All of lore.kernel.org
 help / color / mirror / Atom feed
From: Rob Herring <robherring2@gmail.com>
To: David Daney <ddaney.cavm@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>,
	linux-mips@linux-mips.org, ralf@linux-mips.org,
	devicetree-discuss@lists.ozlabs.org, grant.likely@secretlab.ca,
	linux-kernel@vger.kernel.org,
	David Daney <david.daney@cavium.com>
Subject: Re: [PATCH 3/8] MIPS: Octeon: Add irq_create_of_mapping() and GPIO interrupts.
Date: Fri, 11 Nov 2011 14:41:11 -0600	[thread overview]
Message-ID: <4EBD8867.30205@gmail.com> (raw)
In-Reply-To: <4EBD7C5D.7080606@gmail.com>

David,

On 11/11/2011 01:49 PM, David Daney wrote:
> On 11/11/2011 08:08 AM, Rob Herring wrote:
>> On 11/10/2011 08:21 PM, ddaney.cavm@gmail.com wrote:
>>> From: David Daney<david.daney@cavium.com>
>>>
>>> This is needed for Octeon to use the Device Tree.
>>>
>>> The GPIO interrupts are configured based on Device Tree properties
>>>
>>> Signed-off-by: David Daney<david.daney@cavium.com>
>>> ---
>>>   arch/mips/cavium-octeon/octeon-irq.c |  188
>>> +++++++++++++++++++++++++++++++++-
>>>   1 files changed, 187 insertions(+), 1 deletions(-)
>>>
>>> diff --git a/arch/mips/cavium-octeon/octeon-irq.c
>>> b/arch/mips/cavium-octeon/octeon-irq.c
>>> index ffd4ae6..bb10546 100644
>>> --- a/arch/mips/cavium-octeon/octeon-irq.c
>>> +++ b/arch/mips/cavium-octeon/octeon-irq.c
>>> @@ -8,11 +8,14 @@
>>>
>>>   #include<linux/interrupt.h>
>>>   #include<linux/bitops.h>
>>> +#include<linux/module.h>
>>>   #include<linux/percpu.h>
>>> +#include<linux/of_irq.h>
>>>   #include<linux/irq.h>
>>>   #include<linux/smp.h>
>>>
>>>   #include<asm/octeon/octeon.h>
>>> +#include<asm/octeon/cvmx-gpio-defs.h>
>>>
>>>   static DEFINE_RAW_SPINLOCK(octeon_irq_ciu0_lock);
>>>   static DEFINE_RAW_SPINLOCK(octeon_irq_ciu1_lock);
>>> @@ -58,6 +61,95 @@ static void __init octeon_irq_set_ciu_mapping(int
>>> irq, int line, int bit,
>>>       octeon_irq_ciu_to_irq[line][bit] = irq;
>>>   }
>>>
>>> +static unsigned int octeon_irq_gpio_mapping(struct device_node
>>> *controller,
>>> +                        const u32 *intspec,
>>> +                        unsigned int intsize)
>>> +{
>>> +    struct of_irq oirq;
>>> +    int i;
>>> +    unsigned int irq = 0;
>>> +    unsigned int type;
>>> +    unsigned int ciu = 0, bit = 0;
>>> +    unsigned int pin = be32_to_cpup(intspec);
>>> +    unsigned int trigger = be32_to_cpup(intspec + 1);
>>> +    bool set_edge_handler = false;
>>> +
>>> +    if (pin>= 16)
>>> +        goto err;
>>> +    i = of_irq_map_one(controller, 0,&oirq);
>>> +    if (i)
>>> +        goto err;
>>> +    if (oirq.size != 2)
>>> +        goto err_put;
>>> +
>>> +    ciu = oirq.specifier[0];
>>> +    bit = oirq.specifier[1] + pin;
>>> +
>>> +    if (ciu>= 8 || bit>= 64)
>>> +        goto err_put;
>>> +
>>> +    irq = octeon_irq_ciu_to_irq[ciu][bit];
>>> +    if (!irq)
>>> +        goto err_put;
>>> +
>>> +    switch (trigger&  0xf) {
>>> +    case 1:
>>> +        type = IRQ_TYPE_EDGE_RISING;
>>> +        set_edge_handler = true;
>>> +        break;
>>> +    case 2:
>>> +        type = IRQ_TYPE_EDGE_FALLING;
>>> +        set_edge_handler = true;
>>> +        break;
>>> +    case 4:
>>> +        type = IRQ_TYPE_LEVEL_HIGH;
>>> +        break;
>>> +    case 8:
>>> +        type = IRQ_TYPE_LEVEL_LOW;
>>> +        break;
>>> +    default:
>>> +        pr_err("Error: Invalid irq trigger specification: %x\n",
>>> +               trigger);
>>> +        type = IRQ_TYPE_LEVEL_LOW;
>>> +        break;
>>> +    }
>>> +
>>> +    irq_set_irq_type(irq, type);
>>> +
>>> +    if (set_edge_handler)
>>> +        __irq_set_handler(irq, handle_edge_irq, 0, NULL);
>>> +
>>> +err_put:
>>> +    of_node_put(oirq.controller);
>>> +err:
>>> +    return irq;
>>> +}
>>> +
>>> +/*
>>> + * irq_create_of_mapping - Hook to resolve OF irq specifier into a
>>> Linux irq#
>>> + *
>>> + * Octeon irq maps are a pair of indexes.  The first selects either
>>> + * ciu0 or ciu1, the second is the bit within the ciu register.
>>> + */
>>> +unsigned int irq_create_of_mapping(struct device_node *controller,
>>> +                   const u32 *intspec, unsigned int intsize)
>>> +{
>>> +    unsigned int irq = 0;
>>> +    unsigned int ciu, bit;
>>> +
>>> +    if (of_device_is_compatible(controller, "cavium,octeon-3860-gpio"))
>>> +        return octeon_irq_gpio_mapping(controller, intspec, intsize);
>>> +
>>> +    ciu = be32_to_cpup(intspec);
>>> +    bit = be32_to_cpup(intspec + 1);
>>> +
>>> +    if (ciu<  8&&  bit<  64)
>>> +        irq = octeon_irq_ciu_to_irq[ciu][bit];
>>> +
>>> +    return irq;
>>> +}
>>> +EXPORT_SYMBOL_GPL(irq_create_of_mapping);
>>
>> Have you looked at irq_domains (kernel/irq/irqdomain.c)? That is what
>> you should be using for your (gpio) interrupt controller and then use
>> the common irq_create_of_mapping.
>>
> 
> Unfortunatly, although a good idea, kernel/irq/irqdomain.c makes a bunch
> of assumptions that don't hold for Octeon.  We may be able to improve it
> so that it flexible enough to suit us.
> 
> 
> Here are the problems I see:
> 
> 1) It is assumed that there is some sort of linear correspondence
> between 'hwirq' and 'irq', and that the range of valid values is
> contiguous.

That's not true if you implement .to_irq for your domain.

> 2) It is assumed that the concepts of nr_irq, irq_base and hwirq_base
> have easy to determin values and you can do iteration over their ranges
> by adding indexes to the bases.

That's true for hwirq numbering, but not for Linux irq numbering.
irq_base is only used if you don't provide .to_irq. hwirq_base is just
to allow a linear range that doesn't start with 0.

> I think we can fix this by adding iteration helper functions to struct
> irq_domain.  If these are present, we would just ignore the irq_base,
> nr_irq and hwirq_base elements of the structure.
> irq_domain_for_each_hwirq() and irq_domain_for_each_irq() would be
> modified to use the iteration helper functions.

Expanding irqdomain is certainly the intention and right direction.

It seems to me you just need something different than an irq. controller
with a linear range of irq lines. Perhaps sparsely allocated hwirq's?
Can you describe the h/w?

Also, you could have a different domain per irqdesc if you really needed to.

Rob

WARNING: multiple messages have this Message-ID (diff)
From: Rob Herring <robherring2-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
To: David Daney <ddaney.cavm-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
Cc: linux-mips-6z/3iImG2C8G8FEW9MqTrA@public.gmane.org,
	David Daney <david.daney-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>,
	devicetree-discuss-uLR06cmDAlY/bJ5BZ2RsiQ@public.gmane.org,
	linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
	ralf-6z/3iImG2C8G8FEW9MqTrA@public.gmane.org,
	Thomas Gleixner <tglx-hfZtesqFncYOwBW4kG4KsQ@public.gmane.org>
Subject: Re: [PATCH 3/8] MIPS: Octeon: Add irq_create_of_mapping() and GPIO interrupts.
Date: Fri, 11 Nov 2011 14:41:11 -0600	[thread overview]
Message-ID: <4EBD8867.30205@gmail.com> (raw)
In-Reply-To: <4EBD7C5D.7080606-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>

David,

On 11/11/2011 01:49 PM, David Daney wrote:
> On 11/11/2011 08:08 AM, Rob Herring wrote:
>> On 11/10/2011 08:21 PM, ddaney.cavm-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org wrote:
>>> From: David Daney<david.daney-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
>>>
>>> This is needed for Octeon to use the Device Tree.
>>>
>>> The GPIO interrupts are configured based on Device Tree properties
>>>
>>> Signed-off-by: David Daney<david.daney-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>
>>> ---
>>>   arch/mips/cavium-octeon/octeon-irq.c |  188
>>> +++++++++++++++++++++++++++++++++-
>>>   1 files changed, 187 insertions(+), 1 deletions(-)
>>>
>>> diff --git a/arch/mips/cavium-octeon/octeon-irq.c
>>> b/arch/mips/cavium-octeon/octeon-irq.c
>>> index ffd4ae6..bb10546 100644
>>> --- a/arch/mips/cavium-octeon/octeon-irq.c
>>> +++ b/arch/mips/cavium-octeon/octeon-irq.c
>>> @@ -8,11 +8,14 @@
>>>
>>>   #include<linux/interrupt.h>
>>>   #include<linux/bitops.h>
>>> +#include<linux/module.h>
>>>   #include<linux/percpu.h>
>>> +#include<linux/of_irq.h>
>>>   #include<linux/irq.h>
>>>   #include<linux/smp.h>
>>>
>>>   #include<asm/octeon/octeon.h>
>>> +#include<asm/octeon/cvmx-gpio-defs.h>
>>>
>>>   static DEFINE_RAW_SPINLOCK(octeon_irq_ciu0_lock);
>>>   static DEFINE_RAW_SPINLOCK(octeon_irq_ciu1_lock);
>>> @@ -58,6 +61,95 @@ static void __init octeon_irq_set_ciu_mapping(int
>>> irq, int line, int bit,
>>>       octeon_irq_ciu_to_irq[line][bit] = irq;
>>>   }
>>>
>>> +static unsigned int octeon_irq_gpio_mapping(struct device_node
>>> *controller,
>>> +                        const u32 *intspec,
>>> +                        unsigned int intsize)
>>> +{
>>> +    struct of_irq oirq;
>>> +    int i;
>>> +    unsigned int irq = 0;
>>> +    unsigned int type;
>>> +    unsigned int ciu = 0, bit = 0;
>>> +    unsigned int pin = be32_to_cpup(intspec);
>>> +    unsigned int trigger = be32_to_cpup(intspec + 1);
>>> +    bool set_edge_handler = false;
>>> +
>>> +    if (pin>= 16)
>>> +        goto err;
>>> +    i = of_irq_map_one(controller, 0,&oirq);
>>> +    if (i)
>>> +        goto err;
>>> +    if (oirq.size != 2)
>>> +        goto err_put;
>>> +
>>> +    ciu = oirq.specifier[0];
>>> +    bit = oirq.specifier[1] + pin;
>>> +
>>> +    if (ciu>= 8 || bit>= 64)
>>> +        goto err_put;
>>> +
>>> +    irq = octeon_irq_ciu_to_irq[ciu][bit];
>>> +    if (!irq)
>>> +        goto err_put;
>>> +
>>> +    switch (trigger&  0xf) {
>>> +    case 1:
>>> +        type = IRQ_TYPE_EDGE_RISING;
>>> +        set_edge_handler = true;
>>> +        break;
>>> +    case 2:
>>> +        type = IRQ_TYPE_EDGE_FALLING;
>>> +        set_edge_handler = true;
>>> +        break;
>>> +    case 4:
>>> +        type = IRQ_TYPE_LEVEL_HIGH;
>>> +        break;
>>> +    case 8:
>>> +        type = IRQ_TYPE_LEVEL_LOW;
>>> +        break;
>>> +    default:
>>> +        pr_err("Error: Invalid irq trigger specification: %x\n",
>>> +               trigger);
>>> +        type = IRQ_TYPE_LEVEL_LOW;
>>> +        break;
>>> +    }
>>> +
>>> +    irq_set_irq_type(irq, type);
>>> +
>>> +    if (set_edge_handler)
>>> +        __irq_set_handler(irq, handle_edge_irq, 0, NULL);
>>> +
>>> +err_put:
>>> +    of_node_put(oirq.controller);
>>> +err:
>>> +    return irq;
>>> +}
>>> +
>>> +/*
>>> + * irq_create_of_mapping - Hook to resolve OF irq specifier into a
>>> Linux irq#
>>> + *
>>> + * Octeon irq maps are a pair of indexes.  The first selects either
>>> + * ciu0 or ciu1, the second is the bit within the ciu register.
>>> + */
>>> +unsigned int irq_create_of_mapping(struct device_node *controller,
>>> +                   const u32 *intspec, unsigned int intsize)
>>> +{
>>> +    unsigned int irq = 0;
>>> +    unsigned int ciu, bit;
>>> +
>>> +    if (of_device_is_compatible(controller, "cavium,octeon-3860-gpio"))
>>> +        return octeon_irq_gpio_mapping(controller, intspec, intsize);
>>> +
>>> +    ciu = be32_to_cpup(intspec);
>>> +    bit = be32_to_cpup(intspec + 1);
>>> +
>>> +    if (ciu<  8&&  bit<  64)
>>> +        irq = octeon_irq_ciu_to_irq[ciu][bit];
>>> +
>>> +    return irq;
>>> +}
>>> +EXPORT_SYMBOL_GPL(irq_create_of_mapping);
>>
>> Have you looked at irq_domains (kernel/irq/irqdomain.c)? That is what
>> you should be using for your (gpio) interrupt controller and then use
>> the common irq_create_of_mapping.
>>
> 
> Unfortunatly, although a good idea, kernel/irq/irqdomain.c makes a bunch
> of assumptions that don't hold for Octeon.  We may be able to improve it
> so that it flexible enough to suit us.
> 
> 
> Here are the problems I see:
> 
> 1) It is assumed that there is some sort of linear correspondence
> between 'hwirq' and 'irq', and that the range of valid values is
> contiguous.

That's not true if you implement .to_irq for your domain.

> 2) It is assumed that the concepts of nr_irq, irq_base and hwirq_base
> have easy to determin values and you can do iteration over their ranges
> by adding indexes to the bases.

That's true for hwirq numbering, but not for Linux irq numbering.
irq_base is only used if you don't provide .to_irq. hwirq_base is just
to allow a linear range that doesn't start with 0.

> I think we can fix this by adding iteration helper functions to struct
> irq_domain.  If these are present, we would just ignore the irq_base,
> nr_irq and hwirq_base elements of the structure.
> irq_domain_for_each_hwirq() and irq_domain_for_each_irq() would be
> modified to use the iteration helper functions.

Expanding irqdomain is certainly the intention and right direction.

It seems to me you just need something different than an irq. controller
with a linear range of irq lines. Perhaps sparsely allocated hwirq's?
Can you describe the h/w?

Also, you could have a different domain per irqdesc if you really needed to.

Rob

  reply	other threads:[~2011-11-11 20:41 UTC|newest]

Thread overview: 24+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-11-11  2:21 [PATCH 0/8] of/MIPS/i2c/net: Convert OCTEON to use device-tree ddaney.cavm
2011-11-11  2:21 ` ddaney.cavm
2011-11-11  2:21 ` ddaney.cavm
2011-11-11  2:21 ` [PATCH 1/8] MIPS: Octeon: Add device tree source files ddaney.cavm
2011-11-11  2:21 ` [PATCH 2/8] MIPS: Prune some target specific code out of prom.c ddaney.cavm
2011-11-11  2:21   ` ddaney.cavm-Re5JQEeQqe8AvxtiuMwx3w
2011-11-11  2:21 ` [PATCH 3/8] MIPS: Octeon: Add irq_create_of_mapping() and GPIO interrupts ddaney.cavm
2011-11-11 16:08   ` Rob Herring
2011-11-11 16:08     ` Rob Herring
2011-11-11 19:49     ` David Daney
2011-11-11 19:49       ` David Daney
2011-11-11 20:41       ` Rob Herring [this message]
2011-11-11 20:41         ` Rob Herring
2011-11-11 20:45         ` David Daney
2011-11-11 20:45           ` David Daney
2011-11-11  2:22 ` [PATCH 4/8] MIPS: Octeon: Initialize and fixup device tree ddaney.cavm
2011-11-11  2:22   ` ddaney.cavm-Re5JQEeQqe8AvxtiuMwx3w
     [not found] ` <1320978124-13042-1-git-send-email-ddaney.cavm-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2011-11-11  2:22   ` [PATCH 5/8] i2c: Convert i2c-octeon.c to use " ddaney.cavm-Re5JQEeQqe8AvxtiuMwx3w
2011-11-11  2:22     ` ddaney.cavm
2011-11-11  2:22 ` [PATCH 6/8] netdev: mdio-octeon.c: Convert " ddaney.cavm
2011-11-11  2:22   ` ddaney.cavm-Re5JQEeQqe8AvxtiuMwx3w
2011-11-11  2:22 ` [PATCH 7/8] netdev: octeon_mgmt: " ddaney.cavm
2011-11-11  2:22 ` [PATCH 8/8] staging: octeon_ethernet: " ddaney.cavm
2011-11-27  2:04   ` Greg KH

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4EBD8867.30205@gmail.com \
    --to=robherring2@gmail.com \
    --cc=david.daney@cavium.com \
    --cc=ddaney.cavm@gmail.com \
    --cc=devicetree-discuss@lists.ozlabs.org \
    --cc=grant.likely@secretlab.ca \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mips@linux-mips.org \
    --cc=ralf@linux-mips.org \
    --cc=tglx@linutronix.de \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.