cellular-infrastructure
/
osmo-bts
mirror of https://gerrit.osmocom.org/osmo-bts
9
0
Fork 0
Code Issues Releases Wiki Activity

part 2 of: fix SAPIs for handover, osmo-bts-sysmo 

Change-Id: I30e2ca003a3b15766d9d9dcf529fb2157523758f
This commit is contained in:
Neels Hofmeyr 2020-11-04 13:31:41 +00:00 committed by laforge
parent f733cf8263
commit 1c05ef15ec
1 changed files with 20 additions and 21 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

41
src/osmo-bts-sysmo/oml.c
View File

@ -657,10 +657,6 @@ static const struct sapi_dir pdtch_sapis[] = {
#endif
};
static const struct sapi_dir ho_sapis[] = {
{ GsmL1_Sapi_Rach, GsmL1_Dir_RxUplink },
};
struct lchan_sapis {
const struct sapi_dir *sapis;
unsigned int num_sapis;
@ -693,11 +689,6 @@ static const struct lchan_sapis sapis_for_lchan[_GSM_LCHAN_MAX] = {
},
};
static const struct lchan_sapis sapis_for_ho = {
.sapis = ho_sapis,
.num_sapis = ARRAY_SIZE(ho_sapis),
};
static int mph_send_activate_req(struct gsm_lchan *lchan, struct sapi_cmd *cmd);
static int mph_send_deactivate_req(struct gsm_lchan *lchan, struct sapi_cmd *cmd);
static int mph_send_config_ciphering(struct gsm_lchan *lchan, struct sapi_cmd *cmd);
@ -1176,10 +1167,10 @@ int lchan_activate(struct gsm_lchan *lchan)
"%s Trying to activate lchan, but commands in queue\n",
gsm_lchan_name(lchan));
/* override the regular SAPIs if this is the first hand-over
* related activation of the LCHAN */
/* For handover, always start the main channel immediately. lchan->want_dl_sacch_active indicates whether dl
* SACCH should be activated. Also, for HO, start the RACH SAPI. */
if (lchan->ho.active == HANDOVER_ENABLED)
s4l = &sapis_for_ho;
enqueue_sapi_act_cmd(lchan, GsmL1_Sapi_Rach, GsmL1_Dir_RxUplink);
for (i = 0; i < s4l->num_sapis; i++) {
int sapi = s4l->sapis[i].sapi;
@ -1192,6 +1183,11 @@ int lchan_activate(struct gsm_lchan *lchan)
fl1h->alive_prim_cnt = 0;
osmo_timer_schedule(&fl1h->alive_timer, 5, 0);
}
/* For handover, possibly postpone activating the dl SACCH until the HO RACH is received. */
if (sapi == GsmL1_Sapi_Sacch && dir == GsmL1_Dir_TxDownlink
&& !lchan->want_dl_sacch_active)
continue;
enqueue_sapi_act_cmd(lchan, sapi, dir);
}
@ -1868,9 +1864,6 @@ int l1if_rsl_chan_act(struct gsm_lchan *lchan)
*/
int l1if_rsl_chan_mod(struct gsm_lchan *lchan)
{
const struct lchan_sapis *s4l = &sapis_for_lchan[lchan->type];
unsigned int i;
if (lchan->ho.active == HANDOVER_NONE)
return -1;
@ -1880,12 +1873,18 @@ int l1if_rsl_chan_mod(struct gsm_lchan *lchan)
/* Give up listening to RACH bursts */
release_sapi_ul_rach(lchan);
/* Activate the normal SAPIs */
for (i = 0; i < s4l->num_sapis; i++) {
int sapi = s4l->sapis[i].sapi;
int dir = s4l->sapis[i].dir;
enqueue_sapi_act_cmd(lchan, sapi, dir);
}
/* All the normal SAPIs have already been activated, only DL SACCH may still be missing.
*
* Note: theoretically, it would only be necessary to activate the DL SACCH when it is not active yet. With
* repeated HO RACH received, we shouldn't need to re-send the SAPI activation every time. However, tests with
* sysmoBTS show that when sending this SAPI activation only once, the lchan will release some seconds after a
* handover, with error messages indicating "Lost SACCH block, faking meas reports and ms pwr". When re-sending
* the SAPI activation for every RACH received, the problem goes away.
* Before introducing lchan->want_dl_sacch_active, this code here would activate all SAPIs for the main channel,
* which would also repeat for each RACH received. Now we only activate DL SACCH here.
*/
if (lchan->want_dl_sacch_active)
enqueue_sapi_act_cmd(lchan, GsmL1_Sapi_Sacch, GsmL1_Dir_TxDownlink);
return 0;
}