On Wed, Jun 05, 2019 at 09:54:13AM +0300, Talel Shenhar wrote: > The FIC supports either a (single) wired output, or generation of an MSI-X > interrupt per input (for cases where it is embedded in a PCIe device, > hence, allowing the PCIe drivers to call this API). > This patch introduces the support for allowing the configuration of MSI-X > instead of a wire interrupt. > > Signed-off-by: Talel Shenhar <talel@xxxxxxxxxx> > --- > drivers/irqchip/irq-al-fic.c | 48 +++++++++++++++++++++++++++++++++++++++--- > include/linux/irqchip/al-fic.h | 2 ++ > 2 files changed, 47 insertions(+), 3 deletions(-) > > diff --git a/drivers/irqchip/irq-al-fic.c b/drivers/irqchip/irq-al-fic.c > index d881d42..e49b912 100644 > --- a/drivers/irqchip/irq-al-fic.c > +++ b/drivers/irqchip/irq-al-fic.c > @@ -19,6 +19,7 @@ > #define AL_FIC_MASK 0x10 > #define AL_FIC_CONTROL 0x28 > > +#define CONTROL_AUTO_CLEAR BIT(2) > #define CONTROL_TRIGGER_RISING BIT(3) > #define CONTROL_MASK_MSI_X BIT(5) > > @@ -193,9 +194,11 @@ struct irq_domain *al_fic_wire_get_domain(struct al_fic *fic) > } > EXPORT_SYMBOL_GPL(al_fic_wire_get_domain); > > -static void al_fic_hw_init(struct al_fic *fic) > +static void al_fic_hw_init(struct al_fic *fic, > + int use_msi) > { > - u32 control = CONTROL_MASK_MSI_X; > + u32 control = (use_msi ? (CONTROL_AUTO_CLEAR | CONTROL_TRIGGER_RISING) : > + CONTROL_MASK_MSI_X); > > /* mask out all interrupts */ > writel(0xFFFFFFFF, fic->base + AL_FIC_MASK); > @@ -240,7 +243,7 @@ struct al_fic *al_fic_wire_init(struct device_node *node, > fic->parent_irq = parent_irq; > fic->name = (name ?: "al-fic-wire"); > > - al_fic_hw_init(fic); > + al_fic_hw_init(fic, false); > > ret = al_fic_register(node, fic); > if (ret) { > @@ -260,6 +263,45 @@ struct al_fic *al_fic_wire_init(struct device_node *node, > EXPORT_SYMBOL_GPL(al_fic_wire_init); > > /** > + * al_fic_msi_x_init() - initialize and configure fic in msi-x mode > + * @base: mmio to fic register > + * @name: name of the fic > + * > + * This API will configure the fic hardware to to work in msi-x mode. > + * msi-x fic is to be configured for fics that are embedded inside AL PCIE EP. > + * Those kind of fic are aware of the fact that they live inside PCIE and > + * familiar with the MSI-X table which is configured as part of > + * pci_enable_msix_range() and friends. > + * Interrupt can be generated based on a positive edge or level - configuration > + * is to be determined based on connected hardware to this fic. > + * > + * Returns pointer to fic context or ERR_PTR in case of error. > + */ > +struct al_fic *al_fic_msi_x_init(void __iomem *base, > + const char *name) > +{ > + struct al_fic *fic; > + > + if (!base) > + return ERR_PTR(-EINVAL); > + > + fic = kzalloc(sizeof(*fic), GFP_KERNEL); > + if (!fic) > + return ERR_PTR(-ENOMEM); > + > + fic->base = base; > + fic->name = (name ?: "al-fic-full-fledged"); > + > + al_fic_hw_init(fic, true); > + > + pr_debug("%s initialized successfully in Full-Fledged mode\n", > + fic->name); > + > + return fic; > +} > +EXPORT_SYMBOL_GPL(al_fic_msi_x_init); > + > +/** > * al_fic_cleanup() - free all resources allocated by fic > * @fic: pointer to fic context > * > diff --git a/include/linux/irqchip/al-fic.h b/include/linux/irqchip/al-fic.h > index 0833749..a2e89ff 100644 > --- a/include/linux/irqchip/al-fic.h > +++ b/include/linux/irqchip/al-fic.h > @@ -16,6 +16,8 @@ struct al_fic *al_fic_wire_init(struct device_node *node, > void __iomem *base, > const char *name, > unsigned int parent_irq); > +struct al_fic *al_fic_msi_x_init(void __iomem *base, > + const char *name); Who uses this new function? thanks, greg k-h