[v9,05/11] x86/PCI: add pci_mmconfig_insert()/delete() for PCI root bridge hotplug

Introduce pci_mmconfig_insert()/pci_mmconfig_delete(), which will be used
to update MMCFG information when supporting PCI root bridge hotplug.

[bhelgaas: KERN_INFO, not KERN_ERR, for missing ACPI PNP0C01/02 reservation]
Signed-off-by: Jiang Liu <liuj97@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
---
 arch/x86/include/asm/pci_x86.h |    4 +
 arch/x86/pci/mmconfig-shared.c |  213 ++++++++++++++++++++++++++++++++--------
 2 files changed, 177 insertions(+), 40 deletions(-)

On Thu, Jun 21, 2012 at 2:36 AM, Jiang Liu <jiang.liu@huawei.com> wrote:
> Introduce pci_mmconfig_insert()/pci_mmconfig_delete(), which will be used
> to update MMCFG information when supporting PCI root bridge hotplug.
>
> [bhelgaas: KERN_INFO, not KERN_ERR, for missing ACPI PNP0C01/02 reservation]
> Signed-off-by: Jiang Liu <liuj97@gmail.com>
> Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
> ---
>  arch/x86/include/asm/pci_x86.h |    4 +
>  arch/x86/pci/mmconfig-shared.c |  213 ++++++++++++++++++++++++++++++++--------
>  2 files changed, 177 insertions(+), 40 deletions(-)

this one is way to big, you may chop it to two or three.

>
> diff --git a/arch/x86/include/asm/pci_x86.h b/arch/x86/include/asm/pci_x86.h
> index df898ce..af5018f 100644
> --- a/arch/x86/include/asm/pci_x86.h
> +++ b/arch/x86/include/asm/pci_x86.h
> @@ -137,6 +137,10 @@ extern int __init pci_mmcfg_arch_init(void);
>  extern void __init pci_mmcfg_arch_free(void);
>  extern int __devinit pci_mmcfg_arch_map(struct pci_mmcfg_region *cfg);
>  extern void pci_mmcfg_arch_unmap(struct pci_mmcfg_region *cfg);
> +extern int __devinit pci_mmconfig_insert(struct device *dev,
> +                                        u16 seg, u8 start,
> +                                        u8 end, phys_addr_t addr);
> +extern int pci_mmconfig_delete(u16 seg, u8 start, u8 end);
>  extern struct pci_mmcfg_region *pci_mmconfig_lookup(int segment, int bus);
>
>  extern struct list_head pci_mmcfg_list;
> diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c
> index 0ac97d5..19fc42b 100644
> --- a/arch/x86/pci/mmconfig-shared.c
> +++ b/arch/x86/pci/mmconfig-shared.c
> @@ -27,6 +27,8 @@
>
>  /* Indicate if the mmcfg resources have been placed into the resource table. */
>  static int __initdata pci_mmcfg_resources_inserted;
> +static bool pci_mmcfg_running_state;
> +static bool pci_mmcfg_arch_init_failed;
>  static DEFINE_MUTEX(pci_mmcfg_lock);
>
>  LIST_HEAD(pci_mmcfg_list);
> @@ -91,10 +93,6 @@ static __devinit struct pci_mmcfg_region *pci_mmconfig_alloc(int segment,
>                 "PCI MMCONFIG %04x [bus %02x-%02x]", segment, start, end);
>        res->name = new->name;
>
> -       printk(KERN_INFO PREFIX "MMCONFIG for domain %04x [bus %02x-%02x] at "
> -              "%pR (base %#lx)\n", segment, start, end, &new->res,
> -              (unsigned long) addr);
> -
>        return new;
>  }
>
> @@ -108,6 +106,11 @@ static __init struct pci_mmcfg_region *pci_mmconfig_add(int segment, int start,
>                mutex_lock(&pci_mmcfg_lock);
>                list_add_sorted(new);
>                mutex_unlock(&pci_mmcfg_lock);
> +
> +               printk(KERN_INFO PREFIX
> +                      "MMCONFIG for domain %04x [bus %02x-%02x] at %pR "
> +                      "(base %#lx)\n",
> +                      segment, start, end, &new->res, (unsigned long)addr);
>        }
>
>        return new;
> @@ -375,14 +378,15 @@ static void __init pci_mmcfg_insert_resources(void)
>        struct pci_mmcfg_region *cfg;
>
>        list_for_each_entry(cfg, &pci_mmcfg_list, list)
> -               insert_resource(&iomem_resource, &cfg->res);
> +               if (!cfg->res.parent)
> +                       insert_resource(&iomem_resource, &cfg->res);
>
>        /* Mark that the resources have been inserted. */
>        pci_mmcfg_resources_inserted = 1;
>  }
>
> -static acpi_status __init check_mcfg_resource(struct acpi_resource *res,
> -                                             void *data)
> +static acpi_status __devinit check_mcfg_resource(struct acpi_resource *res,
> +                                                void *data)
>  {
>        struct resource *mcfg_res = data;
>        struct acpi_resource_address64 address;
> @@ -418,8 +422,8 @@ static acpi_status __init check_mcfg_resource(struct acpi_resource *res,
>        return AE_OK;
>  }
>
> -static acpi_status __init find_mboard_resource(acpi_handle handle, u32 lvl,
> -               void *context, void **rv)
> +static acpi_status __devinit find_mboard_resource(acpi_handle handle, u32 lvl,
> +                                                 void *context, void **rv)
>  {
>        struct resource *mcfg_res = context;
>
> @@ -432,7 +436,7 @@ static acpi_status __init find_mboard_resource(acpi_handle handle, u32 lvl,
>        return AE_OK;
>  }
>
> -static int __init is_acpi_reserved(u64 start, u64 end, unsigned not_used)
> +static int __devinit is_acpi_reserved(u64 start, u64 end, unsigned not_used)
>  {
>        struct resource mcfg_res;
>
> @@ -451,13 +455,15 @@ static int __init is_acpi_reserved(u64 start, u64 end, unsigned not_used)
>
>  typedef int (*check_reserved_t)(u64 start, u64 end, unsigned type);
>
> -static int __init is_mmconf_reserved(check_reserved_t is_reserved,
> -                                   struct pci_mmcfg_region *cfg, int with_e820)
> +static int __ref is_mmconf_reserved(check_reserved_t is_reserved,
> +                                   struct pci_mmcfg_region *cfg,
> +                                   struct device *dev, int with_e820)
>  {
>        u64 addr = cfg->res.start;
>        u64 size = resource_size(&cfg->res);
>        u64 old_size = size;
> -       int valid = 0, num_buses;
> +       int num_buses;
> +       char *method = with_e820 ? "E820" : "ACPI motherboard resources";
>
>        while (!is_reserved(addr, addr + size, E820_RESERVED)) {
>                size >>= 1;
> @@ -465,49 +471,75 @@ static int __init is_mmconf_reserved(check_reserved_t is_reserved,
>                        break;
>        }
>
> -       if (size >= (16UL<<20) || size == old_size) {
> -               printk(KERN_INFO PREFIX "MMCONFIG at %pR reserved in %s\n",
> -                      &cfg->res,
> -                      with_e820 ? "E820" : "ACPI motherboard resources");
> -               valid = 1;
> -
> -               if (old_size != size) {
> -                       /* update end_bus */
> -                       cfg->end_bus = cfg->start_bus + ((size>>20) - 1);
> -                       num_buses = cfg->end_bus - cfg->start_bus + 1;
> -                       cfg->res.end = cfg->res.start +
> -                           PCI_MMCFG_BUS_OFFSET(num_buses) - 1;
> -                       snprintf(cfg->name, PCI_MMCFG_RESOURCE_NAME_LEN,
> -                                "PCI MMCONFIG %04x [bus %02x-%02x]",
> -                                cfg->segment, cfg->start_bus, cfg->end_bus);
> +       if (size < (16UL<<20) && size != old_size)
> +               return 0;
> +
> +       if (dev)
> +               dev_info(dev, "MMCONFIG at %pR reserved in %s\n",
> +                        &cfg->res, method);
> +       else
> +               printk(KERN_INFO PREFIX
> +                      "MMCONFIG at %pR reserved in %s\n",
> +                      &cfg->res, method);
> +
> +       if (old_size != size) {
> +               /* update end_bus */
> +               cfg->end_bus = cfg->start_bus + ((size>>20) - 1);
> +               num_buses = cfg->end_bus - cfg->start_bus + 1;
> +               cfg->res.end = cfg->res.start +
> +                   PCI_MMCFG_BUS_OFFSET(num_buses) - 1;
> +               snprintf(cfg->name, PCI_MMCFG_RESOURCE_NAME_LEN,
> +                        "PCI MMCONFIG %04x [bus %02x-%02x]",
> +                        cfg->segment, cfg->start_bus, cfg->end_bus);
> +
> +               if (dev)
> +                       dev_info(dev,
> +                               "MMCONFIG "
> +                               "at %pR (base %#lx) (size reduced!)\n",
> +                               &cfg->res, (unsigned long) cfg->address);
> +               else
>                        printk(KERN_INFO PREFIX
> -                              "MMCONFIG for %04x [bus%02x-%02x] "
> -                              "at %pR (base %#lx) (size reduced!)\n",
> -                              cfg->segment, cfg->start_bus, cfg->end_bus,
> -                              &cfg->res, (unsigned long) cfg->address);
> -               }
> +                               "MMCONFIG for %04x [bus%02x-%02x] "
> +                               "at %pR (base %#lx) (size reduced!)\n",
> +                               cfg->segment, cfg->start_bus, cfg->end_bus,
> +                               &cfg->res, (unsigned long) cfg->address);
>        }
>
> -       return valid;
> +       return 1;
>  }
>
> -static int __devinit pci_mmcfg_check_reserved(struct pci_mmcfg_region *cfg,
> -                                             int early)
> +static int __ref pci_mmcfg_check_reserved(struct device *dev,
> +                 struct pci_mmcfg_region *cfg, int early)
>  {
>        if (!early && !acpi_disabled) {
> -               if (is_mmconf_reserved(is_acpi_reserved, cfg, 0))
> +               if (is_mmconf_reserved(is_acpi_reserved, cfg, dev, 0))
>                        return 1;
> +
> +               if (dev)
> +                       dev_info(dev, FW_INFO
> +                                "MMCONFIG at %pR not reserved in "
> +                                "ACPI motherboard resources\n",
> +                                &cfg->res);
>                else
> -                       printk(KERN_ERR FW_BUG PREFIX
> +                       printk(KERN_INFO FW_INFO PREFIX
>                               "MMCONFIG at %pR not reserved in "
>                               "ACPI motherboard resources\n",
>                               &cfg->res);
>        }
>
> +       /*
> +        * e820_all_mapped() is marked as __init.
> +        * All entries from ACPI MCFG table have been checked at boot time.
> +        * For MCFG information constructed from hotpluggable host bridge's
> +        * _CBA method, just assume it's reserved.
> +        */
> +       if (pci_mmcfg_running_state)
> +               return 1;
> +
>        /* Don't try to do this check unless configuration
>           type 1 is available. how about type 2 ?*/
>        if (raw_pci_ops)
> -               return is_mmconf_reserved(e820_all_mapped, cfg, 1);
> +               return is_mmconf_reserved(e820_all_mapped, cfg, dev, 1);
>
>        return 0;
>  }
> @@ -517,7 +549,7 @@ static void __init pci_mmcfg_reject_broken(int early)
>        struct pci_mmcfg_region *cfg;
>
>        list_for_each_entry(cfg, &pci_mmcfg_list, list) {
> -               if (pci_mmcfg_check_reserved(cfg, early) == 0) {
> +               if (pci_mmcfg_check_reserved(NULL, cfg, early) == 0) {
>                        printk(KERN_INFO PREFIX "not using MMCONFIG\n");
>                        free_all_mmcfg();
>                        return;
> @@ -641,6 +673,7 @@ static void __init __pci_mmcfg_init(int early)
>                 * the architecture mmcfg setup could not initialize.
>                 */
>                pci_mmcfg_resources_inserted = 1;
> +               pci_mmcfg_arch_init_failed = true;
>        }
>  }
>
> @@ -656,6 +689,8 @@ void __init pci_mmcfg_late_init(void)
>
>  static int __init pci_mmcfg_late_insert_resources(void)
>  {
> +       pci_mmcfg_running_state = true;
> +
>        /*
>         * If resources are already inserted or we are not using MMCONFIG,
>         * don't insert the resources.
> @@ -681,3 +716,101 @@ static int __init pci_mmcfg_late_insert_resources(void)
>  * with other system resources.
>  */
>  late_initcall(pci_mmcfg_late_insert_resources);
> +
> +/* Add MMCFG information for host bridges */
> +int __devinit pci_mmconfig_insert(struct device *dev,
> +                                 u16 seg, u8 start, u8 end,
> +                                 phys_addr_t addr)
> +{
> +       int rc;
> +       struct resource *tmp = NULL;
> +       struct pci_mmcfg_region *cfg;
> +
> +       if (!(pci_probe & PCI_PROBE_MMCONF) || pci_mmcfg_arch_init_failed)
> +               return -ENODEV;
> +
> +       if (start > end)
> +               return -EINVAL;
> +
> +       mutex_lock(&pci_mmcfg_lock);
> +       cfg = pci_mmconfig_lookup(seg, start);
> +       if (cfg) {
> +               if (cfg->end_bus < end)
> +                       dev_info(dev, FW_INFO
> +                                "MMCONFIG for "
> +                                "domain %04x [bus %02x-%02x] "
> +                                "only partially covers this bridge\n",
> +                                 cfg->segment, cfg->start_bus, cfg->end_bus);
> +               mutex_unlock(&pci_mmcfg_lock);
> +               return -EEXIST;
> +       }
> +
> +       if (!addr) {
> +               mutex_unlock(&pci_mmcfg_lock);
> +               return -EINVAL;
> +       }
> +
> +       rc = -EBUSY;
> +       cfg = pci_mmconfig_alloc(seg, start, end, addr);
> +       if (cfg == NULL) {
> +               dev_warn(dev, "fail to add MMCONFIG (out of memory)\n");
> +               rc = -ENOMEM;
> +       } else if (!pci_mmcfg_check_reserved(dev, cfg, 0)) {
> +               dev_warn(dev, FW_BUG "MMCONFIG %pR isn't reserved\n",
> +                        &cfg->res);
> +       } else {
> +               /* Insert resource if it's not in boot stage */
> +               if (pci_mmcfg_running_state)
> +                       tmp = insert_resource_conflict(&iomem_resource,
> +                                                      &cfg->res);
> +
> +               if (tmp) {
> +                       dev_warn(dev,
> +                                "MMCONFIG %pR conflicts with "
> +                                "%s %pR\n",
> +                                &cfg->res, tmp->name, tmp);
> +               } else if (pci_mmcfg_arch_map(cfg)) {
> +                       dev_warn(dev, "fail to map MMCONFIG %pR.\n",
> +                                &cfg->res);
> +               } else {
> +                       list_add_sorted(cfg);
> +                       dev_info(dev, "MMCONFIG at %pR (base %#lx)\n",
> +                                &cfg->res, (unsigned long)addr);
> +                       cfg = NULL;
> +                       rc = 0;
> +               }
> +       }
> +
> +       if (cfg) {
> +               if (cfg->res.parent)
> +                       release_resource(&cfg->res);
> +               kfree(cfg);
> +       }
> +
> +       mutex_unlock(&pci_mmcfg_lock);
> +
> +       return rc;
> +}
> +
> +/* Delete MMCFG information for host bridges */
> +int pci_mmconfig_delete(u16 seg, u8 start, u8 end)
> +{
> +       struct pci_mmcfg_region *cfg;
> +
> +       mutex_lock(&pci_mmcfg_lock);
> +       list_for_each_entry_rcu(cfg, &pci_mmcfg_list, list)
> +               if (cfg->segment == seg && cfg->start_bus == start &&
> +                   cfg->end_bus == end) {
> +                       list_del_rcu(&cfg->list);
> +                       synchronize_rcu();
> +                       pci_mmcfg_arch_unmap(cfg);
> +                       if (cfg->res.parent)
> +                               release_resource(&cfg->res);
> +                       mutex_unlock(&pci_mmcfg_lock);
> +                       kfree(cfg);
> +                       return 0;
> +               }
> +       mutex_unlock(&pci_mmcfg_lock);
> +
> +       return -ENOENT;
> +}
> --
> 1.7.1
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-pci" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html