Re: [PATCH] tests/acceptance: Test PMON with Loongson-3A1000 CPU

["Philippe Mathieu-Daudé" <philmd@redhat.com>]

On 1/12/21 3:05 PM, Wainer dos Santos Moschetta wrote:
> Hi,
> 
> On 1/11/21 11:07 PM, Jiaxun Yang wrote:
>> Test booting of PMON bootloader on loongson3-virt platform.
>>
>> $ (venv) AVOCADO_ALLOW_UNTRUSTED_CODE=1 \
>>      avocado --show=app,console \
>>        run -t machine:loongson3-virt tests/acceptance
>> Fetching asset from
>> tests/acceptance/machine_mips_loongson3v.py:MipsLoongson3v.test_pmon_serial_console
>>
>> JOB ID     : 8e202b3727847c9104d0d3d6546ed225d35f6706
>> JOB LOG    :
>> /home/flygoat/avocado/job-results/job-2021-01-12T10.02-8e202b3/job.log
>>   (1/1)
>> tests/acceptance/machine_mips_loongson3v.py:MipsLoongson3v.test_pmon_serial_console: 
>> console: PMON2000 MIPS Initializing. Standby...
>> console: 00000000Jump to 9fc
> <snip>
>> Signed-off-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
>> ---
>>   MAINTAINERS                                 |  1 +
>>   tests/acceptance/machine_mips_loongson3v.py | 39 +++++++++++++++++++++
>>   2 files changed, 40 insertions(+)
>>   create mode 100644 tests/acceptance/machine_mips_loongson3v.py
> 
> Allow me to use this new test as an example to start a discussion about
> the organization of the acceptance files.
> 
> The mips64le tests currently are:
> 
> $ ./venv/bin/avocado list -t arch:mips64el acceptance/
> 
> INSTRUMENTED
> acceptance/boot_linux_console.py:BootLinuxConsole.test_mips64el_malta
> INSTRUMENTED
> acceptance/boot_linux_console.py:BootLinuxConsole.test_mips64el_fuloong2e
> INSTRUMENTED
> acceptance/boot_linux_console.py:BootLinuxConsole.test_mips64el_malta_5KEc_cpio
> 
> INSTRUMENTED
> acceptance/linux_ssh_mips_malta.py:LinuxSSH.test_mips_malta64el_kernel3_2_0
> INSTRUMENTED
> acceptance/machine_mips_loongson3v.py:MipsLoongson3v.test_pmon_serial_console
> 
> INSTRUMENTED
> acceptance/machine_mips_malta.py:MaltaMachineFramebuffer.test_mips_malta_i6400_framebuffer_logo_1core
> 
> INSTRUMENTED
> acceptance/machine_mips_malta.py:MaltaMachineFramebuffer.test_mips_malta_i6400_framebuffer_logo_7cores
> 
> INSTRUMENTED
> acceptance/machine_mips_malta.py:MaltaMachineFramebuffer.test_mips_malta_i6400_framebuffer_logo_8cores
> 
> INSTRUMENTED
> acceptance/replay_kernel.py:ReplayKernelNormal.test_mips64el_malta
> INSTRUMENTED
> acceptance/replay_kernel.py:ReplayKernelSlow.test_mips64el_malta_5KEc_cpio
> 
> Most of them are simple "boot linux (or firmware) and check the console"
> tests. The replay_kernel.py contain tests for a given feature and happen
> to be testing on mips64el as well. So on tests/acceptance directory
> we've got boot tests spread across files and mixed with "generic"
> feature tests.
> 
> I think we should find a home for those boot tests. Maybe throw them all
> in a sub-directory called "boot_tests", or in arch-oriented directories
> (boot_tests/mips64el, boot_tests/x86_64, ...) which would make easier to
> reference them in the MAINTAINERS file.
> 
> Any thought?

[thread hijack...]

A tests might be multi-arch. If you think it is easier to have the
tests sorted in subfolders, go ahead :) Where is a test is not a
problem. The problem we have to solve is how to relate a test with
its maintainers / developers interested in it.

We once said what really matters are Avocado tags. This might be an
underused feature of Avocado. Maybe what we need is a script to
relate test tags with MAINTAINERS?

Let's say we use the 'A' tag for that:

-- >8 --
diff --git a/MAINTAINERS b/MAINTAINERS
index cb0656aec3d..a484d429d78 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -646,6 +646,7 @@ M: Anup Patel <anup.patel@wdc.com>
 M: Alistair Francis <Alistair.Francis@wdc.com>
 L: qemu-riscv@nongnu.org
 S: Maintained
+A: device:goldfish_rtc
 F: hw/rtc/goldfish_rtc.c
 F: include/hw/rtc/goldfish_rtc.h

@@ -1160,6 +1161,11 @@ Loongson-3 virtual platforms
 M: Huacai Chen <chenhuacai@kernel.org>
 R: Jiaxun Yang <jiaxun.yang@flygoat.com>
 S: Maintained
+A: machine:loongson3-virt
+A: cpu:Loongson-3A1000
+A: device:liointc
+A: device:goldfish_rtc
 F: hw/intc/loongson_liointc.c
 F: hw/mips/loongson3_bootp.c
 F: hw/mips/loongson3_bootp.h
---

If this test fails, it should list the maintainers of the
loongson3-virt machine and the goldfish RTC.

Maintainers add their tags of interest, and can Ack when a
developer add a tag to their MAINTAINERS entry.

Thought? :)