Change in osmo-bts[master]: l1sap: add repeated uplink SACCH

[laforge]

laforge has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-bts/+/21185 )

Change subject: l1sap: add repeated uplink SACCH
......................................................................


Patch Set 4:

(1 comment)

https://gerrit.osmocom.org/c/osmo-bts/+/21185/4/src/osmo-bts-trx/sched_lchan_xcch.c 
File src/osmo-bts-trx/sched_lchan_xcch.c:

https://gerrit.osmocom.org/c/osmo-bts/+/21185/4/src/osmo-bts-trx/sched_lchan_xcch.c@37 
PS4, Line 37: /* Add two arrays of sbits */
            : static void add_sbits(sbit_t *current, const sbit_t *previous)
so this is a saturated addition implementation.  I have the non-scientific feeling that what we'd actually want is an average of the two bursts, i.e. "add both soft bits and divide by 0.5".   If both input variables will be int8_t, then by defniton the "add two and divide by half" can never exceed the scaale and you will never saturate.  As you are doing arithmetic in "int" math anyway, there shouldn't be any overflow issues either.

I'm happy to defer to tnt or Hoernchen, who have more insight into the theoretical background of our soft-bit convolutional decoder than I have.



-- 
To view, visit https://gerrit.osmocom.org/c/osmo-bts/+/21185
To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings

Gerrit-Project: osmo-bts
Gerrit-Branch: master
Gerrit-Change-Id: I7e4cc33cc010866e41e3b594351a7f7bf93e08ac
Gerrit-Change-Number: 21185
Gerrit-PatchSet: 4
Gerrit-Owner: dexter <pmaier at sysmocom.de>
Gerrit-Reviewer: Hoernchen <ewild at sysmocom.de>
Gerrit-Reviewer: Jenkins Builder
Gerrit-Reviewer: fixeria <vyanitskiy at sysmocom.de>
Gerrit-Reviewer: tnt <tnt at 246tNt.com>
Gerrit-CC: laforge <laforge at osmocom.org>
Gerrit-Comment-Date: Thu, 26 Nov 2020 10:15:25 +0000
Gerrit-HasComments: Yes
Gerrit-Has-Labels: No
Gerrit-MessageType: comment
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osmocom.org/pipermail/gerrit-log/attachments/20201126/e0b5746f/attachment.htm>