From: Thomas Gleixner on
On Sun, 21 Mar 2010, Yinghai Lu wrote:

> Eric pointed out that radix tree version of irq_to_desc will magnify delay on
> the path of handle_irq.
>
> use vector_desc to reduce the calling of irq_to_desc.
>
> next step: need to change all ack, mask, umask, eoi for all irq_chip to take irq_desc

That's not relevant for this change.

>
> -typedef int vector_irq_t[NR_VECTORS];
> -DECLARE_PER_CPU(vector_irq_t, vector_irq);
> -extern void setup_vector_irq(int cpu);
> +typedef struct irq_desc *vector_desc_t[NR_VECTORS];

Why do we need that typedef ? Please use plain struct irq_desc *

> +DECLARE_PER_CPU(vector_desc_t, vector_desc);
> +extern void setup_vector_desc(int cpu);
....
> void destroy_irq(unsigned int irq)
> {
> unsigned long flags;
> + struct irq_desc *desc;
> + struct irq_cfg *cfg;
>
> dynamic_irq_cleanup_keep_chip_data(irq);
>
> free_irte(irq);
> raw_spin_lock_irqsave(&vector_lock, flags);
> - __clear_irq_vector(irq, get_irq_chip_data(irq));
> + desc = irq_to_desc(irq);
> + cfg = desc->chip_data;
> + __clear_irq_vector(desc, cfg);

__clear_irq_vector(desc, desc->chip_data);

should be sufficient, right ?

> raw_spin_unlock_irqrestore(&vector_lock, flags);
> }
>
> @@ -3377,6 +3376,7 @@ void destroy_irq(unsigned int irq)
> static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq,
> struct msi_msg *msg, u8 hpet_id)
> {
> + struct irq_desc *desc;
> struct irq_cfg *cfg;
> int err;
> unsigned dest;
> @@ -3384,8 +3384,9 @@ static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq,
> if (disable_apic)
> return -ENXIO;
>
> - cfg = irq_cfg(irq);
> - err = assign_irq_vector(irq, cfg, apic->target_cpus());
> + desc = irq_to_desc(irq);
> + cfg = desc->chip_data;
> + err = assign_irq_vector(desc, cfg, apic->target_cpus());

Ditto

> if (err)
> return err;
>
> @@ -3876,14 +3877,16 @@ static struct irq_chip ht_irq_chip = {
>
> int arch_setup_ht_irq(unsigned int irq, struct pci_dev *dev)
> {
> + struct irq_desc *desc;
> struct irq_cfg *cfg;
> int err;
>
> if (disable_apic)
> return -ENXIO;
>
> - cfg = irq_cfg(irq);
> - err = assign_irq_vector(irq, cfg, apic->target_cpus());
> + desc = irq_to_desc(irq);
> + cfg = desc->chip_data;
> + err = assign_irq_vector(desc, cfg, apic->target_cpus());

Ditto

> if (!err) {
> struct ht_irq_msg msg;
> unsigned dest;
> diff --git a/arch/x86/kernel/irq.c b/arch/x86/kernel/irq.c
> index 91fd0c7..f71625c 100644
> --- a/arch/x86/kernel/irq.c
> +++ b/arch/x86/kernel/irq.c
> @@ -229,19 +229,19 @@ unsigned int __irq_entry do_IRQ(struct pt_regs *regs)
>
> /* high bit used in ret_from_ code */
> unsigned vector = ~regs->orig_ax;
> - unsigned irq;
> + struct irq_desc *desc;
>
> exit_idle();
> irq_enter();
>
> - irq = __get_cpu_var(vector_irq)[vector];
> + desc = __get_cpu_var(vector_desc)[vector];
>
> - if (!handle_irq(irq, regs)) {
> + if (!handle_irq(desc, regs)) {
> ack_APIC_irq();
>
> if (printk_ratelimit())
> - pr_emerg("%s: %d.%d No irq handler for vector (irq %d)\n",
> - __func__, smp_processor_id(), vector, irq);
> + pr_emerg("%s: %d.%d No irq handler for vector\n",

That printk is confusing. It's not lacking an irq handler. The
vector is simply not assigned.

Thanks,

tglx
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: Eric W. Biederman on
Thomas Gleixner <tglx(a)linutronix.de> writes:

> On Sun, 21 Mar 2010, Yinghai Lu wrote:
>
>> Eric pointed out that radix tree version of irq_to_desc will magnify delay on
>> the path of handle_irq.
>>
>> use vector_desc to reduce the calling of irq_to_desc.
>>
>> next step: need to change all ack, mask, umask, eoi for all irq_chip to take irq_desc
>
> That's not relevant for this change.
>
>>
>> -typedef int vector_irq_t[NR_VECTORS];
>> -DECLARE_PER_CPU(vector_irq_t, vector_irq);
>> -extern void setup_vector_irq(int cpu);
>> +typedef struct irq_desc *vector_desc_t[NR_VECTORS];
>
> Why do we need that typedef ? Please use plain struct irq_desc *

Well at least originally DECLARE_PER_CPU chocked when given a complex
type. Does:
DECLARE_PER_CPU(struct irq_desc *[NR_VECTORS], vector_desc);
work?


>> +DECLARE_PER_CPU(vector_desc_t, vector_desc);
>> +extern void setup_vector_desc(int cpu);
> ...
>> void destroy_irq(unsigned int irq)
>> {
>> unsigned long flags;
>> + struct irq_desc *desc;
>> + struct irq_cfg *cfg;
>>
>> dynamic_irq_cleanup_keep_chip_data(irq);
>>
>> free_irte(irq);
>> raw_spin_lock_irqsave(&vector_lock, flags);
>> - __clear_irq_vector(irq, get_irq_chip_data(irq));
>> + desc = irq_to_desc(irq);
>> + cfg = desc->chip_data;
>> + __clear_irq_vector(desc, cfg);
>
> __clear_irq_vector(desc, desc->chip_data);
>
> should be sufficient, right ?

You want to deliberately loose a modicum of type safety?


>
>> raw_spin_unlock_irqrestore(&vector_lock, flags);
>> }
>>
>> @@ -3377,6 +3376,7 @@ void destroy_irq(unsigned int irq)
>> static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq,
>> struct msi_msg *msg, u8 hpet_id)
>> {
>> + struct irq_desc *desc;
>> struct irq_cfg *cfg;
>> int err;
>> unsigned dest;
>> @@ -3384,8 +3384,9 @@ static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq,
>> if (disable_apic)
>> return -ENXIO;
>>
>> - cfg = irq_cfg(irq);
>> - err = assign_irq_vector(irq, cfg, apic->target_cpus());
>> + desc = irq_to_desc(irq);
>> + cfg = desc->chip_data;
>> + err = assign_irq_vector(desc, cfg, apic->target_cpus());
>
> Ditto
>
>> if (err)
>> return err;
>>
>> @@ -3876,14 +3877,16 @@ static struct irq_chip ht_irq_chip = {
>>
>> int arch_setup_ht_irq(unsigned int irq, struct pci_dev *dev)
>> {
>> + struct irq_desc *desc;
>> struct irq_cfg *cfg;
>> int err;
>>
>> if (disable_apic)
>> return -ENXIO;
>>
>> - cfg = irq_cfg(irq);
>> - err = assign_irq_vector(irq, cfg, apic->target_cpus());
>> + desc = irq_to_desc(irq);
>> + cfg = desc->chip_data;
>> + err = assign_irq_vector(desc, cfg, apic->target_cpus());
>
> Ditto
>
>> if (!err) {
>> struct ht_irq_msg msg;
>> unsigned dest;
>> diff --git a/arch/x86/kernel/irq.c b/arch/x86/kernel/irq.c
>> index 91fd0c7..f71625c 100644
>> --- a/arch/x86/kernel/irq.c
>> +++ b/arch/x86/kernel/irq.c
>> @@ -229,19 +229,19 @@ unsigned int __irq_entry do_IRQ(struct pt_regs *regs)
>>
>> /* high bit used in ret_from_ code */
>> unsigned vector = ~regs->orig_ax;
>> - unsigned irq;
>> + struct irq_desc *desc;
>>
>> exit_idle();
>> irq_enter();
>>
>> - irq = __get_cpu_var(vector_irq)[vector];
>> + desc = __get_cpu_var(vector_desc)[vector];
>>
>> - if (!handle_irq(irq, regs)) {
>> + if (!handle_irq(desc, regs)) {
>> ack_APIC_irq();
>>
>> if (printk_ratelimit())
>> - pr_emerg("%s: %d.%d No irq handler for vector (irq %d)\n",
>> - __func__, smp_processor_id(), vector, irq);
>> + pr_emerg("%s: %d.%d No irq handler for vector\n",
>
> That printk is confusing. It's not lacking an irq handler. The
> vector is simply not assigned.

Long evolution. Do you have a suggestion of better wording?

Eric
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: Thomas Gleixner on
> >> -typedef int vector_irq_t[NR_VECTORS];
> >> -DECLARE_PER_CPU(vector_irq_t, vector_irq);
> >> -extern void setup_vector_irq(int cpu);
> >> +typedef struct irq_desc *vector_desc_t[NR_VECTORS];
> >
> > Why do we need that typedef ? Please use plain struct irq_desc *
>
> Well at least originally DECLARE_PER_CPU chocked when given a complex
> type. Does:
> DECLARE_PER_CPU(struct irq_desc *[NR_VECTORS], vector_desc);
> work?

Hmm, I thought that was fixed, but I might be wrong as usual.

>
> >> +DECLARE_PER_CPU(vector_desc_t, vector_desc);
> >> +extern void setup_vector_desc(int cpu);
> > ...
> >> void destroy_irq(unsigned int irq)
> >> {
> >> unsigned long flags;
> >> + struct irq_desc *desc;
> >> + struct irq_cfg *cfg;
> >>
> >> dynamic_irq_cleanup_keep_chip_data(irq);
> >>
> >> free_irte(irq);
> >> raw_spin_lock_irqsave(&vector_lock, flags);
> >> - __clear_irq_vector(irq, get_irq_chip_data(irq));
> >> + desc = irq_to_desc(irq);
> >> + cfg = desc->chip_data;
> >> + __clear_irq_vector(desc, cfg);
> >
> > __clear_irq_vector(desc, desc->chip_data);
> >
> > should be sufficient, right ?
>
> You want to deliberately loose a modicum of type safety?

I really have a hard time to see how assigning a void pointer to a
struct irq_cfg pointer is anymore type safe than using the void
pointer as for the function argument right away.

> >> if (printk_ratelimit())
> >> - pr_emerg("%s: %d.%d No irq handler for vector (irq %d)\n",
> >> - __func__, smp_processor_id(), vector, irq);
> >> + pr_emerg("%s: %d.%d No irq handler for vector\n",
> >
> > That printk is confusing. It's not lacking an irq handler. The
> > vector is simply not assigned.
>
> Long evolution. Do you have a suggestion of better wording?

You mean hysterical raisins. Ok, how about:

pr_emerg("irq: %d.d irq vector not assigned\n", ...);

Thanks,

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