On Tue, Jul 6, 2021 at 6:18 AM Huacai Chen <chenhuacai@xxxxxxxxxxx> wrote:
>
> Loongson64 based systems are PC-like systems which use PCI/PCIe as its
> I/O bus, This patch adds the PCI host controller support for LoongArch.
>
> Signed-off-by: Huacai Chen <chenhuacai@xxxxxxxxxxx>
> ---
> arch/loongarch/include/asm/pci.h | 124 ++++++++++++++++++
> arch/loongarch/pci/acpi.c | 194 ++++++++++++++++++++++++++++
> arch/loongarch/pci/mmconfig.c | 105 +++++++++++++++
> arch/loongarch/pci/pci-loongson.c | 130 +++++++++++++++++++
> arch/loongarch/pci/pci.c | 207 ++++++++++++++++++++++++++++++
PCI controller support should generally live in drivers/pci/controller/ and
get reviewed by the subsystem maintainers.
> +/*
> + * This file essentially defines the interface between board specific
> + * PCI code and LoongArch common PCI code. Should potentially put into
> + * include/asm/pci.h file.
> + */
> +
> +#include <linux/ioport.h>
> +#include <linux/list.h>
> +
> +extern const struct pci_ops *__read_mostly loongarch_pci_ops;
There is already an abstraction for this in the common code, don't add another.
> +/*
> + * Each pci channel is a top-level PCI bus seem by CPU. A machine with
> + * multiple PCI channels may have multiple PCI host controllers or a
> + * single controller supporting multiple channels.
> + */
> +struct pci_controller {
> + struct list_head list;
> + struct pci_bus *bus;
> + struct device_node *of_node;
> +
> + struct pci_ops *pci_ops;
> + struct resource *mem_resource;
> + unsigned long mem_offset;
> + struct resource *io_resource;
> + unsigned long io_offset;
> + unsigned long io_map_base;
> + struct resource *busn_resource;
> +
> + unsigned int node;
> + unsigned int index;
> + unsigned int need_domain_info;
> +#ifdef CONFIG_ACPI
> + struct acpi_device *companion;
> +#endif
> + phys_addr_t mcfg_addr;
> +};
> +
> +extern void pcibios_add_root_resources(struct list_head *resources);
> +
> +extern phys_addr_t mcfg_addr_init(int domain);
> +
> +#ifdef CONFIG_PCI_DOMAINS
> +static inline void set_pci_need_domain_info(struct pci_controller *hose,
> + int need_domain_info)
> +{
> + hose->need_domain_info = need_domain_info;
> +}
> +#endif /* CONFIG_PCI_DOMAINS */
Just use PCI_DOMAINS unconditionally
> +
> +/*
> + * Can be used to override the logic in pci_scan_bus for skipping
> + * already-configured bus numbers - to be used for buggy BIOSes
> + * or architectures with incomplete PCI setup by the loader
> + */
> +static inline unsigned int pcibios_assign_all_busses(void)
> +{
> + return 1;
> +}
Since you use ACPI, the BIOS should be responsible for assigning the
buses, otherwise the ACPI data may be mismatched with the PCI
device locations that the kernel sees.
> +#define PCIBIOS_MIN_IO 0
I think this means PCI devices can reuse ports that are reserved
for ISA devices. Since you claim to support ISA, I think this should
be 0x1000
> +
> +int raw_pci_read(unsigned int domain, unsigned int bus, unsigned int devfn,
> + int reg, int len, u32 *val)
> +{
> + struct pci_bus *bus_tmp = pci_find_bus(domain, bus);
> +
> + if (bus_tmp)
> + return bus_tmp->ops->read(bus_tmp, devfn, reg, len, val);
> + return -EINVAL;
> +}
> +
> +int raw_pci_write(unsigned int domain, unsigned int bus, unsigned int devfn,
> + int reg, int len, u32 val)
> +{
> + struct pci_bus *bus_tmp = pci_find_bus(domain, bus);
> +
> + if (bus_tmp)
> + return bus_tmp->ops->write(bus_tmp, devfn, reg, len, val);
> + return -EINVAL;
> +}
This looks like you copied from arch/arm64. I think the code really
needs to be generalized more. Maybe move the arm64 implementation
to drivers/acpi/ so it can be shared with loongarch?
> +/*
> + * We need to avoid collisions with `mirrored' VGA ports
> + * and other strange ISA hardware, so we always want the
> + * addresses to be allocated in the 0x000-0x0ff region
> + * modulo 0x400.
> + *
> + * Why? Because some silly external IO cards only decode
> + * the low 10 bits of the IO address. The 0x00-0xff region
> + * is reserved for motherboard devices that decode all 16
> + * bits, so it's ok to allocate at, say, 0x2800-0x28ff,
> + * but we want to try to avoid allocating at 0x2900-0x2bff
> + * which might have be mirrored at 0x0100-0x03ff..
> + */
> +resource_size_t
> +pcibios_align_resource(void *data, const struct resource *res,
> + resource_size_t size, resource_size_t align)
> +{
> + struct pci_dev *dev = data;
> + struct pci_controller *hose = dev->sysdata;
> + resource_size_t start = res->start;
> +
> + if (res->flags & IORESOURCE_IO) {
> + /* Make sure we start at our min on all hoses */
> + if (start < PCIBIOS_MIN_IO + hose->io_resource->start)
> + start = PCIBIOS_MIN_IO + hose->io_resource->start;
> +
> + /*
> + * Put everything into 0x00-0xff region modulo 0x400
> + */
> + if (start & 0x300)
> + start = (start + 0x3ff) & ~0x3ff;
> + } else if (res->flags & IORESOURCE_MEM) {
> + /* Make sure we start at our min on all hoses */
> + if (start < PCIBIOS_MIN_MEM)
> + start = PCIBIOS_MIN_MEM;
> + }
> +
> + return start;
> +}
Same here, please don't add another copy of this function.
Arnd
