diff mbox series

selftest/vDSO: Fix cross build for the random tests

Message ID 20240830-vdso-chacha-build-v1-1-78f93d2a142f@kernel.org (mailing list archive)
State New
Headers show
Series selftest/vDSO: Fix cross build for the random tests | expand

Commit Message

Mark Brown Aug. 30, 2024, 2:06 p.m. UTC
Unlike the check for the standalone x86 test the check for building the
vDSO getrandom and chacaha tests looks at the architecture for the host
rather than the architecture for the target when deciding if they should
be built. Since the chacha test includes some assembler code this means
that cross building with x86 as either the target or host is broken. Use
a check for ARCH instead.

Fixes: 4920a2590e91 ("selftests/vDSO: add tests for vgetrandom")
Signed-off-by: Mark Brown <broonie@kernel.org>
---
The x86_64 build is still broken for me because nothing installs
tools/arch/x86_64/vdso/vgetrandom-chacha.S (I beleive it's supposed to
be copied from ./arch/x86/entry/vdso/vgetrandom-chacha.S but I don't see
how?) but this at least fixes all the other architectures.
---
 tools/testing/selftests/vDSO/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)


---
base-commit: 985bf40edf4343dcb04c33f58b40b4a85c1776d4
change-id: 20240830-vdso-chacha-build-8d3789bf695c

Best regards,

Comments

Jason A. Donenfeld Aug. 30, 2024, 3:30 p.m. UTC | #1
On Fri, Aug 30, 2024 at 03:06:35PM +0100, Mark Brown wrote:
> Unlike the check for the standalone x86 test the check for building the
> vDSO getrandom and chacaha tests looks at the architecture for the host
> rather than the architecture for the target when deciding if they should
> be built. Since the chacha test includes some assembler code this means
> that cross building with x86 as either the target or host is broken. Use
> a check for ARCH instead.
> 
> Fixes: 4920a2590e91 ("selftests/vDSO: add tests for vgetrandom")
> Signed-off-by: Mark Brown <broonie@kernel.org>

Thanks for the patch. Seems reasonable; I'll queue it up.

> ---
> The x86_64 build is still broken for me because nothing installs
> tools/arch/x86_64/vdso/vgetrandom-chacha.S (I beleive it's supposed to
> be copied from ./arch/x86/entry/vdso/vgetrandom-chacha.S but I don't see
> how?) but this at least fixes all the other architectures.

There should be a symlink installed for that. Are you using this tree?

https://git.kernel.org/pub/scm/linux/kernel/git/crng/random.git

That's where all these fixups are going for 6.12. (And yea, there are a
lot.)
Jason A. Donenfeld Aug. 30, 2024, 3:34 p.m. UTC | #2
On Fri, Aug 30, 2024 at 03:06:35PM +0100, Mark Brown wrote:
> Unlike the check for the standalone x86 test the check for building the
> vDSO getrandom and chacaha tests looks at the architecture for the host
> rather than the architecture for the target when deciding if they should
> be built. Since the chacha test includes some assembler code this means
> that cross building with x86 as either the target or host is broken. Use
> a check for ARCH instead.
> 
> Fixes: 4920a2590e91 ("selftests/vDSO: add tests for vgetrandom")
> Signed-off-by: Mark Brown <broonie@kernel.org>
> ---
> The x86_64 build is still broken for me because nothing installs
> tools/arch/x86_64/vdso/vgetrandom-chacha.S (I beleive it's supposed to
> be copied from ./arch/x86/entry/vdso/vgetrandom-chacha.S but I don't see
> how?) but this at least fixes all the other architectures.
> ---
>  tools/testing/selftests/vDSO/Makefile | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/tools/testing/selftests/vDSO/Makefile b/tools/testing/selftests/vDSO/Makefile
> index e21e78aae24d..7fb59310718c 100644
> --- a/tools/testing/selftests/vDSO/Makefile
> +++ b/tools/testing/selftests/vDSO/Makefile
> @@ -10,7 +10,7 @@ ifeq ($(ARCH),$(filter $(ARCH),x86 x86_64))
>  TEST_GEN_PROGS += vdso_standalone_test_x86
>  endif
>  TEST_GEN_PROGS += vdso_test_correctness
> -ifeq ($(uname_M),x86_64)
> +ifeq ($(ARCH),$(filter $(ARCH),x86_64))

Actually... this doesn't work. Because:

ARCH ?= $(shell echo $(uname_M) | sed -e s/i.86/x86/ -e s/x86_64/x86/)

The x86_64 gets turned into an x86. It's a little trickier when
considering subarch. I'll send a followup here and not queue this.
Mark Brown Aug. 30, 2024, 3:42 p.m. UTC | #3
On Fri, Aug 30, 2024 at 05:30:07PM +0200, Jason A. Donenfeld wrote:
> On Fri, Aug 30, 2024 at 03:06:35PM +0100, Mark Brown wrote:

> > The x86_64 build is still broken for me because nothing installs
> > tools/arch/x86_64/vdso/vgetrandom-chacha.S (I beleive it's supposed to
> > be copied from ./arch/x86/entry/vdso/vgetrandom-chacha.S but I don't see
> > how?) but this at least fixes all the other architectures.

> There should be a symlink installed for that. Are you using this tree?

> https://git.kernel.org/pub/scm/linux/kernel/git/crng/random.git

> That's where all these fixups are going for 6.12. (And yea, there are a
> lot.)

I was using -next, hopefully it's getting merged in there (I see the
master branch is included).  This is also breaking in for example
kernelci:

https://storage.kernelci.org/next/master/next-20240830/x86_64/x86_64_defconfig/gcc-12/logs/kselftest.log

in the same way.
Mark Brown Aug. 30, 2024, 3:43 p.m. UTC | #4
On Fri, Aug 30, 2024 at 05:34:13PM +0200, Jason A. Donenfeld wrote:

> > +ifeq ($(ARCH),$(filter $(ARCH),x86_64))

> Actually... this doesn't work. Because:

> ARCH ?= $(shell echo $(uname_M) | sed -e s/i.86/x86/ -e s/x86_64/x86/)

> The x86_64 gets turned into an x86. It's a little trickier when
> considering subarch. I'll send a followup here and not queue this.

Ah, right - I originally had copied the line above with both x86 and
x86_64 when I tested and did a last minute edit during final review
after I'd switched away.
Jason A. Donenfeld Aug. 30, 2024, 3:58 p.m. UTC | #5
On Fri, Aug 30, 2024 at 5:42 PM Mark Brown <broonie@kernel.org> wrote:
>
> On Fri, Aug 30, 2024 at 05:30:07PM +0200, Jason A. Donenfeld wrote:
> > On Fri, Aug 30, 2024 at 03:06:35PM +0100, Mark Brown wrote:
>
> > > The x86_64 build is still broken for me because nothing installs
> > > tools/arch/x86_64/vdso/vgetrandom-chacha.S (I beleive it's supposed to
> > > be copied from ./arch/x86/entry/vdso/vgetrandom-chacha.S but I don't see
> > > how?) but this at least fixes all the other architectures.
>
> > There should be a symlink installed for that. Are you using this tree?
>
> > https://git.kernel.org/pub/scm/linux/kernel/git/crng/random.git
>
> > That's where all these fixups are going for 6.12. (And yea, there are a
> > lot.)
>
> I was using -next, hopefully it's getting merged in there (I see the
> master branch is included).  This is also breaking in for example
> kernelci:
>
> https://storage.kernelci.org/next/master/next-20240830/x86_64/x86_64_defconfig/gcc-12/logs/kselftest.log
>
> in the same way.

Ahh yea it looks like that is running into the same thing as your v1,
which the v2 I posted fixes via SRCARCH:

make[4]: *** No rule to make target
'/tmp/kci/linux/tools/testing/selftests/../../../tools/arch/x86_64/vdso/vgetrandom-chacha.S',
needed by '/tmp/kci/linux/build/kselftest/vDSO/vdso_test_chacha'.
Stop.

Here it's looking for arch/x86_64/ instead of arch/x86/, because
$(ARCH) was used to interpolate the path instead of $(SRCARCH).
diff mbox series

Patch

diff --git a/tools/testing/selftests/vDSO/Makefile b/tools/testing/selftests/vDSO/Makefile
index e21e78aae24d..7fb59310718c 100644
--- a/tools/testing/selftests/vDSO/Makefile
+++ b/tools/testing/selftests/vDSO/Makefile
@@ -10,7 +10,7 @@  ifeq ($(ARCH),$(filter $(ARCH),x86 x86_64))
 TEST_GEN_PROGS += vdso_standalone_test_x86
 endif
 TEST_GEN_PROGS += vdso_test_correctness
-ifeq ($(uname_M),x86_64)
+ifeq ($(ARCH),$(filter $(ARCH),x86_64))
 TEST_GEN_PROGS += vdso_test_getrandom
 TEST_GEN_PROGS += vdso_test_chacha
 endif