Multislot allocation failures and defects

Holger Hans Peter Freyther hfreyther at sysmocom.de
Wed Jan 15 09:59:35 UTC 2014


On Sat, Jan 04, 2014 at 04:19:34PM +0100, Andreas Eversberg wrote:
> dear holger,

dear andreas,

> 
> i added serveral fixes that showed up with your test code. i have pushed
> them to the jolly/allocation-fixes branch. in also includes a fix for
> the missing incrementation of 'i' in select_first_ts() and
> select_ul_slots().

I merged your changes and the crazy test that tested all combinations
started to work so I went ahead an merged your change. What Daniel (and
the jenkins) noticed is that I didn't update the test result. I asked
him to update the test result but he pointed out that there appears to
be a genuine regression:

@@ -409,7 +409,6 @@
 PDCH[5] is first common for DL
 PDCH[5] is used for UL
 PDCH[6] is used for UL
-PDCH[7] is used for UL
 PDCH[5] is control_ts for UL
 PDCH[5] is first common for UL
 PDCH[5] is used for DL
@@ -420,7 +419,6 @@
 Testing jolly example
 PDCH[1] is used for UL
 PDCH[2] is used for UL
-PDCH[3] is used for UL
 PDCH[1] is control_ts for UL
 PDCH[1] is first common for UL
 PDCH[1] is used for DL

so it appears that the multislot algorithm started to not assign/use
the last timeslot. Could you please have a look at that?

> i have no fix for the USF problem. if the first_common_ts on concurrent
> TBFs is different, we should reject the TBF by sending a Packet Access
> Reject, but this message is also not implemented.

I will add a todo to the code.

-- 
- Holger Freyther <hfreyther at sysmocom.de>       http://www.sysmocom.de/
=======================================================================
* sysmocom - systems for mobile communications GmbH
* Schivelbeiner Str. 5
* 10439 Berlin, Germany
* Sitz / Registered office: Berlin, HRB 134158 B
* Geschaeftsfuehrer / Managing Directors: Holger Freyther, Harald Welte





More information about the osmocom-net-gprs mailing list