Commit de337ee3 authored by Marc Zyngier's avatar Marc Zyngier

irqchip/gic-v2m: Add PCI Multi-MSI support

We'd never implemented Multi-MSI support with GICv2m, because
it is weird and clunky, and you'd think people would rather use
MSI-X.

Turns out there is still plenty of devices out there that rely
on Multi-MSI. Oh well, let's teach that trick to the v2m widget,
it is not a big deal anyway.
Signed-off-by: default avatarMarc Zyngier <marc.zyngier@arm.com>
parent 95a25625
...@@ -94,7 +94,7 @@ static struct irq_chip gicv2m_msi_irq_chip = { ...@@ -94,7 +94,7 @@ static struct irq_chip gicv2m_msi_irq_chip = {
static struct msi_domain_info gicv2m_msi_domain_info = { static struct msi_domain_info gicv2m_msi_domain_info = {
.flags = (MSI_FLAG_USE_DEF_DOM_OPS | MSI_FLAG_USE_DEF_CHIP_OPS | .flags = (MSI_FLAG_USE_DEF_DOM_OPS | MSI_FLAG_USE_DEF_CHIP_OPS |
MSI_FLAG_PCI_MSIX), MSI_FLAG_PCI_MSIX | MSI_FLAG_MULTI_PCI_MSI),
.chip = &gicv2m_msi_irq_chip, .chip = &gicv2m_msi_irq_chip,
}; };
...@@ -155,18 +155,12 @@ static int gicv2m_irq_gic_domain_alloc(struct irq_domain *domain, ...@@ -155,18 +155,12 @@ static int gicv2m_irq_gic_domain_alloc(struct irq_domain *domain,
return 0; return 0;
} }
static void gicv2m_unalloc_msi(struct v2m_data *v2m, unsigned int hwirq) static void gicv2m_unalloc_msi(struct v2m_data *v2m, unsigned int hwirq,
int nr_irqs)
{ {
int pos;
pos = hwirq - v2m->spi_start;
if (pos < 0 || pos >= v2m->nr_spis) {
pr_err("Failed to teardown msi. Invalid hwirq %d\n", hwirq);
return;
}
spin_lock(&v2m_lock); spin_lock(&v2m_lock);
__clear_bit(pos, v2m->bm); bitmap_release_region(v2m->bm, hwirq - v2m->spi_start,
get_count_order(nr_irqs));
spin_unlock(&v2m_lock); spin_unlock(&v2m_lock);
} }
...@@ -174,13 +168,13 @@ static int gicv2m_irq_domain_alloc(struct irq_domain *domain, unsigned int virq, ...@@ -174,13 +168,13 @@ static int gicv2m_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
unsigned int nr_irqs, void *args) unsigned int nr_irqs, void *args)
{ {
struct v2m_data *v2m = NULL, *tmp; struct v2m_data *v2m = NULL, *tmp;
int hwirq, offset, err = 0; int hwirq, offset, i, err = 0;
spin_lock(&v2m_lock); spin_lock(&v2m_lock);
list_for_each_entry(tmp, &v2m_nodes, entry) { list_for_each_entry(tmp, &v2m_nodes, entry) {
offset = find_first_zero_bit(tmp->bm, tmp->nr_spis); offset = bitmap_find_free_region(tmp->bm, tmp->nr_spis,
if (offset < tmp->nr_spis) { get_count_order(nr_irqs));
__set_bit(offset, tmp->bm); if (offset >= 0) {
v2m = tmp; v2m = tmp;
break; break;
} }
...@@ -192,16 +186,21 @@ static int gicv2m_irq_domain_alloc(struct irq_domain *domain, unsigned int virq, ...@@ -192,16 +186,21 @@ static int gicv2m_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
hwirq = v2m->spi_start + offset; hwirq = v2m->spi_start + offset;
err = gicv2m_irq_gic_domain_alloc(domain, virq, hwirq); for (i = 0; i < nr_irqs; i++) {
if (err) { err = gicv2m_irq_gic_domain_alloc(domain, virq + i, hwirq + i);
gicv2m_unalloc_msi(v2m, hwirq); if (err)
return err; goto fail;
}
irq_domain_set_hwirq_and_chip(domain, virq, hwirq, irq_domain_set_hwirq_and_chip(domain, virq + i, hwirq + i,
&gicv2m_irq_chip, v2m); &gicv2m_irq_chip, v2m);
}
return 0; return 0;
fail:
irq_domain_free_irqs_parent(domain, virq, nr_irqs);
gicv2m_unalloc_msi(v2m, hwirq, get_count_order(nr_irqs));
return err;
} }
static void gicv2m_irq_domain_free(struct irq_domain *domain, static void gicv2m_irq_domain_free(struct irq_domain *domain,
...@@ -210,8 +209,7 @@ static void gicv2m_irq_domain_free(struct irq_domain *domain, ...@@ -210,8 +209,7 @@ static void gicv2m_irq_domain_free(struct irq_domain *domain,
struct irq_data *d = irq_domain_get_irq_data(domain, virq); struct irq_data *d = irq_domain_get_irq_data(domain, virq);
struct v2m_data *v2m = irq_data_get_irq_chip_data(d); struct v2m_data *v2m = irq_data_get_irq_chip_data(d);
BUG_ON(nr_irqs != 1); gicv2m_unalloc_msi(v2m, d->hwirq, nr_irqs);
gicv2m_unalloc_msi(v2m, d->hwirq);
irq_domain_free_irqs_parent(domain, virq, nr_irqs); irq_domain_free_irqs_parent(domain, virq, nr_irqs);
} }
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment