next-20160222 build: 5 failures 9 warnings (next-20160222)

next-20160222 build: 5 failures 9 warnings (next-20160222)

[Mark Brown]

On Mon, Feb 22, 2016 at 08:36:46AM +0000, Build bot for Mark Brown wrote:

Today's -next fails to build on arm64 due to:

> 	arm64-allnoconfig
> ../arch/arm64/include/asm/kvm_host.h:338:2: error: implicit declaration of function 'kvm_call_hyp' [-Werror=implicit-function-declaration]
> 
> 	arm64-allmodconfig
> ../arch/arm64/include/asm/kvm_host.h:338:2: error: implicit declaration of function 'kvm_call_hyp' [-Werror=implicit-function-declaration]

> 	arm64-defconfig
> ../arch/arm64/include/asm/kvm_host.h:338:2: error: implicit declaration of function 'kvm_call_hyp' [-Werror=implicit-function-declaration]

triggered by the merge of the kvm-arm tree with -next today, it looks
like changes in that tree and arm64 aren't playing nicely with each
other (bisect came down to the merge and visual inspection of the
changes didn't immediately show anything else though the issue is
obvious in the code).
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20160222/c72925ef/attachment.sig>

next-20160222 build: 5 failures 9 warnings (next-20160222)

[Marc Zyngier]

On 22/02/16 10:08, Mark Brown wrote:
> On Mon, Feb 22, 2016 at 08:36:46AM +0000, Build bot for Mark Brown wrote:
> 
> Today's -next fails to build on arm64 due to:
> 
>> 	arm64-allnoconfig
>> ../arch/arm64/include/asm/kvm_host.h:338:2: error: implicit declaration of function 'kvm_call_hyp' [-Werror=implicit-function-declaration]
>>
>> 	arm64-allmodconfig
>> ../arch/arm64/include/asm/kvm_host.h:338:2: error: implicit declaration of function 'kvm_call_hyp' [-Werror=implicit-function-declaration]
> 
>> 	arm64-defconfig
>> ../arch/arm64/include/asm/kvm_host.h:338:2: error: implicit declaration of function 'kvm_call_hyp' [-Werror=implicit-function-declaration]
> 
> triggered by the merge of the kvm-arm tree with -next today, it looks
> like changes in that tree and arm64 aren't playing nicely with each
> other (bisect came down to the merge and visual inspection of the
> changes didn't immediately show anything else though the issue is
> obvious in the code).

I'm about to push the following patch on top of the KVM tree:

diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h
index 43688d9..31fe7d6 100644
--- a/arch/arm64/include/asm/kvm_host.h
+++ b/arch/arm64/include/asm/kvm_host.h
@@ -333,11 +333,6 @@ static inline void __cpu_init_hyp_mode(phys_addr_t boot_pgd_ptr,
 		     hyp_stack_ptr, vector_ptr);
 }
 
-static inline void __cpu_init_stage2(void)
-{
-	kvm_call_hyp(__init_stage2_translation);
-}
-
 static inline void kvm_arch_hardware_disable(void) {}
 static inline void kvm_arch_hardware_unsetup(void) {}
 static inline void kvm_arch_sync_events(struct kvm *kvm) {}
@@ -349,4 +344,11 @@ void kvm_arm_setup_debug(struct kvm_vcpu *vcpu);
 void kvm_arm_clear_debug(struct kvm_vcpu *vcpu);
 void kvm_arm_reset_debug_ptr(struct kvm_vcpu *vcpu);
 
+/* #define kvm_call_hyp(f, ...) __kvm_call_hyp(kvm_ksym_ref(f), ##__VA_ARGS__) */
+
+static inline void __cpu_init_stage2(void)
+{
+	kvm_call_hyp(__init_stage2_translation);
+}
+
 #endif /* __ARM64_KVM_HOST_H__ */

This will create a new conflict, but will make the resolution 100% clear.

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...

next-20160222 build: 5 failures 9 warnings (next-20160222)

[Brian Starkey]

Hi Mark,

On Mon, Feb 22, 2016 at 10:41:42AM +0000, Marc Zyngier wrote:
>I'm about to push the following patch on top of the KVM tree:
>
>diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h
>index 43688d9..31fe7d6 100644
>--- a/arch/arm64/include/asm/kvm_host.h
>+++ b/arch/arm64/include/asm/kvm_host.h
>@@ -333,11 +333,6 @@ static inline void __cpu_init_hyp_mode(phys_addr_t boot_pgd_ptr,
> 		     hyp_stack_ptr, vector_ptr);
> }
>
>-static inline void __cpu_init_stage2(void)
>-{
>-	kvm_call_hyp(__init_stage2_translation);
>-}
>-
> static inline void kvm_arch_hardware_disable(void) {}
> static inline void kvm_arch_hardware_unsetup(void) {}
> static inline void kvm_arch_sync_events(struct kvm *kvm) {}
>@@ -349,4 +344,11 @@ void kvm_arm_setup_debug(struct kvm_vcpu *vcpu);
> void kvm_arm_clear_debug(struct kvm_vcpu *vcpu);
> void kvm_arm_reset_debug_ptr(struct kvm_vcpu *vcpu);
>
>+/* #define kvm_call_hyp(f, ...) __kvm_call_hyp(kvm_ksym_ref(f), ##__VA_ARGS__) */
>+
>+static inline void __cpu_init_stage2(void)
>+{
>+	kvm_call_hyp(__init_stage2_translation);
>+}
>+
> #endif /* __ARM64_KVM_HOST_H__ */
>
>This will create a new conflict, but will make the resolution 100% clear.

I applied this, but arm64 defconfig wouldn't build until I uncommented
the '#define kvm_call_hyp(...' line.

Is that right?

Thanks,
Brian

>
>Thanks,
>
>	M.
>-- 
>Jazz is not dead. It just smells funny...
>
>_______________________________________________
>linux-arm-kernel mailing list
>linux-arm-kernel at lists.infradead.org
>http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
>

next-20160222 build: 5 failures 9 warnings (next-20160222)

[Marc Zyngier]

On 22/02/16 11:52, Brian Starkey wrote:
> Hi Mark,
> 
> On Mon, Feb 22, 2016 at 10:41:42AM +0000, Marc Zyngier wrote:
>> I'm about to push the following patch on top of the KVM tree:
>>
>> diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h
>> index 43688d9..31fe7d6 100644
>> --- a/arch/arm64/include/asm/kvm_host.h
>> +++ b/arch/arm64/include/asm/kvm_host.h
>> @@ -333,11 +333,6 @@ static inline void __cpu_init_hyp_mode(phys_addr_t boot_pgd_ptr,
>> 		     hyp_stack_ptr, vector_ptr);
>> }
>>
>> -static inline void __cpu_init_stage2(void)
>> -{
>> -	kvm_call_hyp(__init_stage2_translation);
>> -}
>> -
>> static inline void kvm_arch_hardware_disable(void) {}
>> static inline void kvm_arch_hardware_unsetup(void) {}
>> static inline void kvm_arch_sync_events(struct kvm *kvm) {}
>> @@ -349,4 +344,11 @@ void kvm_arm_setup_debug(struct kvm_vcpu *vcpu);
>> void kvm_arm_clear_debug(struct kvm_vcpu *vcpu);
>> void kvm_arm_reset_debug_ptr(struct kvm_vcpu *vcpu);
>>
>> +/* #define kvm_call_hyp(f, ...) __kvm_call_hyp(kvm_ksym_ref(f), ##__VA_ARGS__) */
>> +
>> +static inline void __cpu_init_stage2(void)
>> +{
>> +	kvm_call_hyp(__init_stage2_translation);
>> +}
>> +
>> #endif /* __ARM64_KVM_HOST_H__ */
>>
>> This will create a new conflict, but will make the resolution 100% clear.
> 
> I applied this, but arm64 defconfig wouldn't build until I uncommented
> the '#define kvm_call_hyp(...' line.
> 
> Is that right?

This patch is intended as a merge indication so that when you do the
merge with the arm64 tree, it conflicts and you can resolve it by going
the right thing (merging the non-comment version of this line).

So yes, uncommenting it is the right thing to do.

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...

next-20160222 build: 5 failures 9 warnings (next-20160222)

[Brian Starkey]

On Mon, Feb 22, 2016 at 12:00:40PM +0000, Marc Zyngier wrote:
>On 22/02/16 11:52, Brian Starkey wrote:
>>> This will create a new conflict, but will make the resolution 100% 
>>> clear.
>>
>> I applied this, but arm64 defconfig wouldn't build until I uncommented
>> the '#define kvm_call_hyp(...' line.
>>
>> Is that right?
>
>This patch is intended as a merge indication so that when you do the
>merge with the arm64 tree, it conflicts and you can resolve it by going
>the right thing (merging the non-comment version of this line).

Right, now I understand "make the resolution 100% clear" - I guess 100%
clear isn't clear enough for me :-(

I've not seen that used as notation before, thanks for explaining.

-Brian

>
>So yes, uncommenting it is the right thing to do.
>
>Thanks,
>
>	M.
>-- 
>Jazz is not dead. It just smells funny...
>

next-20160222 build: 5 failures 9 warnings (next-20160222)

[Marc Zyngier]

On 22/02/16 12:07, Brian Starkey wrote:
> On Mon, Feb 22, 2016 at 12:00:40PM +0000, Marc Zyngier wrote:
>> On 22/02/16 11:52, Brian Starkey wrote:
>>>> This will create a new conflict, but will make the resolution 100% 
>>>> clear.
>>>
>>> I applied this, but arm64 defconfig wouldn't build until I uncommented
>>> the '#define kvm_call_hyp(...' line.
>>>
>>> Is that right?
>>
>> This patch is intended as a merge indication so that when you do the
>> merge with the arm64 tree, it conflicts and you can resolve it by going
>> the right thing (merging the non-comment version of this line).
> 
> Right, now I understand "make the resolution 100% clear" - I guess 100%
> clear isn't clear enough for me :-(

It is not meant to be clear to you, but to whoever is actually performing the merge:

$ git checkout arm64/for-next/core
Note: checking out 'arm64/for-next/core'.
$ git merge kvmarm-master/next
Auto-merging virt/kvm/arm/hyp/vgic-v2-sr.c
Auto-merging virt/kvm/arm/hyp/timer-sr.c
Auto-merging arch/arm64/mm/fault.c
Removing arch/arm64/kvm/hyp/hyp.h
Auto-merging arch/arm64/kvm/hyp/debug-sr.c
Auto-merging arch/arm64/kvm/hyp.S
CONFLICT (content): Merge conflict in arch/arm64/kvm/hyp.S
Auto-merging arch/arm64/kernel/image.h
Auto-merging arch/arm64/kernel/head.S
Auto-merging arch/arm64/kernel/debug-monitors.c
Auto-merging arch/arm64/kernel/cpufeature.c
CONFLICT (content): Merge conflict in arch/arm64/kernel/cpufeature.c
Auto-merging arch/arm64/include/asm/kvm_host.h
CONFLICT (content): Merge conflict in arch/arm64/include/asm/kvm_host.h
Auto-merging arch/arm64/include/asm/kvm_asm.h
Auto-merging arch/arm64/include/asm/cpufeature.h
CONFLICT (content): Merge conflict in arch/arm64/include/asm/cpufeature.h
Auto-merging arch/arm64/Kconfig
Removing arch/arm/kvm/interrupts_head.S
Auto-merging arch/arm/kvm/arm.c
CONFLICT (content): Merge conflict in arch/arm/kvm/arm.c
Auto-merging arch/arm/include/asm/kvm_asm.h
Automatic merge failed; fix conflicts and then commit the result.
maz at approximate:~/Work/arm-platforms$ git diff arch/arm64/include/asm/kvm_host.h
diff --cc arch/arm64/include/asm/kvm_host.h
index e3d67ff,31fe7d6..0000000
--- a/arch/arm64/include/asm/kvm_host.h
+++ b/arch/arm64/include/asm/kvm_host.h
@@@ -343,6 -344,11 +344,15 @@@ void kvm_arm_setup_debug(struct kvm_vcp
  void kvm_arm_clear_debug(struct kvm_vcpu *vcpu);
  void kvm_arm_reset_debug_ptr(struct kvm_vcpu *vcpu);
  
++<<<<<<< HEAD
 +#define kvm_call_hyp(f, ...) __kvm_call_hyp(kvm_ksym_ref(f), ##__VA_ARGS__)
++=======
+ /* #define kvm_call_hyp(f, ...) __kvm_call_hyp(kvm_ksym_ref(f), ##__VA_ARGS__) */
+ 
+ static inline void __cpu_init_stage2(void)
+ {
+       kvm_call_hyp(__init_stage2_translation);
+ }
++>>>>>>> kvmarm-master/next
  
  #endif /* __ARM64_KVM_HOST_H__ */

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...

next-20160222 build: 5 failures 9 warnings (next-20160222)

[Stephen Rothwell]

Hi Marc,

On Mon, 22 Feb 2016 12:44:15 +0000 Marc Zyngier <marc.zyngier@arm.com> wrote:
>
> On 22/02/16 12:07, Brian Starkey wrote:
> > On Mon, Feb 22, 2016 at 12:00:40PM +0000, Marc Zyngier wrote:  
> >> On 22/02/16 11:52, Brian Starkey wrote:  
> >>>> This will create a new conflict, but will make the resolution 100% 
> >>>> clear.  
> >>>
> >>> I applied this, but arm64 defconfig wouldn't build until I uncommented
> >>> the '#define kvm_call_hyp(...' line.
> >>>
> >>> Is that right?  
> >>
> >> This patch is intended as a merge indication so that when you do the
> >> merge with the arm64 tree, it conflicts and you can resolve it by going
> >> the right thing (merging the non-comment version of this line).  
> > 
> > Right, now I understand "make the resolution 100% clear" - I guess 100%
> > clear isn't clear enough for me :-(  
> 
> It is not meant to be clear to you, but to whoever is actually performing the merge:

Excellent, thanks for that - hopefully I will get it right tomorrow.

-- 
Cheers,
Stephen Rothwell

next-20160222 build: 5 failures 9 warnings (next-20160222)

[Mark Brown]

On Mon, Feb 22, 2016 at 10:41:42AM +0000, Marc Zyngier wrote:
> On 22/02/16 10:08, Mark Brown wrote:
> > On Mon, Feb 22, 2016 at 08:36:46AM +0000, Build bot for Mark Brown wrote:

> > Today's -next fails to build on arm64 due to:

> >> 	arm64-allnoconfig
> >> ../arch/arm64/include/asm/kvm_host.h:338:2: error: implicit declaration of function 'kvm_call_hyp' [-Werror=implicit-function-declaration]

> I'm about to push the following patch on top of the KVM tree:

> This will create a new conflict, but will make the resolution 100% clear.

That didn't work out unfortunately, the error is still there today.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20160224/d80a1d16/attachment.sig>

next-20160222 build: 5 failures 9 warnings (next-20160222)

[Stephen Rothwell]

Hi Mark,

On Wed, 24 Feb 2016 12:28:42 +0900 Mark Brown <broonie@kernel.org> wrote:
>
> That didn't work out unfortunately, the error is still there today.

Yeah, the patch didn't hit -next yesterday and I missed the opportunity
to fix in manually.  However, it should be fixed in today's linux-next
(out in a few hours, hopefully).

-- 
Cheers,
Stephen Rothwell

next-20160222 build: 5 failures 9 warnings (next-20160222)

[Mark Brown]

On Wed, Feb 24, 2016 at 02:48:59PM +1100, Stephen Rothwell wrote:
> On Wed, 24 Feb 2016 12:28:42 +0900 Mark Brown <broonie@kernel.org> wrote:

> > That didn't work out unfortunately, the error is still there today.

> Yeah, the patch didn't hit -next yesterday and I missed the opportunity
> to fix in manually.  However, it should be fixed in today's linux-next
> (out in a few hours, hopefully).

Ah, good - it was just too late rather than not working.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20160224/4b822083/attachment.sig>

next-20160222 build: 5 failures 9 warnings (next-20160222)

[Mark Brown]

On Mon, Feb 22, 2016 at 08:36:46AM +0000, Build bot for Mark Brown wrote:

Today's -next fails to build an arm allmodconfig due to:

> 	arm-allmodconfig
> ../drivers/net/ethernet/ti/netcp_core.c:1846:31: error: invalid type argument of '->' (have 'struct tc_to_netdev')
> ../drivers/net/ethernet/ti/netcp_core.c:1851:35: error: invalid type argument of '->' (have 'struct tc_to_netdev')
> ../drivers/net/ethernet/ti/netcp_core.c:1855:8: error: invalid type argument of '->' (have 'struct tc_to_netdev')
> ../drivers/net/ethernet/ti/netcp_core.c:1856:28: error: invalid type argument of '->' (have 'struct tc_to_netdev')
> ../drivers/net/ethernet/ti/netcp_core.c:1857:21: error: invalid type argument of '->' (have 'struct tc_to_netdev')

caused by commit 16e5cc647173a (net: rework setup_tc ndo op to consume
general tc operand) in the net-next tree which looks to have a missing *
in the rewrite of the function arguments for that driver.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20160222/87ece996/attachment-0001.sig>

next-20160222 build: 5 failures 9 warnings (next-20160222)

[Mark Brown]

On Mon, Feb 22, 2016 at 08:36:46AM +0000, Build bot for Mark Brown wrote:

For the past few days -next has been failing to build an arm
allmodconfig due to:

> 	arm-allmodconfig
> ../arch/arm/kernel/hyp-stub.S:168: Error: selected processor does not support ARM mode `ubfx r7,r7,#28,#4'
> ../arch/arm/kernel/hyp-stub.S:176: Error: selected processor does not support ARM mode `isb'

caused by e59941b9b3817c (ARM: 8527/1: virt: enable GICv3 system
registers) in the ARM tree.  My hot take is that it looks like
CONFIG_ARM_GIC_V3 needs tighter dependencies so it can't get enabled for
a v6 system but I could be wrong.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20160222/9736c9c3/attachment.sig>

next-20160222 build: 5 failures 9 warnings (next-20160222)

[Jean-Philippe Brucker]

Mark,

On Mon, Feb 22, 2016 at 07:25:42PM +0900, Mark Brown wrote:
> On Mon, Feb 22, 2016 at 08:36:46AM +0000, Build bot for Mark Brown wrote:
> 
> For the past few days -next has been failing to build an arm
> allmodconfig due to:
> 
> > 	arm-allmodconfig
> > ../arch/arm/kernel/hyp-stub.S:168: Error: selected processor does not support ARM mode `ubfx r7,r7,#28,#4'
> > ../arch/arm/kernel/hyp-stub.S:176: Error: selected processor does not support ARM mode `isb'
> 
> caused by e59941b9b3817c (ARM: 8527/1: virt: enable GICv3 system
> registers) in the ARM tree.  My hot take is that it looks like
> CONFIG_ARM_GIC_V3 needs tighter dependencies so it can't get enabled for
> a v6 system but I could be wrong.

Thanks for the report. A fix is already in Russell's patch system [1],
and should be applied shortly.

Thanks,
Jean-Philippe

[1] http://www.arm.linux.org.uk/developer/patches/viewpatch.php?id=8534/1