Actions
Feature #967
openThink on error traces "signatures"
Start date:
03/18/2011
Due date:
% Done:
0%
Estimated time:
Published in build:
Description
By the analogy with #562 an another interesting and rather complex task is a comparison of unsafes coming from different techniques and models and with time. It seems that we may simplify a bit the given task. For example we may build some error trace "signature" and compare them aromatically (both traditional error trace analysis and manual examining of such "signatures" should also be presented). For my mind such a "signature" may be just a full function call stack like following:
entry_point; { ldv_initialize; sm_lm78_init; { lm78_isa_found; { __request_region; printk; } i2c_add_driver; { i2c_register_driver; } } nondet_int; nondet_int; lm78_i2c_detect; { dev_get_drvdata; { } i2c_check_functionality; { i2c_get_functionality; { (mem_4 ); //Hmm... } } mutex_lock; { } i2c_smbus_read_byte_data; lm78_alias_detect; { dev_get_drvdata; { } lm78_read_value; { mutex_lock; { ldv_blast_assert; { } } } } } }
More intellectual "signatures" may be developed. I think that this the task for error trace visualizer and then for stats server.
Updated by Evgeny Novikov over 13 years ago
- Status changed from New to Open
- Assignee changed from Pavel Shved to Evgeny Novikov
- Priority changed from Normal to High
This is a part of future knowledge base.
Updated by Evgeny Novikov about 13 years ago
- Status changed from Open to Resolved
The result of thought is placed here (especially, see section: "Критерии сравнения трасс ошибок").
Updated by Alexey Khoroshilov about 13 years ago
- Status changed from Resolved to Open
It should be copied to the bug or to our wiki before closing.
Actions