[PATCH 0/2] kunit: qemu_configs: Add or1k / openrisc configuration

[PATCH 0/2] kunit: qemu_configs: Add or1k / openrisc configuration

[Thomas Weißschuh]

Add a basic configuration to run kunit tests on or1k / openrisc.

Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
---
Thomas Weißschuh (2):
      gen_compile_commands: Ignore libgcc.a
      kunit: qemu_configs: Add or1k / openrisc configuration

 scripts/clang-tools/gen_compile_commands.py |  2 ++
 tools/testing/kunit/qemu_configs/or1k.py    | 18 ++++++++++++++++++
 2 files changed, 20 insertions(+)
---
base-commit: 254f49634ee16a731174d2ae34bc50bd5f45e731
change-id: 20260427-kunit-or1k-00035c13d70e

Best regards,
--  
Thomas Weißschuh <linux@weissschuh.net>

[PATCH 1/2] gen_compile_commands: Ignore libgcc.a

[Thomas Weißschuh]

Some architectures link libgcc.a from the toolchain into the kernel.
gen_compile_commands trie to read the kbuild .cmd files of its
constituent object files, which are not available.

Flat out ignore libgcc.a, as it is not built as part of the kernel
anyways.

Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
---
 scripts/clang-tools/gen_compile_commands.py | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/scripts/clang-tools/gen_compile_commands.py b/scripts/clang-tools/gen_compile_commands.py
index 96e6e46ad1a7..8d14b81efd73 100755
--- a/scripts/clang-tools/gen_compile_commands.py
+++ b/scripts/clang-tools/gen_compile_commands.py
@@ -201,6 +201,8 @@ def main():
         # Modules are listed in modules.order.
         if os.path.isdir(path):
             cmdfiles = cmdfiles_in_dir(path)
+        elif os.path.basename(path) == 'libgcc.a':
+            cmdfiles = []
         elif path.endswith('.a'):
             cmdfiles = cmdfiles_for_a(path, ar)
         elif path.endswith('modules.order'):

-- 
2.53.0

Re: [PATCH 1/2] gen_compile_commands: Ignore libgcc.a

[Nathan Chancellor]

On Mon, 27 Apr 2026 18:04:56 +0200, Thomas Weißschuh <linux@weissschuh.net> wrote:
> Some architectures link libgcc.a from the toolchain into the kernel.
> gen_compile_commands trie to read the kbuild .cmd files of its
> constituent object files, which are not available.
> 
> Flat out ignore libgcc.a, as it is not built as part of the kernel
> anyways.
> 
> [...]

Acked-by: Nathan Chancellor <nathan@kernel.org>

-- 
Cheers,
Nathan

[PATCH 2/2] kunit: qemu_configs: Add or1k / openrisc configuration

[Thomas Weißschuh]

Add a basic configuration to run kunit tests on or1k / openrisc.

Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
---
 tools/testing/kunit/qemu_configs/or1k.py | 18 ++++++++++++++++++
 1 file changed, 18 insertions(+)

diff --git a/tools/testing/kunit/qemu_configs/or1k.py b/tools/testing/kunit/qemu_configs/or1k.py
new file mode 100644
index 000000000000..dfbbad0f9076
--- /dev/null
+++ b/tools/testing/kunit/qemu_configs/or1k.py
@@ -0,0 +1,18 @@
+# SPDX-License-Identifier: GPL-2.0-only
+from ..qemu_config import QemuArchParams
+
+QEMU_ARCH = QemuArchParams(linux_arch='openrisc',
+			   kconfig='''
+CONFIG_SERIAL_8250=y
+CONFIG_SERIAL_8250_CONSOLE=y
+CONFIG_SERIAL_OF_PLATFORM=y
+CONFIG_POWER_RESET=y
+CONFIG_POWER_RESET_SYSCON=y
+''',
+			   qemu_arch='or1k',
+			   kernel_path='vmlinux',
+			   kernel_command_line='console=ttyS0',
+			   extra_qemu_params=[
+					    '-machine', 'virt',
+                        '-m', '512',
+			  ])

-- 
2.53.0

Re: [PATCH 2/2] kunit: qemu_configs: Add or1k / openrisc configuration

[David Gow]

Le 28/04/2026 à 12:04 AM, 'Thomas Weißschuh' via KUnit Development a écrit :
> Add a basic configuration to run kunit tests on or1k / openrisc.
> 
> Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
> ---

Thanks very much. This works well here (assuming 
CONFIG_KUNIT_FAULT_TESTS=n, and either a long timeout, or slow tests 
skipped, as qemu-or1k is not fast on my machine).

Reviewed-by: David Gow <david@davidgow.net>

Do we want to split the series and take this patch via kselftest/kunit, 
and the generate_compile_commands one separately, or just pick both up 
in the same tree?

Cheers,
-- David

>   tools/testing/kunit/qemu_configs/or1k.py | 18 ++++++++++++++++++
>   1 file changed, 18 insertions(+)
> 
> diff --git a/tools/testing/kunit/qemu_configs/or1k.py b/tools/testing/kunit/qemu_configs/or1k.py
> new file mode 100644
> index 000000000000..dfbbad0f9076
> --- /dev/null
> +++ b/tools/testing/kunit/qemu_configs/or1k.py
> @@ -0,0 +1,18 @@
> +# SPDX-License-Identifier: GPL-2.0-only
> +from ..qemu_config import QemuArchParams
> +
> +QEMU_ARCH = QemuArchParams(linux_arch='openrisc',
> +			   kconfig='''
> +CONFIG_SERIAL_8250=y
> +CONFIG_SERIAL_8250_CONSOLE=y
> +CONFIG_SERIAL_OF_PLATFORM=y
> +CONFIG_POWER_RESET=y
> +CONFIG_POWER_RESET_SYSCON=y
> +''',
> +			   qemu_arch='or1k',
> +			   kernel_path='vmlinux',
> +			   kernel_command_line='console=ttyS0',
> +			   extra_qemu_params=[
> +					    '-machine', 'virt',
> +                        '-m', '512',
> +			  ])
> 

Re: [PATCH 2/2] kunit: qemu_configs: Add or1k / openrisc configuration

[Thomas Weißschuh]

On 2026-04-29 18:10:58+0800, David Gow wrote:
> Le 28/04/2026 à 12:04 AM, 'Thomas Weißschuh' via KUnit Development a écrit :
> > Add a basic configuration to run kunit tests on or1k / openrisc.
> > 
> > Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
> > ---
> 
> Thanks very much. This works well here (assuming CONFIG_KUNIT_FAULT_TESTS=n,
> and either a long timeout, or slow tests skipped, as qemu-or1k is not fast
> on my machine).
> 
> Reviewed-by: David Gow <david@davidgow.net>

Thanks!

> Do we want to split the series and take this patch via kselftest/kunit, and
> the generate_compile_commands one separately, or just pick both up in the
> same tree?

They should go through the same tree as otherwise kunit.py will fail to
build the or1k architecture, as it always tries to generate
compile_commands.json.

(...)

Re: [PATCH 2/2] kunit: qemu_configs: Add or1k / openrisc configuration

[Nathan Chancellor]

On 2026-04-29 13:14:42+02:00, Thomas Weißschuh wrote:
> On 2026-04-29 18:10:58+0800, David Gow wrote:
> > Do we want to split the series and take this patch via kselftest/kunit, and
> > the generate_compile_commands one separately, or just pick both up in the
> > same tree?
> 
> They should go through the same tree as otherwise kunit.py will fail to
> build the or1k architecture, as it always tries to generate
> compile_commands.json.

I don't anticipate any conflicts from the gen_compile_commands.py
change, so I think you can just take both through whatever tree you
prefer. I'll provide my Ack shortly.