LU-18229 ldlm: BL_AST|CANCELLING lock still can be batched

LU-18229 ldlm: BL_AST|CANCELLING lock still can be batched

The current code makes BL_AST locks which are also CANCELLING to be
cancelled individually (one lh per RPC), because they are already in
the l_bl_list. This still could be optimised.

A small cleanup in mdc_rename(): ldlm_cli_cancel_list() is already
called by mdc_prep_elc_req()->ldlm_prep_elc_req(), aligned with other
mdc_prep_elc_req() calls.

Lustre-change: https://review.whamcloud.com/56389
Lustre-commit: 1c9115f919109936576193a4d20baeccedb8be41

HPE-bug-id: LUS-12470
Fixes: b65374d9 ("LU-16285 ldlm: send the cancel RPC asap")
Signed-off-by: Frederick Dilger <fdilger@whamcloud.com>
Signed-off-by: Vitaly Fertman <vitaly.fertman@hpe.com>
Change-Id: I218d38bc56a885845c48a3c982840b35b132f213
Reviewed-by: Andriy Skulysh <andriy.skulysh@hpe.com>
Reviewed-by: Alexander Boyko <alexander.boyko@hpe.com>
Reviewed-on: https://review.whamcloud.com/c/fs/lustre-release/+/58770
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>