From 98655a9bd08290d6a674e76f00e4faf47d0970a3 Mon Sep 17 00:00:00 2001
From: Ankur Bansal <ankban@codeaurora.org>
Date: Sat, 17 Feb 2018 15:55:58 +0530
Subject: [PATCH] lrdp_v2 : memusage : Fix memory usages parsing
In case of task list corruption, traverse the task list in both
forward and backward direction to get memory usages by all the
task.
Change-Id: Ia6f1e05e1f82841a9cda29a0e27a6759f0c9b079
---
linux-ramdump-parser-v2/parsers/memusage.py | 44 ++++++++++++++++++++-
1 file changed, 42 insertions(+), 2 deletions(-)
diff --git a/linux-ramdump-parser-v2/parsers/memusage.py b/linux-ramdump-parser-v2/parsers/memusage.py
index af808dd..965f761 100644
--- a/linux-ramdump-parser-v2/parsers/memusage.py
+++ b/linux-ramdump-parser-v2/parsers/memusage.py
@@ -1,4 +1,4 @@
-# Copyright (c) 2016-2017 The Linux Foundation. All rights reserved.
+# Copyright (c) 2016-2018 The Linux Foundation. All rights reserved.
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 and
@@ -30,6 +30,7 @@ def do_dump_process_memory(ramdump):
offset_tasks = ramdump.field_offset('struct task_struct', 'tasks')
offset_comm = ramdump.field_offset('struct task_struct', 'comm')
offset_signal = ramdump.field_offset('struct task_struct', 'signal')
+ prev_offset = ramdump.field_offset('struct list_head','prev')
offset_adj = ramdump.field_offset('struct signal_struct', 'oom_score_adj')
offset_thread_group = ramdump.field_offset(
'struct task_struct', 'thread_group')
@@ -72,6 +73,45 @@ def do_dump_process_memory(ramdump):
next_task = ramdump.read_word(init_next_task)
if next_task is None:
+ init_next_task = init_addr + offset_tasks
+ init_next_task = init_next_task + prev_offset
+ init_next_task = ramdump.read_word(init_next_task)
+ init_thread_group = init_next_task - offset_tasks \
+ + offset_thread_group
+ while True:
+ init_next_task = init_next_task + prev_offset
+ orig_init_next_task = init_next_task
+ task_struct = init_thread_group - offset_thread_group
+ next_thread_comm = task_struct + offset_comm
+ thread_task_name = cleanupString(
+ ramdump.read_cstring(next_thread_comm, 16))
+ next_thread_pid = task_struct + offset_pid
+ thread_task_pid = ramdump.read_int(next_thread_pid)
+ signal_struct = ramdump.read_word(task_struct + offset_signal)
+ next_task = ramdump.read_word(init_next_task)
+ if next_task is None:
+ break
+ if (next_task == init_next_task and
+ next_task != orig_init_next_task):
+ break
+ if next_task in seen_tasks:
+ break
+ seen_tasks.add(next_task)
+ init_next_task = next_task
+ init_thread_group = init_next_task - offset_tasks\
+ + offset_thread_group
+ if init_next_task == orig_init_next_task:
+ break
+
+ if signal_struct == 0 or signal_struct is None:
+ continue
+ adj = ramdump.read_u16(signal_struct + offset_adj)
+ if adj & 0x8000:
+ adj = adj - 0x10000
+ rss, swap = get_rss(ramdump, task_struct)
+ if rss != 0:
+ task_info.append([thread_task_name, thread_task_pid, rss,
+ swap, rss + swap, adj])
break
if (next_task == init_next_task and
@@ -87,7 +127,7 @@ def do_dump_process_memory(ramdump):
if init_next_task == orig_init_next_task:
break
- if signal_struct == 0 or signal_struct == None :
+ if signal_struct == 0 or signal_struct is None:
continue
adj = ramdump.read_u16(signal_struct + offset_adj)
--
GitLab