[RFC 0/2] Experimental kunit test for signal context handling

Benjamin Berg benjamin at sipsolutions.net
Fri Jun 27 02:28:58 PDT 2025


Hi,

On Fri, 2025-06-27 at 11:22 +0200, Thomas Weißschuh wrote:
> On Fri, Jun 27, 2025 at 10:48:30AM +0200, Benjamin Berg wrote:
> > Hi,
> > 
> > On Fri, 2025-06-27 at 10:43 +0200, Thomas Weißschuh wrote:
> > > On Fri, Jun 27, 2025 at 09:29:12AM +0200, Benjamin Berg wrote:
> > > > > > [SNIP]
> > > 
> > > Or we gate the testcases behind BROKEN/EXPORT on UML in the beginning.
> > > 
> > > Does this sound reasonable?
> > 
> > Not sure how to gate them, but I do like the idea of being able to
> > enable them using only minor modifications or even a Kconfig option to
> > see what happens on UM.
> 
> With my WIP x86 enablement commit as example [0].
> 
> 
> Possibility 1, I *drop* these two hunks, you can readd them:
> 
> diff --git a/arch/x86/um/Kconfig b/arch/x86/um/Kconfig
> index 986045d5e63854..f2aeaf1e8e6d9a 100644
> --- a/arch/x86/um/Kconfig
> +++ b/arch/x86/um/Kconfig
> @@ -3,6 +3,7 @@
>  menu "Host processor type and features"
>  
>  source "arch/x86/Kconfig.cpu"
> +source "arch/x86/tests/Kconfig"
>  
>  endmenu
>  
> diff --git a/arch/x86/um/Makefile b/arch/x86/um/Makefile
> index b42c31cd2390cb..e71bb244afac9d 100644
> --- a/arch/x86/um/Makefile
> +++ b/arch/x86/um/Makefile
> @@ -13,7 +13,7 @@ obj-y = bugs_$(BITS).o delay.o fault.o \
>  	ptrace.o ptrace_$(BITS).o ptrace_user.o setjmp_$(BITS).o signal.o \
>  	stub_segv.o \
>  	sys_call_table_$(BITS).o sysrq_$(BITS).o tls_$(BITS).o \
> -	mem_$(BITS).o subarch.o os-Linux/
> +	mem_$(BITS).o subarch.o os-Linux/ ../tests/
>  
>  ifeq ($(CONFIG_X86_32),y)
> 
> 
> Possibility 2, the kconfig looks like this:
> 
> config X86_UAPI_TEST
>         tristate "x86 UAPI tests" if !KUNIT_ALL_TESTS
>         depends on KUNIT_UAPI
> 	depends on !UML || BROKEN
>         default KUNIT_ALL_TESTS
> 
> You enable CONFIG_BROKEN and can run the tests.

I think I prefer option 2. Not a strong opinion, but it shows that they
should be working just fine UML.

Benjamin

> 
> Of course as soon as all tests work on UML, this all goes away and they
> stay enabled.
> 
> [0] https://git.kernel.org/pub/scm/linux/kernel/git/thomas.weissschuh/linux.git/commit/?h=kunit-kselftests-integration&id=14bf93f09eaa328a18e9190f46fa4ee697378617
> 




More information about the linux-um mailing list