GCR (osmo_store16be/osmo_store32be) - OpenBSC

12 Jan 2023

Hi All,

I was working a few months back on some LCLS implementation with 
bts-loop and I got it working, including that the pbx can use SIP 
Re-INVITES to get in and out of the audio loop during calls.

However, I only ever tested it with one call, that is with two phones. 
When I tested it on alive site, immediately there where problems so I 
reverted and left it. Looking at it again now, there's a very obvious 
problem which is that osmo-msc generates the same GCR for (almost) every 
call. I had noticed this before but for some reason I thought I had seen 
it generating different GCR for a second simultaneous call, but no.

Anyway, the above is extraneous info to the question:

Could somebody take a look at this:

https://gitea.osmocom.org/cellular-infrastructure/osmo-msc/src/branch/maste…

where we have:

     osmo_store32be(trans->callref, lcls->gcr.cr);
     osmo_store16be(use_lac ? trans->msc_a->via_cell.lai.lac : 
trans->msc_a->via_cell.cell_identity, lcls->gcr.cr + 3);

Now, If I change the order, such that would seem logical:

     osmo_store16be(use_lac ? trans->msc_a->via_cell.lai.lac : 
trans->msc_a->via_cell.cell_identity, lcls->gcr.cr + 3);
     osmo_store32be(trans->callref, lcls->gcr.cr);

Then I get a different GCR, reflecting the trans->callref for each call. 
But am I then maybe overwriting the LAC/CI ?

Would seem to make sense, but I'm just not sure if that is all there is 
to it, as I don't really grok osmo_store_xxxx

     for(i = 0; i < n; q[i] = (x >> ((n - 1 - i) * 8)) & 0xFF, i++);

Is it just a simple order error, and is everything OK with lcls->gcr.cr 
+ 3 as the pointer *p passed to osmo_store32be_ext() ?

  given that:

struct osmo_lcls *lcls;

where:

struct osmo_lcls {

     struct osmo_gcr_parsed gcr;

};

struct osmo_gcr_parsed {

     uint8_t cr[5];

};

I guess I'm still not really 100% on the char/uint8_t thing and 
advancing pointers.

Thanks!

k.