Re: [PATCH 2/4] irqdomain: augment add_simple() to allocate descs

2012-11-26 Thread Grant Likely
On Thu, 27 Sep 2012 08:41:53 -0500, Rob Herring  wrote:
> On 09/27/2012 08:13 AM, Linus Walleij wrote:
> > ---
> >  kernel/irq/irqdomain.c | 21 -
> >  1 file changed, 16 insertions(+), 5 deletions(-)
> > 
> > diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c
> > index 49a7772..a0655b6 100644
> > --- a/kernel/irq/irqdomain.c
> > +++ b/kernel/irq/irqdomain.c
> > @@ -148,7 +148,8 @@ static unsigned int irq_domain_legacy_revmap(struct 
> > irq_domain *domain,
> >   * @host_data: Controller private data pointer
> >   *
> >   * Allocates a legacy irq_domain if irq_base is positive or a linear
> > - * domain otherwise.
> > + * domain otherwise. For the legacy domain, IRQ descriptors will also
> > + * be allocated.
> >   *
> >   * This is intended to implement the expected behaviour for most
> >   * interrupt controllers which is that a linear mapping should
> > @@ -162,11 +163,21 @@ struct irq_domain *irq_domain_add_simple(struct 
> > device_node *of_node,
> >  const struct irq_domain_ops *ops,
> >  void *host_data)
> >  {
> > -   if (first_irq > 0)
> > -   return irq_domain_add_legacy(of_node, size, first_irq, 0,
> > +   if (first_irq > 0) {
> > +   int irq_base;
> > +
> > +   irq_base = irq_alloc_descs(first_irq, 0, size, numa_node_id());
> 
> This will always fail for !SPARSE_IRQ, so we probably don't want warn in
> that case. So perhaps something like this:
> 
> int irq_base = first_irq;
> 
> if (is_enabled(CONFIG_SPARSE_IRQ)) {
>   irq_base = irq_alloc_descs(first_irq, 0, size, numa_node_id());
>   if (irq_base < 0) {
>   WARN(1, "Cannot allocate irq_descs @ IRQ%d, assuming 
> pre-allocated\n",
>first_irq);
>   irq_base = first_irq;
>   }
> }

Also, warn seems a little over-loud in this case. If the controller is
legitimately claiming irqs in the reserved range, then dumping out a
backtrace is kind of nasty.

g.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/


Re: [PATCH 2/4] irqdomain: augment add_simple() to allocate descs

2012-11-26 Thread Grant Likely
On Thu, 27 Sep 2012 08:41:53 -0500, Rob Herring robherri...@gmail.com wrote:
 On 09/27/2012 08:13 AM, Linus Walleij wrote:
  ---
   kernel/irq/irqdomain.c | 21 -
   1 file changed, 16 insertions(+), 5 deletions(-)
  
  diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c
  index 49a7772..a0655b6 100644
  --- a/kernel/irq/irqdomain.c
  +++ b/kernel/irq/irqdomain.c
  @@ -148,7 +148,8 @@ static unsigned int irq_domain_legacy_revmap(struct 
  irq_domain *domain,
* @host_data: Controller private data pointer
*
* Allocates a legacy irq_domain if irq_base is positive or a linear
  - * domain otherwise.
  + * domain otherwise. For the legacy domain, IRQ descriptors will also
  + * be allocated.
*
* This is intended to implement the expected behaviour for most
* interrupt controllers which is that a linear mapping should
  @@ -162,11 +163,21 @@ struct irq_domain *irq_domain_add_simple(struct 
  device_node *of_node,
   const struct irq_domain_ops *ops,
   void *host_data)
   {
  -   if (first_irq  0)
  -   return irq_domain_add_legacy(of_node, size, first_irq, 0,
  +   if (first_irq  0) {
  +   int irq_base;
  +
  +   irq_base = irq_alloc_descs(first_irq, 0, size, numa_node_id());
 
 This will always fail for !SPARSE_IRQ, so we probably don't want warn in
 that case. So perhaps something like this:
 
 int irq_base = first_irq;
 
 if (is_enabled(CONFIG_SPARSE_IRQ)) {
   irq_base = irq_alloc_descs(first_irq, 0, size, numa_node_id());
   if (irq_base  0) {
   WARN(1, Cannot allocate irq_descs @ IRQ%d, assuming 
 pre-allocated\n,
first_irq);
   irq_base = first_irq;
   }
 }

Also, warn seems a little over-loud in this case. If the controller is
legitimately claiming irqs in the reserved range, then dumping out a
backtrace is kind of nasty.

g.
--
To unsubscribe from this list: send the line unsubscribe linux-kernel in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/


Re: [PATCH 2/4] irqdomain: augment add_simple() to allocate descs

2012-09-27 Thread Paul Mundt
On Thu, Sep 27, 2012 at 03:13:45PM +0200, Linus Walleij wrote:
> + if (first_irq > 0) {
> + int irq_base;
> +
> + irq_base = irq_alloc_descs(first_irq, 0, size, numa_node_id());
> + if (irq_base < 0) {
> + WARN(1, "Cannot allocate irq_descs @ IRQ%d, assuming 
> pre-allocated\n",
> +  first_irq);
> + irq_base = first_irq;
> + }
> + return irq_domain_add_legacy(of_node, size, irq_base, 0,
>ops, host_data);

If first_irq is specified you presumably want irq_alloc_desc_at()
behaviour, so you should use irq_alloc_descs(first_irq, first_irq, ...),
similar to what irq_create_strict_mappings() does.

Also don't use numa_node_id() for this, of_node_to_nid() handles both the
OF and non-OF cases.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/


Re: [PATCH 2/4] irqdomain: augment add_simple() to allocate descs

2012-09-27 Thread Rob Herring
On 09/27/2012 08:13 AM, Linus Walleij wrote:
> From: Linus Walleij 
> 
> Currently we rely on all IRQ chip instances to dynamically
> allocate their IRQ descriptors unless they use the linear
> IRQ domain. So for irqdomain_add_legacy() and
> irqdomain_add_simple() the caller need to make sure that
> descriptors are allocated.
> 
> Let's slightly augment the yet unused irqdomain_add_simple()
> to also allocate descriptors as a means to simplify usage
> and avoid code duplication throughout the kernel.
> 
> We warn if descriptors cannot be allocated, e.g. if a
> platform has the bad habit of hogging descriptors at boot
> time.
> 
> Cc: Rob Herring 
> Cc: Thomas Gleixner 
> Cc: Grant Likely 
> Cc: Russell King 
> Cc: Lee Jones 
> Signed-off-by: Linus Walleij 
> ---
> Rob/Grant/Thomas if you ACK this I will take it through the
> pinctrl tree since I introduce its only user right there.

You really should have Grant's ack here.

> ---
>  kernel/irq/irqdomain.c | 21 -
>  1 file changed, 16 insertions(+), 5 deletions(-)
> 
> diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c
> index 49a7772..a0655b6 100644
> --- a/kernel/irq/irqdomain.c
> +++ b/kernel/irq/irqdomain.c
> @@ -148,7 +148,8 @@ static unsigned int irq_domain_legacy_revmap(struct 
> irq_domain *domain,
>   * @host_data: Controller private data pointer
>   *
>   * Allocates a legacy irq_domain if irq_base is positive or a linear
> - * domain otherwise.
> + * domain otherwise. For the legacy domain, IRQ descriptors will also
> + * be allocated.
>   *
>   * This is intended to implement the expected behaviour for most
>   * interrupt controllers which is that a linear mapping should
> @@ -162,11 +163,21 @@ struct irq_domain *irq_domain_add_simple(struct 
> device_node *of_node,
>const struct irq_domain_ops *ops,
>void *host_data)
>  {
> - if (first_irq > 0)
> - return irq_domain_add_legacy(of_node, size, first_irq, 0,
> + if (first_irq > 0) {
> + int irq_base;
> +
> + irq_base = irq_alloc_descs(first_irq, 0, size, numa_node_id());

This will always fail for !SPARSE_IRQ, so we probably don't want warn in
that case. So perhaps something like this:

int irq_base = first_irq;

if (is_enabled(CONFIG_SPARSE_IRQ)) {
irq_base = irq_alloc_descs(first_irq, 0, size, numa_node_id());
if (irq_base < 0) {
WARN(1, "Cannot allocate irq_descs @ IRQ%d, assuming 
pre-allocated\n",
 first_irq);
irq_base = first_irq;
}
}

> + if (irq_base < 0) {
> + WARN(1, "Cannot allocate irq_descs @ IRQ%d, assuming 
> pre-allocated\n",
> +  first_irq);
> + irq_base = first_irq;
> + }
> + return irq_domain_add_legacy(of_node, size, irq_base, 0,
>ops, host_data);
> - else
> - return irq_domain_add_linear(of_node, size, ops, host_data);
> + }
> +
> + /* A linear domain is the default */
> + return irq_domain_add_linear(of_node, size, ops, host_data);
>  }
>  
>  /**
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/


Re: [PATCH 2/4] irqdomain: augment add_simple() to allocate descs

2012-09-27 Thread Rob Herring
On 09/27/2012 08:13 AM, Linus Walleij wrote:
 From: Linus Walleij linus.wall...@linaro.org
 
 Currently we rely on all IRQ chip instances to dynamically
 allocate their IRQ descriptors unless they use the linear
 IRQ domain. So for irqdomain_add_legacy() and
 irqdomain_add_simple() the caller need to make sure that
 descriptors are allocated.
 
 Let's slightly augment the yet unused irqdomain_add_simple()
 to also allocate descriptors as a means to simplify usage
 and avoid code duplication throughout the kernel.
 
 We warn if descriptors cannot be allocated, e.g. if a
 platform has the bad habit of hogging descriptors at boot
 time.
 
 Cc: Rob Herring rob.herr...@calxeda.com
 Cc: Thomas Gleixner t...@linutronix.de
 Cc: Grant Likely grant.lik...@secretlab.ca
 Cc: Russell King li...@arm.linux.org.uk
 Cc: Lee Jones lee.jo...@linaro.org
 Signed-off-by: Linus Walleij linus.wall...@linaro.org
 ---
 Rob/Grant/Thomas if you ACK this I will take it through the
 pinctrl tree since I introduce its only user right there.

You really should have Grant's ack here.

 ---
  kernel/irq/irqdomain.c | 21 -
  1 file changed, 16 insertions(+), 5 deletions(-)
 
 diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c
 index 49a7772..a0655b6 100644
 --- a/kernel/irq/irqdomain.c
 +++ b/kernel/irq/irqdomain.c
 @@ -148,7 +148,8 @@ static unsigned int irq_domain_legacy_revmap(struct 
 irq_domain *domain,
   * @host_data: Controller private data pointer
   *
   * Allocates a legacy irq_domain if irq_base is positive or a linear
 - * domain otherwise.
 + * domain otherwise. For the legacy domain, IRQ descriptors will also
 + * be allocated.
   *
   * This is intended to implement the expected behaviour for most
   * interrupt controllers which is that a linear mapping should
 @@ -162,11 +163,21 @@ struct irq_domain *irq_domain_add_simple(struct 
 device_node *of_node,
const struct irq_domain_ops *ops,
void *host_data)
  {
 - if (first_irq  0)
 - return irq_domain_add_legacy(of_node, size, first_irq, 0,
 + if (first_irq  0) {
 + int irq_base;
 +
 + irq_base = irq_alloc_descs(first_irq, 0, size, numa_node_id());

This will always fail for !SPARSE_IRQ, so we probably don't want warn in
that case. So perhaps something like this:

int irq_base = first_irq;

if (is_enabled(CONFIG_SPARSE_IRQ)) {
irq_base = irq_alloc_descs(first_irq, 0, size, numa_node_id());
if (irq_base  0) {
WARN(1, Cannot allocate irq_descs @ IRQ%d, assuming 
pre-allocated\n,
 first_irq);
irq_base = first_irq;
}
}

 + if (irq_base  0) {
 + WARN(1, Cannot allocate irq_descs @ IRQ%d, assuming 
 pre-allocated\n,
 +  first_irq);
 + irq_base = first_irq;
 + }
 + return irq_domain_add_legacy(of_node, size, irq_base, 0,
ops, host_data);
 - else
 - return irq_domain_add_linear(of_node, size, ops, host_data);
 + }
 +
 + /* A linear domain is the default */
 + return irq_domain_add_linear(of_node, size, ops, host_data);
  }
  
  /**
 
--
To unsubscribe from this list: send the line unsubscribe linux-kernel in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/


Re: [PATCH 2/4] irqdomain: augment add_simple() to allocate descs

2012-09-27 Thread Paul Mundt
On Thu, Sep 27, 2012 at 03:13:45PM +0200, Linus Walleij wrote:
 + if (first_irq  0) {
 + int irq_base;
 +
 + irq_base = irq_alloc_descs(first_irq, 0, size, numa_node_id());
 + if (irq_base  0) {
 + WARN(1, Cannot allocate irq_descs @ IRQ%d, assuming 
 pre-allocated\n,
 +  first_irq);
 + irq_base = first_irq;
 + }
 + return irq_domain_add_legacy(of_node, size, irq_base, 0,
ops, host_data);

If first_irq is specified you presumably want irq_alloc_desc_at()
behaviour, so you should use irq_alloc_descs(first_irq, first_irq, ...),
similar to what irq_create_strict_mappings() does.

Also don't use numa_node_id() for this, of_node_to_nid() handles both the
OF and non-OF cases.
--
To unsubscribe from this list: send the line unsubscribe linux-kernel in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/