[patch] single confirm (phonebook read range)

andrzej zaborowski balrogg at gmail.com
Sat Oct 27 23:26:27 CEST 2007

On 27/10/2007, Jim Rayner <openmoko at jrayner.com> wrote:
> Regarding stability issues with gsmd, the keep alive timer has also
> been removed in this patch since (in the current form in the openmoko
> svn) the timer fires without taking into account what actions gsmd is
> engaged in. With keep alive removed I have run gsmd for long periods
> of testing with no problems.

I believe the firing midway through a command has no relation to your
stability issue, and removing the timer is certainly not a solution
(and especially not commenting out code, comments are not meant for
disabling code). In my experience the timeout was caused by two timers
being created instead of one - this should be fixed by the patches
already present in bugzilla.

Afterall the timer just adds the command at the end of the queue and
gives it 30 seconds to finish, this should be enough even when gsmd is
very busy.

More information about the gsmd-devel mailing list