Re: Locking API testsuite: Failures

Previous thread: [PATCH] USB: add USB test and measurement class driver by Greg KH on Tuesday, August 26, 2008 - 5:05 pm. (17 messages)

Next thread: Re: [RFC PATCH -v2] percpu_counters: make fbc->count read atomic on 32 bit architecture by Andrew Morton on Tuesday, August 26, 2008 - 5:26 pm. (1 message)
From: Peter Teoh
Date: Tuesday, August 26, 2008 - 5:27 pm

[    0.000999] | Locking API testsuite:
[    0.000999] ----------------------------------------------------------------------------
[    0.000999]                                  | spin |wlock |rlock
|mutex | wsem | rsem |
[    0.000999]
--------------------------------------------------------------------------
[    0.000999]                      A-A deadlock:failed|failed|  ok
|failed|failed|failed|
[    0.000999]                  A-B-B-A deadlock:failed|failed|  ok
|failed|failed|failed|
[    0.000999]              A-B-B-C-C-A deadlock:failed|failed|  ok
|failed|failed|failed|
[    0.000999]              A-B-C-A-B-C deadlock:failed|failed|  ok
|failed|failed|failed|
[    0.000999]          A-B-B-C-C-D-D-A deadlock:failed|failed|  ok
|failed|failed|failed|
[    0.000999]          A-B-C-D-B-D-D-A deadlock:failed|failed|  ok
|failed|failed|failed|
[    0.000999]          A-B-C-D-B-C-D-A deadlock:failed|failed|  ok
|failed|failed|failed|
[    0.000999]                     double
unlock:failed|failed|failed|failed|failed|failed|
[    0.000999]                   initialize
held:failed|failed|failed|failed|failed|failed|
[    0.000999]                  bad unlock order:  ok  |  ok  |  ok  |
 ok  |  ok  |  ok  |
[    0.000999]
--------------------------------------------------------------------------
[    0.000999]               recursive read-lock:             |  ok  |
            |failed|
[    0.000999]            recursive read-lock #2:             |  ok  |
            |failed|
[    0.000999]             mixed read-write-lock:             |failed|
            |failed|
[    0.000999]             mixed write-read-lock:             |failed|
            |failed|
[    0.000999]
--------------------------------------------------------------------------
[    0.000999]      hard-irqs-on + irq-safe-A/12:failed|failed|  ok  |
[    0.000999]      soft-irqs-on + irq-safe-A/12:failed|failed|  ok  |
[    0.000999]      hard-irqs-on + irq-safe-A/21:failed|failed|  ok  |
[    0.000999]  ...
From: Peter Zijlstra
Date: Tuesday, August 26, 2008 - 6:58 pm

Probably that you forgot to enable lockdep ;-)

In which case the error cases fail to produce an error, and the test
fails, but its an expected error, otherwise it would have shouted much
more verbose.



--

From: Peter Teoh
Date: Tuesday, August 26, 2008 - 11:32 pm

Thanks for the reply.


Several questions:

1.   I noticed there is a Kconfig.debug with this entry:

config DEBUG_LOCKING_API_SELFTESTS
        bool "Locking API boot-time self-tests"
        depends on DEBUG_KERNEL
        help
          Say Y here if you want the kernel to run a short self-test during
          bootup. The self-test checks whether common types of locking bugs
          are detected by debugging mechanisms or not. (if you disable
          lock debugging then those bugs wont be detected of course.)
          The following locking APIs are covered: spinlocks, rwlocks,
          mutexes and rwsems.

How do I use this Kconfig.debug, just overwriting the standard
lib/Kconfig?   I overwrote it and got the following errors:

scripts/kconfig/conf -o arch/x86/Kconfig
file kernel/trace/Kconfig already scanned?
make[1]: *** [oldconfig] Error 1
make: *** [oldconfig] Error 2

2.   So does it make sense to  compile with
CONFIG_DEBUG_LOCKING_API_SELFTESTS without CONFIG_LOCKDEP?
Should the lib/Kconfig indicate such a dependency?

3.   My current .config (which gave rise to the failure errors) has
CONFIG_LOCKDEP_SUPPORT=y, and CONFIG_DEBUG_LOCKING_API_SELFTESTS=y.
Question is what is the different between LOCKDEP_SUPPORT and LOCKDEP
alone?




-- 
Regards,
Peter Teoh
--

From: Ingo Molnar
Date: Tuesday, August 26, 2008 - 11:45 pm

You should have read this bit of the self-test printout:

[    0.004000] 133 out of 218 testcases failed, as expected. |

there was no unexpected failure. A real test-case failure is displayed 
as: 'FAILED', and you'll get a verbose printout about the unexpected 

no, that messes up the kernel. Kconfig.debug is a separate, unique 

yes. It shows the kinds of bugs we dont find when PROVE_LOCKING is 
disabled. It also shows that the test-cases are working as expected.

	Ingo
--

Previous thread: [PATCH] USB: add USB test and measurement class driver by Greg KH on Tuesday, August 26, 2008 - 5:05 pm. (17 messages)

Next thread: Re: [RFC PATCH -v2] percpu_counters: make fbc->count read atomic on 32 bit architecture by Andrew Morton on Tuesday, August 26, 2008 - 5:26 pm. (1 message)