Commit eee82ce591 for frr

commit eee82ce591fade0d083477bb1ef39770668c3ae1
Author: Enke Chen <enchen@paloaltonetworks.com>
Date:   Tue Nov 3 12:37:16 2026 -0800

    bgpd: fix "clear bgp soft out" to bypass duplicate suppression

    When "clear bgp * soft out" is executed, updates should be re-sent
    to peers even if they are duplicates of previously sent updates.
    This requires passing force=true to bgp_announce_route() so that
    SUBGRP_STATUS_FORCE_UPDATES is set, which bypasses the duplicate
    suppression check.

    The bug was introduced in commit e1a32ec1c5 ("bgpd: bgp_announce_route
    should know if we should force the update or not") when the force
    parameter was added. All existing callers were converted to pass
    false, but peer_clear_soft() should have been set to true.

    Fixes: e1a32ec1c5 ("bgpd: bgp_announce_route should know if we should force the update or not")
    Signed-off-by: Enke Chen <enchen@paloaltonetworks.com>

diff --git a/bgpd/bgpd.c b/bgpd/bgpd.c
index e5775f0152..c5f9d53238 100644
--- a/bgpd/bgpd.c
+++ b/bgpd/bgpd.c
@@ -9484,7 +9484,7 @@ int peer_clear_soft(struct peer *peer, afi_t afi, safi_t safi,
 			UNSET_FLAG(paf->subgroup->sflags,
 				   SUBGRP_STATUS_DEFAULT_ORIGINATE);

-		bgp_announce_route(peer, afi, safi, false);
+		bgp_announce_route(peer, afi, safi, true);
 	}

 	if (stype == BGP_CLEAR_SOFT_IN_ORF_PREFIX) {