseaudio locks up after cycling USB audio device

It appears to be the case that the attached debug info hasn't got a backtrace, so here's one generated from GDB:

(gdb) bt
#0 0xb80e1430 in __kernel_vsyscall ()
#1 0xb7c633bb in write () from /lib/tls/i686/cmov/libc.so.6
#2 0xb7bfcb3c in _IO_file_write () from /lib/tls/i686/cmov/libc.so.6
#3 0xb7bfc9f6 in _IO_file_xsputn () from /lib/tls/i686/cmov/libc.so.6
#4 0xb7bd6827 in ?? () from /lib/tls/i686/cmov/libc.so.6
#5 0xb7bd1c3e in vfprintf () from /lib/tls/i686/cmov/libc.so.6
#6 0xb7c8a157 in __fprintf_chk () from /lib/tls/i686/cmov/libc.so.6
#7 0xb8060650 in pa_log_levelv_meta () from /usr/lib/libpulsecore.so.5
#8 0xb8060a10 in pa_log_level_meta () from /usr/lib/libpulsecore.so.5
#9 0xb77da3dd in pa_alsa_set_hw_params () from /usr/lib/pulse-0.9/modules/libalsa-util.so
#10 0xb77db8f9 in pa_alsa_open_by_device_id () from /usr/lib/pulse-0.9/modules/libalsa-util.so
#11 0xb77cd932 in module_alsa_sink_LTX_pa__init () from /usr/lib/pulse-0.9/modules/module-alsa-sink.so
#12 0xb8069def in pa_module_load () from /usr/lib/libpulsecore.so.5
#13 0xb7a2acb6 in ?? () from /usr/lib/pulse-0.9/modules/module-hal-detect.so
#14 0xb7a2b328 in ?? () from /usr/lib/pulse-0.9/modules/module-hal-detect.so
#15 0xb8045e0c in pa_mainloop_dispatch () from /usr/lib/libpulsecore.so.5
#16 0xb80460a1 in pa_mainloop_iterate () from /usr/lib/libpulsecore.so.5
#17 0xb8046164 in pa_mainloop_run () from /usr/lib/libpulsecore.so.5
#18 0x0805118d in main ()

Just committed a fix from git to my pa branch

This bug was fixed in the package pulseaudio - 0.9.13-2ubuntu5

---------------
pulseaudio (0.9.13-2ubuntu5) jaunty; urgency=low

  [ Daniel T Chen ]
  * Add fixes from git:
    - 0015_use_fionread.patch,
    - 0016_add_pa-source-message-get-latency.patch,
    - 0017_allow_dev_dsp_w-ok.patch,
    - 0018_fix_return_val_pa-frame-aligned.patch,
    - 0019_fix_macro_pa-unlikely.patch,
    - 0020_catch_driver_errors.patch (LP: #312373),
    - 0021_convert_to_pa-bool-t.patch,
    - 0022_fix_return_val_dump-resample-methods.patch,
    - 0023_use_gdbm-nolock.patch,
    - 0024_dont_mix_front-center_into_rear.patch,
    - 0025_drop_cap-nice_properly.patch,
    - 0027_handle_multicast_sdp_with_same_ip_ttl_as_rtp.patch,
    - 0028_prefer_mixer_controls_with_vols.patch (LP: #281605),
    - 0029_add_minor_fixes.patch
  * Forward-port workaround from hardy branch for source/sink and
    suspend-on-idle race (LP: #203654):
    - 0026_work_around_suspend-on-idle_source_sink_race.patch
  * debian/control: Drop padevchooser(Recommends) and pavucontrol
    (Suggests) completely - functionality has been subsumed by
    jaunty's gnome-volume-control.
  * debian/pulseaudio.init:
    - Make initscript more informative in the default case of per-user
      sessions (LP: #259522),
    - Make initscript retry (until three-second timeout) if stopping
      fails when restarting (LP: #244414),
    - LSB {Required-*,Should-*} should specify hal instead of dbus,
      since hal is required (and already requires dbus) (LP: #244679)
  * Previous upload contains patches to close:
    LP: #193491, #268891, #295164, #298301

  [ Luke Yelavich ]
  * 0008_regen_autotools.patch: Pull some libtool 2.2 updates from git,
    and regenerate autotools glue.
  * debian/rules: Add --enable-static to force creation of libpulse.a,
    a side-effect of messing with the autotools files. This shouldn't
    be needed for the 0.9.14 release.

 -- Luke Yelavich <email address hidden> Fri, 09 Jan 2009 17:05:38 +1100