trouble building 2.6.38: 'IRQF_FORCE_RESUME' undeclared

trouble building 2.6.38: 'IRQF_FORCE_RESUME' undeclared

[Daniel Stodden]

Iirc I derived this from a prior commit from either Ian or Jeremy.

Didn't really check it out, so there's very likely a more correct fix
than what's attached here.

/local/exp/dns/scratch/xenbits/xen-unstable.hg/linux-2.6-pvops.git/arch/x86/xen/time.c: In function 'xen_setup_timer':
/local/exp/dns/scratch/xenbits/xen-unstable.hg/linux-2.6-pvops.git/arch/x86/xen/time.c:402: error: 'IRQF_FORCE_RESUME' undeclared (first use in this function)
/local/exp/dns/scratch/xenbits/xen-unstable.hg/linux-2.6-pvops.git/arch/x86/xen/time.c:402: error: (Each undeclared identifier is reported only once
/local/exp/dns/scratch/xenbits/xen-unstable.hg/linux-2.6-pvops.git/arch/x86/xen/time.c:402: error: for each function it appears in.)

This reproducible?

Thanks,
Daniel

[PATCH] kernel: Fix pvops build on 2.6.38.

[Daniel Stodden]

---
 arch/x86/xen/time.c  |    2 +-
 drivers/xen/events.c |    2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/x86/xen/time.c b/arch/x86/xen/time.c
index 2e2d370..1f48076 100644
--- a/arch/x86/xen/time.c
+++ b/arch/x86/xen/time.c
@@ -399,7 +399,7 @@ void xen_setup_timer(int cpu)
 	irq = bind_virq_to_irqhandler(VIRQ_TIMER, cpu, xen_timer_interrupt,
 				      IRQF_DISABLED|IRQF_PERCPU|
 				      IRQF_NOBALANCING|IRQF_TIMER|
-				      IRQF_FORCE_RESUME,
+				      IRQF_NO_SUSPEND,
 				      name, NULL);
 
 	evt = &per_cpu(xen_clock_events, cpu);
diff --git a/drivers/xen/events.c b/drivers/xen/events.c
index 916d9a8..d4251ad 100644
--- a/drivers/xen/events.c
+++ b/drivers/xen/events.c
@@ -1026,7 +1026,7 @@ int bind_ipi_to_irqhandler(enum ipi_vector ipi,
 	if (irq < 0)
 		return irq;
 
-	irqflags |= IRQF_NO_SUSPEND | IRQF_FORCE_RESUME;
+	irqflags |= IRQF_NO_SUSPEND;
 	retval = request_irq(irq, handler, irqflags, devname, dev_id);
 	if (retval != 0) {
 		unbind_from_irq(irq);
-- 
1.7.0.4

Re: trouble building 2.6.38: 'IRQF_FORCE_RESUME' undeclared

[Konrad Rzeszutek Wilk]

On Thu, Mar 03, 2011 at 12:52:54PM -0800, Daniel Stodden wrote:
> Iirc I derived this from a prior commit from either Ian or Jeremy.
> 
> Didn't really check it out, so there's very likely a more correct fix
> than what's attached here.
> 
> /local/exp/dns/scratch/xenbits/xen-unstable.hg/linux-2.6-pvops.git/arch/x86/xen/time.c: In function 'xen_setup_timer':
> /local/exp/dns/scratch/xenbits/xen-unstable.hg/linux-2.6-pvops.git/arch/x86/xen/time.c:402: error: 'IRQF_FORCE_RESUME' undeclared (first use in this function)
> /local/exp/dns/scratch/xenbits/xen-unstable.hg/linux-2.6-pvops.git/arch/x86/xen/time.c:402: error: (Each undeclared identifier is reported only once
> /local/exp/dns/scratch/xenbits/xen-unstable.hg/linux-2.6-pvops.git/arch/x86/xen/time.c:402: error: for each function it appears in.)
> 

Uh, what git tree is that? 

> This reproducible?
> 
> Thanks,
> Daniel
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel

Re: [PATCH] kernel: Fix pvops build on 2.6.38.

[Ian Campbell]

On Thu, 2011-03-03 at 20:52 +0000, Daniel Stodden wrote:
> ---
>  arch/x86/xen/time.c  |    2 +-
>  drivers/xen/events.c |    2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/x86/xen/time.c b/arch/x86/xen/time.c
> index 2e2d370..1f48076 100644
> --- a/arch/x86/xen/time.c
> +++ b/arch/x86/xen/time.c
> @@ -399,7 +399,7 @@ void xen_setup_timer(int cpu)
>  	irq = bind_virq_to_irqhandler(VIRQ_TIMER, cpu, xen_timer_interrupt,
>  				      IRQF_DISABLED|IRQF_PERCPU|
>  				      IRQF_NOBALANCING|IRQF_TIMER|
> -				      IRQF_FORCE_RESUME,
> +				      IRQF_NO_SUSPEND,
>  				      name, NULL);
>  
>  	evt = &per_cpu(xen_clock_events, cpu);
> diff --git a/drivers/xen/events.c b/drivers/xen/events.c
> index 916d9a8..d4251ad 100644
> --- a/drivers/xen/events.c
> +++ b/drivers/xen/events.c
> @@ -1026,7 +1026,7 @@ int bind_ipi_to_irqhandler(enum ipi_vector ipi,
>  	if (irq < 0)
>  		return irq;
>  
> -	irqflags |= IRQF_NO_SUSPEND | IRQF_FORCE_RESUME;
> +	irqflags |= IRQF_NO_SUSPEND;
>  	retval = request_irq(irq, handler, irqflags, devname, dev_id);
>  	if (retval != 0) {
>  		unbind_from_irq(irq);

These flags aren't interchangeable like this, they are dependent on each
other. IRQF_NO_SUSPEND does what it says but IRQF_FORCE_RESUME means
"resume even if IRQF_NO_SUSPEND".

IIRC we want both for IPIs but VIRQ_TIMER needs neither, since the
correct flags are included in IRQF_TIMER already. Nothing else on the
Xen side should be using either flag AFAIK.

However, I don't know of any tree which doesn't have this correct
already. Both Jeremy and Konrad's branches with 2.6.38 in the name are
correct.

I think you need to figure out where you got this tree from (and make
sure you are up to date with it) so we can tell if it is in error or
not.

Ian.