Re: [PATCH 2/6] irqchip/armada-370-xp: Implement SoC Error interrupts

[Marc Zyngier <maz@kernel.org>]

On Fri, 06 May 2022 19:30:51 +0100,
Pali Rohár <pali@kernel.org> wrote:
> 
> On Friday 06 May 2022 19:19:46 Marc Zyngier wrote:
> > On Fri, 06 May 2022 14:40:25 +0100,
> > Pali Rohár <pali@kernel.org> wrote:
> > > 
> > > MPIC IRQ 4 is used as SoC Error Summary interrupt and provides access to
> > > another hierarchy of SoC Error interrupts. Implement a new IRQ chip and
> > > domain for accessing this IRQ hierarchy.
> > > 
> > > Signed-off-by: Pali Rohár <pali@kernel.org>
> > > ---
> > >  drivers/irqchip/irq-armada-370-xp.c | 213 +++++++++++++++++++++++++++-
> > >  1 file changed, 210 insertions(+), 3 deletions(-)
> > > 
> > > diff --git a/drivers/irqchip/irq-armada-370-xp.c b/drivers/irqchip/irq-armada-370-xp.c
> > > index ebd76ea1c69b..71578b65f5c8 100644
> > > --- a/drivers/irqchip/irq-armada-370-xp.c
> > > +++ b/drivers/irqchip/irq-armada-370-xp.c
> > > @@ -117,6 +117,8 @@
> > >  /* Registers relative to main_int_base */
> > >  #define ARMADA_370_XP_INT_CONTROL		(0x00)
> > >  #define ARMADA_370_XP_SW_TRIG_INT_OFFS		(0x04)
> > > +#define ARMADA_370_XP_INT_SOC_ERR_0_CAUSE_OFFS	(0x20)
> > > +#define ARMADA_370_XP_INT_SOC_ERR_1_CAUSE_OFFS	(0x24)
> > >  #define ARMADA_370_XP_INT_SET_ENABLE_OFFS	(0x30)
> > >  #define ARMADA_370_XP_INT_CLEAR_ENABLE_OFFS	(0x34)
> > >  #define ARMADA_370_XP_INT_SOURCE_CTL(irq)	(0x100 + irq*4)
> > > @@ -130,6 +132,8 @@
> > >  #define ARMADA_370_XP_CPU_INTACK_OFFS		(0x44)
> > >  #define ARMADA_370_XP_INT_SET_MASK_OFFS		(0x48)
> > >  #define ARMADA_370_XP_INT_CLEAR_MASK_OFFS	(0x4C)
> > > +#define ARMADA_370_XP_INT_SOC_ERR_0_MASK_OFF	(0x50)
> > > +#define ARMADA_370_XP_INT_SOC_ERR_1_MASK_OFF	(0x54)
> > >  #define ARMADA_370_XP_INT_FABRIC_MASK_OFFS	(0x54)
> > >  #define ARMADA_370_XP_INT_CAUSE_PERF(cpu)	(1 << cpu)
> > >  
> > > @@ -146,6 +150,8 @@
> > >  static void __iomem *per_cpu_int_base;
> > >  static void __iomem *main_int_base;
> > >  static struct irq_domain *armada_370_xp_mpic_domain;
> > > +static struct irq_domain *armada_370_xp_soc_err_domain;
> > > +static unsigned int soc_err_irq_num_regs;
> > >  static u32 doorbell_mask_reg;
> > >  static int parent_irq;
> > >  #ifdef CONFIG_PCI_MSI
> > > @@ -156,6 +162,8 @@ static DEFINE_MUTEX(msi_used_lock);
> > >  static phys_addr_t msi_doorbell_addr;
> > >  #endif
> > >  
> > > +static void armada_370_xp_soc_err_irq_unmask(struct irq_data *d);
> > > +
> > >  static inline bool is_percpu_irq(irq_hw_number_t irq)
> > >  {
> > >  	if (irq <= ARMADA_370_XP_MAX_PER_CPU_IRQS)
> > > @@ -509,6 +517,27 @@ static void armada_xp_mpic_reenable_percpu(void)
> > >  		armada_370_xp_irq_unmask(data);
> > >  	}
> > >  
> > > +	/* Re-enable per-CPU SoC Error interrupts that were enabled before suspend */
> > > +	for (irq = 0; irq < soc_err_irq_num_regs * 32; irq++) {
> > > +		struct irq_data *data;
> > > +		int virq;
> > > +
> > > +		virq = irq_linear_revmap(armada_370_xp_soc_err_domain, irq);
> > > +		if (virq == 0)
> > > +			continue;
> > > +
> > > +		data = irq_get_irq_data(virq);
> > > +
> > > +		if (!irq_percpu_is_enabled(virq))
> > > +			continue;
> > > +
> > > +		armada_370_xp_soc_err_irq_unmask(data);
> > > +	}
> > 
> > So you do this loop and all these lookups, both here and in the resume
> > function (duplicated code!) just to be able to call the unmask
> > function?  This would be better served by two straight writes of the
> > mask register, which you'd conveniently save on suspend.
> > 
> > Yes, you have only duplicated the existing logic. But surely there is
> > something better to do.
> 
> Yes, I just used existing logic.
> 
> I'm not rewriting driver or doing big refactor of it, as this is not in
> the scope of the PCIe AER interrupt support.

Fair enough. By the same logic, I'm not taking any change to the
driver until it is put in a better shape. Your call.

> > > +
> > > +	/* Unmask summary SoC Error Interrupt */
> > > +	if (soc_err_irq_num_regs > 0)
> > > +		writel(4, per_cpu_int_base + ARMADA_370_XP_INT_CLEAR_MASK_OFFS);
> > > +
> > >  	ipi_resume();
> > >  }
> > >  
> > > @@ -546,8 +575,8 @@ static struct irq_chip armada_370_xp_irq_chip = {
> > >  static int armada_370_xp_mpic_irq_map(struct irq_domain *h,
> > >  				      unsigned int virq, irq_hw_number_t hw)
> > >  {
> > > -	/* IRQs 0 and 1 cannot be mapped, they are handled internally */
> > > -	if (hw <= 1)
> > > +	/* IRQs 0, 1 and 4 cannot be mapped, they are handled internally */
> > > +	if (hw <= 1 || hw == 4)
> > >  		return -EINVAL;
> > >  
> > >  	armada_370_xp_irq_mask(irq_get_irq_data(virq));
> > > @@ -577,6 +606,99 @@ static const struct irq_domain_ops armada_370_xp_mpic_irq_ops = {
> > >  	.xlate = irq_domain_xlate_onecell,
> > >  };
> > >  
> > > +static DEFINE_RAW_SPINLOCK(armada_370_xp_soc_err_lock);
> > > +
> > > +static void armada_370_xp_soc_err_irq_mask(struct irq_data *d)
> > > +{
> > > +	irq_hw_number_t hwirq = irqd_to_hwirq(d);
> > > +	u32 reg, mask;
> > > +
> > > +	reg = hwirq >= 32 ? ARMADA_370_XP_INT_SOC_ERR_1_MASK_OFF
> > > +			  : ARMADA_370_XP_INT_SOC_ERR_0_MASK_OFF;
> > > +
> > > +	raw_spin_lock(&armada_370_xp_soc_err_lock);
> > > +	mask = readl(per_cpu_int_base + reg);
> > > +	mask &= ~BIT(hwirq % 32);
> > > +	writel(mask, per_cpu_int_base + reg);
> > > +	raw_spin_unlock(&armada_370_xp_soc_err_lock);
> > > +}
> > > +
> > > +static void armada_370_xp_soc_err_irq_unmask(struct irq_data *d)
> > > +{
> > > +	irq_hw_number_t hwirq = irqd_to_hwirq(d);
> > > +	u32 reg, mask;
> > > +
> > > +	reg = hwirq >= 32 ? ARMADA_370_XP_INT_SOC_ERR_1_MASK_OFF
> > > +			  : ARMADA_370_XP_INT_SOC_ERR_0_MASK_OFF;
> > > +
> > > +	raw_spin_lock(&armada_370_xp_soc_err_lock);
> > > +	mask = readl(per_cpu_int_base + reg);
> > > +	mask |= BIT(hwirq % 32);
> > > +	writel(mask, per_cpu_int_base + reg);
> > > +	raw_spin_unlock(&armada_370_xp_soc_err_lock);
> > > +}
> > > +
> > > +static int armada_370_xp_soc_err_irq_mask_on_cpu(void *par)
> > > +{
> > > +	struct irq_data *d = par;
> > > +	armada_370_xp_soc_err_irq_mask(d);
> > > +	return 0;
> > > +}
> > > +
> > > +static int armada_370_xp_soc_err_irq_unmask_on_cpu(void *par)
> > > +{
> > > +	struct irq_data *d = par;
> > > +	armada_370_xp_soc_err_irq_unmask(d);
> > > +	return 0;
> > > +}
> > > +
> > > +static int armada_xp_soc_err_irq_set_affinity(struct irq_data *d,
> > > +					      const struct cpumask *mask,
> > > +					      bool force)
> > > +{
> > > +	unsigned int cpu;
> > > +
> > > +	cpus_read_lock();
> > > +
> > > +	/* First disable IRQ on all cores */
> > > +	for_each_online_cpu(cpu)
> > > +		smp_call_on_cpu(cpu, armada_370_xp_soc_err_irq_mask_on_cpu, d, true);
> > > +
> > > +	/* Select a single core from the affinity mask which is online */
> > > +	cpu = cpumask_any_and(mask, cpu_online_mask);
> > > +	smp_call_on_cpu(cpu, armada_370_xp_soc_err_irq_unmask_on_cpu, d, true);
> > > +
> > > +	cpus_read_unlock();
> > > +
> > > +	irq_data_update_effective_affinity(d, cpumask_of(cpu));
> > > +
> > > +	return IRQ_SET_MASK_OK;
> > > +}
> > 
> > Aren't these per-CPU interrupts anyway? What does it mean to set their
> > affinity? /me rolls eyes...
> 
> Yes, they are per-CPU interrupts. But to mask or unmask particular
> interrupt for specific CPU is possible only from that CPU. CPU 0 just
> cannot move interrupt from CPU 0 to CPU 1. CPU 0 can only mask that
> interrupt and CPU 1 has to unmask it.

And that's no different form other per-CPU interrupts that have the
exact same requirements. NAK to this sort of hacks.

	M.

-- 
Without deviation from the norm, progress is not possible.