blob: 876e50e9880a339cc0b2f74f4a96d66ccc82ea5a [file] [log] [blame]
/*
* Copyright (C) 2019 The Android Open Source Project
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
-- note: These queries do comparisons based on raw_ftrace_entries.id by treating it as if it was equivalent to the temporal timestamp.
-- in practice, the ID of raw_ftrace_entries is based on its order in the ftrace buffer [and on the same cpu its equivalent].
-- we can always resort raw_ftrace_entries to ensure id order matches timestamp order. We should rarely need to compare by timestamp directly.
-- accessing 'floats' is inferior as they are harder to index, and will result in slower queries.
--
-- Naming convention note: '_fid' corresponds to 'raw_ftrace_entry.id'.
DROP VIEW IF EXISTS start_process_ui_threads;
-- Map of started process names to their UI thread's TID (as returned by gettid).
CREATE VIEW IF NOT EXISTS start_process_ui_threads AS
WITH
start_proc_tids AS (
SELECT sp.raw_ftrace_entry_id AS start_proc_fid,
sp.atrace_pid AS atrace_pid,
sp.process_name AS process_name,
--MIN(nc.raw_ftrace_entry_id) as next_comm_fid,
nc.raw_ftrace_entry_id AS next_comm_fid,
nc.next_pid as next_pid,
nc.next_comm as next_comm,
SUBSTR(sp.process_name, -15) AS cut -- why -15? See TASK_MAX in kernel, the sched_switch name is truncated to 16 bytes.
FROM start_procs AS sp,
sched_switch_next_comm_pids AS nc
WHERE sp.process_name LIKE '%' || nc.next_comm -- kernel truncates the sched_switch::next_comm event, so we must match the prefix of the full name.
--WHERE SUBSTR(sp.process_name, -16) == nc.next_comm
--WHERE cut == nc.next_comm
),
start_proc_tids_filtered AS (
SELECT *
FROM start_proc_tids
WHERE next_comm_fid > start_proc_fid -- safeguard that avoids choosing "earlier" sched_switch before process was even started.
--ORDER BY start_proc_fid, next_comm_fid
),
start_proc_all_threads AS (
SELECT DISTINCT
start_proc_fid, -- this is the ftrace entry of the system server 'Start proc: $process_name'. only need this to join for timestamp.
process_name, -- this is the '$process_name' from the system server entry.
-- next up we have all the possible thread IDs as parsed from sched_switch that corresponds most closest to the start proc.
next_pid AS ui_thread_tpid, -- sched_switch.next_pid. This can be any of the threads in that process, it's not necessarily the main UI thread yet.
next_comm,
MIN(next_comm_fid) AS next_comm_fid -- don't pick the 'later' next_comm_fid because it could correspond to another app start.
FROM start_proc_tids_filtered
GROUP BY start_proc_fid, ui_thread_tpid
),
activity_thread_mains AS (
SELECT * FROM tracing_mark_write_split WHERE atrace_message = 'ActivityThreadMain'
),
start_proc_ui_threads AS (
SELECT start_proc_fid,
process_name,
ui_thread_tpid,
next_comm,
next_comm_fid,
atm.raw_ftrace_entry_id as atm_fid,
atm.atrace_pid as atm_ui_thread_tid
FROM start_proc_all_threads AS spt,
activity_thread_mains AS atm
WHERE atm.atrace_pid == spt.ui_thread_tpid AND atm.raw_ftrace_entry_id > spt.start_proc_fid -- Ensure we ignore earlier ActivityThreadMains prior to their Start proc.
),
start_proc_ui_threads_filtered AS (
SELECT start_proc_fid,
process_name, -- e.g. 'com.android.settings'
--ui_thread_tpid,
--next_comm,
--next_comm_fid,
MIN(atm_fid) AS atm_fid,
atm_ui_thread_tid -- equivalent to gettid() for the process's UI thread.
FROM start_proc_ui_threads
GROUP BY start_proc_fid, atm_ui_thread_tid -- find the temporally closest ActivityTaskMain to a "Start proc: $process_name"
)
SELECT * FROM start_proc_ui_threads_filtered;
SELECT * FROM start_process_ui_threads;