[LTP] Bug in memory controller test - memctl test #4

[LTP] Bug in memory controller test - memctl test #4

[Mike Gahagan]

[-- Attachment #1: Type: text/plain, Size: 1432 bytes --]

When running the Memory Controller test # 4, I'm seeing this behavior below.

TEST 4: MEMORY CONTROLLER TESTING
RUNNING SETUP.....
TEST STARTED: Please avoid using system while this test executes
/mnt/tests/kernel/distribution/ltp/20090731/ltp-full-20090731/testcases/bin/run_memctl_test.sh: 
line 327: [: 62955520: unary operator expected
TINFO   Memory Resource Controller: stat check test fails in first run
/mnt/tests/kernel/distribution/ltp/20090731/ltp-full-20090731/testcases/bin/run_memctl_test.sh: 
line 364: [: 78704640: unary operator expected
TFAIL   Memory Resource Controller: stat check test FAILED
*** infinite loop here, script exists memctl_test01 remains running ***
^Cincrementing stop

The bash error is due to the 'grep -w "active"' not matching anything so 
one of the variables isn't getting set like it is supposed to. I believe 
the test was written with an old version of the memory controller what 
put a field called memory.stat.active in the file memory.stat. The 
current memory controller I am testing appears to have replaced this 
field with "rss". There are fields showing anonymous and file backed 
pages in memory.stat, but I don't think they are ever going to match the 
memory.usage.in.bytes the test is looking for. Rss always matches and I 
believe that is what we should update the check in the test to look for. 
The attached patch fixes test #4 to compare memory.usage.in.bytes with rss.

[-- Attachment #2: run_memctl_test.patch --]
[-- Type: text/plain, Size: 912 bytes --]

--- ltp-full-20090731/testcases/kernel/controllers/memctl/run_memctl_test.sh	2008-05-26 07:26:44.000000000 -0400
+++ run_memctl_test.sh.new	2009-08-24 14:17:12.475154507 -0400
@@ -323,7 +323,7 @@
 		check_mem_allocated;# $MEM_TOTAL;
 		# now we can check the memory usage from both files
 		USAGE_FROM_STAT=`cat /dev/memctl/group_1/memory.stat \
-					| grep -w "active" | cut -d" " -f2`;
+					| grep -w "rss" | cut -d" " -f2`;
 		if [ $USAGE_FROM_USAGE_IN_BYTES -eq $USAGE_FROM_STAT ]
 		then
 			echo "memory usage from memory.usage_in_bytes= $USAGE_FROM_USAGE_IN_BYTES";
@@ -360,7 +360,7 @@
 
 		# now we can check the memory usage from both files
 		USAGE_FROM_STAT=`cat /dev/memctl/group_1/memory.stat \
-					| grep -w "active" | cut -d" " -f2`;
+					| grep -w "rss" | cut -d" " -f2`;
 		if [ $USAGE_FROM_USAGE_IN_BYTESGRP1_MEMUSAGE -eq $USAGE_FROM_STAT ] \
 							&& [ $FIRST_STEP_PASS -eq 1 ]
 		then

[-- Attachment #3: Type: text/plain, Size: 355 bytes --]

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july

[-- Attachment #4: Type: text/plain, Size: 155 bytes --]

_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] Bug in memory controller test - memctl test #4

[Subrata Modak]

On Mon, 2009-08-24 at 14:32 -0400, Mike Gahagan wrote: 
> When running the Memory Controller test # 4, I'm seeing this behavior below.
> 
> TEST 4: MEMORY CONTROLLER TESTING
> RUNNING SETUP.....
> TEST STARTED: Please avoid using system while this test executes
> /mnt/tests/kernel/distribution/ltp/20090731/ltp-full-20090731/testcases/bin/run_memctl_test.sh: 
> line 327: [: 62955520: unary operator expected
> TINFO   Memory Resource Controller: stat check test fails in first run
> /mnt/tests/kernel/distribution/ltp/20090731/ltp-full-20090731/testcases/bin/run_memctl_test.sh: 
> line 364: [: 78704640: unary operator expected
> TFAIL   Memory Resource Controller: stat check test FAILED
> *** infinite loop here, script exists memctl_test01 remains running ***
> ^Cincrementing stop
> 
> The bash error is due to the 'grep -w "active"' not matching anything so 
> one of the variables isn't getting set like it is supposed to. I believe 
> the test was written with an old version of the memory controller what 
> put a field called memory.stat.active in the file memory.stat. The 

Correct. Somebody needs to revisit these tests, as the ABIs has changed.
Thanks for pointing that out. However, your patch has the following
problem:

1) Missing Signed-off-by:
2) missing header for unified diff at line 3 of patch
patching file testcases/kernel/controllers/memctl/run_memctl_test.sh

Please look into them when you send a patch next time ;-)

Regards--
Subrata

> current memory controller I am testing appears to have replaced this 
> field with "rss". There are fields showing anonymous and file backed 
> pages in memory.stat, but I don't think they are ever going to match the 
> memory.usage.in.bytes the test is looking for. Rss always matches and I 
> believe that is what we should update the check in the test to look for. 
> The attached patch fixes test #4 to compare memory.usage.in.bytes with rss.
> ------------------------------------------------------------------------------
> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
> trial. Simplify your report design, integration and deployment - and focus on 
> what you do best, core application coding. Discover what's new with 
> Crystal Reports now.  http://p.sf.net/sfu/bobj-july
> _______________________________________________ Ltp-list mailing list Ltp-list@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ltp-list


------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] Bug in memory controller test - memctl test #4 - PATCH repost

[Mike Gahagan]

[-- Attachment #1: Type: text/plain, Size: 2277 bytes --]

On 08/25/2009 05:58 AM, Subrata Modak wrote:
> On Mon, 2009-08-24 at 14:32 -0400, Mike Gahagan wrote:
>> When running the Memory Controller test # 4, I'm seeing this behavior below.
>>
>> TEST 4: MEMORY CONTROLLER TESTING
>> RUNNING SETUP.....
>> TEST STARTED: Please avoid using system while this test executes
>> /mnt/tests/kernel/distribution/ltp/20090731/ltp-full-20090731/testcases/bin/run_memctl_test.sh:
>> line 327: [: 62955520: unary operator expected
>> TINFO   Memory Resource Controller: stat check test fails in first run
>> /mnt/tests/kernel/distribution/ltp/20090731/ltp-full-20090731/testcases/bin/run_memctl_test.sh:
>> line 364: [: 78704640: unary operator expected
>> TFAIL   Memory Resource Controller: stat check test FAILED
>> *** infinite loop here, script exists memctl_test01 remains running ***
>> ^Cincrementing stop
>>
>> The bash error is due to the 'grep -w "active"' not matching anything so
>> one of the variables isn't getting set like it is supposed to. I believe
>> the test was written with an old version of the memory controller what
>> put a field called memory.stat.active in the file memory.stat. The
>
> Correct. Somebody needs to revisit these tests, as the ABIs has changed.
> Thanks for pointing that out. However, your patch has the following
> problem:
>
> 1) Missing Signed-off-by:
> 2) missing header for unified diff at line 3 of patch
> patching file testcases/kernel/controllers/memctl/run_memctl_test.sh
>
> Please look into them when you send a patch next time ;-)
>
> Regards--
> Subrata
>
>> current memory controller I am testing appears to have replaced this
>> field with "rss". There are fields showing anonymous and file backed
>> pages in memory.stat, but I don't think they are ever going to match the
>> memory.usage.in.bytes the test is looking for. Rss always matches and I
>> believe that is what we should update the check in the test to look for.
>> The attached patch fixes test #4 to compare memory.usage.in.bytes with rss.
>> ------------------------------------------------------------------------------

Hello,

Here is a resubmission of my memctl test patch. It should clear up the 
issues noted above. No functional change from the previous patch.

Signed-off-by: Mike Gahagan <mgahagan@redhat.com>


[-- Attachment #2: run_memctl_test.patch --]
[-- Type: text/plain, Size: 1207 bytes --]

 run_memctl_test.sh |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
diff -uNr ltp-full-20090731/testcases/kernel/controllers/memctl/run_memctl_test.sh ltp-full-20090731.new/testcases/kernel/controllers/memctl/run_memctl_test.sh
--- ltp-full-20090731/testcases/kernel/controllers/memctl/run_memctl_test.sh	2009-08-25 16:21:47.298273822 -0400
+++ ltp-full-20090731.new/testcases/kernel/controllers/memctl/run_memctl_test.sh	2009-08-25 16:41:57.606201495 -0400
@@ -323,7 +323,7 @@
 		check_mem_allocated;# $MEM_TOTAL;
 		# now we can check the memory usage from both files
 		USAGE_FROM_STAT=`cat /dev/memctl/group_1/memory.stat \
-					| grep -w "active" | cut -d" " -f2`;
+					| grep -w "rss" | cut -d" " -f2`;
 		if [ $USAGE_FROM_USAGE_IN_BYTES -eq $USAGE_FROM_STAT ]
 		then
 			echo "memory usage from memory.usage_in_bytes= $USAGE_FROM_USAGE_IN_BYTES";
@@ -360,7 +360,7 @@
 
 		# now we can check the memory usage from both files
 		USAGE_FROM_STAT=`cat /dev/memctl/group_1/memory.stat \
-					| grep -w "active" | cut -d" " -f2`;
+					| grep -w "rss" | cut -d" " -f2`;
 		if [ $USAGE_FROM_USAGE_IN_BYTESGRP1_MEMUSAGE -eq $USAGE_FROM_STAT ] \
 							&& [ $FIRST_STEP_PASS -eq 1 ]
 		then

[-- Attachment #3: Type: text/plain, Size: 355 bytes --]

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july

[-- Attachment #4: Type: text/plain, Size: 155 bytes --]

_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] Bug in memory controller test - memctl test #4 - PATCH repost

[Subrata Modak]

Hi,

On Wed, 2009-08-26 at 12:01 -0400, Mike Gahagan wrote: 
> On 08/25/2009 05:58 AM, Subrata Modak wrote:
> > On Mon, 2009-08-24 at 14:32 -0400, Mike Gahagan wrote:
> >> When running the Memory Controller test # 4, I'm seeing this behavior below.
> >>
> >> TEST 4: MEMORY CONTROLLER TESTING
> >> RUNNING SETUP.....
> >> TEST STARTED: Please avoid using system while this test executes
> >> /mnt/tests/kernel/distribution/ltp/20090731/ltp-full-20090731/testcases/bin/run_memctl_test.sh:
> >> line 327: [: 62955520: unary operator expected
> >> TINFO   Memory Resource Controller: stat check test fails in first run
> >> /mnt/tests/kernel/distribution/ltp/20090731/ltp-full-20090731/testcases/bin/run_memctl_test.sh:
> >> line 364: [: 78704640: unary operator expected
> >> TFAIL   Memory Resource Controller: stat check test FAILED
> >> *** infinite loop here, script exists memctl_test01 remains running ***
> >> ^Cincrementing stop
> >>
> >> The bash error is due to the 'grep -w "active"' not matching anything so
> >> one of the variables isn't getting set like it is supposed to. I believe
> >> the test was written with an old version of the memory controller what
> >> put a field called memory.stat.active in the file memory.stat. The
> >
> > Correct. Somebody needs to revisit these tests, as the ABIs has changed.

Would be able to spend some time here ?

> > Thanks for pointing that out. However, your patch has the following
> > problem:
> >
> > 1) Missing Signed-off-by:
> > 2) missing header for unified diff at line 3 of patch
> > patching file testcases/kernel/controllers/memctl/run_memctl_test.sh
> >
> > Please look into them when you send a patch next time ;-)
> >
> > Regards--
> > Subrata
> >
> >> current memory controller I am testing appears to have replaced this
> >> field with "rss". There are fields showing anonymous and file backed
> >> pages in memory.stat, but I don't think they are ever going to match the
> >> memory.usage.in.bytes the test is looking for. Rss always matches and I
> >> believe that is what we should update the check in the test to look for.
> >> The attached patch fixes test #4 to compare memory.usage.in.bytes with rss.
> >> ------------------------------------------------------------------------------
> 
> Hello,
> 
> Here is a resubmission of my memctl test patch. It should clear up the 
> issues noted above. No functional change from the previous patch.
> 
> Signed-off-by: Mike Gahagan <mgahagan@redhat.com>

Thanks. I already merged your patch.

Regards--
Subrata

> 
> ------------------------------------------------------------------------------
> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
> trial. Simplify your report design, integration and deployment - and focus on 
> what you do best, core application coding. Discover what's new with 
> Crystal Reports now.  http://p.sf.net/sfu/bobj-july
> _______________________________________________ Ltp-list mailing list Ltp-list@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ltp-list


------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] Bug in memory controller test - memctl test #4 - PATCH repost

[Mike Gahagan]

On 08/27/2009 06:19 AM, Subrata Modak wrote:
> Hi,
>
> On Wed, 2009-08-26 at 12:01 -0400, Mike Gahagan wrote:
>> On 08/25/2009 05:58 AM, Subrata Modak wrote:
>>> On Mon, 2009-08-24 at 14:32 -0400, Mike Gahagan wrote:
>>>> When running the Memory Controller test # 4, I'm seeing this behavior below.
>>>>
>>>> TEST 4: MEMORY CONTROLLER TESTING
>>>> RUNNING SETUP.....
>>>> TEST STARTED: Please avoid using system while this test executes
>>>> /mnt/tests/kernel/distribution/ltp/20090731/ltp-full-20090731/testcases/bin/run_memctl_test.sh:
>>>> line 327: [: 62955520: unary operator expected
>>>> TINFO   Memory Resource Controller: stat check test fails in first run
>>>> /mnt/tests/kernel/distribution/ltp/20090731/ltp-full-20090731/testcases/bin/run_memctl_test.sh:
>>>> line 364: [: 78704640: unary operator expected
>>>> TFAIL   Memory Resource Controller: stat check test FAILED
>>>> *** infinite loop here, script exists memctl_test01 remains running ***
>>>> ^Cincrementing stop
>>>>
>>>> The bash error is due to the 'grep -w "active"' not matching anything so
>>>> one of the variables isn't getting set like it is supposed to. I believe
>>>> the test was written with an old version of the memory controller what
>>>> put a field called memory.stat.active in the file memory.stat. The
>>>
>>> Correct. Somebody needs to revisit these tests, as the ABIs has changed.
>
> Would be able to spend some time here ?
>

Sure, I would be glad to. So far this was the only issue I have found 
(or the only really obvious one at least). I plan on picking up and 
running the recently checked in new tests as well, probably in the next 
week or so.



>>> Thanks for pointing that out. However, your patch has the following
>>> problem:
>>>
>>> 1) Missing Signed-off-by:
>>> 2) missing header for unified diff at line 3 of patch
>>> patching file testcases/kernel/controllers/memctl/run_memctl_test.sh
>>>
>>> Please look into them when you send a patch next time ;-)
>>>
>>> Regards--
>>> Subrata
>>>
>>>> current memory controller I am testing appears to have replaced this
>>>> field with "rss". There are fields showing anonymous and file backed
>>>> pages in memory.stat, but I don't think they are ever going to match the
>>>> memory.usage.in.bytes the test is looking for. Rss always matches and I
>>>> believe that is what we should update the check in the test to look for.
>>>> The attached patch fixes test #4 to compare memory.usage.in.bytes with rss.
>>>> ------------------------------------------------------------------------------
>>
>> Hello,
>>
>> Here is a resubmission of my memctl test patch. It should clear up the
>> issues noted above. No functional change from the previous patch.
>>
>> Signed-off-by: Mike Gahagan<mgahagan@redhat.com>
>
> Thanks. I already merged your patch.
>

Thank you !



------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] Bug in memory controller test - memctl test #4 - PATCH repost

[Subrata Modak]

On Fri, 2009-08-28 at 11:07 -0400, Mike Gahagan wrote: 
> On 08/27/2009 06:19 AM, Subrata Modak wrote:
> > Hi,
> >
> > On Wed, 2009-08-26 at 12:01 -0400, Mike Gahagan wrote:
> >> On 08/25/2009 05:58 AM, Subrata Modak wrote:
> >>> On Mon, 2009-08-24 at 14:32 -0400, Mike Gahagan wrote:
> >>>> When running the Memory Controller test # 4, I'm seeing this behavior below.
> >>>>
> >>>> TEST 4: MEMORY CONTROLLER TESTING
> >>>> RUNNING SETUP.....
> >>>> TEST STARTED: Please avoid using system while this test executes
> >>>> /mnt/tests/kernel/distribution/ltp/20090731/ltp-full-20090731/testcases/bin/run_memctl_test.sh:
> >>>> line 327: [: 62955520: unary operator expected
> >>>> TINFO   Memory Resource Controller: stat check test fails in first run
> >>>> /mnt/tests/kernel/distribution/ltp/20090731/ltp-full-20090731/testcases/bin/run_memctl_test.sh:
> >>>> line 364: [: 78704640: unary operator expected
> >>>> TFAIL   Memory Resource Controller: stat check test FAILED
> >>>> *** infinite loop here, script exists memctl_test01 remains running ***
> >>>> ^Cincrementing stop
> >>>>
> >>>> The bash error is due to the 'grep -w "active"' not matching anything so
> >>>> one of the variables isn't getting set like it is supposed to. I believe
> >>>> the test was written with an old version of the memory controller what
> >>>> put a field called memory.stat.active in the file memory.stat. The
> >>>
> >>> Correct. Somebody needs to revisit these tests, as the ABIs has changed.
> >
> > Would be able to spend some time here ?
> >
> 
> Sure, I would be glad to. So far this was the only issue I have found 
> (or the only really obvious one at least). I plan on picking up and 
> running the recently checked in new tests as well, probably in the next 
> week or so.

Great. Thanks for this interest.

Regards--
Subrata

> 
> 
> 
> >>> Thanks for pointing that out. However, your patch has the following
> >>> problem:
> >>>
> >>> 1) Missing Signed-off-by:
> >>> 2) missing header for unified diff at line 3 of patch
> >>> patching file testcases/kernel/controllers/memctl/run_memctl_test.sh
> >>>
> >>> Please look into them when you send a patch next time ;-)
> >>>
> >>> Regards--
> >>> Subrata
> >>>
> >>>> current memory controller I am testing appears to have replaced this
> >>>> field with "rss". There are fields showing anonymous and file backed
> >>>> pages in memory.stat, but I don't think they are ever going to match the
> >>>> memory.usage.in.bytes the test is looking for. Rss always matches and I
> >>>> believe that is what we should update the check in the test to look for.
> >>>> The attached patch fixes test #4 to compare memory.usage.in.bytes with rss.
> >>>> ------------------------------------------------------------------------------
> >>
> >> Hello,
> >>
> >> Here is a resubmission of my memctl test patch. It should clear up the
> >> issues noted above. No functional change from the previous patch.
> >>
> >> Signed-off-by: Mike Gahagan<mgahagan@redhat.com>
> >
> > Thanks. I already merged your patch.
> >
> 
> Thank you !
> 
> 


------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list