[Bug 1003] GSM modem is not powered down when Linux is shut down

bugzilla-daemon at bugzilla.openmoko.org bugzilla-daemon at bugzilla.openmoko.org
Sun Nov 18 08:56:00 CET 2007


http://bugzilla.openmoko.org/cgi-bin/bugzilla/show_bug.cgi?id=1003

mwester at dls.net changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |



------- Additional Comments From mwester at dls.net  2007-11-18 08:55 -------
The API for power on/off the GSM modem using gsmd is a necessary feature, but it
does not really resolve the issue reported by this bug.  Specifically, the patch
and the K34modem script will *ONLY* accomplish the desired end effect *IFF* gsmd
is running.

The principle of least surprise requires that the system make every effort to
power off the GSM modem at shutdown, especially when that shutdown is due to the
failure or abnormal behavior of the software on the device.  It is quite
reasonable and rational for a user to assume that if they were able to initiate
a power-off, and that the screen went blank, that the phone is truly *OFF*.  By
making the power-off dependent upon gsmd, we really haven't solved this problem
for any of the cases where gsmd has failed, or if gsmd has lost communications
with the modem, or any of a huge number of failure modes.

I applaud the addition of the power-off ability to gsmd, but I think gsmd cannot
be trusted in the event that the user powers off the phone because of some sort
of problem with the phone's behavior.

I remain firmly convinced that at shutdown, we must:
A) Reset the modem
B) issue the AT at POFF command
and do so with as little reliance on external processes, libraries, and
especially daemons as is possible, as we simply cannot assume that those things
are available or functional at shutdown.




------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.




More information about the buglog mailing list