Comment 3 for bug 137744

network-manager (0.6.5-0ubuntu11) gutsy; urgency=low

  * debian/patches/25_lp90267-dont-tear-down-upped-interfaces.patch: remove
        redundant nm_info output previously send to syslog/daemon.log
        (LP: #137744).
  * debian/patches/41r_wpa_control_custom_timeout.patch: add a state timeout
       to wpa control Private struct and add functions to set/unset this custom
       timeout; in turn implement this feature and make wpa_ctrl obey this
       custom timeout.
  * debian/patches/41r_wpa_control_explicit_TIMEOUT_reply_fixup.patch: wpa_ctrl
       returns bogus reply string which may lead to access of not-initialized
       memory in case an operation times out. We fix this by explicitly setting
       reply to TIMEOUT[CLI] on wpa_ctrl side before returning.
  * debian/patches/41s_wpa_ctrl_socket_non_block.patch: setup wpa_ctrl_socket as
       a real NON_BLOCK socket to prevent one potential reason for stale states
       in network-manager (deadlocks).
  * debian/patches/41t_nm_device_wireless_index_ctrl_sockets_by_run_count.patch:
       to prevent new supplicant sockets to not conflict with sockets that might
       still exist from previous run we now append a run_count index to _all_
       socket control paths. This run_count is incremented in supplicant_cleanup
       and is mod 10 to prevent any overflow when having more than 32-bit runs.
  * debian/patches/41q_retry_supplicant_cleanup_startup_if_ctrl_interface_connect_failes.patch:
       another try to terminate wpasupplicant more graceful: new shutdown
       sequence sends disable_network, ap_scan 0, interface_remove, terminate.
  * debian/patches/41u_custom_timeout_for_some_wpa_ctrl_operations.patch: we now
       use a custom timeout for ctrl operations: INTERFACE_ADD, DISABLE_NETWORK,
       TERMINATE; we take care that this timeout is unset after operations are
       through.
  * debian/patches/series: add all these great new patches to quilt _series_.
  * debian/patches/41d_ipw3945_turn_off_essid_in_stage1.patch,
      41l_enable_ipw3945_reset_essid.patch,series: drop from patchset - because
      we have a real fix in ipw3945 driver now.
  * debian/patches/41n_graceful_supplicant_shutdown.patch: update patch in turn of
      dropping the patches above.

 -- Alexander Sack <email address hidden> Mon, 10 Sep 2007 13:06:24 +0200