v6.6.33-rc1 build failure

v6.6.33-rc1 build failure

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 1100 bytes --]

I'm not seeing a test mail for v6.6.33-rc1 but it's in the stable-rc git
and I'm seeing build failures in the KVM selftests for arm64 with it:

/usr/bin/ld: /build/stage/build-work/kselftest/kvm/aarch64/vgic_init.o: in funct
ion `test_v2_uaccess_cpuif_no_vcpus':
/build/stage/linux/tools/testing/selftests/kvm/aarch64/vgic_init.c:388:(.text+0x
1234): undefined reference to `FIELD_PREP'
/usr/bin/ld: /build/stage/linux/tools/testing/selftests/kvm/aarch64/vgic_init.c:
388:(.text+0x1244): undefined reference to `FIELD_PREP'
/usr/bin/ld: /build/stage/linux/tools/testing/selftests/kvm/aarch64/vgic_init.c:
393:(.text+0x12a4): undefined reference to `FIELD_PREP'
/usr/bin/ld: /build/stage/linux/tools/testing/selftests/kvm/aarch64/vgic_init.c:
393:(.text+0x12b4): undefined reference to `FIELD_PREP'
/usr/bin/ld: /build/stage/linux/tools/testing/selftests/kvm/aarch64/vgic_init.c:
398:(.text+0x1308): undefined reference to `FIELD_PREP'

due to 12237178b318fb3 ("KVM: selftests: Add test for uaccesses to
non-existent vgic-v2 CPUIF") which was backported from
160933e330f4c5a13931d725a4d952a4b9aefa71.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: v6.6.33-rc1 build failure

[Oliver Upton]

Hey,

Thanks for flagging this broonie.

On Tue, Jun 04, 2024 at 02:35:20PM +0100, Mark Brown wrote:
> I'm not seeing a test mail for v6.6.33-rc1 but it's in the stable-rc git
> and I'm seeing build failures in the KVM selftests for arm64 with it:
> 
> /usr/bin/ld: /build/stage/build-work/kselftest/kvm/aarch64/vgic_init.o: in funct
> ion `test_v2_uaccess_cpuif_no_vcpus':
> /build/stage/linux/tools/testing/selftests/kvm/aarch64/vgic_init.c:388:(.text+0x
> 1234): undefined reference to `FIELD_PREP'
> /usr/bin/ld: /build/stage/linux/tools/testing/selftests/kvm/aarch64/vgic_init.c:
> 388:(.text+0x1244): undefined reference to `FIELD_PREP'
> /usr/bin/ld: /build/stage/linux/tools/testing/selftests/kvm/aarch64/vgic_init.c:
> 393:(.text+0x12a4): undefined reference to `FIELD_PREP'
> /usr/bin/ld: /build/stage/linux/tools/testing/selftests/kvm/aarch64/vgic_init.c:
> 393:(.text+0x12b4): undefined reference to `FIELD_PREP'
> /usr/bin/ld: /build/stage/linux/tools/testing/selftests/kvm/aarch64/vgic_init.c:
> 398:(.text+0x1308): undefined reference to `FIELD_PREP'
> 
> due to 12237178b318fb3 ("KVM: selftests: Add test for uaccesses to
> non-existent vgic-v2 CPUIF") which was backported from
> 160933e330f4c5a13931d725a4d952a4b9aefa71.

Yeah, so this is likely because commit 0359c946b131 ("tools headers arm64:
Update sysreg.h with kernel sources") upstream got this test to indirectly
include bitfield.h. We should *not* backport the patch that fixes it,
though.

Let's either squash in the following, or just drop the offending patch
altogether.

-- 
Thanks,
Oliver

From 5a957ab6d80f4fcb4b1f2bcbd5b999fde003bffd Mon Sep 17 00:00:00 2001
From: Oliver Upton <oliver.upton@linux.dev>
Date: Tue, 4 Jun 2024 18:36:29 +0000
Subject: [PATCH] fixup! KVM: selftests: Add test for uaccesses to non-existent
 vgic-v2 CPUIF

commit 0359c946b131 ("tools headers arm64: Update sysreg.h with kernel
sources") upstream indirectly included bitfield.h, which is a dependency
of this patch. Work around the situation locally by directly including
bitfield.h in the test.

Signed-off-by: Oliver Upton <oliver.upton@linux.dev>
---
 tools/testing/selftests/kvm/aarch64/vgic_init.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/tools/testing/selftests/kvm/aarch64/vgic_init.c b/tools/testing/selftests/kvm/aarch64/vgic_init.c
index ca917c71ff60..4ac4d3ea976e 100644
--- a/tools/testing/selftests/kvm/aarch64/vgic_init.c
+++ b/tools/testing/selftests/kvm/aarch64/vgic_init.c
@@ -6,6 +6,7 @@
  */
 #define _GNU_SOURCE
 #include <linux/kernel.h>
+#include <linux/bitfield.h>
 #include <sys/syscall.h>
 #include <asm/kvm.h>
 #include <asm/kvm_para.h>
-- 
2.45.1.467.gbab1589fc0-goog

Re: v6.6.33-rc1 build failure

[Greg Kroah-Hartman]

On Tue, Jun 04, 2024 at 11:41:29AM -0700, Oliver Upton wrote:
> Hey,
> 
> Thanks for flagging this broonie.
> 
> On Tue, Jun 04, 2024 at 02:35:20PM +0100, Mark Brown wrote:
> > I'm not seeing a test mail for v6.6.33-rc1 but it's in the stable-rc git
> > and I'm seeing build failures in the KVM selftests for arm64 with it:
> > 
> > /usr/bin/ld: /build/stage/build-work/kselftest/kvm/aarch64/vgic_init.o: in funct
> > ion `test_v2_uaccess_cpuif_no_vcpus':
> > /build/stage/linux/tools/testing/selftests/kvm/aarch64/vgic_init.c:388:(.text+0x
> > 1234): undefined reference to `FIELD_PREP'
> > /usr/bin/ld: /build/stage/linux/tools/testing/selftests/kvm/aarch64/vgic_init.c:
> > 388:(.text+0x1244): undefined reference to `FIELD_PREP'
> > /usr/bin/ld: /build/stage/linux/tools/testing/selftests/kvm/aarch64/vgic_init.c:
> > 393:(.text+0x12a4): undefined reference to `FIELD_PREP'
> > /usr/bin/ld: /build/stage/linux/tools/testing/selftests/kvm/aarch64/vgic_init.c:
> > 393:(.text+0x12b4): undefined reference to `FIELD_PREP'
> > /usr/bin/ld: /build/stage/linux/tools/testing/selftests/kvm/aarch64/vgic_init.c:
> > 398:(.text+0x1308): undefined reference to `FIELD_PREP'
> > 
> > due to 12237178b318fb3 ("KVM: selftests: Add test for uaccesses to
> > non-existent vgic-v2 CPUIF") which was backported from
> > 160933e330f4c5a13931d725a4d952a4b9aefa71.
> 
> Yeah, so this is likely because commit 0359c946b131 ("tools headers arm64:
> Update sysreg.h with kernel sources") upstream got this test to indirectly
> include bitfield.h. We should *not* backport the patch that fixes it,
> though.
> 
> Let's either squash in the following, or just drop the offending patch
> altogether.
> 
> -- 
> Thanks,
> Oliver
> 
> >From 5a957ab6d80f4fcb4b1f2bcbd5b999fde003bffd Mon Sep 17 00:00:00 2001
> From: Oliver Upton <oliver.upton@linux.dev>
> Date: Tue, 4 Jun 2024 18:36:29 +0000
> Subject: [PATCH] fixup! KVM: selftests: Add test for uaccesses to non-existent
>  vgic-v2 CPUIF
> 
> commit 0359c946b131 ("tools headers arm64: Update sysreg.h with kernel
> sources") upstream indirectly included bitfield.h, which is a dependency
> of this patch. Work around the situation locally by directly including
> bitfield.h in the test.
> 
> Signed-off-by: Oliver Upton <oliver.upton@linux.dev>
> ---
>  tools/testing/selftests/kvm/aarch64/vgic_init.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/tools/testing/selftests/kvm/aarch64/vgic_init.c b/tools/testing/selftests/kvm/aarch64/vgic_init.c
> index ca917c71ff60..4ac4d3ea976e 100644
> --- a/tools/testing/selftests/kvm/aarch64/vgic_init.c
> +++ b/tools/testing/selftests/kvm/aarch64/vgic_init.c
> @@ -6,6 +6,7 @@
>   */
>  #define _GNU_SOURCE
>  #include <linux/kernel.h>
> +#include <linux/bitfield.h>
>  #include <sys/syscall.h>
>  #include <asm/kvm.h>
>  #include <asm/kvm_para.h>
> -- 
> 2.45.1.467.gbab1589fc0-goog
> 

Thanks, I've merged this in now.

greg k-h