[yocto-autobuilder2][PATCH 1/2] config.py: include qemux86-ptest-fast (32 bit) in a-quick

[yocto-autobuilder2][PATCH 1/2] config.py: include qemux86-ptest-fast (32 bit) in a-quick

[Alexander Kanavin]

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
---
 config.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/config.py b/config.py
index 3c67351..e638099 100644
--- a/config.py
+++ b/config.py
@@ -82,7 +82,7 @@ trigger_builders_wait_shared = [
 ]
 
 trigger_builders_wait_quick = trigger_builders_wait_shared + [
-    "oe-selftest", "qemux86-64-ptest-fast", "qemuarm64-ptest-fast"
+    "oe-selftest", "qemux86-64-ptest-fast", "qemuarm64-ptest-fast", "qemux86-ptest-fast"
 ]
 
 trigger_builders_wait_full = trigger_builders_wait_shared + [
@@ -117,7 +117,7 @@ builders_others = [
     "oe-selftest-arm",
     "metrics",
     "qemuriscv32", "qemuriscv64", "qemuriscv64-ptest", "qemuppc64",
-    "qemux86-ptest", "qemux86-ptest-fast",
+    "qemux86-ptest",
     "auh"
 ]
 
-- 
2.38.1

[yocto-autobuilder2][PATCH 2/2] config.py: include 32 bit qemux86-ptest (full) into a-full

[Alexander Kanavin]

The manually triggered test result is here:
https://autobuilder.yoctoproject.org/typhoon/#/builders/141/builds/1

The execution time is similar to the 64 bit x86 full ptest, as valgrind has been excluded
from the 32 bit set, so I believe we can afford to run this in the standard
matrix for master-next tests.

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
---
 config.py | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/config.py b/config.py
index e638099..711dd98 100644
--- a/config.py
+++ b/config.py
@@ -88,7 +88,7 @@ trigger_builders_wait_quick = trigger_builders_wait_shared + [
 trigger_builders_wait_full = trigger_builders_wait_shared + [
     "qemumips-alt", "edgerouter-alt", "qemuppc-alt", "qemux86-world-alt",
     "oe-selftest-ubuntu", "oe-selftest-debian", "oe-selftest-fedora", "oe-selftest-centos",
-    "qemux86-64-ptest", "qemux86-64-ltp", "qemuarm64-ptest", "qemuarm64-ltp", 
+    "qemux86-64-ptest", "qemux86-64-ltp", "qemuarm64-ptest", "qemuarm64-ltp", "qemux86-ptest",
     "meta-intel", "meta-arm", "meta-aws", "meta-agl-core", "meta-virt"
 ]
 
@@ -117,7 +117,6 @@ builders_others = [
     "oe-selftest-arm",
     "metrics",
     "qemuriscv32", "qemuriscv64", "qemuriscv64-ptest", "qemuppc64",
-    "qemux86-ptest",
     "auh"
 ]
 
-- 
2.38.1

Re: [yocto] [yocto-autobuilder2][PATCH 2/2] config.py: include 32 bit qemux86-ptest (full) into a-full

[Richard Purdie]

On Tue, 2022-12-13 at 16:40 +0100, Alexander Kanavin wrote:
> The manually triggered test result is here:
> https://autobuilder.yoctoproject.org/typhoon/#/builders/141/builds/1
> 
> The execution time is similar to the 64 bit x86 full ptest, as valgrind has been excluded
> from the 32 bit set, so I believe we can afford to run this in the standard
> matrix for master-next tests.
> 
> Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
> ---
>  config.py | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/config.py b/config.py
> index e638099..711dd98 100644
> --- a/config.py
> +++ b/config.py
> @@ -88,7 +88,7 @@ trigger_builders_wait_quick = trigger_builders_wait_shared + [
>  trigger_builders_wait_full = trigger_builders_wait_shared + [
>      "qemumips-alt", "edgerouter-alt", "qemuppc-alt", "qemux86-world-alt",
>      "oe-selftest-ubuntu", "oe-selftest-debian", "oe-selftest-fedora", "oe-selftest-centos",
> -    "qemux86-64-ptest", "qemux86-64-ltp", "qemuarm64-ptest", "qemuarm64-ltp", 
> +    "qemux86-64-ptest", "qemux86-64-ltp", "qemuarm64-ptest", "qemuarm64-ltp", "qemux86-ptest",
>      "meta-intel", "meta-arm", "meta-aws", "meta-agl-core", "meta-virt"
>  ]
>  
> @@ -117,7 +117,6 @@ builders_others = [
>      "oe-selftest-arm",
>      "metrics",
>      "qemuriscv32", "qemuriscv64", "qemuriscv64-ptest", "qemuppc64",
> -    "qemux86-ptest",
>      "auh"
>  ]

I'm not sure I agree. Yes, we can do it but I'm not sure we should.

This test doesn't seem to be one which is showing regressions, given we
have a low number of issues despite never having run it. The day to day
patch review struggles with the volume of tests we're running without
adding the load this test needs to a-full.

We need to focus the tests on those which catch issues creeping into
the project and I'm not convinced this is one of them. Running this
test once or twice a month is probably all we need?

Cheers,

Richard

Re: [yocto] [yocto-autobuilder2][PATCH 2/2] config.py: include 32 bit qemux86-ptest (full) into a-full

[Alexander Kanavin]

On Tue, 13 Dec 2022 at 23:56, Richard Purdie
<richard.purdie@linuxfoundation.org> wrote:

> This test doesn't seem to be one which is showing regressions, given we
> have a low number of issues despite never having run it. The day to day
> patch review struggles with the volume of tests we're running without
> adding the load this test needs to a-full.
>
> We need to focus the tests on those which catch issues creeping into
> the project and I'm not convinced this is one of them. Running this
> test once or twice a month is probably all we need?

I don't have a strong opinion here. But what about qemux86-ptest-fast,
should that be in a-quick only, a-quick+a-full, or neither?

Alex