laforge at gnumonks.org
Sun Apr 11 07:43:26 CEST 2010
thanks for debugging this.
On Sat, Apr 10, 2010 at 11:24:11PM +0200, Sylvain Munaut wrote:
> Reverting commit f5284ae1cf8babc1567b33f469e20a66a73fcd9e
> ipa: Reduce the throttling of the IPA msges
> This code used to be a sleep, it was changed to be a timer by Andreas
> but this timer does not seem to have any use. When doing the sw load
> this timer is increasing the upload time dramatically, reduce it to
> make it work faster.
> fixes the problem for me.
That makes complete sense. In fact, the reason why that throttling was
in place ('does not seem to have any use') was exactly to make the
initialization work at all. Initially I was working with a GSM900 unit,
and the initialization never worked without those sleeps.
This probably means that we're still doing something wrong with regard
to the object state transitions. I never really understood why 12.21
has to be so horribly complex with all those objects and their
dependencies, as well as administrative, operational and a third stage.
It sounds like a stupid exercise in object oriented data model serving
nothing but the purpose of advertising the perceived marvels of object
I for my part think reverting the patch is the way to go, until somebody
has debugged this issue. Is that ok with you, zecke?
- Harald Welte <laforge at gnumonks.org> http://laforge.gnumonks.org/
"Privacy in residential applications is a desirable marketing option."
(ETSI EN 300 175-7 Ch. A6)
More information about the OpenBSC