[dts] [PATCH] tests/kni: fix failures for kni

[dts] [PATCH] tests/kni: fix failures for kni

[Jianbo Liu]

The tests need to enable KNI ethtool, and recomple DPDK.
Also bind to thread 0 of each core for function tests.

Signed-off-by: Jianbo Liu <jianbo.liu@linaro.org>
---
 tests/TestSuite_kni.py | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/tests/TestSuite_kni.py b/tests/TestSuite_kni.py
index acd6230..93203f5 100644
--- a/tests/TestSuite_kni.py
+++ b/tests/TestSuite_kni.py
@@ -52,8 +52,8 @@ packet_sizes_routing = [64, 256]
 
 ports_cores_template = '\(P([0123]),(C\{\d.\d.\d\}),(C\{\d.\d.\d\}),(C\{\d.\d.\d\}),?(C\{\d.\d.\d\})?\),?'
 
-default_1_port_cores_config = '(P0,C{1.0.0},C{1.1.0},C{1.0.1})'
-default_2_port_cores_config = '(P0,C{1.0.0},C{1.1.0},C{1.0.1}),(P1,C{1.2.0},C{1.3.0},C{1.2.1})'
+default_1_port_cores_config = '(P0,C{1.0.0},C{1.1.0},C{1.2.0})'
+default_2_port_cores_config = '(P0,C{1.0.0},C{1.1.0},C{1.2.0}),(P1,C{1.3.0},C{1.4.0},C{1.5.0})'
 
 stress_test_iterations = 50
 stress_test_random_iterations = 50
@@ -258,6 +258,9 @@ class TestKni(TestCase):
         self.verify('no brctl' not in out,
                     "The linux tool brctl is needed to run this test suite")
 
+        self.dut.send_expect("sed -i -e 's/KNI_KMOD_ETHTOOL=n$/KNI_KMOD_ETHTOOL=y/' config/common_base", "# ", 30)
+        self.dut.build_install_dpdk(self.target)
+
         out = self.dut.send_expect("make -C ./examples/kni/", "# ", 5)
         self.verify('Error' not in out, "Compilation failed")
 
@@ -1207,5 +1210,9 @@ class TestKni(TestCase):
         """
         Run after each test suite.
         """
+
+        self.dut.send_expect("sed -i -e 's/KNI_KMOD_ETHTOOL=y$/KNI_KMOD_ETHTOOL=n/' config/common_base", "# ", 30)
+        self.dut.build_install_dpdk(self.target)
+
         self.dut.kill_all()
         self.dut.send_expect("rmmod rte_kni", "# ", 10)
-- 
1.9.1

Re: [dts] [PATCH] tests/kni: fix failures for kni

[Liu, Yong]

Hi Jianbo,
Is there any particular reason for not chose hyper-threading cores?
There're some performance cases are depending on hyper-threading cores. 
If hyper-threading is not enabled on the platform, those cases won't be able to run. 
Could you please clear the reason?

Thanks,
Marvin

> -----Original Message-----
> From: dts [mailto:dts-bounces@dpdk.org] On Behalf Of Jianbo Liu
> Sent: Thursday, August 17, 2017 1:23 PM
> To: dts@dpdk.org; herbert.guan@arm.com
> Cc: Jianbo Liu <jianbo.liu@linaro.org>
> Subject: [dts] [PATCH] tests/kni: fix failures for kni
> 
> The tests need to enable KNI ethtool, and recomple DPDK.
> Also bind to thread 0 of each core for function tests.
> 
> Signed-off-by: Jianbo Liu <jianbo.liu@linaro.org>
> ---
>  tests/TestSuite_kni.py | 11 +++++++++--
>  1 file changed, 9 insertions(+), 2 deletions(-)
> 
> diff --git a/tests/TestSuite_kni.py b/tests/TestSuite_kni.py
> index acd6230..93203f5 100644
> --- a/tests/TestSuite_kni.py
> +++ b/tests/TestSuite_kni.py
> @@ -52,8 +52,8 @@ packet_sizes_routing = [64, 256]
> 
>  ports_cores_template =
> '\(P([0123]),(C\{\d.\d.\d\}),(C\{\d.\d.\d\}),(C\{\d.\d.\d\}),?(C\{\d.\d.\d
> \})?\),?'	
> 
> -default_1_port_cores_config = '(P0,C{1.0.0},C{1.1.0},C{1.0.1})'
> -default_2_port_cores_config =
> '(P0,C{1.0.0},C{1.1.0},C{1.0.1}),(P1,C{1.2.0},C{1.3.0},C{1.2.1})'
> +default_1_port_cores_config = '(P0,C{1.0.0},C{1.1.0},C{1.2.0})'
> +default_2_port_cores_config =
> '(P0,C{1.0.0},C{1.1.0},C{1.2.0}),(P1,C{1.3.0},C{1.4.0},C{1.5.0})'
> 
>  stress_test_iterations = 50
>  stress_test_random_iterations = 50
> @@ -258,6 +258,9 @@ class TestKni(TestCase):
>          self.verify('no brctl' not in out,
>                      "The linux tool brctl is needed to run this test
> suite")
> 
> +        self.dut.send_expect("sed -i -e
> 's/KNI_KMOD_ETHTOOL=n$/KNI_KMOD_ETHTOOL=y/' config/common_base", "# ", 30)
> +        self.dut.build_install_dpdk(self.target)
> +
>          out = self.dut.send_expect("make -C ./examples/kni/", "# ", 5)
>          self.verify('Error' not in out, "Compilation failed")
> 
> @@ -1207,5 +1210,9 @@ class TestKni(TestCase):
>          """
>          Run after each test suite.
>          """
> +
> +        self.dut.send_expect("sed -i -e
> 's/KNI_KMOD_ETHTOOL=y$/KNI_KMOD_ETHTOOL=n/' config/common_base", "# ", 30)
> +        self.dut.build_install_dpdk(self.target)
> +
>          self.dut.kill_all()
>          self.dut.send_expect("rmmod rte_kni", "# ", 10)
> --
> 1.9.1

Re: [dts] [PATCH] tests/kni: fix failures for kni

[Jianbo Liu]

Hi Marvin,

On 17 August 2017 at 20:40, Liu, Yong <yong.liu@intel.com> wrote:
> Hi Jianbo,
> Is there any particular reason for not chose hyper-threading cores?
> There're some performance cases are depending on hyper-threading cores.
> If hyper-threading is not enabled on the platform, those cases won't be able to run.
> Could you please clear the reason?
>

I think my patch only changed the function testing cases, not performance ones.
Besides, there is no hyper-threading for the arm64 server.

Thanks!
Jianbo

> Thanks,
> Marvin
>
>> -----Original Message-----
>> From: dts [mailto:dts-bounces@dpdk.org] On Behalf Of Jianbo Liu
>> Sent: Thursday, August 17, 2017 1:23 PM
>> To: dts@dpdk.org; herbert.guan@arm.com
>> Cc: Jianbo Liu <jianbo.liu@linaro.org>
>> Subject: [dts] [PATCH] tests/kni: fix failures for kni
>>
>> The tests need to enable KNI ethtool, and recomple DPDK.
>> Also bind to thread 0 of each core for function tests.
>>
>> Signed-off-by: Jianbo Liu <jianbo.liu@linaro.org>
>> ---
>>  tests/TestSuite_kni.py | 11 +++++++++--
>>  1 file changed, 9 insertions(+), 2 deletions(-)
>>
>> diff --git a/tests/TestSuite_kni.py b/tests/TestSuite_kni.py
>> index acd6230..93203f5 100644
>> --- a/tests/TestSuite_kni.py
>> +++ b/tests/TestSuite_kni.py
>> @@ -52,8 +52,8 @@ packet_sizes_routing = [64, 256]
>>
>>  ports_cores_template =
>> '\(P([0123]),(C\{\d.\d.\d\}),(C\{\d.\d.\d\}),(C\{\d.\d.\d\}),?(C\{\d.\d.\d
>> \})?\),?'
>>
>> -default_1_port_cores_config = '(P0,C{1.0.0},C{1.1.0},C{1.0.1})'
>> -default_2_port_cores_config =
>> '(P0,C{1.0.0},C{1.1.0},C{1.0.1}),(P1,C{1.2.0},C{1.3.0},C{1.2.1})'
>> +default_1_port_cores_config = '(P0,C{1.0.0},C{1.1.0},C{1.2.0})'
>> +default_2_port_cores_config =
>> '(P0,C{1.0.0},C{1.1.0},C{1.2.0}),(P1,C{1.3.0},C{1.4.0},C{1.5.0})'
>>
>>  stress_test_iterations = 50
>>  stress_test_random_iterations = 50
>> @@ -258,6 +258,9 @@ class TestKni(TestCase):
>>          self.verify('no brctl' not in out,
>>                      "The linux tool brctl is needed to run this test
>> suite")
>>
>> +        self.dut.send_expect("sed -i -e
>> 's/KNI_KMOD_ETHTOOL=n$/KNI_KMOD_ETHTOOL=y/' config/common_base", "# ", 30)
>> +        self.dut.build_install_dpdk(self.target)
>> +
>>          out = self.dut.send_expect("make -C ./examples/kni/", "# ", 5)
>>          self.verify('Error' not in out, "Compilation failed")
>>
>> @@ -1207,5 +1210,9 @@ class TestKni(TestCase):
>>          """
>>          Run after each test suite.
>>          """
>> +
>> +        self.dut.send_expect("sed -i -e
>> 's/KNI_KMOD_ETHTOOL=y$/KNI_KMOD_ETHTOOL=n/' config/common_base", "# ", 30)
>> +        self.dut.build_install_dpdk(self.target)
>> +
>>          self.dut.kill_all()
>>          self.dut.send_expect("rmmod rte_kni", "# ", 10)
>> --
>> 1.9.1
>

Re: [dts] [PATCH] tests/kni: fix failures for kni

[Liu, Yong]

Hi Jianbo,

> -----Original Message-----
> From: Jianbo Liu [mailto:jianbo.liu@linaro.org]
> Sent: Friday, August 18, 2017 9:56 AM
> To: Liu, Yong <yong.liu@intel.com>
> Cc: dts@dpdk.org; herbert.guan@arm.com
> Subject: Re: [dts] [PATCH] tests/kni: fix failures for kni
> 
> Hi Marvin,
> 
> On 17 August 2017 at 20:40, Liu, Yong <yong.liu@intel.com> wrote:
> > Hi Jianbo,
> > Is there any particular reason for not chose hyper-threading cores?
> > There're some performance cases are depending on hyper-threading cores.
> > If hyper-threading is not enabled on the platform, those cases won't be
> able to run.
> > Could you please clear the reason?
> >
> 
> I think my patch only changed the function testing cases, not performance
> ones.
> Besides, there is no hyper-threading for the arm64 server.
> 

There're still some cases which required for hyper-threading cores remaining in DTS.
I'm looking up for one method to handle those cases. Anyway, thanks a lot and patch merged.

> Thanks!
> Jianbo
> 
> > Thanks,
> > Marvin
> >