Partition Driver Test ( case 5103 )

Thread view:

Enable EMail Alerts

Start New Thread

Thread rating:

ykliu - 27 Oct 2006 09:24 GMT

Hello everyone,
I test case 5103 and have a failed result. After I trace the source
code, this case separate into 3 parts and do 3 kinds of test.
1. write partitions
2. write and verify partitions
3. verify partitions
Our test passed first test successfully, but failed at second. And this
fail do the same thing as first, It's so strange. Does someone know
what happened here ?

<TESTCASE ID=5103>
*** vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
*** TEST STARTING
***
*** Test Name: R/W and bounds check disk partition data
*** Test ID: 5103
*** Library Path: \windows\temp\msparttest.dll
*** Command Line: -zorch
*** Random Seed: 2033
*** Thread Count: 1
*** vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
BEGIN TEST:
Testing disk 0
Opening storage device "MSFlash"...
Opening storage device "MSFlash"...
Opened store 0x0fd1af46
Formatting store 0x0fd1af46
Getting store info for store 0x0fd1af46
Store opened and formatted successfully

================ Partition three parts ================
Requesting partition "TestPart1" of size 10602 sectors
Created partition "TestPart1" of size 10602 sectors
Requesting partition "TestPart2" of size 10601 sectors
Created partition "TestPart2" of size 10601 sectors
Requesting partition "TestPart3" of size 10601 sectors
Created partition "TestPart3" of size 10601 sectors

================== Write Partitions ==================

Opened partition TestPart1 on store 0x0fd1af46 as 0x6ff69646
Writing known data to partition 0x6ff69646
Opened partition TestPart2 on store 0x0fd1af46 as 0xcfb14f9a
Writing known data to partition 0xcfb14f9a
Opened partition TestPart3 on store 0x0fd1af46 as 0x6fad468a
Writing known data to partition 0x6fad468a

============== write and verify partitions ==============

Opened partition TestPart1 on store 0x0fd1af46 as 0x6ff90b0a
Writing known data to partition 0x6ff90b0a
FAIL in c:\mlqa\private\test\drivers\partitionmgr\api\msparttest\
diskhelp.cpp at line 540:
ReadWritePartition()
FAIL in c:\mlqa\private\test\drivers\partitionmgr\api\msparttest\
diskhelp.cpp at line 718:
WritePartitionData()
FAIL in c:\mlqa\private\test\drivers\partitionmgr\api\msparttest\
testproc.cpp at line 1334:
WriteAndVerifyPartition()
END TEST: "R/W and bounds check disk partition data", FAILED
</TESTCASE RESULT="FAILED">

Reply to this Message

Vadim Yushprakh - 28 Oct 2006 01:30 GMT

Hi,

It is strange that the same test function is failing during the second pass
over the same partition. According to your log,

> Writing known data to partition 0x6ff90b0a
> FAIL in co:\melba\private\test\drivers\partitionmgr\api\msparttest\
> diskhelp.cpp at line 540:
> ReadWritePartition()

the failure occurred in ReadWritePartion(). In the code for this function,
we print a failure explanation for all errors except the following:

diskhelp.cpp:842

__try
{
if(!DeviceIoControl(hPart, ctlCode, (PBYTE)&sgReq, sizeof(sgReq),
NULL, 0, &cBytes, NULL))
{
goto Error;
}

}
__except(EXCEPTION_EXECUTE_HANDLER)
...

So it seems that your test is failing in DeviceIoControl(). I would
recommend that you breakpoint at this part of the test code, and investigate
why the disk write is failing in the second pass.

Please let me know if you have any questions!

Thank You,

--Vadim Yushprakh

> Hello everyone,
> I test case 5103 and have a failed result. After I trace the source
[quoted text clipped - 58 lines]
> END TEST: "R/W and bounds check disk partition data", FAILED
> </TESTCASE RESULT="FAILED">

Reply to this Message

ykliu - 02 Nov 2006 07:58 GMT

Hello Vadim,

I know that the problem is in the code where you say. The difference
of them is the hpart address. I have done this test several times and
the error was the same as what I have pasted. Does it have a possible
that the handles of these three parts don't close yet? And this situation
makes the second time to do the function "OpenPartition" failed ?
According to the code, I don't support what I have guessed. If the
function "OpenPartition" fail, the test will stop. By the way, I can't
rebuild the msparttest.dll from the private code. Probably you could
give me a new one which have a message throw out during the test.

Thanks

YiKaiLiu

----- Original Message -----
From: "Vadim Yushprakh" <vadimy@microsoft.com>
Newsgroups: microsoft.public.windowsce.testtools
Sent: Saturday, October 28, 2006 8:30 AM
Subject: Re: Partition Driver Test ( case 5103 )

> Hi,
>
[quoted text clipped - 33 lines]
>
> --Vadim Yushprakh

Reply to this Message

Vadim Yushprakh - 08 Nov 2006 00:25 GMT

YiKaiLiu,

I did some more research on this issue and I was able to reproduce the
problem you have encountered. There is a known problem in the flash driver,
in which the compactor thread cannot run if there are no dirty sectors left
on the partition. Since our test touches every sector during the first pass
only once, none of them are marked as dirty. During the second pass when we
write again to the first sector, the compactor thread can't find any dirty
sectors to perform the compaction. This is why the test is failing to
write.

I would recommend skipping test case 5103 in your test pass. I apologize for
the problem.

Thank You!

--Vadim Yushprakh

> Hello Vadim,
>
[quoted text clipped - 55 lines]
>>
>> --Vadim Yushprakh

Reply to this Message

Tanveer123 - 03 Oct 2008 12:41 GMT

Hi

I am running the test case 5103 & 5106 on sd card, it is still in the process
of executing 5103. It takes a lot of time in executing the above two test
cases. Though it didnt fail but then it didnt finish completion.

Can any one explain what the above two test cases do? How much time it should
typically take? Does this depend on platform? One of the statements of the
output of the testcase I got is "All tests passed on disk3" , after which
again it starts opening the partitions and writing known data into them.

It is running since two days alas not completed

Plzz provide some info on this
TIA

>YiKaiLiu,
>
[quoted text clipped - 19 lines]
>>>
>>> --Vadim Yushprakh

Reply to this Message