[mynewt-nimble] branch master updated: nimble/ll: Adjust start time of the first anchor point of connection

[ja...@apache.org]

This is an automated email from the ASF dual-hosted git repository.

janc pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/mynewt-nimble.git


The following commit(s) were added to refs/heads/master by this push:
     new 35a5603  nimble/ll: Adjust start time of the first anchor point of connection
35a5603 is described below

commit 35a560341c294c0d626a7d5fcd17126f9e2e1b17
Author: Yeming Li <62...@qq.com>
AuthorDate: Fri Dec 31 15:40:18 2021 +0800

    nimble/ll: Adjust start time of the first anchor point of connection
    
    Set the start_time of the first anchor point of new connection to
    entry->end_time+1.
    
    For the 32768 Hz crystal in nrf chip, 1 tick is 30.517us.
    The connection state machine use anchor point to store the cpu
    ticks and anchor_point_usec to store the remainder. Therefore,
    to compensate the inaccuracy of the crystal, the ticks of
    anchor_point will be add with 1 once the value of
    anchor_point_usec exceed 31. If two connections have same
    connection interval, the time difference between the two start
    of schedule item will decreased 1, which lead to an overlap.
    To prevent this from happenning, we set the start_time of sch
    to 1 cpu tick after the end_time of entry.
---
 nimble/controller/src/ble_ll_sched.c | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)

diff --git a/nimble/controller/src/ble_ll_sched.c b/nimble/controller/src/ble_ll_sched.c
index fe47dbc..cd60014 100644
--- a/nimble/controller/src/ble_ll_sched.c
+++ b/nimble/controller/src/ble_ll_sched.c
@@ -237,7 +237,19 @@ ble_ll_sched_insert(struct ble_ll_sched_item *sch, uint32_t max_delay,
                 }
             } else {
                 preempt_first = NULL;
-                sch->start_time = entry->end_time;
+                /*
+                 * For the 32768 Hz crystal in nrf chip, 1 tick is 30.517us.
+                 * The connection state machine use anchor point to store the
+                 * cpu ticks and anchor_point_usec to store the remainder.
+                 * Therefore, to compensate the inaccuracy of the crystal, the
+                 * ticks of anchor_point will be add with 1 once the value of
+                 * anchor_point_usec exceed 31. If two connections have same
+                 * connection interval, the time difference between the two
+                 * start of schedule item will decreased 1, which lead to
+                 * an overlap. To prevent this from happenning, we set the
+                 * start_time of sch to 1 cpu tick after the end_time of entry.
+                 */
+                sch->start_time = entry->end_time + 1;
 
                 if ((max_delay == 0) || CPUTIME_GEQ(sch->start_time,
                                                     max_start_time)) {