From the functionality point of view this series might be split into the
following logic parts: 1. Make MMCONFIG code arch-agnostic which allows all architectures to collect PCI config regions and used when necessary. 2. Move non-arch specific bits to the core code. 3. Use MMCONFIG code and implement generic ACPI based PCI host controller driver. 4. Enable above driver on ARM64
Patches has been built on top of 4.4 and can be found here: git@github.com:semihalf-nowicki-tomasz/linux.git (pci-acpi-v3)
NOTE, this patch set depends on Matthew's patches: http://www.spinics.net/lists/linux-pci/msg45950.html https://github.com/Vality/linux/tree/pci-fixes
This has been tested on Cavium ThunderX server and QEMU. Any help in reviewing and testing is very appreciated.
v2 -> v3 - fix legacy IRQ assigning and IO ports registration - remove reference to arch specific companion device for ia64 - move ACPI PCI host controller driver to pci_root.c - drop generic domain assignment for x86 and ia64 as I am not able to run all necessary test variants - drop patch which cleaned legacy IRQ assignment since it belongs to Mathew's series: https://patchwork.ozlabs.org/patch/557504/ - extend MCFG quirk code - rebased to 4.4
v1 -> v2 - moved non-arch specific piece of code to dirver/acpi/ directory - fixed IO resource handling - introduced PCI config accessors quirks matching - moved ACPI_COMPANION_SET to generic code
Liu Jiang (1): ACPI, PCI: Refine the way to handle translation_offset for ACPI resources
Tomasz Nowicki (20): x86, pci: Reorder logic of pci_mmconfig_insert() function x86, pci, acpi: Move arch-agnostic MMCONFIG (aka ECAM) and ACPI code out of arch/x86/ directory pci, acpi, mcfg: Provide generic implementation of MCFG code initialization. x86, pci: mmconfig_{32,64}.c code refactoring - remove code duplication. x86, pci, ecam: mmconfig_64.c becomes default implementation for ECAM driver. XEN / PCI: Remove the dependence on arch x86 when PCI_MMCONFIG=y pci, acpi, mcfg: Provide default RAW ACPI PCI config space accessors. arm64, acpi: Use empty PCI config space accessors from mcfg.c file. pci, acpi, ecam: Add flag to indicate whether ECAM region was hot added or not. x86, pci: Cleanup platform specific MCFG data using previously added ECAM hot_added flag. pci, acpi: Move ACPI host bridge device companion assignment to core code. x86, ia64, pci: Remove ACPI companion device from platform specific data. pci, acpi: Provide generic way to assign bus domain number. x86, ia64: Include acpi_pci_{add|remove}_bus to the default pcibios_{add|remove}_bus implementation. acpi, mcfg: Implement two calls that might be used to inject/remove MCFG region. x86, acpi, pci: Use equivalent function introduced in previous patch. acpi, mcfg: Add default PCI config accessors implementation and initial support for related quirks. pci, acpi: Support for ACPI based generic PCI host controller init pci, acpi: Match PCI config space accessors against platfrom specific quirks. arm64, pci, acpi: Start using ACPI based PCI host bridge driver for ARM64.
arch/arm64/Kconfig | 7 ++ arch/arm64/kernel/pci.c | 25 ---- arch/ia64/hp/common/sba_iommu.c | 2 +- arch/ia64/include/asm/pci.h | 1 - arch/ia64/pci/pci.c | 52 +++----- arch/ia64/sn/kernel/io_acpi_init.c | 4 +- arch/x86/Kconfig | 4 + arch/x86/include/asm/pci.h | 3 - arch/x86/include/asm/pci_x86.h | 28 +---- arch/x86/pci/acpi.c | 41 ++---- arch/x86/pci/common.c | 10 -- arch/x86/pci/irq.c | 10 -- arch/x86/pci/mmconfig-shared.c | 250 ++++++------------------------------- arch/x86/pci/mmconfig_32.c | 11 +- arch/x86/pci/mmconfig_64.c | 67 +--------- arch/x86/pci/numachip.c | 1 + drivers/acpi/Kconfig | 5 + drivers/acpi/Makefile | 1 + drivers/acpi/mcfg.c | 201 +++++++++++++++++++++++++++++ drivers/acpi/pci_root.c | 136 +++++++++++++++++++- drivers/acpi/resource.c | 12 +- drivers/pci/Kconfig | 10 ++ drivers/pci/Makefile | 5 + drivers/pci/ecam.c | 234 ++++++++++++++++++++++++++++++++++ drivers/pci/pci.c | 29 ++++- drivers/pci/probe.c | 5 + drivers/xen/pci.c | 7 +- include/acpi/acpi_bus.h | 1 + include/asm-generic/vmlinux.lds.h | 7 ++ include/linux/acpi.h | 2 + include/linux/ecam.h | 62 +++++++++ include/linux/pci-acpi.h | 17 +++ 32 files changed, 802 insertions(+), 448 deletions(-) create mode 100644 drivers/acpi/mcfg.c create mode 100644 drivers/pci/ecam.c create mode 100644 include/linux/ecam.h
This patch is the first step for MMCONFIG refactoring process.
Code that uses pci_mmcfg_lock will be moved to common file and become accessible for all architectures. pci_mmconfig_insert() cannot be moved so easily since it is mixing generic mmconfig code with x86 specific logic inside of mutual exclusive block guarded by pci_mmcfg_lock.
To get rid of that constraint, we reorder actions as follow: 1. sanity check for mmconfig region presence, if we already have such region it doesn't make snese to alloc new mmconfig list entry 2. mmconfig entry allocation, no need to lock 3. insertion to iomem_resource has its own lock, no need to wrap it into mutex 4. insertion to mmconfig list can be done as the final step in separate function (candidate for further refactoring) and needs another mmconfig lookup to avoid race condition.
Signed-off-by: Tomasz Nowicki tomasz.nowicki@linaro.org Tested-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com Tested-by: Jeremy Linton jeremy.linton@arm.com --- arch/x86/pci/mmconfig-shared.c | 101 +++++++++++++++++++++++------------------ 1 file changed, 58 insertions(+), 43 deletions(-)
diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index dd30b7e..c8bb9b0 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -720,6 +720,38 @@ static int __init pci_mmcfg_late_insert_resources(void) */ late_initcall(pci_mmcfg_late_insert_resources);
+static int pci_mmconfig_inject(struct pci_mmcfg_region *cfg) +{ + struct pci_mmcfg_region *cfg_conflict; + int err = 0; + + mutex_lock(&pci_mmcfg_lock); + cfg_conflict = pci_mmconfig_lookup(cfg->segment, cfg->start_bus); + if (cfg_conflict) { + if (cfg_conflict->end_bus < cfg->end_bus) + pr_info(FW_INFO "MMCONFIG for " + "domain %04x [bus %02x-%02x] " + "only partially covers this bridge\n", + cfg_conflict->segment, cfg_conflict->start_bus, + cfg_conflict->end_bus); + err = -EEXIST; + goto out; + } + + if (pci_mmcfg_arch_map(cfg)) { + pr_warn("fail to map MMCONFIG %pR.\n", &cfg->res); + err = -ENOMEM; + goto out; + } else { + list_add_sorted(cfg); + pr_info("MMCONFIG at %pR (base %#lx)\n", + &cfg->res, (unsigned long)cfg->address); + } +out: + mutex_unlock(&pci_mmcfg_lock); + return err; +} + /* Add MMCFG information for host bridges */ int pci_mmconfig_insert(struct device *dev, u16 seg, u8 start, u8 end, phys_addr_t addr) @@ -734,63 +766,46 @@ int pci_mmconfig_insert(struct device *dev, u16 seg, u8 start, u8 end, if (start > end) return -EINVAL;
- mutex_lock(&pci_mmcfg_lock); + rcu_read_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); + rcu_read_unlock(); + if (cfg) return -EEXIST; - }
- if (!addr) { - mutex_unlock(&pci_mmcfg_lock); + if (!addr) return -EINVAL; - }
- rc = -EBUSY; cfg = pci_mmconfig_alloc(seg, start, end, addr); - if (cfg == NULL) { + if (!cfg) { dev_warn(dev, "fail to add MMCONFIG (out of memory)\n"); - rc = -ENOMEM; - } else if (!pci_mmcfg_check_reserved(dev, cfg, 0)) { + return -ENOMEM; + } + + rc = -EBUSY; + 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; - } + goto error; }
- if (cfg) { - if (cfg->res.parent) - release_resource(&cfg->res); - kfree(cfg); + /* 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); + goto error; }
- mutex_unlock(&pci_mmcfg_lock); + rc = pci_mmconfig_inject(cfg); + if (!rc) + return 0;
+error: + if (cfg->res.parent) + release_resource(&cfg->res); + kfree(cfg); return rc; }
ECAM standard and MCFG table are architecture independent and it makes sense to share common code across all architectures. Both are going to corresponding files - ecam.c and mcfg.c
While we are here, rename pci_parse_mcfg to acpi_parse_mcfg. We already have acpi_parse_mcfg prototype which is used nowhere. At the same time, we need pci_parse_mcfg been global so acpi_parse_mcfg can be used perfectly here.
Signed-off-by: Tomasz Nowicki tn@semihalf.com Tested-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com Tested-by: Jeremy Linton jeremy.linton@arm.com --- arch/x86/Kconfig | 3 + arch/x86/include/asm/pci_x86.h | 21 ----- arch/x86/pci/acpi.c | 1 + arch/x86/pci/mmconfig-shared.c | 205 +---------------------------------------- arch/x86/pci/mmconfig_32.c | 1 + arch/x86/pci/mmconfig_64.c | 1 + arch/x86/pci/numachip.c | 1 + drivers/acpi/Makefile | 1 + drivers/acpi/mcfg.c | 59 ++++++++++++ drivers/pci/Kconfig | 7 ++ drivers/pci/Makefile | 5 + drivers/pci/ecam.c | 175 +++++++++++++++++++++++++++++++++++ drivers/xen/pci.c | 1 + include/linux/acpi.h | 2 + include/linux/ecam.h | 37 ++++++++ 15 files changed, 299 insertions(+), 221 deletions(-) create mode 100644 drivers/acpi/mcfg.c create mode 100644 drivers/pci/ecam.c create mode 100644 include/linux/ecam.h
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index db3622f..350bd52 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -128,6 +128,7 @@ config X86 select HAVE_MIXED_BREAKPOINTS_REGS select HAVE_OPROFILE select HAVE_OPTPROBES + select HAVE_PCI_ECAM select HAVE_PCSPKR_PLATFORM select HAVE_PERF_EVENTS select HAVE_PERF_EVENTS_NMI @@ -2365,6 +2366,7 @@ config PCI_DIRECT
config PCI_MMCONFIG def_bool y + select PCI_ECAM depends on X86_32 && PCI && (ACPI || SFI) && (PCI_GOMMCONFIG || PCI_GOANY)
config PCI_OLPC @@ -2382,6 +2384,7 @@ config PCI_DOMAINS
config PCI_MMCONFIG bool "Support mmconfig PCI config space access" + select PCI_ECAM depends on X86_64 && PCI && ACPI
config PCI_CNB20LE_QUIRK diff --git a/arch/x86/include/asm/pci_x86.h b/arch/x86/include/asm/pci_x86.h index 16fd8e9..c1c0f37 100644 --- a/arch/x86/include/asm/pci_x86.h +++ b/arch/x86/include/asm/pci_x86.h @@ -123,33 +123,12 @@ extern int pci_legacy_init(void); extern int pcibios_fixup_irq(struct pci_dev *dev, u8 pin);
/* pci-mmconfig.c */ - -/* "PCI MMCONFIG %04x [bus %02x-%02x]" */ -#define PCI_MMCFG_RESOURCE_NAME_LEN (22 + 4 + 2 + 2) - -struct pci_mmcfg_region { - struct list_head list; - struct resource res; - u64 address; - char __iomem *virt; - u16 segment; - u8 start_bus; - u8 end_bus; - char name[PCI_MMCFG_RESOURCE_NAME_LEN]; -}; - extern int __init pci_mmcfg_arch_init(void); extern void __init pci_mmcfg_arch_free(void); extern int pci_mmcfg_arch_map(struct pci_mmcfg_region *cfg); extern void pci_mmcfg_arch_unmap(struct pci_mmcfg_region *cfg); extern int 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; - -#define PCI_MMCFG_BUS_OFFSET(bus) ((bus) << 20)
/* * AMD Fam10h CPUs are buggy, and cannot access MMIO config space diff --git a/arch/x86/pci/acpi.c b/arch/x86/pci/acpi.c index dda4bc1..64caf2b 100644 --- a/arch/x86/pci/acpi.c +++ b/arch/x86/pci/acpi.c @@ -5,6 +5,7 @@ #include <linux/dmi.h> #include <linux/slab.h> #include <linux/pci-acpi.h> +#include <linux/ecam.h> #include <asm/numa.h> #include <asm/pci_x86.h>
diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index c8bb9b0..ce2c2e4 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -18,6 +18,7 @@ #include <linux/slab.h> #include <linux/mutex.h> #include <linux/rculist.h> +#include <linux/ecam.h> #include <asm/e820.h> #include <asm/pci_x86.h> #include <asm/acpi.h> @@ -27,103 +28,6 @@ /* Indicate if the mmcfg resources have been placed into the resource table. */ static bool pci_mmcfg_running_state; static bool pci_mmcfg_arch_init_failed; -static DEFINE_MUTEX(pci_mmcfg_lock); - -LIST_HEAD(pci_mmcfg_list); - -static void __init pci_mmconfig_remove(struct pci_mmcfg_region *cfg) -{ - if (cfg->res.parent) - release_resource(&cfg->res); - list_del(&cfg->list); - kfree(cfg); -} - -static void __init free_all_mmcfg(void) -{ - struct pci_mmcfg_region *cfg, *tmp; - - pci_mmcfg_arch_free(); - list_for_each_entry_safe(cfg, tmp, &pci_mmcfg_list, list) - pci_mmconfig_remove(cfg); -} - -static void list_add_sorted(struct pci_mmcfg_region *new) -{ - struct pci_mmcfg_region *cfg; - - /* keep list sorted by segment and starting bus number */ - list_for_each_entry_rcu(cfg, &pci_mmcfg_list, list) { - if (cfg->segment > new->segment || - (cfg->segment == new->segment && - cfg->start_bus >= new->start_bus)) { - list_add_tail_rcu(&new->list, &cfg->list); - return; - } - } - list_add_tail_rcu(&new->list, &pci_mmcfg_list); -} - -static struct pci_mmcfg_region *pci_mmconfig_alloc(int segment, int start, - int end, u64 addr) -{ - struct pci_mmcfg_region *new; - struct resource *res; - - if (addr == 0) - return NULL; - - new = kzalloc(sizeof(*new), GFP_KERNEL); - if (!new) - return NULL; - - new->address = addr; - new->segment = segment; - new->start_bus = start; - new->end_bus = end; - - res = &new->res; - res->start = addr + PCI_MMCFG_BUS_OFFSET(start); - res->end = addr + PCI_MMCFG_BUS_OFFSET(end + 1) - 1; - res->flags = IORESOURCE_MEM | IORESOURCE_BUSY; - snprintf(new->name, PCI_MMCFG_RESOURCE_NAME_LEN, - "PCI MMCONFIG %04x [bus %02x-%02x]", segment, start, end); - res->name = new->name; - - return new; -} - -static struct pci_mmcfg_region *__init pci_mmconfig_add(int segment, int start, - int end, u64 addr) -{ - struct pci_mmcfg_region *new; - - new = pci_mmconfig_alloc(segment, start, end, addr); - if (new) { - mutex_lock(&pci_mmcfg_lock); - list_add_sorted(new); - mutex_unlock(&pci_mmcfg_lock); - - pr_info(PREFIX - "MMCONFIG for domain %04x [bus %02x-%02x] at %pR " - "(base %#lx)\n", - segment, start, end, &new->res, (unsigned long)addr); - } - - return new; -} - -struct pci_mmcfg_region *pci_mmconfig_lookup(int segment, int bus) -{ - struct pci_mmcfg_region *cfg; - - list_for_each_entry_rcu(cfg, &pci_mmcfg_list, list) - if (cfg->segment == segment && - cfg->start_bus <= bus && bus <= cfg->end_bus) - return cfg; - - return NULL; -}
static const char *__init pci_mmcfg_e7520(void) { @@ -543,8 +447,8 @@ static void __init pci_mmcfg_reject_broken(int early) } }
-static int __init acpi_mcfg_check_entry(struct acpi_table_mcfg *mcfg, - struct acpi_mcfg_allocation *cfg) +int __init acpi_mcfg_check_entry(struct acpi_table_mcfg *mcfg, + struct acpi_mcfg_allocation *cfg) { int year;
@@ -566,50 +470,6 @@ static int __init acpi_mcfg_check_entry(struct acpi_table_mcfg *mcfg, return -EINVAL; }
-static int __init pci_parse_mcfg(struct acpi_table_header *header) -{ - struct acpi_table_mcfg *mcfg; - struct acpi_mcfg_allocation *cfg_table, *cfg; - unsigned long i; - int entries; - - if (!header) - return -EINVAL; - - mcfg = (struct acpi_table_mcfg *)header; - - /* how many config structures do we have */ - free_all_mmcfg(); - entries = 0; - i = header->length - sizeof(struct acpi_table_mcfg); - while (i >= sizeof(struct acpi_mcfg_allocation)) { - entries++; - i -= sizeof(struct acpi_mcfg_allocation); - } - if (entries == 0) { - pr_err(PREFIX "MMCONFIG has no entries\n"); - return -ENODEV; - } - - cfg_table = (struct acpi_mcfg_allocation *) &mcfg[1]; - for (i = 0; i < entries; i++) { - cfg = &cfg_table[i]; - if (acpi_mcfg_check_entry(mcfg, cfg)) { - free_all_mmcfg(); - return -ENODEV; - } - - if (pci_mmconfig_add(cfg->pci_segment, cfg->start_bus_number, - cfg->end_bus_number, cfg->address) == NULL) { - pr_warn(PREFIX "no memory for MCFG entries\n"); - free_all_mmcfg(); - return -ENOMEM; - } - } - - return 0; -} - #ifdef CONFIG_ACPI_APEI extern int (*arch_apei_filter_addr)(int (*func)(__u64 start, __u64 size, void *data), void *data); @@ -668,7 +528,7 @@ void __init pci_mmcfg_early_init(void) if (pci_mmcfg_check_hostbridge()) known_bridge = 1; else - acpi_sfi_table_parse(ACPI_SIG_MCFG, pci_parse_mcfg); + acpi_sfi_table_parse(ACPI_SIG_MCFG, acpi_parse_mcfg); __pci_mmcfg_init(1);
set_apei_filter(); @@ -686,7 +546,7 @@ void __init pci_mmcfg_late_init(void)
/* MMCONFIG hasn't been enabled yet, try again */ if (pci_probe & PCI_PROBE_MASK & ~PCI_PROBE_MMCONF) { - acpi_sfi_table_parse(ACPI_SIG_MCFG, pci_parse_mcfg); + acpi_sfi_table_parse(ACPI_SIG_MCFG, acpi_parse_mcfg); __pci_mmcfg_init(0); } } @@ -720,38 +580,6 @@ static int __init pci_mmcfg_late_insert_resources(void) */ late_initcall(pci_mmcfg_late_insert_resources);
-static int pci_mmconfig_inject(struct pci_mmcfg_region *cfg) -{ - struct pci_mmcfg_region *cfg_conflict; - int err = 0; - - mutex_lock(&pci_mmcfg_lock); - cfg_conflict = pci_mmconfig_lookup(cfg->segment, cfg->start_bus); - if (cfg_conflict) { - if (cfg_conflict->end_bus < cfg->end_bus) - pr_info(FW_INFO "MMCONFIG for " - "domain %04x [bus %02x-%02x] " - "only partially covers this bridge\n", - cfg_conflict->segment, cfg_conflict->start_bus, - cfg_conflict->end_bus); - err = -EEXIST; - goto out; - } - - if (pci_mmcfg_arch_map(cfg)) { - pr_warn("fail to map MMCONFIG %pR.\n", &cfg->res); - err = -ENOMEM; - goto out; - } else { - list_add_sorted(cfg); - pr_info("MMCONFIG at %pR (base %#lx)\n", - &cfg->res, (unsigned long)cfg->address); - } -out: - mutex_unlock(&pci_mmcfg_lock); - return err; -} - /* Add MMCFG information for host bridges */ int pci_mmconfig_insert(struct device *dev, u16 seg, u8 start, u8 end, phys_addr_t addr) @@ -808,26 +636,3 @@ error: kfree(cfg); 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; -} diff --git a/arch/x86/pci/mmconfig_32.c b/arch/x86/pci/mmconfig_32.c index 43984bc..246f135 100644 --- a/arch/x86/pci/mmconfig_32.c +++ b/arch/x86/pci/mmconfig_32.c @@ -12,6 +12,7 @@ #include <linux/pci.h> #include <linux/init.h> #include <linux/rcupdate.h> +#include <linux/ecam.h> #include <asm/e820.h> #include <asm/pci_x86.h>
diff --git a/arch/x86/pci/mmconfig_64.c b/arch/x86/pci/mmconfig_64.c index bea5249..b14fcd3 100644 --- a/arch/x86/pci/mmconfig_64.c +++ b/arch/x86/pci/mmconfig_64.c @@ -10,6 +10,7 @@ #include <linux/acpi.h> #include <linux/bitmap.h> #include <linux/rcupdate.h> +#include <linux/ecam.h> #include <asm/e820.h> #include <asm/pci_x86.h>
diff --git a/arch/x86/pci/numachip.c b/arch/x86/pci/numachip.c index 2e565e6..55fbd18 100644 --- a/arch/x86/pci/numachip.c +++ b/arch/x86/pci/numachip.c @@ -13,6 +13,7 @@ * */
+#include <linux/ecam.h> #include <linux/pci.h> #include <asm/pci_x86.h>
diff --git a/drivers/acpi/Makefile b/drivers/acpi/Makefile index 96b3183..265eb90 100644 --- a/drivers/acpi/Makefile +++ b/drivers/acpi/Makefile @@ -65,6 +65,7 @@ obj-$(CONFIG_ACPI_BUTTON) += button.o obj-$(CONFIG_ACPI_FAN) += fan.o obj-$(CONFIG_ACPI_VIDEO) += video.o obj-$(CONFIG_ACPI_PCI_SLOT) += pci_slot.o +obj-$(CONFIG_PCI_MMCONFIG) += mcfg.o obj-$(CONFIG_ACPI_PROCESSOR) += processor.o obj-y += container.o obj-$(CONFIG_ACPI_THERMAL) += thermal.o diff --git a/drivers/acpi/mcfg.c b/drivers/acpi/mcfg.c new file mode 100644 index 0000000..5ecef20 --- /dev/null +++ b/drivers/acpi/mcfg.c @@ -0,0 +1,59 @@ +/* + * MCFG ACPI table parser. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#include <linux/acpi.h> +#include <linux/ecam.h> + +#include <asm/pci_x86.h> /* Temp hack before refactoring arch-specific calls */ + +#define PREFIX "MCFG: " + +int __init acpi_parse_mcfg(struct acpi_table_header *header) +{ + struct acpi_table_mcfg *mcfg; + struct acpi_mcfg_allocation *cfg_table, *cfg; + unsigned long i; + int entries; + + if (!header) + return -EINVAL; + + mcfg = (struct acpi_table_mcfg *)header; + + /* how many config structures do we have */ + free_all_mmcfg(); + entries = 0; + i = header->length - sizeof(struct acpi_table_mcfg); + while (i >= sizeof(struct acpi_mcfg_allocation)) { + entries++; + i -= sizeof(struct acpi_mcfg_allocation); + } + if (entries == 0) { + pr_err(PREFIX "MCFG table has no entries\n"); + return -ENODEV; + } + + cfg_table = (struct acpi_mcfg_allocation *) &mcfg[1]; + for (i = 0; i < entries; i++) { + cfg = &cfg_table[i]; + if (acpi_mcfg_check_entry(mcfg, cfg)) { + free_all_mmcfg(); + return -ENODEV; + } + + if (pci_mmconfig_add(cfg->pci_segment, cfg->start_bus_number, + cfg->end_bus_number, cfg->address) == NULL) { + pr_warn(PREFIX "no memory for MCFG entries\n"); + free_all_mmcfg(); + return -ENOMEM; + } + } + + return 0; +} diff --git a/drivers/pci/Kconfig b/drivers/pci/Kconfig index 73de4ef..9950248 100644 --- a/drivers/pci/Kconfig +++ b/drivers/pci/Kconfig @@ -26,6 +26,13 @@ config PCI_MSI_IRQ_DOMAIN depends on PCI_MSI select GENERIC_MSI_IRQ_DOMAIN
+config PCI_ECAM + bool "Enhanced Configuration Access Mechanism (ECAM)" + depends on PCI && HAVE_PCI_ECAM + +config HAVE_PCI_ECAM + bool + config PCI_DEBUG bool "PCI Debugging" depends on PCI && DEBUG_KERNEL diff --git a/drivers/pci/Makefile b/drivers/pci/Makefile index 8417f55..c41acf1 100644 --- a/drivers/pci/Makefile +++ b/drivers/pci/Makefile @@ -30,6 +30,11 @@ obj-$(CONFIG_PCI_ATS) += ats.o obj-$(CONFIG_PCI_IOV) += iov.o
# +# Enhanced Configuration Access Mechanism (ECAM) +# +obj-$(CONFIG_PCI_ECAM) += ecam.o + +# # ACPI Related PCI FW Functions # ACPI _DSM provided firmware instance and string name # diff --git a/drivers/pci/ecam.c b/drivers/pci/ecam.c new file mode 100644 index 0000000..d221dba --- /dev/null +++ b/drivers/pci/ecam.c @@ -0,0 +1,175 @@ +/* + * Arch agnostic direct PCI config space access via + * ECAM (Enhanced Configuration Access Mechanism) + * + * Per-architecture code takes care of the mappings, region validation and + * accesses themselves. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + */ + +#include <linux/mutex.h> +#include <linux/rculist.h> +#include <linux/ecam.h> + +#include <asm/io.h> +#include <asm/pci_x86.h> /* Temp hack before refactoring arch-specific calls */ + +#define PREFIX "PCI: " + +static DEFINE_MUTEX(pci_mmcfg_lock); + +LIST_HEAD(pci_mmcfg_list); + +static void __init pci_mmconfig_remove(struct pci_mmcfg_region *cfg) +{ + if (cfg->res.parent) + release_resource(&cfg->res); + list_del(&cfg->list); + kfree(cfg); +} + +void __init free_all_mmcfg(void) +{ + struct pci_mmcfg_region *cfg, *tmp; + + pci_mmcfg_arch_free(); + list_for_each_entry_safe(cfg, tmp, &pci_mmcfg_list, list) + pci_mmconfig_remove(cfg); +} + +void list_add_sorted(struct pci_mmcfg_region *new) +{ + struct pci_mmcfg_region *cfg; + + /* keep list sorted by segment and starting bus number */ + list_for_each_entry_rcu(cfg, &pci_mmcfg_list, list) { + if (cfg->segment > new->segment || + (cfg->segment == new->segment && + cfg->start_bus >= new->start_bus)) { + list_add_tail_rcu(&new->list, &cfg->list); + return; + } + } + list_add_tail_rcu(&new->list, &pci_mmcfg_list); +} + +struct pci_mmcfg_region *pci_mmconfig_alloc(int segment, int start, + int end, u64 addr) +{ + struct pci_mmcfg_region *new; + struct resource *res; + + if (addr == 0) + return NULL; + + new = kzalloc(sizeof(*new), GFP_KERNEL); + if (!new) + return NULL; + + new->address = addr; + new->segment = segment; + new->start_bus = start; + new->end_bus = end; + + res = &new->res; + res->start = addr + PCI_MMCFG_BUS_OFFSET(start); + res->end = addr + PCI_MMCFG_BUS_OFFSET(end + 1) - 1; + res->flags = IORESOURCE_MEM | IORESOURCE_BUSY; + snprintf(new->name, PCI_MMCFG_RESOURCE_NAME_LEN, + "PCI MMCONFIG %04x [bus %02x-%02x]", segment, start, end); + res->name = new->name; + + return new; +} + +struct pci_mmcfg_region *pci_mmconfig_add(int segment, int start, + int end, u64 addr) +{ + struct pci_mmcfg_region *new; + + new = pci_mmconfig_alloc(segment, start, end, addr); + if (new) { + mutex_lock(&pci_mmcfg_lock); + list_add_sorted(new); + mutex_unlock(&pci_mmcfg_lock); + + pr_info(PREFIX + "MMCONFIG for domain %04x [bus %02x-%02x] at %pR " + "(base %#lx)\n", + segment, start, end, &new->res, (unsigned long)addr); + } + + return new; +} + +struct pci_mmcfg_region *pci_mmconfig_lookup(int segment, int bus) +{ + struct pci_mmcfg_region *cfg; + + list_for_each_entry_rcu(cfg, &pci_mmcfg_list, list) + if (cfg->segment == segment && + cfg->start_bus <= bus && bus <= cfg->end_bus) + return cfg; + + return NULL; +} + +/* 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; +} + +int pci_mmconfig_inject(struct pci_mmcfg_region *cfg) +{ + struct pci_mmcfg_region *cfg_conflict; + int err = 0; + + mutex_lock(&pci_mmcfg_lock); + cfg_conflict = pci_mmconfig_lookup(cfg->segment, cfg->start_bus); + if (cfg_conflict) { + if (cfg_conflict->end_bus < cfg->end_bus) + pr_info(FW_INFO "MMCONFIG for " + "domain %04x [bus %02x-%02x] " + "only partially covers this bridge\n", + cfg_conflict->segment, cfg_conflict->start_bus, + cfg_conflict->end_bus); + err = -EEXIST; + goto out; + } + + if (pci_mmcfg_arch_map(cfg)) { + pr_warn("fail to map MMCONFIG %pR.\n", &cfg->res); + err = -ENOMEM; + goto out; + } else { + list_add_sorted(cfg); + pr_info("MMCONFIG at %pR (base %#lx)\n", + &cfg->res, (unsigned long)cfg->address); + + } +out: + mutex_unlock(&pci_mmcfg_lock); + return err; +} diff --git a/drivers/xen/pci.c b/drivers/xen/pci.c index 7494dbe..6785ebb 100644 --- a/drivers/xen/pci.c +++ b/drivers/xen/pci.c @@ -20,6 +20,7 @@ #include <linux/pci.h> #include <linux/acpi.h> #include <linux/pci-acpi.h> +#include <linux/ecam.h> #include <xen/xen.h> #include <xen/interface/physdev.h> #include <xen/interface/xen.h> diff --git a/include/linux/acpi.h b/include/linux/acpi.h index 1991aea..7c603e4 100644 --- a/include/linux/acpi.h +++ b/include/linux/acpi.h @@ -166,6 +166,8 @@ int acpi_table_parse_madt(enum acpi_madt_type id, acpi_tbl_entry_handler handler, unsigned int max_entries); int acpi_parse_mcfg (struct acpi_table_header *header); +int acpi_mcfg_check_entry(struct acpi_table_mcfg *mcfg, + struct acpi_mcfg_allocation *cfg); void acpi_table_print_madt_entry (struct acpi_subtable_header *madt);
/* the following four functions are architecture-dependent */ diff --git a/include/linux/ecam.h b/include/linux/ecam.h new file mode 100644 index 0000000..dec3b52 --- /dev/null +++ b/include/linux/ecam.h @@ -0,0 +1,37 @@ +#ifndef __ECAM_H +#define __ECAM_H +#ifdef __KERNEL__ + +#include <linux/types.h> +#include <linux/acpi.h> + +/* "PCI MMCONFIG %04x [bus %02x-%02x]" */ +#define PCI_MMCFG_RESOURCE_NAME_LEN (22 + 4 + 2 + 2) + +struct pci_mmcfg_region { + struct list_head list; + struct resource res; + u64 address; + char __iomem *virt; + u16 segment; + u8 start_bus; + u8 end_bus; + char name[PCI_MMCFG_RESOURCE_NAME_LEN]; +}; + +struct pci_mmcfg_region *pci_mmconfig_lookup(int segment, int bus); +struct pci_mmcfg_region *pci_mmconfig_alloc(int segment, int start, + int end, u64 addr); +int pci_mmconfig_inject(struct pci_mmcfg_region *cfg); +struct pci_mmcfg_region *pci_mmconfig_add(int segment, int start, + int end, u64 addr); +void list_add_sorted(struct pci_mmcfg_region *new); +void free_all_mmcfg(void); +int pci_mmconfig_delete(u16 seg, u8 start, u8 end); + +extern struct list_head pci_mmcfg_list; + +#define PCI_MMCFG_BUS_OFFSET(bus) ((bus) << 20) + +#endif /* __KERNEL__ */ +#endif /* __ECAM_H */
First function acpi_mcfg_check_entry() does not apply any quirks by default.
Last two functions are required by ACPI subsystem to make PCI config space accessible. Generic code assume to do nothing for early init call but late init call does as follow: - parse MCFG table and add regions to ECAM resource list - map regions - add regions to iomem_resource
Signed-off-by: Tomasz Nowicki tn@semihalf.com Tested-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com Tested-by: Jeremy Linton jeremy.linton@arm.com --- drivers/acpi/mcfg.c | 26 ++++++++++++++++++++++++++ 1 file changed, 26 insertions(+)
diff --git a/drivers/acpi/mcfg.c b/drivers/acpi/mcfg.c index 5ecef20..fad9917 100644 --- a/drivers/acpi/mcfg.c +++ b/drivers/acpi/mcfg.c @@ -57,3 +57,29 @@ int __init acpi_parse_mcfg(struct acpi_table_header *header)
return 0; } + +int __init __weak acpi_mcfg_check_entry(struct acpi_table_mcfg *mcfg, + struct acpi_mcfg_allocation *cfg) +{ + return 0; +} + +void __init __weak pci_mmcfg_early_init(void) +{ + +} + +void __init __weak pci_mmcfg_late_init(void) +{ + struct pci_mmcfg_region *cfg; + + acpi_table_parse(ACPI_SIG_MCFG, acpi_parse_mcfg); + + if (list_empty(&pci_mmcfg_list)) + return; + if (!pci_mmcfg_arch_init()) + free_all_mmcfg(); + + list_for_each_entry(cfg, &pci_mmcfg_list, list) + insert_resource(&iomem_resource, &cfg->res); +}
mmconfig_64.c version is going to be default implementation for low-level operation on mmconfig regions. However, now it initializes raw_pci_ext_ops pointer which is specific for x86 only. Moreover, mmconfig_32.c is doing the same thing at the same time. So lets move it to mmconfig_shared.c so it becomes common for both and mmconfig_64.c turns out to be purely arch agnostic.
Signed-off-by: Tomasz Nowicki tn@semihalf.com Tested-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com Tested-by: Jeremy Linton jeremy.linton@arm.com --- arch/x86/include/asm/pci_x86.h | 5 +++++ arch/x86/pci/mmconfig-shared.c | 10 ++++++++-- arch/x86/pci/mmconfig_32.c | 10 ++-------- arch/x86/pci/mmconfig_64.c | 11 ++--------- 4 files changed, 17 insertions(+), 19 deletions(-)
diff --git a/arch/x86/include/asm/pci_x86.h b/arch/x86/include/asm/pci_x86.h index c1c0f37..0482807 100644 --- a/arch/x86/include/asm/pci_x86.h +++ b/arch/x86/include/asm/pci_x86.h @@ -130,6 +130,11 @@ extern void pci_mmcfg_arch_unmap(struct pci_mmcfg_region *cfg); extern int pci_mmconfig_insert(struct device *dev, u16 seg, u8 start, u8 end, phys_addr_t addr);
+int pci_mmcfg_read(unsigned int seg, unsigned int bus, unsigned int devfn, + int reg, int len, u32 *value); +int pci_mmcfg_write(unsigned int seg, unsigned int bus, unsigned int devfn, + int reg, int len, u32 value); + /* * AMD Fam10h CPUs are buggy, and cannot access MMIO config space * on their northbrige except through the * %eax register. As such, you MUST diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index ce2c2e4..980f304 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -29,6 +29,11 @@ static bool pci_mmcfg_running_state; static bool pci_mmcfg_arch_init_failed;
+const struct pci_raw_ops pci_mmcfg = { + .read = pci_mmcfg_read, + .write = pci_mmcfg_write, +}; + static const char *__init pci_mmcfg_e7520(void) { u32 win; @@ -512,9 +517,10 @@ static void __init __pci_mmcfg_init(int early) } }
- if (pci_mmcfg_arch_init()) + if (pci_mmcfg_arch_init()) { + raw_pci_ext_ops = &pci_mmcfg; pci_probe = (pci_probe & ~PCI_PROBE_MASK) | PCI_PROBE_MMCONF; - else { + } else { free_all_mmcfg(); pci_mmcfg_arch_init_failed = true; } diff --git a/arch/x86/pci/mmconfig_32.c b/arch/x86/pci/mmconfig_32.c index 246f135..2ded56f 100644 --- a/arch/x86/pci/mmconfig_32.c +++ b/arch/x86/pci/mmconfig_32.c @@ -50,7 +50,7 @@ static void pci_exp_set_dev_base(unsigned int base, int bus, int devfn) } }
-static int pci_mmcfg_read(unsigned int seg, unsigned int bus, +int pci_mmcfg_read(unsigned int seg, unsigned int bus, unsigned int devfn, int reg, int len, u32 *value) { unsigned long flags; @@ -89,7 +89,7 @@ err: *value = -1; return 0; }
-static int pci_mmcfg_write(unsigned int seg, unsigned int bus, +int pci_mmcfg_write(unsigned int seg, unsigned int bus, unsigned int devfn, int reg, int len, u32 value) { unsigned long flags; @@ -126,15 +126,9 @@ static int pci_mmcfg_write(unsigned int seg, unsigned int bus, return 0; }
-const struct pci_raw_ops pci_mmcfg = { - .read = pci_mmcfg_read, - .write = pci_mmcfg_write, -}; - int __init pci_mmcfg_arch_init(void) { printk(KERN_INFO "PCI: Using MMCONFIG for extended config space\n"); - raw_pci_ext_ops = &pci_mmcfg; return 1; }
diff --git a/arch/x86/pci/mmconfig_64.c b/arch/x86/pci/mmconfig_64.c index b14fcd3..d0c48eb 100644 --- a/arch/x86/pci/mmconfig_64.c +++ b/arch/x86/pci/mmconfig_64.c @@ -25,7 +25,7 @@ static char __iomem *pci_dev_base(unsigned int seg, unsigned int bus, unsigned i return NULL; }
-static int pci_mmcfg_read(unsigned int seg, unsigned int bus, +int pci_mmcfg_read(unsigned int seg, unsigned int bus, unsigned int devfn, int reg, int len, u32 *value) { char __iomem *addr; @@ -59,7 +59,7 @@ err: *value = -1; return 0; }
-static int pci_mmcfg_write(unsigned int seg, unsigned int bus, +int pci_mmcfg_write(unsigned int seg, unsigned int bus, unsigned int devfn, int reg, int len, u32 value) { char __iomem *addr; @@ -91,11 +91,6 @@ static int pci_mmcfg_write(unsigned int seg, unsigned int bus, return 0; }
-const struct pci_raw_ops pci_mmcfg = { - .read = pci_mmcfg_read, - .write = pci_mmcfg_write, -}; - static void __iomem *mcfg_ioremap(struct pci_mmcfg_region *cfg) { void __iomem *addr; @@ -121,8 +116,6 @@ int __init pci_mmcfg_arch_init(void) return 0; }
- raw_pci_ext_ops = &pci_mmcfg; - return 1; }
Hosts with custom ECAM hooks (like 32bit x86) should select ARCH_HAS_CUSTOM_PCI_ECAM. Otherwise, host will use generic version provided by this patch (like 64bit x86 does).
Note, we leaved x86-specific PCI config accessors in corresponding files.
Signed-off-by: Tomasz Nowicki tn@semihalf.com Tested-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com Tested-by: Jeremy Linton jeremy.linton@arm.com --- arch/x86/Kconfig | 1 + arch/x86/include/asm/pci_x86.h | 4 --- arch/x86/pci/mmconfig_64.c | 55 --------------------------------------- drivers/acpi/mcfg.c | 2 -- drivers/pci/Kconfig | 3 +++ drivers/pci/ecam.c | 59 +++++++++++++++++++++++++++++++++++++++++- include/linux/ecam.h | 6 +++++ 7 files changed, 68 insertions(+), 62 deletions(-)
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 350bd52..102d7d1 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -30,6 +30,7 @@ config X86 select ARCH_HAS_PMEM_API if X86_64 select ARCH_HAS_MMIO_FLUSH select ARCH_HAS_SG_CHAIN + select ARCH_HAS_CUSTOM_PCI_ECAM if X86_32 select ARCH_HAVE_NMI_SAFE_CMPXCHG select ARCH_MIGHT_HAVE_ACPI_PDC if ACPI select ARCH_MIGHT_HAVE_PC_PARPORT diff --git a/arch/x86/include/asm/pci_x86.h b/arch/x86/include/asm/pci_x86.h index 0482807..be091b4 100644 --- a/arch/x86/include/asm/pci_x86.h +++ b/arch/x86/include/asm/pci_x86.h @@ -123,10 +123,6 @@ extern int pci_legacy_init(void); extern int pcibios_fixup_irq(struct pci_dev *dev, u8 pin);
/* pci-mmconfig.c */ -extern int __init pci_mmcfg_arch_init(void); -extern void __init pci_mmcfg_arch_free(void); -extern int pci_mmcfg_arch_map(struct pci_mmcfg_region *cfg); -extern void pci_mmcfg_arch_unmap(struct pci_mmcfg_region *cfg); extern int pci_mmconfig_insert(struct device *dev, u16 seg, u8 start, u8 end, phys_addr_t addr);
diff --git a/arch/x86/pci/mmconfig_64.c b/arch/x86/pci/mmconfig_64.c index d0c48eb..fd356cc 100644 --- a/arch/x86/pci/mmconfig_64.c +++ b/arch/x86/pci/mmconfig_64.c @@ -90,58 +90,3 @@ int pci_mmcfg_write(unsigned int seg, unsigned int bus,
return 0; } - -static void __iomem *mcfg_ioremap(struct pci_mmcfg_region *cfg) -{ - void __iomem *addr; - u64 start, size; - int num_buses; - - start = cfg->address + PCI_MMCFG_BUS_OFFSET(cfg->start_bus); - num_buses = cfg->end_bus - cfg->start_bus + 1; - size = PCI_MMCFG_BUS_OFFSET(num_buses); - addr = ioremap_nocache(start, size); - if (addr) - addr -= PCI_MMCFG_BUS_OFFSET(cfg->start_bus); - return addr; -} - -int __init pci_mmcfg_arch_init(void) -{ - struct pci_mmcfg_region *cfg; - - list_for_each_entry(cfg, &pci_mmcfg_list, list) - if (pci_mmcfg_arch_map(cfg)) { - pci_mmcfg_arch_free(); - return 0; - } - - return 1; -} - -void __init pci_mmcfg_arch_free(void) -{ - struct pci_mmcfg_region *cfg; - - list_for_each_entry(cfg, &pci_mmcfg_list, list) - pci_mmcfg_arch_unmap(cfg); -} - -int pci_mmcfg_arch_map(struct pci_mmcfg_region *cfg) -{ - cfg->virt = mcfg_ioremap(cfg); - if (!cfg->virt) { - pr_err(PREFIX "can't map MMCONFIG at %pR\n", &cfg->res); - return -ENOMEM; - } - - return 0; -} - -void pci_mmcfg_arch_unmap(struct pci_mmcfg_region *cfg) -{ - if (cfg && cfg->virt) { - iounmap(cfg->virt + PCI_MMCFG_BUS_OFFSET(cfg->start_bus)); - cfg->virt = NULL; - } -} diff --git a/drivers/acpi/mcfg.c b/drivers/acpi/mcfg.c index fad9917..745b83e 100644 --- a/drivers/acpi/mcfg.c +++ b/drivers/acpi/mcfg.c @@ -10,8 +10,6 @@ #include <linux/acpi.h> #include <linux/ecam.h>
-#include <asm/pci_x86.h> /* Temp hack before refactoring arch-specific calls */ - #define PREFIX "MCFG: "
int __init acpi_parse_mcfg(struct acpi_table_header *header) diff --git a/drivers/pci/Kconfig b/drivers/pci/Kconfig index 9950248..b2e27c8 100644 --- a/drivers/pci/Kconfig +++ b/drivers/pci/Kconfig @@ -33,6 +33,9 @@ config PCI_ECAM config HAVE_PCI_ECAM bool
+config ARCH_HAS_CUSTOM_PCI_ECAM + bool + config PCI_DEBUG bool "PCI Debugging" depends on PCI && DEBUG_KERNEL diff --git a/drivers/pci/ecam.c b/drivers/pci/ecam.c index d221dba..8a5eef7 100644 --- a/drivers/pci/ecam.c +++ b/drivers/pci/ecam.c @@ -16,7 +16,6 @@ #include <linux/ecam.h>
#include <asm/io.h> -#include <asm/pci_x86.h> /* Temp hack before refactoring arch-specific calls */
#define PREFIX "PCI: "
@@ -24,6 +23,64 @@ static DEFINE_MUTEX(pci_mmcfg_lock);
LIST_HEAD(pci_mmcfg_list);
+#ifndef CONFIG_ARCH_HAS_CUSTOM_PCI_ECAM + +static void __iomem *mcfg_ioremap(struct pci_mmcfg_region *cfg) +{ + void __iomem *addr; + u64 start, size; + int num_buses; + + start = cfg->address + PCI_MMCFG_BUS_OFFSET(cfg->start_bus); + num_buses = cfg->end_bus - cfg->start_bus + 1; + size = PCI_MMCFG_BUS_OFFSET(num_buses); + addr = ioremap_nocache(start, size); + if (addr) + addr -= PCI_MMCFG_BUS_OFFSET(cfg->start_bus); + return addr; +} + +int __init pci_mmcfg_arch_init(void) +{ + struct pci_mmcfg_region *cfg; + + list_for_each_entry(cfg, &pci_mmcfg_list, list) + if (pci_mmcfg_arch_map(cfg)) { + pci_mmcfg_arch_free(); + return 0; + } + + return 1; +} + +void __init pci_mmcfg_arch_free(void) +{ + struct pci_mmcfg_region *cfg; + + list_for_each_entry(cfg, &pci_mmcfg_list, list) + pci_mmcfg_arch_unmap(cfg); +} + +int pci_mmcfg_arch_map(struct pci_mmcfg_region *cfg) +{ + cfg->virt = mcfg_ioremap(cfg); + if (!cfg->virt) { + pr_err(PREFIX "can't map MMCONFIG at %pR\n", &cfg->res); + return -ENOMEM; + } + + return 0; +} + +void pci_mmcfg_arch_unmap(struct pci_mmcfg_region *cfg) +{ + if (cfg && cfg->virt) { + iounmap(cfg->virt + PCI_MMCFG_BUS_OFFSET(cfg->start_bus)); + cfg->virt = NULL; + } +} +#endif + static void __init pci_mmconfig_remove(struct pci_mmcfg_region *cfg) { if (cfg->res.parent) diff --git a/include/linux/ecam.h b/include/linux/ecam.h index dec3b52..813acd1 100644 --- a/include/linux/ecam.h +++ b/include/linux/ecam.h @@ -29,6 +29,12 @@ void list_add_sorted(struct pci_mmcfg_region *new); void free_all_mmcfg(void); int pci_mmconfig_delete(u16 seg, u8 start, u8 end);
+/* Arch specific calls */ +int pci_mmcfg_arch_init(void); +void pci_mmcfg_arch_free(void); +int pci_mmcfg_arch_map(struct pci_mmcfg_region *cfg); +void pci_mmcfg_arch_unmap(struct pci_mmcfg_region *cfg); + extern struct list_head pci_mmcfg_list;
#define PCI_MMCFG_BUS_OFFSET(bus) ((bus) << 20)
In drivers/xen/pci.c, there are arch x86 dependent codes when CONFIG_PCI_MMCONFIG is enabled, since CONFIG_PCI_MMCONFIG depends on ACPI, so this will prevent XEN PCI running on other architectures using ACPI with PCI_MMCONFIG enabled (such as ARM64).
Fortunatly, it can be sloved in a simple way. In drivers/xen/pci.c, the only x86 dependent code is if ((pci_probe & PCI_PROBE_MMCONF) == 0), and it's defined in asm/pci_x86.h, the code means that if the PCI resource is not probed in PCI_PROBE_MMCONF way, just ingnore the xen mcfg init. Actually this is duplicate, because if PCI resource is not probed in PCI_PROBE_MMCONF way, the pci_mmconfig_list will be empty, and the if (list_empty()) after it will do the same job.
So just remove the arch related code and the head file, this will be no functional change for x86, and also makes xen/pci.c usable for other architectures.
Signed-off-by: Hanjun Guo hanjun.guo@linaro.org CC: Konrad Rzeszutek Wilk konrad.wilk@oracle.com CC: Boris Ostrovsky boris.ostrovsky@oracle.com Cc: Stefano Stabellini Stefano.Stabellini@eu.citrix.com Tested-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com Tested-by: Jeremy Linton jeremy.linton@arm.com Reviewed-by: Boris Ostrovsky boris.ostrovsky@oracle.com Acked-by: Stefano Stabellini stefano.stabellini@eu.citrix.com --- drivers/xen/pci.c | 6 ------ 1 file changed, 6 deletions(-)
diff --git a/drivers/xen/pci.c b/drivers/xen/pci.c index 6785ebb..9a8dbe3 100644 --- a/drivers/xen/pci.c +++ b/drivers/xen/pci.c @@ -28,9 +28,6 @@ #include <asm/xen/hypervisor.h> #include <asm/xen/hypercall.h> #include "../pci/pci.h" -#ifdef CONFIG_PCI_MMCONFIG -#include <asm/pci_x86.h> -#endif
static bool __read_mostly pci_seg_supported = true;
@@ -222,9 +219,6 @@ static int __init xen_mcfg_late(void) if (!xen_initial_domain()) return 0;
- if ((pci_probe & PCI_PROBE_MMCONF) == 0) - return 0; - if (list_empty(&pci_mmcfg_list)) return 0;
Lets keep RAW ACPI PCI config space accessors empty by default, since we are note sure if they are necessary accross all archs. Once we sort this out, we can provide generic version or let architectures to overwrite, like now x86.
Suggested-by: Lorenzo Pieralisi lorenzo.pieralisi@arm.com Signed-off-by: Tomasz Nowicki tn@semihalf.com Tested-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com Tested-by: Jeremy Linton jeremy.linton@arm.com --- drivers/acpi/mcfg.c | 21 +++++++++++++++++++++ 1 file changed, 21 insertions(+)
diff --git a/drivers/acpi/mcfg.c b/drivers/acpi/mcfg.c index 745b83e..3e1e7be 100644 --- a/drivers/acpi/mcfg.c +++ b/drivers/acpi/mcfg.c @@ -9,9 +9,30 @@
#include <linux/acpi.h> #include <linux/ecam.h> +#include <linux/pci.h>
#define PREFIX "MCFG: "
+/* + * raw_pci_read/write - raw ACPI PCI config space accessors. + * + * By defauly (__weak) these accessors are empty and should be overwritten + * by architectures which support operations on ACPI PCI_Config regions, + * see osl.c file. + */ + +int __weak raw_pci_read(unsigned int domain, unsigned int bus, + unsigned int devfn, int reg, int len, u32 *val) +{ + return PCIBIOS_DEVICE_NOT_FOUND; +} + +int __weak raw_pci_write(unsigned int domain, unsigned int bus, + unsigned int devfn, int reg, int len, u32 val) +{ + return PCIBIOS_DEVICE_NOT_FOUND; +} + int __init acpi_parse_mcfg(struct acpi_table_header *header) { struct acpi_table_mcfg *mcfg;
We can now use previously prepared empty ACPI RAW accessors and cleanup a bit before adding full support for PCI host bridge driver.
Signed-off-by: Tomasz Nowicki tn@semihalf.com Tested-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com Tested-by: Jeremy Linton jeremy.linton@arm.com --- arch/arm64/Kconfig | 6 ++++++ arch/arm64/kernel/pci.c | 15 --------------- 2 files changed, 6 insertions(+), 15 deletions(-)
diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index 871f217..d65d315 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -92,6 +92,7 @@ config ARM64 select SPARSE_IRQ select SYSCTL_EXCEPTION_TRACE select HAVE_CONTEXT_TRACKING + select HAVE_PCI_ECAM help ARM 64-bit (AArch64) Linux support.
@@ -207,6 +208,11 @@ source "drivers/pci/Kconfig" source "drivers/pci/pcie/Kconfig" source "drivers/pci/hotplug/Kconfig"
+config PCI_MMCONFIG + def_bool y + select PCI_ECAM + depends on ACPI + endmenu
menu "Kernel Features" diff --git a/arch/arm64/kernel/pci.c b/arch/arm64/kernel/pci.c index b3d098b..023b983 100644 --- a/arch/arm64/kernel/pci.c +++ b/arch/arm64/kernel/pci.c @@ -61,21 +61,6 @@ int pcibios_add_device(struct pci_dev *dev) return 0; }
-/* - * raw_pci_read/write - Platform-specific PCI config space access. - */ -int raw_pci_read(unsigned int domain, unsigned int bus, - unsigned int devfn, int reg, int len, u32 *val) -{ - return -ENXIO; -} - -int raw_pci_write(unsigned int domain, unsigned int bus, - unsigned int devfn, int reg, int len, u32 val) -{ - return -ENXIO; -} - #ifdef CONFIG_ACPI /* Root bridge scanning */ struct pci_bus *pci_acpi_scan_root(struct acpi_pci_root *root)
There are two ways we can get ECAM (aka MCFG) regions using ACPI, from MCFG static table and from _CBA method. We cannot remove static regions, however regions coming from _CBA should be removed while removing bridge device.
In the light of above we need flag to mark hot added ECAM entries so that user should use pci_mmconfig_inject while adding regions from _CBA method.
Signed-off-by: Tomasz Nowicki tn@semihalf.com Tested-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com Tested-by: Jeremy Linton jeremy.linton@arm.com --- drivers/pci/ecam.c | 2 ++ include/linux/ecam.h | 1 + 2 files changed, 3 insertions(+)
diff --git a/drivers/pci/ecam.c b/drivers/pci/ecam.c index 8a5eef7..da35b4c 100644 --- a/drivers/pci/ecam.c +++ b/drivers/pci/ecam.c @@ -131,6 +131,7 @@ struct pci_mmcfg_region *pci_mmconfig_alloc(int segment, int start, new->segment = segment; new->start_bus = start; new->end_bus = end; + new->hot_added = false;
res = &new->res; res->start = addr + PCI_MMCFG_BUS_OFFSET(start); @@ -221,6 +222,7 @@ int pci_mmconfig_inject(struct pci_mmcfg_region *cfg) err = -ENOMEM; goto out; } else { + cfg->hot_added = true; list_add_sorted(cfg); pr_info("MMCONFIG at %pR (base %#lx)\n", &cfg->res, (unsigned long)cfg->address); diff --git a/include/linux/ecam.h b/include/linux/ecam.h index 813acd1..e0f322e 100644 --- a/include/linux/ecam.h +++ b/include/linux/ecam.h @@ -17,6 +17,7 @@ struct pci_mmcfg_region { u8 start_bus; u8 end_bus; char name[PCI_MMCFG_RESOURCE_NAME_LEN]; + bool hot_added; };
struct pci_mmcfg_region *pci_mmconfig_lookup(int segment, int bus);
Now that we have hot_added flag and all information in struct acpi_pci_root we need, we can get rid of arch specific mcfg data from struct pci_root_info.
Signed-off-by: Tomasz Nowicki tn@semihalf.com Tested-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com --- arch/x86/pci/acpi.c | 36 ++++++++++++++---------------------- 1 file changed, 14 insertions(+), 22 deletions(-)
diff --git a/arch/x86/pci/acpi.c b/arch/x86/pci/acpi.c index 64caf2b..56714a9 100644 --- a/arch/x86/pci/acpi.c +++ b/arch/x86/pci/acpi.c @@ -12,11 +12,6 @@ struct pci_root_info { struct acpi_pci_root_info common; struct pci_sysdata sd; -#ifdef CONFIG_PCI_MMCONFIG - bool mcfg_added; - u8 start_bus; - u8 end_bus; -#endif };
static bool pci_use_crs = true; @@ -180,16 +175,13 @@ static int check_segment(u16 seg, struct device *dev, char *estr)
static int setup_mcfg_map(struct acpi_pci_root_info *ci) { - int result, seg; - struct pci_root_info *info; + int result, seg, start, end; struct acpi_pci_root *root = ci->root; struct device *dev = &ci->bridge->dev;
- info = container_of(ci, struct pci_root_info, common); - info->start_bus = (u8)root->secondary.start; - info->end_bus = (u8)root->secondary.end; - info->mcfg_added = false; - seg = info->sd.domain; + seg = root->segment; + start = root->secondary.start; + end = root->secondary.end;
/* return success if MMCFG is not in use */ if (raw_pci_ext_ops && raw_pci_ext_ops != &pci_mmcfg) @@ -198,13 +190,11 @@ static int setup_mcfg_map(struct acpi_pci_root_info *ci) if (!(pci_probe & PCI_PROBE_MMCONF)) return check_segment(seg, dev, "MMCONFIG is disabled,");
- result = pci_mmconfig_insert(dev, seg, info->start_bus, info->end_bus, - root->mcfg_addr); + result = pci_mmconfig_insert(dev, seg, start, end, root->mcfg_addr); if (result == 0) { /* enable MMCFG if it hasn't been enabled yet */ if (raw_pci_ext_ops == NULL) raw_pci_ext_ops = &pci_mmcfg; - info->mcfg_added = true; } else if (result != -EEXIST) return check_segment(seg, dev, "fail to add MMCONFIG information,"); @@ -214,14 +204,16 @@ static int setup_mcfg_map(struct acpi_pci_root_info *ci)
static void teardown_mcfg_map(struct acpi_pci_root_info *ci) { - struct pci_root_info *info; + struct acpi_pci_root *root = ci->root; + struct pci_mmcfg_region *cfg;
- info = container_of(ci, struct pci_root_info, common); - if (info->mcfg_added) { - pci_mmconfig_delete(info->sd.domain, - info->start_bus, info->end_bus); - info->mcfg_added = false; - } + cfg = pci_mmconfig_lookup(root->segment, root->secondary.start); + if (!cfg) + return; + + if (cfg->hot_added) + pci_mmconfig_delete(root->segment, root->secondary.start, + root->secondary.end); } #else static int setup_mcfg_map(struct acpi_pci_root_info *ci)
Currently we have two platforms (x86 & ia64) capable of PCI ACPI host bridge initialization. They both use sysdata pill to pass down parent device reference and both relay on NULL parent in pci_create_root_bus() to validate sysdata content.
It looks hacky and prevent us from getting some frimware specific info for PCI host controller e.g. bus domain number. However, it seems we can overcome that blocker by passing down parent device via pci_create_root_bus parameter (as the ACPI device type) and using ACPI_COMPANION_SET in core code for ACPI boot method. ACPI_COMPANION_SET is safe to run for all cases DT, ACPI and DT&ACPI.
Suggested-by: Lorenzo Pieralisi lorenzo.pieralisi@arm.com Signed-off-by: Tomasz Nowicki tn@semihalf.com --- drivers/acpi/pci_root.c | 5 ++++- drivers/pci/probe.c | 2 ++ 2 files changed, 6 insertions(+), 1 deletion(-)
diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c index ae3fe4e..a65c8c2 100644 --- a/drivers/acpi/pci_root.c +++ b/drivers/acpi/pci_root.c @@ -846,7 +846,10 @@ struct pci_bus *acpi_pci_root_create(struct acpi_pci_root *root,
pci_acpi_root_add_resources(info); pci_add_resource(&info->resources, &root->secondary); - bus = pci_create_root_bus(NULL, busnum, ops->pci_ops, + + /* Root bridge device needs to be sure of parent ACPI type */ + ACPI_COMPANION_SET(&device->dev, device); + bus = pci_create_root_bus(&device->dev, busnum, ops->pci_ops, sysdata, &info->resources); if (!bus) goto out_release_info; diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 553a029..cad836f 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -2107,6 +2107,8 @@ struct pci_bus *pci_create_root_bus(struct device *parent, int bus, bridge->dev.parent = parent; bridge->dev.release = pci_release_host_bridge_dev; dev_set_name(&bridge->dev, "pci%04x:%02x", pci_domain_nr(b), bus); + ACPI_COMPANION_SET(&bridge->dev, + parent ? to_acpi_device_node(parent->fwnode) : NULL); error = pcibios_root_bridge_prepare(bridge); if (error) { kfree(bridge);
On Wed, Jan 13, 2016 at 02:20:57PM +0100, Tomasz Nowicki wrote:
Currently we have two platforms (x86 & ia64) capable of PCI ACPI host bridge initialization. They both use sysdata pill to pass down parent
pill ?
device reference and both relay on NULL parent in pci_create_root_bus()
s/relay/rely
to validate sysdata content.
It looks hacky and prevent us from getting some frimware specific
"prevents us from getting some firmware..."
info for PCI host controller e.g. bus domain number. However, it seems we can overcome that blocker by passing down parent
It does not seem, either this patch overcomes the issue and it is accepted, or it does not and it is discarded.
device via pci_create_root_bus parameter (as the ACPI device type)
I do not understand what you mean by "as the ACPI device type".
and using ACPI_COMPANION_SET in core code for ACPI boot method. ACPI_COMPANION_SET is safe to run for all cases DT, ACPI and DT&ACPI.
Suggested-by: Lorenzo Pieralisi lorenzo.pieralisi@arm.com Signed-off-by: Tomasz Nowicki tn@semihalf.com
drivers/acpi/pci_root.c | 5 ++++- drivers/pci/probe.c | 2 ++ 2 files changed, 6 insertions(+), 1 deletion(-)
diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c index ae3fe4e..a65c8c2 100644 --- a/drivers/acpi/pci_root.c +++ b/drivers/acpi/pci_root.c @@ -846,7 +846,10 @@ struct pci_bus *acpi_pci_root_create(struct acpi_pci_root *root, pci_acpi_root_add_resources(info); pci_add_resource(&info->resources, &root->secondary);
- bus = pci_create_root_bus(NULL, busnum, ops->pci_ops,
- /* Root bridge device needs to be sure of parent ACPI type */
- ACPI_COMPANION_SET(&device->dev, device);
I do not understand why the code above is needed, can you elaborate please ?
Thanks, Lorenzo
- bus = pci_create_root_bus(&device->dev, busnum, ops->pci_ops, sysdata, &info->resources); if (!bus) goto out_release_info;
diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 553a029..cad836f 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -2107,6 +2107,8 @@ struct pci_bus *pci_create_root_bus(struct device *parent, int bus, bridge->dev.parent = parent; bridge->dev.release = pci_release_host_bridge_dev; dev_set_name(&bridge->dev, "pci%04x:%02x", pci_domain_nr(b), bus);
- ACPI_COMPANION_SET(&bridge->dev,
error = pcibios_root_bridge_prepare(bridge); if (error) { kfree(bridge);parent ? to_acpi_device_node(parent->fwnode) : NULL);
-- 1.9.1
On 19.01.2016 17:02, Lorenzo Pieralisi wrote:
On Wed, Jan 13, 2016 at 02:20:57PM +0100, Tomasz Nowicki wrote:
[...]
device via pci_create_root_bus parameter (as the ACPI device type)
I do not understand what you mean by "as the ACPI device type".
See below.
and using ACPI_COMPANION_SET in core code for ACPI boot method. ACPI_COMPANION_SET is safe to run for all cases DT, ACPI and DT&ACPI.
Suggested-by: Lorenzo Pieralisi lorenzo.pieralisi@arm.com Signed-off-by: Tomasz Nowicki tn@semihalf.com
drivers/acpi/pci_root.c | 5 ++++- drivers/pci/probe.c | 2 ++ 2 files changed, 6 insertions(+), 1 deletion(-)
diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c index ae3fe4e..a65c8c2 100644 --- a/drivers/acpi/pci_root.c +++ b/drivers/acpi/pci_root.c @@ -846,7 +846,10 @@ struct pci_bus *acpi_pci_root_create(struct acpi_pci_root *root,
pci_acpi_root_add_resources(info); pci_add_resource(&info->resources, &root->secondary);
- bus = pci_create_root_bus(NULL, busnum, ops->pci_ops,
- /* Root bridge device needs to be sure of parent ACPI type */
- ACPI_COMPANION_SET(&device->dev, device);
I do not understand why the code above is needed, can you elaborate please ?
This makes sure that device->dev can be identified as ACPI device, so we can use to_acpi_device_node() and assign companion safely below.
- bus = pci_create_root_bus(&device->dev, busnum, ops->pci_ops, sysdata, &info->resources); if (!bus) goto out_release_info;
diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 553a029..cad836f 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -2107,6 +2107,8 @@ struct pci_bus *pci_create_root_bus(struct device *parent, int bus, bridge->dev.parent = parent; bridge->dev.release = pci_release_host_bridge_dev; dev_set_name(&bridge->dev, "pci%04x:%02x", pci_domain_nr(b), bus);
- ACPI_COMPANION_SET(&bridge->dev,
parent ? to_acpi_device_node(parent->fwnode) : NULL);
pci_create_root_bus() can be called with DT device, then ACPI_COMPANION_SET assign no companion to bridge device.
Thanks, Tomasz
On Wed, Jan 20, 2016 at 12:20:07PM +0100, Tomasz Nowicki wrote:
On 19.01.2016 17:02, Lorenzo Pieralisi wrote:
On Wed, Jan 13, 2016 at 02:20:57PM +0100, Tomasz Nowicki wrote:
[...]
device via pci_create_root_bus parameter (as the ACPI device type)
I do not understand what you mean by "as the ACPI device type".
See below.
and using ACPI_COMPANION_SET in core code for ACPI boot method. ACPI_COMPANION_SET is safe to run for all cases DT, ACPI and DT&ACPI.
Suggested-by: Lorenzo Pieralisi lorenzo.pieralisi@arm.com Signed-off-by: Tomasz Nowicki tn@semihalf.com
drivers/acpi/pci_root.c | 5 ++++- drivers/pci/probe.c | 2 ++ 2 files changed, 6 insertions(+), 1 deletion(-)
diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c index ae3fe4e..a65c8c2 100644 --- a/drivers/acpi/pci_root.c +++ b/drivers/acpi/pci_root.c @@ -846,7 +846,10 @@ struct pci_bus *acpi_pci_root_create(struct acpi_pci_root *root,
pci_acpi_root_add_resources(info); pci_add_resource(&info->resources, &root->secondary);
- bus = pci_create_root_bus(NULL, busnum, ops->pci_ops,
- /* Root bridge device needs to be sure of parent ACPI type */
- ACPI_COMPANION_SET(&device->dev, device);
I do not understand why the code above is needed, can you elaborate please ?
This makes sure that device->dev can be identified as ACPI device, so we can use to_acpi_device_node() and assign companion safely below.
I do not follow. If you refer to the fwnode handle type, that is already set by ACPI core code (before acpi_pci_root_add() is called, in acpi_init_device_object()).
Did you add the code above to solve a real issue you encountered ?
Thanks, Lorenzo
On 20.01.2016 13:38, Lorenzo Pieralisi wrote:
On Wed, Jan 20, 2016 at 12:20:07PM +0100, Tomasz Nowicki wrote:
On 19.01.2016 17:02, Lorenzo Pieralisi wrote:
On Wed, Jan 13, 2016 at 02:20:57PM +0100, Tomasz Nowicki wrote:
[...]
device via pci_create_root_bus parameter (as the ACPI device type)
I do not understand what you mean by "as the ACPI device type".
See below.
and using ACPI_COMPANION_SET in core code for ACPI boot method. ACPI_COMPANION_SET is safe to run for all cases DT, ACPI and DT&ACPI.
Suggested-by: Lorenzo Pieralisi lorenzo.pieralisi@arm.com Signed-off-by: Tomasz Nowicki tn@semihalf.com
drivers/acpi/pci_root.c | 5 ++++- drivers/pci/probe.c | 2 ++ 2 files changed, 6 insertions(+), 1 deletion(-)
diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c index ae3fe4e..a65c8c2 100644 --- a/drivers/acpi/pci_root.c +++ b/drivers/acpi/pci_root.c @@ -846,7 +846,10 @@ struct pci_bus *acpi_pci_root_create(struct acpi_pci_root *root,
pci_acpi_root_add_resources(info); pci_add_resource(&info->resources, &root->secondary);
- bus = pci_create_root_bus(NULL, busnum, ops->pci_ops,
- /* Root bridge device needs to be sure of parent ACPI type */
- ACPI_COMPANION_SET(&device->dev, device);
I do not understand why the code above is needed, can you elaborate please ?
This makes sure that device->dev can be identified as ACPI device, so we can use to_acpi_device_node() and assign companion safely below.
I do not follow. If you refer to the fwnode handle type, that is already set by ACPI core code (before acpi_pci_root_add() is called, in acpi_init_device_object()).
acpi_init_device_object() sets ACPI fwnode handle type only for "device", but not for "device->dev" which is what is passed as an argument to pci_create_root_bus().
Without ACPI_COMPANION_SET(&device->dev, device) here, no one can be sure if we have ACPI device in pci_create_root_bus().
Did you add the code above to solve a real issue you encountered ?
Yes. You can try to run series with this patch:
diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c index 88c9d20..1b74bf6 100644 --- a/drivers/acpi/pci_root.c +++ b/drivers/acpi/pci_root.c @@ -967,7 +967,11 @@ struct pci_bus *acpi_pci_root_create(struct acpi_pci_root *root, pci_add_resource(&info->resources, &root->secondary);
- /* Root bridge device needs to be sure of parent ACPI type */ - ACPI_COMPANION_SET(&device->dev, device); + if (to_acpi_device_node(device->dev.fwnode) == NULL) + pr_err("NON-ACPI TYPE\n"); + else + pr_err("ACPI TYPE\n"); + bus = pci_create_root_bus(&device->dev, busnum, ops->pci_ops, sysdata, &info->resources);
Thanks, Tomasz
On Wed, Jan 20, 2016 at 02:40:08PM +0100, Tomasz Nowicki wrote:
[...]
- /* Root bridge device needs to be sure of parent ACPI type */
- ACPI_COMPANION_SET(&device->dev, device);
I do not understand why the code above is needed, can you elaborate please ?
This makes sure that device->dev can be identified as ACPI device, so we can use to_acpi_device_node() and assign companion safely below.
I do not follow. If you refer to the fwnode handle type, that is already set by ACPI core code (before acpi_pci_root_add() is called, in acpi_init_device_object()).
acpi_init_device_object() sets ACPI fwnode handle type only for "device", but not for "device->dev" which is what is passed as an argument to pci_create_root_bus().
Without ACPI_COMPANION_SET(&device->dev, device) here, no one can be sure if we have ACPI device in pci_create_root_bus().
Ok, got it. The question is whether this should be done in ACPI core instead (is there a reason why it should *not* be done ?), but I now understand your point.
Thanks, Lorenzo
On 20.01.2016 15:22, Lorenzo Pieralisi wrote:
On Wed, Jan 20, 2016 at 02:40:08PM +0100, Tomasz Nowicki wrote:
[...]
- /* Root bridge device needs to be sure of parent ACPI type */
- ACPI_COMPANION_SET(&device->dev, device);
I do not understand why the code above is needed, can you elaborate please ?
This makes sure that device->dev can be identified as ACPI device, so we can use to_acpi_device_node() and assign companion safely below.
I do not follow. If you refer to the fwnode handle type, that is already set by ACPI core code (before acpi_pci_root_add() is called, in acpi_init_device_object()).
acpi_init_device_object() sets ACPI fwnode handle type only for "device", but not for "device->dev" which is what is passed as an argument to pci_create_root_bus().
Without ACPI_COMPANION_SET(&device->dev, device) here, no one can be sure if we have ACPI device in pci_create_root_bus().
Ok, got it. The question is whether this should be done in ACPI core instead (is there a reason why it should *not* be done ?), but I now understand your point.
I am not able to answer that question, but I see lots of changes in this area made by Rafael.
Rafael, can you please shed some light on this?
Thanks, Tomasz
On Wed, Jan 20, 2016 at 03:41:56PM +0100, Tomasz Nowicki wrote:
On 20.01.2016 15:22, Lorenzo Pieralisi wrote:
On Wed, Jan 20, 2016 at 02:40:08PM +0100, Tomasz Nowicki wrote:
[...]
>+ /* Root bridge device needs to be sure of parent ACPI type */ >+ ACPI_COMPANION_SET(&device->dev, device);
I do not understand why the code above is needed, can you elaborate please ?
This makes sure that device->dev can be identified as ACPI device, so we can use to_acpi_device_node() and assign companion safely below.
I do not follow. If you refer to the fwnode handle type, that is already set by ACPI core code (before acpi_pci_root_add() is called, in acpi_init_device_object()).
acpi_init_device_object() sets ACPI fwnode handle type only for "device", but not for "device->dev" which is what is passed as an argument to pci_create_root_bus().
Without ACPI_COMPANION_SET(&device->dev, device) here, no one can be sure if we have ACPI device in pci_create_root_bus().
Ok, got it. The question is whether this should be done in ACPI core instead (is there a reason why it should *not* be done ?), but I now understand your point.
I am not able to answer that question, but I see lots of changes in this area made by Rafael.
I had a further look and I am not sure why the fwnode_handle* in the dev member of the acpi_device is not made to point at the fwnode member of the acpi_device by *default* (acpi_init_device_object() ?), probably because it is never used as such but certainly Rafael knows, so I would ask him to clarify please it is unclear to me.
I do not see anything wrong in what you are doing in this patch.
Rafael, any comments ?
Thanks, Lorenzo
Since PCI core code is setting ACPI companion device for us now, platform specific ACPI companion device setting turns out to be dead now. Therefore we can get rid of it, including related companion reference from PCI sysdata structure.
Signed-off-by: Tomasz Nowicki tn@semihalf.com --- arch/ia64/hp/common/sba_iommu.c | 2 +- arch/ia64/include/asm/pci.h | 1 - arch/ia64/pci/pci.c | 16 ---------------- arch/ia64/sn/kernel/io_acpi_init.c | 4 ++-- arch/x86/include/asm/pci.h | 3 --- arch/x86/pci/acpi.c | 2 -- arch/x86/pci/irq.c | 10 ---------- 7 files changed, 3 insertions(+), 35 deletions(-)
diff --git a/arch/ia64/hp/common/sba_iommu.c b/arch/ia64/hp/common/sba_iommu.c index a6d6190..78e4444 100644 --- a/arch/ia64/hp/common/sba_iommu.c +++ b/arch/ia64/hp/common/sba_iommu.c @@ -1981,7 +1981,7 @@ sba_connect_bus(struct pci_bus *bus) if (PCI_CONTROLLER(bus)->iommu) return;
- handle = acpi_device_handle(PCI_CONTROLLER(bus)->companion); + handle = acpi_device_handle(ACPI_COMPANION(bus->bridge)); if (!handle) return;
diff --git a/arch/ia64/include/asm/pci.h b/arch/ia64/include/asm/pci.h index 07039d1..5050748 100644 --- a/arch/ia64/include/asm/pci.h +++ b/arch/ia64/include/asm/pci.h @@ -65,7 +65,6 @@ extern int pci_mmap_legacy_page_range(struct pci_bus *bus, #define pci_legacy_write platform_pci_legacy_write
struct pci_controller { - struct acpi_device *companion; void *iommu; int segment; int node; /* nearest node with memory or NUMA_NO_NODE for global allocation */ diff --git a/arch/ia64/pci/pci.c b/arch/ia64/pci/pci.c index 8f6ac2f..978d6af 100644 --- a/arch/ia64/pci/pci.c +++ b/arch/ia64/pci/pci.c @@ -301,28 +301,12 @@ struct pci_bus *pci_acpi_scan_root(struct acpi_pci_root *root) }
info->controller.segment = root->segment; - info->controller.companion = device; info->controller.node = acpi_get_node(device->handle); INIT_LIST_HEAD(&info->io_resources); return acpi_pci_root_create(root, &pci_acpi_root_ops, &info->common, &info->controller); }
-int pcibios_root_bridge_prepare(struct pci_host_bridge *bridge) -{ - /* - * We pass NULL as parent to pci_create_root_bus(), so if it is not NULL - * here, pci_create_root_bus() has been called by someone else and - * sysdata is likely to be different from what we expect. Let it go in - * that case. - */ - if (!bridge->dev.parent) { - struct pci_controller *controller = bridge->bus->sysdata; - ACPI_COMPANION_SET(&bridge->dev, controller->companion); - } - return 0; -} - void pcibios_fixup_device_resources(struct pci_dev *dev) { int idx; diff --git a/arch/ia64/sn/kernel/io_acpi_init.c b/arch/ia64/sn/kernel/io_acpi_init.c index 0640739..bcfddc2 100644 --- a/arch/ia64/sn/kernel/io_acpi_init.c +++ b/arch/ia64/sn/kernel/io_acpi_init.c @@ -132,7 +132,7 @@ sn_get_bussoft_ptr(struct pci_bus *bus) struct acpi_resource_vendor_typed *vendor;
- handle = acpi_device_handle(PCI_CONTROLLER(bus)->companion); + handle = acpi_device_handle(ACPI_COMPANION(bus->bridge)); status = acpi_get_vendor_resource(handle, METHOD_NAME__CRS, &sn_uuid, &buffer); if (ACPI_FAILURE(status)) { @@ -360,7 +360,7 @@ sn_acpi_get_pcidev_info(struct pci_dev *dev, struct pcidev_info **pcidev_info, acpi_status status; struct acpi_buffer name_buffer = { ACPI_ALLOCATE_BUFFER, NULL };
- rootbus_handle = acpi_device_handle(PCI_CONTROLLER(dev)->companion); + rootbus_handle = acpi_device_handle(ACPI_COMPANION(dev->bus->bridge)); status = acpi_evaluate_integer(rootbus_handle, METHOD_NAME__SEG, NULL, &segment); if (ACPI_SUCCESS(status)) { diff --git a/arch/x86/include/asm/pci.h b/arch/x86/include/asm/pci.h index 4625943..a98c022 100644 --- a/arch/x86/include/asm/pci.h +++ b/arch/x86/include/asm/pci.h @@ -14,9 +14,6 @@ struct pci_sysdata { int domain; /* PCI domain */ int node; /* NUMA node */ -#ifdef CONFIG_ACPI - struct acpi_device *companion; /* ACPI companion device */ -#endif #ifdef CONFIG_X86_64 void *iommu; /* IOMMU private data */ #endif diff --git a/arch/x86/pci/acpi.c b/arch/x86/pci/acpi.c index 56714a9..286e0f5 100644 --- a/arch/x86/pci/acpi.c +++ b/arch/x86/pci/acpi.c @@ -333,7 +333,6 @@ struct pci_bus *pci_acpi_scan_root(struct acpi_pci_root *root) struct pci_sysdata sd = { .domain = domain, .node = node, - .companion = root->device };
memcpy(bus->sysdata, &sd, sizeof(sd)); @@ -348,7 +347,6 @@ struct pci_bus *pci_acpi_scan_root(struct acpi_pci_root *root) else { info->sd.domain = domain; info->sd.node = node; - info->sd.companion = root->device; bus = acpi_pci_root_create(root, &acpi_pci_root_ops, &info->common, &info->sd); } diff --git a/arch/x86/pci/irq.c b/arch/x86/pci/irq.c index 7032798..cc62226 100644 --- a/arch/x86/pci/irq.c +++ b/arch/x86/pci/irq.c @@ -1153,16 +1153,6 @@ void __init pcibios_irq_init(void)
int pcibios_root_bridge_prepare(struct pci_host_bridge *bridge) { - /* - * We pass NULL as parent to pci_create_root_bus(), so if it is not NULL - * here, pci_create_root_bus() has been called by someone else and - * sysdata is likely to be different from what we expect. Let it go in - * that case. - */ - if (!bridge->dev.parent) { - struct pci_sysdata *sd = bridge->bus->sysdata; - ACPI_COMPANION_SET(&bridge->dev, sd->companion); - } bridge->map_irq = pci_map_irq; return 0; }
As we now have valid PCI host bridge device reference we can introduce code that is going to find its bus domain number using ACPI _SEG method.
Note that _SEG method is optional, therefore _SEG absence means that all PCI buses belong to domain 0.
Signed-off-by: Tomasz Nowicki tn@semihalf.com Reviewed-by: Liviu Dudau Liviu.Dudau@arm.com Tested-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com Tested-by: Jeremy Linton jeremy.linton@arm.com --- drivers/pci/pci.c | 29 +++++++++++++++++++++++++---- 1 file changed, 25 insertions(+), 4 deletions(-)
diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index 7274006..39a985b 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -25,6 +25,7 @@ #include <linux/device.h> #include <linux/pm_runtime.h> #include <linux/pci_hotplug.h> +#include <linux/acpi.h> #include <asm-generic/pci-bridge.h> #include <asm/setup.h> #include <linux/aer.h> @@ -4796,14 +4797,34 @@ void pci_bus_assign_domain_nr(struct pci_bus *bus, struct device *parent) * API and update the use_dt_domains value to keep track of method we * are using to assign domain numbers (use_dt_domains = 0). * + * IF ACPI, we expect non-DT method (use_dt_domains == -1) + * and call _SEG method for corresponding host bridge device. + * If _SEG method does not exist, following ACPI spec (6.5.6) + * all PCI buses belong to domain 0. + * * All other combinations imply we have a platform that is trying - * to mix domain numbers obtained from DT and pci_get_new_domain_nr(), - * which is a recipe for domain mishandling and it is prevented by - * invalidating the domain value (domain = -1) and printing a - * corresponding error. + * to mix domain numbers obtained from DT, ACPI and + * pci_get_new_domain_nr(), which is a recipe for domain mishandling and + * it is prevented by invalidating the domain value (domain = -1) and + * printing a corresponding error. */ + if (domain >= 0 && use_dt_domains) { use_dt_domains = 1; +#ifdef CONFIG_ACPI + } else if (!acpi_disabled && use_dt_domains == -1) { + struct acpi_device *acpi_dev = to_acpi_device(parent); + unsigned long long segment = 0; + acpi_status status; + + status = acpi_evaluate_integer(acpi_dev->handle, + METHOD_NAME__SEG, NULL, + &segment); + if (ACPI_FAILURE(status) && status != AE_NOT_FOUND) + dev_err(&acpi_dev->dev, "can't evaluate _SEG\n"); + + domain = segment; +#endif } else if (domain < 0 && use_dt_domains != 1) { use_dt_domains = 0; domain = pci_get_new_domain_nr();
Hi Tomasz,
On Wed, Jan 13, 2016 at 02:20:59PM +0100, Tomasz Nowicki wrote:
[...]
@@ -4796,14 +4797,34 @@ void pci_bus_assign_domain_nr(struct pci_bus *bus, struct device *parent) * API and update the use_dt_domains value to keep track of method we * are using to assign domain numbers (use_dt_domains = 0). *
* IF ACPI, we expect non-DT method (use_dt_domains == -1)
* and call _SEG method for corresponding host bridge device.
* If _SEG method does not exist, following ACPI spec (6.5.6)
* all PCI buses belong to domain 0.
*
- All other combinations imply we have a platform that is trying
* to mix domain numbers obtained from DT and pci_get_new_domain_nr(),
* which is a recipe for domain mishandling and it is prevented by
* invalidating the domain value (domain = -1) and printing a
* corresponding error.
* to mix domain numbers obtained from DT, ACPI and
* pci_get_new_domain_nr(), which is a recipe for domain mishandling and
* it is prevented by invalidating the domain value (domain = -1) and
*/* printing a corresponding error.
- if (domain >= 0 && use_dt_domains) { use_dt_domains = 1;
+#ifdef CONFIG_ACPI
- } else if (!acpi_disabled && use_dt_domains == -1) {
struct acpi_device *acpi_dev = to_acpi_device(parent);
unsigned long long segment = 0;
acpi_status status;
status = acpi_evaluate_integer(acpi_dev->handle,
METHOD_NAME__SEG, NULL,
&segment);
if (ACPI_FAILURE(status) && status != AE_NOT_FOUND)
dev_err(&acpi_dev->dev, "can't evaluate _SEG\n");
domain = segment;
+#endif
I think you can reshuffle a bit the code to make it easier to follow.
How about this ? (on top of mainline, I just compiled it do not take it verbatim):
-- >8 -- diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index d1a7105..467a316 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -7,6 +7,7 @@ * Copyright 1997 -- 2000 Martin Mares mj@ucw.cz */
+#include <linux/acpi.h> #include <linux/kernel.h> #include <linux/delay.h> #include <linux/init.h> @@ -4769,7 +4770,27 @@ int pci_get_new_domain_nr(void) }
#ifdef CONFIG_PCI_DOMAINS_GENERIC -void pci_bus_assign_domain_nr(struct pci_bus *bus, struct device *parent) +/* Feel free to add this to drivers/acpi as a helper */ +#ifdef CONFIG_ACPI +int pci_bus_acpi_domain_nr(struct device *parent) +{ + struct acpi_device *acpi_dev = to_acpi_device(parent); + unsigned long long segment = 0; + acpi_status status; + + status = acpi_evaluate_integer(acpi_dev->handle, + METHOD_NAME__SEG, NULL, + &segment); + if (ACPI_FAILURE(status) && status != AE_NOT_FOUND) + dev_err(&acpi_dev->dev, "can't evaluate _SEG\n"); + + return segment; +} +#else +int pci_bus_acpi_domain_nr(struct device *parent) { return -1; } +#endif + +int pci_bus_of_domain_nr(struct device *parent) { static int use_dt_domains = -1; int domain = of_get_pci_domain_nr(parent->of_node); @@ -4811,7 +4832,13 @@ void pci_bus_assign_domain_nr(struct pci_bus *bus, struct device *parent) domain = -1; }
- bus->domain_nr = domain; + return domain; +} + +void pci_bus_assign_domain_nr(struct pci_bus *bus, struct device *parent) +{ + bus->domain_nr = acpi_disabled ? pci_bus_of_domain_nr(parent) : + pci_bus_acpi_domain_nr(parent); } #endif #endif
On 21.01.2016 19:22, Lorenzo Pieralisi wrote:
Hi Tomasz,
On Wed, Jan 13, 2016 at 02:20:59PM +0100, Tomasz Nowicki wrote:
[...]
@@ -4796,14 +4797,34 @@ void pci_bus_assign_domain_nr(struct pci_bus *bus, struct device *parent) * API and update the use_dt_domains value to keep track of method we * are using to assign domain numbers (use_dt_domains = 0). *
* IF ACPI, we expect non-DT method (use_dt_domains == -1)
* and call _SEG method for corresponding host bridge device.
* If _SEG method does not exist, following ACPI spec (6.5.6)
* all PCI buses belong to domain 0.
*
- All other combinations imply we have a platform that is trying
* to mix domain numbers obtained from DT and pci_get_new_domain_nr(),
* which is a recipe for domain mishandling and it is prevented by
* invalidating the domain value (domain = -1) and printing a
* corresponding error.
* to mix domain numbers obtained from DT, ACPI and
* pci_get_new_domain_nr(), which is a recipe for domain mishandling and
* it is prevented by invalidating the domain value (domain = -1) and
*/* printing a corresponding error.
- if (domain >= 0 && use_dt_domains) { use_dt_domains = 1;
+#ifdef CONFIG_ACPI
- } else if (!acpi_disabled && use_dt_domains == -1) {
struct acpi_device *acpi_dev = to_acpi_device(parent);
unsigned long long segment = 0;
acpi_status status;
status = acpi_evaluate_integer(acpi_dev->handle,
METHOD_NAME__SEG, NULL,
&segment);
if (ACPI_FAILURE(status) && status != AE_NOT_FOUND)
dev_err(&acpi_dev->dev, "can't evaluate _SEG\n");
domain = segment;
+#endif
I think you can reshuffle a bit the code to make it easier to follow.
How about this ? (on top of mainline, I just compiled it do not take it verbatim):
-- >8 -- diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index d1a7105..467a316 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -7,6 +7,7 @@
- Copyright 1997 -- 2000 Martin Mares mj@ucw.cz
*/
+#include <linux/acpi.h> #include <linux/kernel.h> #include <linux/delay.h> #include <linux/init.h> @@ -4769,7 +4770,27 @@ int pci_get_new_domain_nr(void) }
#ifdef CONFIG_PCI_DOMAINS_GENERIC -void pci_bus_assign_domain_nr(struct pci_bus *bus, struct device *parent) +/* Feel free to add this to drivers/acpi as a helper */ +#ifdef CONFIG_ACPI +int pci_bus_acpi_domain_nr(struct device *parent) +{
- struct acpi_device *acpi_dev = to_acpi_device(parent);
- unsigned long long segment = 0;
- acpi_status status;
- status = acpi_evaluate_integer(acpi_dev->handle,
METHOD_NAME__SEG, NULL,
&segment);
- if (ACPI_FAILURE(status) && status != AE_NOT_FOUND)
dev_err(&acpi_dev->dev, "can't evaluate _SEG\n");
- return segment;
+} +#else +int pci_bus_acpi_domain_nr(struct device *parent) { return -1; } +#endif
+int pci_bus_of_domain_nr(struct device *parent) { static int use_dt_domains = -1; int domain = of_get_pci_domain_nr(parent->of_node); @@ -4811,7 +4832,13 @@ void pci_bus_assign_domain_nr(struct pci_bus *bus, struct device *parent) domain = -1; }
- bus->domain_nr = domain;
- return domain;
+}
+void pci_bus_assign_domain_nr(struct pci_bus *bus, struct device *parent) +{
- bus->domain_nr = acpi_disabled ? pci_bus_of_domain_nr(parent) :
} #endif #endifpci_bus_acpi_domain_nr(parent);
It looks nice, I will integrate your patch. Thanks!
Tomasz
On Thu, Jan 21, 2016 at 07:38:51PM +0100, Tomasz Nowicki wrote:
[...]
I think you can reshuffle a bit the code to make it easier to follow.
How about this ? (on top of mainline, I just compiled it do not take it verbatim):
-- >8 -- diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index d1a7105..467a316 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -7,6 +7,7 @@
- Copyright 1997 -- 2000 Martin Mares mj@ucw.cz
*/
+#include <linux/acpi.h> #include <linux/kernel.h> #include <linux/delay.h> #include <linux/init.h> @@ -4769,7 +4770,27 @@ int pci_get_new_domain_nr(void) }
#ifdef CONFIG_PCI_DOMAINS_GENERIC -void pci_bus_assign_domain_nr(struct pci_bus *bus, struct device *parent) +/* Feel free to add this to drivers/acpi as a helper */ +#ifdef CONFIG_ACPI +int pci_bus_acpi_domain_nr(struct device *parent) +{
- struct acpi_device *acpi_dev = to_acpi_device(parent);
- unsigned long long segment = 0;
- acpi_status status;
- status = acpi_evaluate_integer(acpi_dev->handle,
METHOD_NAME__SEG, NULL,
&segment);
- if (ACPI_FAILURE(status) && status != AE_NOT_FOUND)
dev_err(&acpi_dev->dev, "can't evaluate _SEG\n");
- return segment;
+} +#else +int pci_bus_acpi_domain_nr(struct device *parent) { return -1; } +#endif
+int pci_bus_of_domain_nr(struct device *parent) { static int use_dt_domains = -1; int domain = of_get_pci_domain_nr(parent->of_node); @@ -4811,7 +4832,13 @@ void pci_bus_assign_domain_nr(struct pci_bus *bus, struct device *parent) domain = -1; }
- bus->domain_nr = domain;
- return domain;
+}
+void pci_bus_assign_domain_nr(struct pci_bus *bus, struct device *parent) +{
- bus->domain_nr = acpi_disabled ? pci_bus_of_domain_nr(parent) :
pci_bus_acpi_domain_nr(parent);
} #endif #endif
It looks nice, I will integrate your patch. Thanks!
While at it, as I mentioned, please move the ACPI helper function to drivers/acpi, thanks.
Lorenzo
Arches in subject are the only one who use pcibios_{add|remove}_bus hooks and implement it in the same way. Moreover ARM64 is going to do the same. So it seams that acpi_pci_{add|remove}_bus is generic enough to be default option for pcibios_{add|remove}_bus hooks. Also, it is always safe to run acpi_pci_{add|remove}_bus as they have empty stubs for !ACPI case and return if ACPI has been switched off in run time.
After all we can remove x86 and ia64 pcibios_{add|remove}_bus implementation.
Signed-off-by: Tomasz Nowicki tn@semihalf.com --- arch/ia64/pci/pci.c | 10 ---------- arch/x86/pci/common.c | 10 ---------- drivers/pci/probe.c | 3 +++ 3 files changed, 3 insertions(+), 20 deletions(-)
diff --git a/arch/ia64/pci/pci.c b/arch/ia64/pci/pci.c index 978d6af..be4c9ef 100644 --- a/arch/ia64/pci/pci.c +++ b/arch/ia64/pci/pci.c @@ -358,16 +358,6 @@ void pcibios_fixup_bus(struct pci_bus *b) platform_pci_fixup_bus(b); }
-void pcibios_add_bus(struct pci_bus *bus) -{ - acpi_pci_add_bus(bus); -} - -void pcibios_remove_bus(struct pci_bus *bus) -{ - acpi_pci_remove_bus(bus); -} - void pcibios_set_master (struct pci_dev *dev) { /* No special bus mastering setup handling */ diff --git a/arch/x86/pci/common.c b/arch/x86/pci/common.c index eccd4d9..ed3236d 100644 --- a/arch/x86/pci/common.c +++ b/arch/x86/pci/common.c @@ -171,16 +171,6 @@ void pcibios_fixup_bus(struct pci_bus *b) pcibios_fixup_device_resources(dev); }
-void pcibios_add_bus(struct pci_bus *bus) -{ - acpi_pci_add_bus(bus); -} - -void pcibios_remove_bus(struct pci_bus *bus) -{ - acpi_pci_remove_bus(bus); -} - /* * Only use DMI information to set this if nothing was passed * on the kernel command line (which was parsed earlier). diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index cad836f..2fbf840 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -12,6 +12,7 @@ #include <linux/slab.h> #include <linux/module.h> #include <linux/cpumask.h> +#include <linux/pci-acpi.h> #include <linux/pci-aspm.h> #include <linux/aer.h> #include <linux/acpi.h> @@ -2067,10 +2068,12 @@ int __weak pcibios_root_bridge_prepare(struct pci_host_bridge *bridge)
void __weak pcibios_add_bus(struct pci_bus *bus) { + acpi_pci_add_bus(bus); }
void __weak pcibios_remove_bus(struct pci_bus *bus) { + acpi_pci_remove_bus(bus); }
struct pci_bus *pci_create_root_bus(struct device *parent, int bus,
On Wed, Jan 13, 2016 at 02:21:00PM +0100, Tomasz Nowicki wrote:
Arches in subject are the only one who use pcibios_{add|remove}_bus hooks
"x86 and ia64 are the only arches that implement..."
and implement it in the same way. Moreover ARM64 is going to do the same.
"implement them"
So it seams that acpi_pci_{add|remove}_bus is generic enough to be default
s/seams/seems
option for pcibios_{add|remove}_bus hooks. Also, it is always safe to run acpi_pci_{add|remove}_bus as they have empty stubs for !ACPI case and return if ACPI has been switched off in run time.
After all we can remove x86 and ia64 pcibios_{add|remove}_bus implementation.
I think you should rephrase the commit log and fix the typos.
Signed-off-by: Tomasz Nowicki tn@semihalf.com
arch/ia64/pci/pci.c | 10 ---------- arch/x86/pci/common.c | 10 ---------- drivers/pci/probe.c | 3 +++ 3 files changed, 3 insertions(+), 20 deletions(-)
Patch seems fine to me, I do not see why moving the acpi calls into generic code would create issues (and why it has not been done before) as you mentioned in the commit log, so:
Reviewed-by: Lorenzo Pieralisi lorenzo.pieralisi@arm.com
diff --git a/arch/ia64/pci/pci.c b/arch/ia64/pci/pci.c index 978d6af..be4c9ef 100644 --- a/arch/ia64/pci/pci.c +++ b/arch/ia64/pci/pci.c @@ -358,16 +358,6 @@ void pcibios_fixup_bus(struct pci_bus *b) platform_pci_fixup_bus(b); } -void pcibios_add_bus(struct pci_bus *bus) -{
- acpi_pci_add_bus(bus);
-}
-void pcibios_remove_bus(struct pci_bus *bus) -{
- acpi_pci_remove_bus(bus);
-}
void pcibios_set_master (struct pci_dev *dev) { /* No special bus mastering setup handling */ diff --git a/arch/x86/pci/common.c b/arch/x86/pci/common.c index eccd4d9..ed3236d 100644 --- a/arch/x86/pci/common.c +++ b/arch/x86/pci/common.c @@ -171,16 +171,6 @@ void pcibios_fixup_bus(struct pci_bus *b) pcibios_fixup_device_resources(dev); } -void pcibios_add_bus(struct pci_bus *bus) -{
- acpi_pci_add_bus(bus);
-}
-void pcibios_remove_bus(struct pci_bus *bus) -{
- acpi_pci_remove_bus(bus);
-}
/*
- Only use DMI information to set this if nothing was passed
- on the kernel command line (which was parsed earlier).
diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index cad836f..2fbf840 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -12,6 +12,7 @@ #include <linux/slab.h> #include <linux/module.h> #include <linux/cpumask.h> +#include <linux/pci-acpi.h> #include <linux/pci-aspm.h> #include <linux/aer.h> #include <linux/acpi.h> @@ -2067,10 +2068,12 @@ int __weak pcibios_root_bridge_prepare(struct pci_host_bridge *bridge) void __weak pcibios_add_bus(struct pci_bus *bus) {
- acpi_pci_add_bus(bus);
} void __weak pcibios_remove_bus(struct pci_bus *bus) {
- acpi_pci_remove_bus(bus);
} struct pci_bus *pci_create_root_bus(struct device *parent, int bus, -- 1.9.1
On 01/14/2016 05:33 PM, Lorenzo Pieralisi wrote:
On Wed, Jan 13, 2016 at 02:21:00PM +0100, Tomasz Nowicki wrote:
Arches in subject are the only one who use pcibios_{add|remove}_bus hooks
"x86 and ia64 are the only arches that implement..."
and implement it in the same way. Moreover ARM64 is going to do the same.
"implement them"
So it seams that acpi_pci_{add|remove}_bus is generic enough to be default
s/seams/seems
option for pcibios_{add|remove}_bus hooks. Also, it is always safe to run acpi_pci_{add|remove}_bus as they have empty stubs for !ACPI case and return if ACPI has been switched off in run time.
After all we can remove x86 and ia64 pcibios_{add|remove}_bus implementation.
I think you should rephrase the commit log and fix the typos.
yes, I will.
Signed-off-by: Tomasz Nowicki tn@semihalf.com
arch/ia64/pci/pci.c | 10 ---------- arch/x86/pci/common.c | 10 ---------- drivers/pci/probe.c | 3 +++ 3 files changed, 3 insertions(+), 20 deletions(-)
Patch seems fine to me, I do not see why moving the acpi calls into generic code would create issues (and why it has not been done before) as you mentioned in the commit log, so:
Reviewed-by: Lorenzo Pieralisi lorenzo.pieralisi@arm.com
Thanks, Tomasz
Lets abstract two calls which allow to inject and remove MCFG regions which may come from DSDT table. These calls will be used for x86 and ARM64 PCI host bridge driver in the later patches.
Signed-off-by: Tomasz Nowicki tn@semihalf.com --- drivers/acpi/mcfg.c | 38 ++++++++++++++++++++++++++++++++++++++ include/linux/pci-acpi.h | 9 +++++++++ 2 files changed, 47 insertions(+)
diff --git a/drivers/acpi/mcfg.c b/drivers/acpi/mcfg.c index 3e1e7be..dca4c4e 100644 --- a/drivers/acpi/mcfg.c +++ b/drivers/acpi/mcfg.c @@ -10,6 +10,7 @@ #include <linux/acpi.h> #include <linux/ecam.h> #include <linux/pci.h> +#include <linux/pci-acpi.h>
#define PREFIX "MCFG: "
@@ -77,6 +78,43 @@ int __init acpi_parse_mcfg(struct acpi_table_header *header) return 0; }
+int pci_mmcfg_setup_map(struct acpi_pci_root_info *ci) +{ + struct pci_mmcfg_region *cfg; + struct acpi_pci_root *root; + int seg, start, end, err; + + root = ci->root; + seg = root->segment; + start = root->secondary.start; + end = root->secondary.end; + + cfg = pci_mmconfig_lookup(seg, start); + if (cfg) + return 0; + + cfg = pci_mmconfig_alloc(seg, start, end, root->mcfg_addr); + if (!cfg) + return -ENOMEM; + + err = pci_mmconfig_inject(cfg); + return err; +} + +void pci_mmcfg_teardown_map(struct acpi_pci_root_info *ci) +{ + struct acpi_pci_root *root = ci->root; + struct pci_mmcfg_region *cfg; + + cfg = pci_mmconfig_lookup(root->segment, root->secondary.start); + if (!cfg) + return; + + if (cfg->hot_added) + pci_mmconfig_delete(root->segment, root->secondary.start, + root->secondary.end); +} + int __init __weak acpi_mcfg_check_entry(struct acpi_table_mcfg *mcfg, struct acpi_mcfg_allocation *cfg) { diff --git a/include/linux/pci-acpi.h b/include/linux/pci-acpi.h index 89ab057..c277415 100644 --- a/include/linux/pci-acpi.h +++ b/include/linux/pci-acpi.h @@ -79,6 +79,15 @@ extern struct pci_bus *acpi_pci_root_create(struct acpi_pci_root *root, void acpi_pci_add_bus(struct pci_bus *bus); void acpi_pci_remove_bus(struct pci_bus *bus);
+#ifdef CONFIG_PCI_MMCONFIG +int pci_mmcfg_setup_map(struct acpi_pci_root_info *ci); +void pci_mmcfg_teardown_map(struct acpi_pci_root_info *ci); +#else +static inline int pci_mmcfg_setup_map(struct acpi_pci_root_info *ci) +{ return 0; } +static inline void pci_mmcfg_teardown_map(struct acpi_pci_root_info *ci) { } +#endif + #ifdef CONFIG_ACPI_PCI_SLOT void acpi_pci_slot_init(void); void acpi_pci_slot_enumerate(struct pci_bus *bus);
Signed-off-by: Tomasz Nowicki tn@semihalf.com --- arch/x86/pci/acpi.c | 20 +------------------- 1 file changed, 1 insertion(+), 19 deletions(-)
diff --git a/arch/x86/pci/acpi.c b/arch/x86/pci/acpi.c index 286e0f5..c4d79d2 100644 --- a/arch/x86/pci/acpi.c +++ b/arch/x86/pci/acpi.c @@ -201,29 +201,11 @@ static int setup_mcfg_map(struct acpi_pci_root_info *ci)
return 0; } - -static void teardown_mcfg_map(struct acpi_pci_root_info *ci) -{ - struct acpi_pci_root *root = ci->root; - struct pci_mmcfg_region *cfg; - - cfg = pci_mmconfig_lookup(root->segment, root->secondary.start); - if (!cfg) - return; - - if (cfg->hot_added) - pci_mmconfig_delete(root->segment, root->secondary.start, - root->secondary.end); -} #else static int setup_mcfg_map(struct acpi_pci_root_info *ci) { return 0; } - -static void teardown_mcfg_map(struct acpi_pci_root_info *ci) -{ -} #endif
static int pci_acpi_root_get_node(struct acpi_pci_root *root) @@ -251,7 +233,7 @@ static int pci_acpi_root_init_info(struct acpi_pci_root_info *ci)
static void pci_acpi_root_release_info(struct acpi_pci_root_info *ci) { - teardown_mcfg_map(ci); + pci_mmcfg_teardown_map(ci); kfree(container_of(ci, struct pci_root_info, common)); }
We use generic accessors from access.c by default. However, we already know platforms that need special handling while accessing to PCI config space. These platforms will need different accessors set matched against platform ID, domain, bus touple. Therefore we are going to add (in future) DECLARE_ACPI_MCFG_FIXUP which will register platform specific custom accessors. For now we let pci_mcfg_get_ops takes domain and bus arguments and left some space for matching algorithm.
Signed-off-by: Tomasz Nowicki tn@semihalf.com --- drivers/acpi/mcfg.c | 30 ++++++++++++++++++++++++++++++ include/linux/pci-acpi.h | 8 ++++++++ 2 files changed, 38 insertions(+)
diff --git a/drivers/acpi/mcfg.c b/drivers/acpi/mcfg.c index dca4c4e..dfc2d14 100644 --- a/drivers/acpi/mcfg.c +++ b/drivers/acpi/mcfg.c @@ -34,6 +34,36 @@ int __weak raw_pci_write(unsigned int domain, unsigned int bus, return PCIBIOS_DEVICE_NOT_FOUND; }
+void __iomem * +pci_mcfg_dev_base(struct pci_bus *bus, unsigned int devfn, int offset) +{ + struct pci_mmcfg_region *cfg; + + cfg = pci_mmconfig_lookup(pci_domain_nr(bus), bus->number); + if (cfg && cfg->virt) + return cfg->virt + + (PCI_MMCFG_BUS_OFFSET(bus->number) | (devfn << 12)) + + offset; + return NULL; +} + +/* Default generic PCI config accessors */ +static struct pci_ops default_pci_mcfg_ops = { + .map_bus = pci_mcfg_dev_base, + .read = pci_generic_config_read, + .write = pci_generic_config_write, +}; + +struct pci_ops *pci_mcfg_get_ops(struct acpi_pci_root *root) +{ + /* + * TODO: Match against platform specific quirks and return + * corresponding PCI config space accessor set. + */ + + return &default_pci_mcfg_ops; +} + int __init acpi_parse_mcfg(struct acpi_table_header *header) { struct acpi_table_mcfg *mcfg; diff --git a/include/linux/pci-acpi.h b/include/linux/pci-acpi.h index c277415..45807a6 100644 --- a/include/linux/pci-acpi.h +++ b/include/linux/pci-acpi.h @@ -82,10 +82,18 @@ void acpi_pci_remove_bus(struct pci_bus *bus); #ifdef CONFIG_PCI_MMCONFIG int pci_mmcfg_setup_map(struct acpi_pci_root_info *ci); void pci_mmcfg_teardown_map(struct acpi_pci_root_info *ci); +struct pci_ops *pci_mcfg_get_ops(struct acpi_pci_root *root); +void __iomem * +pci_mcfg_dev_base(struct pci_bus *bus, unsigned int devfn, int offset); #else static inline int pci_mmcfg_setup_map(struct acpi_pci_root_info *ci) { return 0; } static inline void pci_mmcfg_teardown_map(struct acpi_pci_root_info *ci) { } +static inline struct pci_ops *pci_mcfg_get_ops(struct acpi_pci_root *root) +{ return NULL; } +static inline void __iomem * +pci_mcfg_dev_base(struct pci_bus *bus, unsigned int devfn, int offset) +{ return NULL; } #endif
#ifdef CONFIG_ACPI_PCI_SLOT
From: Liu Jiang jiang.liu@linux.intel.com
Some architectures, such as IA64 and ARM64, have no instructions to directly access PCI IO ports, so they map PCI IO ports into PCI MMIO address space. Typically PCI host bridges on those architectures take the responsibility to map (translate) PCI IO port transactions into Memory-Mapped IO transactions. ACPI specification provides support of such a usage case by using resource translation_offset.
But current ACPI resource parsing interface isn't neutral enough, it still has some special logic for IA64. So refine the ACPI resource parsing interface and IA64 code to neutrally handle translation_offset by: 1) ACPI resource parsing interface doesn't do any translation, it just save the translation_offset to be used by arch code. 2) Arch code will do the mapping(translation) based on arch specific information. Typically it does: 2.a) Translate per PCI domain IO port address space into system global IO port address space. 2.b) Setup MMIO address mapping for IO ports. void handle_io_resource(struct resource_entry *io_entry) { struct resource *mmio_res;
mmio_res = kzalloc(sizeof(*mmio_res), GFP_KERNEL); mmio_res->flags = IORESOURCE_MEM; mmio_res->start = io_entry->offset + io_entry->res->start; mmio_res->end = io_entry->offset + io_entry->res->end; insert_resource(&iomem_resource, mmio_res)
base = map_to_system_ioport_address(entry); io_entry->offset = base; io_entry->res->start += base; io_entry->res->end += base; }
Signed-off-by: Jiang Liu jiang.liu@linux.intel.com --- arch/ia64/pci/pci.c | 26 ++++++++++++++++---------- drivers/acpi/resource.c | 12 +++++------- 2 files changed, 21 insertions(+), 17 deletions(-)
diff --git a/arch/ia64/pci/pci.c b/arch/ia64/pci/pci.c index be4c9ef..c75356b 100644 --- a/arch/ia64/pci/pci.c +++ b/arch/ia64/pci/pci.c @@ -154,7 +154,7 @@ static int add_io_space(struct device *dev, struct pci_root_info *info, struct resource_entry *iospace; struct resource *resource, *res = entry->res; char *name; - unsigned long base, min, max, base_port; + unsigned long base_mmio, base_port; unsigned int sparse = 0, space_nr, len;
len = strlen(info->common.name) + 32; @@ -172,12 +172,10 @@ static int add_io_space(struct device *dev, struct pci_root_info *info, goto free_resource;
name = (char *)(iospace + 1); - min = res->start - entry->offset; - max = res->end - entry->offset; - base = __pa(io_space[space_nr].mmio_base); + base_mmio = __pa(io_space[space_nr].mmio_base); base_port = IO_SPACE_BASE(space_nr); snprintf(name, len, "%s I/O Ports %08lx-%08lx", info->common.name, - base_port + min, base_port + max); + base_port + res->start, base_port + res->end);
/* * The SDM guarantees the legacy 0-64K space is sparse, but if the @@ -190,19 +188,27 @@ static int add_io_space(struct device *dev, struct pci_root_info *info, resource = iospace->res; resource->name = name; resource->flags = IORESOURCE_MEM; - resource->start = base + (sparse ? IO_SPACE_SPARSE_ENCODING(min) : min); - resource->end = base + (sparse ? IO_SPACE_SPARSE_ENCODING(max) : max); + resource->start = base_mmio; + resource->end = base_mmio; + if (sparse) { + resource->start += IO_SPACE_SPARSE_ENCODING(res->start); + resource->end += IO_SPACE_SPARSE_ENCODING(res->end); + } else { + resource->start += res->start; + resource->end += res->end; + } if (insert_resource(&iomem_resource, resource)) { dev_err(dev, "can't allocate host bridge io space resource %pR\n", resource); goto free_resource; } + resource_list_add_tail(iospace, &info->io_resources);
+ /* Adjust base of original IO port resource descriptor */ entry->offset = base_port; - res->start = min + base_port; - res->end = max + base_port; - resource_list_add_tail(iospace, &info->io_resources); + res->start += base_port; + res->end += base_port;
return 0;
diff --git a/drivers/acpi/resource.c b/drivers/acpi/resource.c index cdc5c25..6578f68 100644 --- a/drivers/acpi/resource.c +++ b/drivers/acpi/resource.c @@ -190,8 +190,7 @@ static bool acpi_decode_space(struct resource_win *win, { u8 iodec = attr->granularity == 0xfff ? ACPI_DECODE_10 : ACPI_DECODE_16; bool wp = addr->info.mem.write_protect; - u64 len = attr->address_length; - u64 start, end, offset = 0; + u64 len = attr->address_length, offset = 0; struct resource *res = &win->res;
/* @@ -215,14 +214,13 @@ static bool acpi_decode_space(struct resource_win *win, else if (attr->translation_offset) pr_debug("ACPI: translation_offset(%lld) is invalid for non-bridge device.\n", attr->translation_offset); - start = attr->minimum + offset; - end = attr->maximum + offset;
win->offset = offset; - res->start = start; - res->end = end; + res->start = attr->minimum; + res->end = attr->maximum; if (sizeof(resource_size_t) < sizeof(u64) && - (offset != win->offset || start != res->start || end != res->end)) { + (offset != win->offset || attr->minimum != res->start || + attr->maximum != res->end)) { pr_warn("acpi resource window ([%#llx-%#llx] ignored, not CPU addressable)\n", attr->minimum, attr->maximum); return false;
Gerry, Tomasz,
On Wed, Jan 13, 2016 at 02:21:04PM +0100, Tomasz Nowicki wrote:
From: Liu Jiang jiang.liu@linux.intel.com
Some architectures, such as IA64 and ARM64, have no instructions to directly access PCI IO ports, so they map PCI IO ports into PCI MMIO address space. Typically PCI host bridges on those architectures take the responsibility to map (translate) PCI IO port transactions into Memory-Mapped IO transactions. ACPI specification provides support of such a usage case by using resource translation_offset.
But current ACPI resource parsing interface isn't neutral enough, it still has some special logic for IA64. So refine the ACPI resource parsing interface and IA64 code to neutrally handle translation_offset by:
- ACPI resource parsing interface doesn't do any translation, it just save the translation_offset to be used by arch code.
- Arch code will do the mapping(translation) based on arch specific information. Typically it does:
2.a) Translate per PCI domain IO port address space into system global IO port address space. 2.b) Setup MMIO address mapping for IO ports. void handle_io_resource(struct resource_entry *io_entry) { struct resource *mmio_res;
mmio_res = kzalloc(sizeof(*mmio_res), GFP_KERNEL); mmio_res->flags = IORESOURCE_MEM; mmio_res->start = io_entry->offset + io_entry->res->start; mmio_res->end = io_entry->offset + io_entry->res->end; insert_resource(&iomem_resource, mmio_res)
base = map_to_system_ioport_address(entry); io_entry->offset = base; io_entry->res->start += base; io_entry->res->end += base; }
Signed-off-by: Jiang Liu jiang.liu@linux.intel.com
arch/ia64/pci/pci.c | 26 ++++++++++++++++---------- drivers/acpi/resource.c | 12 +++++------- 2 files changed, 21 insertions(+), 17 deletions(-)
I still do not understand what TranslationType argument means in the ACPI specifications Resource descriptors (if you do not check it in generic code that info is not propagated to arches through the resources so basically we ignore it - I do not think we can do anything else given that on ia64 it is missing from ACPI tables), but this patch makes sense to me (it needs testing/reviewing for ia64 though):
Reviewed-by: Lorenzo Pieralisi lorenzo.pieralisi@arm.com
diff --git a/arch/ia64/pci/pci.c b/arch/ia64/pci/pci.c index be4c9ef..c75356b 100644 --- a/arch/ia64/pci/pci.c +++ b/arch/ia64/pci/pci.c @@ -154,7 +154,7 @@ static int add_io_space(struct device *dev, struct pci_root_info *info, struct resource_entry *iospace; struct resource *resource, *res = entry->res; char *name;
- unsigned long base, min, max, base_port;
- unsigned long base_mmio, base_port; unsigned int sparse = 0, space_nr, len;
len = strlen(info->common.name) + 32; @@ -172,12 +172,10 @@ static int add_io_space(struct device *dev, struct pci_root_info *info, goto free_resource; name = (char *)(iospace + 1);
- min = res->start - entry->offset;
- max = res->end - entry->offset;
- base = __pa(io_space[space_nr].mmio_base);
- base_mmio = __pa(io_space[space_nr].mmio_base); base_port = IO_SPACE_BASE(space_nr); snprintf(name, len, "%s I/O Ports %08lx-%08lx", info->common.name,
base_port + min, base_port + max);
base_port + res->start, base_port + res->end);
/* * The SDM guarantees the legacy 0-64K space is sparse, but if the @@ -190,19 +188,27 @@ static int add_io_space(struct device *dev, struct pci_root_info *info, resource = iospace->res; resource->name = name; resource->flags = IORESOURCE_MEM;
- resource->start = base + (sparse ? IO_SPACE_SPARSE_ENCODING(min) : min);
- resource->end = base + (sparse ? IO_SPACE_SPARSE_ENCODING(max) : max);
- resource->start = base_mmio;
- resource->end = base_mmio;
- if (sparse) {
resource->start += IO_SPACE_SPARSE_ENCODING(res->start);
resource->end += IO_SPACE_SPARSE_ENCODING(res->end);
- } else {
resource->start += res->start;
resource->end += res->end;
- } if (insert_resource(&iomem_resource, resource)) { dev_err(dev, "can't allocate host bridge io space resource %pR\n", resource); goto free_resource; }
- resource_list_add_tail(iospace, &info->io_resources);
- /* Adjust base of original IO port resource descriptor */ entry->offset = base_port;
- res->start = min + base_port;
- res->end = max + base_port;
- resource_list_add_tail(iospace, &info->io_resources);
- res->start += base_port;
- res->end += base_port;
return 0; diff --git a/drivers/acpi/resource.c b/drivers/acpi/resource.c index cdc5c25..6578f68 100644 --- a/drivers/acpi/resource.c +++ b/drivers/acpi/resource.c @@ -190,8 +190,7 @@ static bool acpi_decode_space(struct resource_win *win, { u8 iodec = attr->granularity == 0xfff ? ACPI_DECODE_10 : ACPI_DECODE_16; bool wp = addr->info.mem.write_protect;
- u64 len = attr->address_length;
- u64 start, end, offset = 0;
- u64 len = attr->address_length, offset = 0; struct resource *res = &win->res;
/* @@ -215,14 +214,13 @@ static bool acpi_decode_space(struct resource_win *win, else if (attr->translation_offset) pr_debug("ACPI: translation_offset(%lld) is invalid for non-bridge device.\n", attr->translation_offset);
- start = attr->minimum + offset;
- end = attr->maximum + offset;
win->offset = offset;
- res->start = start;
- res->end = end;
- res->start = attr->minimum;
- res->end = attr->maximum; if (sizeof(resource_size_t) < sizeof(u64) &&
(offset != win->offset || start != res->start || end != res->end)) {
(offset != win->offset || attr->minimum != res->start ||
pr_warn("acpi resource window ([%#llx-%#llx] ignored, not CPU addressable)\n", attr->minimum, attr->maximum); return false;attr->maximum != res->end)) {
-- 1.9.1
Gerry,
On Wed, Jan 13, 2016 at 02:21:04PM +0100, Tomasz Nowicki wrote:
From: Liu Jiang jiang.liu@linux.intel.com
Some architectures, such as IA64 and ARM64, have no instructions to directly access PCI IO ports, so they map PCI IO ports into PCI MMIO address space. Typically PCI host bridges on those architectures take the responsibility to map (translate) PCI IO port transactions into Memory-Mapped IO transactions. ACPI specification provides support of such a usage case by using resource translation_offset.
But current ACPI resource parsing interface isn't neutral enough, it still has some special logic for IA64. So refine the ACPI resource parsing interface and IA64 code to neutrally handle translation_offset by:
- ACPI resource parsing interface doesn't do any translation, it just save the translation_offset to be used by arch code.
- Arch code will do the mapping(translation) based on arch specific information. Typically it does:
2.a) Translate per PCI domain IO port address space into system global IO port address space. 2.b) Setup MMIO address mapping for IO ports.
This patch fixes IO space handling on IA64 and should go in as a fix.
IA64 PCI IO space is currently broken (Hanjun tested this on an IA64 box).
The first broken commit is:
3772aea7d6f3 ("ia64/PCI/ACPI: Use common ACPI resource parsing interface for host bridge")
because acpi core code checks (in acpi_dev_ioresource_flags()) the resource.end>=0x10003, which fails on ia64 - currently resource.end is set in acpi_decode_space() to:
AddressMaximum + AddressTranslation
where AddressTranslation is the CPU physical address mapping IO space on IA64, the >=0x10003 check in acpi_dev_ioresource_flags always triggers and the IO resource is then disabled.
Do you want me to re-send this patch as a fix, with updated commit log ?
Thanks, Lorenzo
void handle_io_resource(struct resource_entry *io_entry) { struct resource *mmio_res;
mmio_res = kzalloc(sizeof(*mmio_res), GFP_KERNEL); mmio_res->flags = IORESOURCE_MEM; mmio_res->start = io_entry->offset + io_entry->res->start; mmio_res->end = io_entry->offset + io_entry->res->end; insert_resource(&iomem_resource, mmio_res)
base = map_to_system_ioport_address(entry); io_entry->offset = base; io_entry->res->start += base; io_entry->res->end += base; }
Signed-off-by: Jiang Liu jiang.liu@linux.intel.com
arch/ia64/pci/pci.c | 26 ++++++++++++++++---------- drivers/acpi/resource.c | 12 +++++------- 2 files changed, 21 insertions(+), 17 deletions(-)
diff --git a/arch/ia64/pci/pci.c b/arch/ia64/pci/pci.c index be4c9ef..c75356b 100644 --- a/arch/ia64/pci/pci.c +++ b/arch/ia64/pci/pci.c @@ -154,7 +154,7 @@ static int add_io_space(struct device *dev, struct pci_root_info *info, struct resource_entry *iospace; struct resource *resource, *res = entry->res; char *name;
- unsigned long base, min, max, base_port;
- unsigned long base_mmio, base_port; unsigned int sparse = 0, space_nr, len;
len = strlen(info->common.name) + 32; @@ -172,12 +172,10 @@ static int add_io_space(struct device *dev, struct pci_root_info *info, goto free_resource; name = (char *)(iospace + 1);
- min = res->start - entry->offset;
- max = res->end - entry->offset;
- base = __pa(io_space[space_nr].mmio_base);
- base_mmio = __pa(io_space[space_nr].mmio_base); base_port = IO_SPACE_BASE(space_nr); snprintf(name, len, "%s I/O Ports %08lx-%08lx", info->common.name,
base_port + min, base_port + max);
base_port + res->start, base_port + res->end);
/* * The SDM guarantees the legacy 0-64K space is sparse, but if the @@ -190,19 +188,27 @@ static int add_io_space(struct device *dev, struct pci_root_info *info, resource = iospace->res; resource->name = name; resource->flags = IORESOURCE_MEM;
- resource->start = base + (sparse ? IO_SPACE_SPARSE_ENCODING(min) : min);
- resource->end = base + (sparse ? IO_SPACE_SPARSE_ENCODING(max) : max);
- resource->start = base_mmio;
- resource->end = base_mmio;
- if (sparse) {
resource->start += IO_SPACE_SPARSE_ENCODING(res->start);
resource->end += IO_SPACE_SPARSE_ENCODING(res->end);
- } else {
resource->start += res->start;
resource->end += res->end;
- } if (insert_resource(&iomem_resource, resource)) { dev_err(dev, "can't allocate host bridge io space resource %pR\n", resource); goto free_resource; }
- resource_list_add_tail(iospace, &info->io_resources);
- /* Adjust base of original IO port resource descriptor */ entry->offset = base_port;
- res->start = min + base_port;
- res->end = max + base_port;
- resource_list_add_tail(iospace, &info->io_resources);
- res->start += base_port;
- res->end += base_port;
return 0; diff --git a/drivers/acpi/resource.c b/drivers/acpi/resource.c index cdc5c25..6578f68 100644 --- a/drivers/acpi/resource.c +++ b/drivers/acpi/resource.c @@ -190,8 +190,7 @@ static bool acpi_decode_space(struct resource_win *win, { u8 iodec = attr->granularity == 0xfff ? ACPI_DECODE_10 : ACPI_DECODE_16; bool wp = addr->info.mem.write_protect;
- u64 len = attr->address_length;
- u64 start, end, offset = 0;
- u64 len = attr->address_length, offset = 0; struct resource *res = &win->res;
/* @@ -215,14 +214,13 @@ static bool acpi_decode_space(struct resource_win *win, else if (attr->translation_offset) pr_debug("ACPI: translation_offset(%lld) is invalid for non-bridge device.\n", attr->translation_offset);
- start = attr->minimum + offset;
- end = attr->maximum + offset;
win->offset = offset;
- res->start = start;
- res->end = end;
- res->start = attr->minimum;
- res->end = attr->maximum; if (sizeof(resource_size_t) < sizeof(u64) &&
(offset != win->offset || start != res->start || end != res->end)) {
(offset != win->offset || attr->minimum != res->start ||
pr_warn("acpi resource window ([%#llx-%#llx] ignored, not CPU addressable)\n", attr->minimum, attr->maximum); return false;attr->maximum != res->end)) {
-- 1.9.1
On Tue, Jan 19, 2016 at 12:20:26PM +0000, Lorenzo Pieralisi wrote:
Gerry,
On Wed, Jan 13, 2016 at 02:21:04PM +0100, Tomasz Nowicki wrote:
From: Liu Jiang jiang.liu@linux.intel.com
Some architectures, such as IA64 and ARM64, have no instructions to directly access PCI IO ports, so they map PCI IO ports into PCI MMIO address space. Typically PCI host bridges on those architectures take the responsibility to map (translate) PCI IO port transactions into Memory-Mapped IO transactions. ACPI specification provides support of such a usage case by using resource translation_offset.
But current ACPI resource parsing interface isn't neutral enough, it still has some special logic for IA64. So refine the ACPI resource parsing interface and IA64 code to neutrally handle translation_offset by:
- ACPI resource parsing interface doesn't do any translation, it just save the translation_offset to be used by arch code.
- Arch code will do the mapping(translation) based on arch specific information. Typically it does:
2.a) Translate per PCI domain IO port address space into system global IO port address space. 2.b) Setup MMIO address mapping for IO ports.
This patch fixes IO space handling on IA64 and should go in as a fix.
IA64 PCI IO space is currently broken (Hanjun tested this on an IA64 box).
The first broken commit is:
3772aea7d6f3 ("ia64/PCI/ACPI: Use common ACPI resource parsing interface for host bridge")
because acpi core code checks (in acpi_dev_ioresource_flags()) the resource.end>=0x10003, which fails on ia64 - currently resource.end is set in acpi_decode_space() to:
AddressMaximum + AddressTranslation
where AddressTranslation is the CPU physical address mapping IO space on IA64, the >=0x10003 check in acpi_dev_ioresource_flags always triggers and the IO resource is then disabled.
Do you want me to re-send this patch as a fix, with updated commit log ?
Two more points to discuss here. IA64, in its prepare_resources() callback calls acpi_pci_probe_root_resources() in turn. That function parses the _CRS and validate the resources (acpi_pci_root_validate_resources()).
That does not make sense IMO, because IA64 changes the IO port resources after calling acpi_pci_probe_root_resources(), hence the validation carried out in acpi_pci_probe_root_resources(), at least for IO ports seems wrong (what's the point of validating the CRS against ioport_resource if we change the _CRS IO space resources afterwards ?).
Second point: we are aware that by removing the offset addition in acpi_decode_space(), if for any reason on x86 or IA64 a resource has that offset !=0 (speaking in terms of memory resources for instance) we are in trouble. Jiang mentioned that on x86 and IA64 offset is always 0x0 for memory resources, but just want to make sure we are all aware of this potential pitfall.
Comments appreciated, it is time to a) fix IA64 and b) get this _CRS parsing consolidation done.
Thanks, Lorenzo
On Mon, 2016-01-25 at 09:52 +0000, Lorenzo Pieralisi wrote:
On Tue, Jan 19, 2016 at 12:20:26PM +0000, Lorenzo Pieralisi wrote:
Gerry,
On Wed, Jan 13, 2016 at 02:21:04PM +0100, Tomasz Nowicki wrote:
From: Liu Jiang jiang.liu@linux.intel.com
Some architectures, such as IA64 and ARM64, have no instructions to directly access PCI IO ports, so they map PCI IO ports into PCI MMIO address space. Typically PCI host bridges on those architectures take the responsibility to map (translate) PCI IO port transactions into Memory-Mapped IO transactions. ACPI specification provides support of such a usage case by using resource translation_offset.
But current ACPI resource parsing interface isn't neutral enough, it still has some special logic for IA64. So refine the ACPI resource parsing interface and IA64 code to neutrally handle translation_offset by:
- ACPI resource parsing interface doesn't do any translation, it just
save the translation_offset to be used by arch code. 2) Arch code will do the mapping(translation) based on arch specific information. Typically it does: 2.a) Translate per PCI domain IO port address space into system global IO port address space. 2.b) Setup MMIO address mapping for IO ports.
This patch fixes IO space handling on IA64 and should go in as a fix.
IA64 PCI IO space is currently broken (Hanjun tested this on an IA64 box).
The first broken commit is:
3772aea7d6f3 ("ia64/PCI/ACPI: Use common ACPI resource parsing interface for host bridge")
because acpi core code checks (in acpi_dev_ioresource_flags()) the resource.end>=0x10003, which fails on ia64 - currently resource.end is set in acpi_decode_space() to:
AddressMaximum + AddressTranslation
where AddressTranslation is the CPU physical address mapping IO space on IA64, the >=0x10003 check in acpi_dev_ioresource_flags always triggers and the IO resource is then disabled.
Do you want me to re-send this patch as a fix, with updated commit log ?
Two more points to discuss here. IA64, in its prepare_resources() callback calls acpi_pci_probe_root_resources() in turn. That function parses the _CRS and validate the resources (acpi_pci_root_validate_resources()).
That does not make sense IMO, because IA64 changes the IO port resources after calling acpi_pci_probe_root_resources(), hence the validation carried out in acpi_pci_probe_root_resources(), at least for IO ports seems wrong (what's the point of validating the CRS against ioport_resource if we change the _CRS IO space resources afterwards ?).
I agree that acpi_pci_root_validate_resources() is doing the wrong thing generally. The current code (without this patch series) is checking CPU bus address range for PCI IO window against ioport_resource which is based on ioport cookies made up by the kernel. This patch takes out the addition of the translation offset so the check is now PCI bus io address range validated against the cookies in ioport_resource. I think that works by chance because all ia64 PCI segments use the same PCI bus io address range based at zero. So the PCI bus io addresses will always look valid against ia64 IO space zero in the ioport_resource list. The PCI ioports actually get installed after ia64 changes the resource to hold the cookie. Interestingly, ia64 also installs an iomem resource for the CPU bus address of the ioport range window. The generic ACPI PCI host code should probably do the same in pci_acpi_root_prepare_resource(). And at some point, it might make sense to consolidate the ia64 ioport cookie handling with that done by the devicetree interfaces used by the generic ACPI PCI host code.
Second point: we are aware that by removing the offset addition in acpi_decode_space(), if for any reason on x86 or IA64 a resource has that offset !=0 (speaking in terms of memory resources for instance) we are in trouble. Jiang mentioned that on x86 and IA64 offset is always 0x0 for memory resources, but just want to make sure we are all aware of this potential pitfall.
I think that has always been the case. At least for x86 which doesn't appear to have ever used the offset in _CRS. In any case, it still leaves us with the same problem where acpi_pci_root_validate_resources() is validating PCI bus addresses against something different (cookies/cpu iospace for ioport, CPU bus for iomem).
Comments appreciated, it is time to a) fix IA64 and b) get this _CRS parsing consolidation done.
Thanks, Lorenzo
Hi Lorenzo,
On 2016/1/19 20:20, Lorenzo Pieralisi wrote:
Gerry,
On Wed, Jan 13, 2016 at 02:21:04PM +0100, Tomasz Nowicki wrote:
From: Liu Jiang jiang.liu@linux.intel.com
Some architectures, such as IA64 and ARM64, have no instructions to directly access PCI IO ports, so they map PCI IO ports into PCI MMIO address space. Typically PCI host bridges on those architectures take the responsibility to map (translate) PCI IO port transactions into Memory-Mapped IO transactions. ACPI specification provides support of such a usage case by using resource translation_offset.
But current ACPI resource parsing interface isn't neutral enough, it still has some special logic for IA64. So refine the ACPI resource parsing interface and IA64 code to neutrally handle translation_offset by:
- ACPI resource parsing interface doesn't do any translation, it just save the translation_offset to be used by arch code.
- Arch code will do the mapping(translation) based on arch specific information. Typically it does:
2.a) Translate per PCI domain IO port address space into system global IO port address space. 2.b) Setup MMIO address mapping for IO ports.
This patch fixes IO space handling on IA64 and should go in as a fix.
IA64 PCI IO space is currently broken (Hanjun tested this on an IA64 box).
The first broken commit is:
3772aea7d6f3 ("ia64/PCI/ACPI: Use common ACPI resource parsing interface for host bridge")
because acpi core code checks (in acpi_dev_ioresource_flags()) the resource.end>=0x10003, which fails on ia64 - currently resource.end is set in acpi_decode_space() to:
AddressMaximum + AddressTranslation
where AddressTranslation is the CPU physical address mapping IO space on IA64, the >=0x10003 check in acpi_dev_ioresource_flags always triggers and the IO resource is then disabled.
Do you want me to re-send this patch as a fix, with updated commit log ?
I talked to Gerry offline, he is busy these days, he said it's pretty OK for you to resend this patch and fix the problem.
Thanks Hanjun
Because of two patch series: 1. Jiang Liu's common interface to support PCI host controller init 2. MMCONFIG refactoring (part of this patch set) now we can think about generic ACPI based PCI host controller init implementation out of arch/ directory.
These calls use information from MCFG table (PCI config space regions) and _CRS method (IO/irq resources) to initialize PCI hostbridge.
TBD: We are still not sure whether we should reassign resources after PCI bus enumeration or trust firmware to do all that work for us properly.
Signed-off-by: Tomasz Nowicki tn@semihalf.com Signed-off-by: Hanjun Guo hanjun.guo@linaro.org Signed-off-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com CC: Arnd Bergmann arnd@arndb.de CC: Catalin Marinas catalin.marinas@arm.com CC: Liviu Dudau Liviu.Dudau@arm.com CC: Lorenzo Pieralisi Lorenzo.Pieralisi@arm.com CC: Will Deacon will.deacon@arm.com Tested-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com Tested-by: Jeremy Linton jeremy.linton@arm.com --- drivers/acpi/Kconfig | 5 ++ drivers/acpi/pci_root.c | 131 ++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 136 insertions(+)
diff --git a/drivers/acpi/Kconfig b/drivers/acpi/Kconfig index c3664be..e315061 100644 --- a/drivers/acpi/Kconfig +++ b/drivers/acpi/Kconfig @@ -335,6 +335,11 @@ config ACPI_PCI_SLOT i.e., segment/bus/device/function tuples, with physical slots in the system. If you are unsure, say N.
+config ACPI_PCI_HOST_GENERIC + bool "Generic ACPI PCI host controller" + help + Say Y here if you want to support generic ACPI PCI host controller. + config X86_PM_TIMER bool "Power Management Timer Support" if EXPERT depends on X86 diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c index a65c8c2..d483e2a 100644 --- a/drivers/acpi/pci_root.c +++ b/drivers/acpi/pci_root.c @@ -24,6 +24,7 @@ #include <linux/init.h> #include <linux/types.h> #include <linux/mutex.h> +#include <linux/of_address.h> #include <linux/pm.h> #include <linux/pm_runtime.h> #include <linux/pci.h> @@ -514,6 +515,136 @@ static void negotiate_os_control(struct acpi_pci_root *root, int *no_aspm) } }
+#ifdef CONFIG_ACPI_PCI_HOST_GENERIC +static int pcibios_map_irq(struct pci_dev *dev, u8 slot, u8 pin) +{ + if (pci_dev_msi_enabled(dev)) + return 0; + + acpi_pci_irq_enable(dev); + return dev->irq; +} + +static void pci_mcfg_release_info(struct acpi_pci_root_info *ci) +{ + pci_mmcfg_teardown_map(ci); + kfree(ci); +} + +static int pci_acpi_root_prepare_resources(struct acpi_pci_root_info *ci) +{ + struct list_head *list = &ci->resources; + struct acpi_device *device = ci->bridge; + struct resource_entry *entry, *tmp; + unsigned long flags; + int ret; + + flags = IORESOURCE_IO | IORESOURCE_MEM; + ret = acpi_dev_get_resources(device, list, + acpi_dev_filter_resource_type_cb, + (void *)flags); + if (ret < 0) { + dev_warn(&device->dev, + "failed to parse _CRS method, error code %d\n", ret); + return ret; + } else if (ret == 0) + dev_dbg(&device->dev, + "no IO and memory resources present in _CRS\n"); + + resource_list_for_each_entry_safe(entry, tmp, &ci->resources) { + resource_size_t cpu_addr, length; + struct resource *res = entry->res; + + if (entry->res->flags & IORESOURCE_DISABLED) + resource_list_destroy_entry(entry); + else + res->name = ci->name; + + /* PCI -> CPU space translation */ + cpu_addr = res->start + entry->offset; + length = res->end - res->start + 1; + + if (res->flags & IORESOURCE_MEM) { + res->start = cpu_addr; + res->end = cpu_addr + length - 1; + } else if (res->flags & IORESOURCE_IO) { + resource_size_t pci_addr = res->start; + unsigned long port; + + if (pci_register_io_range(cpu_addr, length)) { + resource_list_destroy_entry(entry); + continue; + } + + port = pci_address_to_pio(cpu_addr); + if (port == (unsigned long)-1) { + resource_list_destroy_entry(entry); + continue; + } + + res->start = port; + res->end = port + length - 1; + entry->offset = port - pci_addr; + + if (pci_remap_iospace(res, cpu_addr) < 0) + resource_list_destroy_entry(entry); + } + } + return ret; +} + +static struct acpi_pci_root_ops acpi_pci_root_ops = { + .init_info = pci_mmcfg_setup_map, + .release_info = pci_mcfg_release_info, + .prepare_resources = pci_acpi_root_prepare_resources, +}; + +/* Root bridge scanning */ +struct pci_bus *pci_acpi_scan_root(struct acpi_pci_root *root) +{ + int node = acpi_get_node(root->device->handle); + int domain = root->segment; + int busnum = root->secondary.start; + struct acpi_pci_root_info *info; + struct pci_host_bridge *bridge; + struct pci_bus *bus, *child; + + if (domain && !pci_domains_supported) { + pr_warn("PCI %04x:%02x: multiple domains not supported.\n", + domain, busnum); + return NULL; + } + + info = kzalloc_node(sizeof(*info), GFP_KERNEL, node); + if (!info) { + dev_err(&root->device->dev, + "pci_bus %04x:%02x: ignored (out of memory)\n", + domain, busnum); + return NULL; + } + + acpi_pci_root_ops.pci_ops = pci_mcfg_get_ops(root); + bus = acpi_pci_root_create(root, &acpi_pci_root_ops, info, root); + if (!bus) + return NULL; + + bridge = pci_find_host_bridge(bus); + bridge->map_irq = pcibios_map_irq; + + pci_bus_size_bridges(bus); + pci_bus_assign_resources(bus); + + /* + * After the PCI-E bus has been walked and all devices discovered, + * configure any settings of the fabric that might be necessary. + */ + list_for_each_entry(child, &bus->children, node) + pcie_bus_configure_settings(child); + + return bus; +} +#endif /* CONFIG_ARCH_PCI_HOST_GENERIC_ACPI */ + static int acpi_pci_root_add(struct acpi_device *device, const struct acpi_device_id *not_used) {
Hi Tomasz,
On 2016/1/13 21:21, Tomasz Nowicki wrote:
Because of two patch series:
- Jiang Liu's common interface to support PCI host controller init
- MMCONFIG refactoring (part of this patch set)
now we can think about generic ACPI based PCI host controller init implementation out of arch/ directory.
These calls use information from MCFG table (PCI config space regions) and _CRS method (IO/irq resources) to initialize PCI hostbridge.
TBD: We are still not sure whether we should reassign resources after PCI bus enumeration or trust firmware to do all that work for us properly.
Signed-off-by: Tomasz Nowicki tn@semihalf.com Signed-off-by: Hanjun Guo hanjun.guo@linaro.org Signed-off-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com CC: Arnd Bergmann arnd@arndb.de CC: Catalin Marinas catalin.marinas@arm.com CC: Liviu Dudau Liviu.Dudau@arm.com CC: Lorenzo Pieralisi Lorenzo.Pieralisi@arm.com CC: Will Deacon will.deacon@arm.com Tested-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com Tested-by: Jeremy Linton jeremy.linton@arm.com
drivers/acpi/Kconfig | 5 ++ drivers/acpi/pci_root.c | 131 ++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 136 insertions(+)
diff --git a/drivers/acpi/Kconfig b/drivers/acpi/Kconfig index c3664be..e315061 100644 --- a/drivers/acpi/Kconfig +++ b/drivers/acpi/Kconfig @@ -335,6 +335,11 @@ config ACPI_PCI_SLOT i.e., segment/bus/device/function tuples, with physical slots in the system. If you are unsure, say N.
+config ACPI_PCI_HOST_GENERIC
- bool "Generic ACPI PCI host controller"
- help
Say Y here if you want to support generic ACPI PCI host controller.
Since x86 and IA64 will not use ACPI_PCI_HOST_GENERIC in this patch, it should be defined as
config ACPI_PCI_HOST_GENERIC bool
and select for ARCHs, or will default y on x86 and IA64 too.
Compiling the kernel on a IA64 machine, will send out the result when it's ready.
Thanks Hanjun
On 15.01.2016 10:57, Hanjun Guo wrote:
Hi Tomasz,
On 2016/1/13 21:21, Tomasz Nowicki wrote:
Because of two patch series:
- Jiang Liu's common interface to support PCI host controller init
- MMCONFIG refactoring (part of this patch set)
now we can think about generic ACPI based PCI host controller init implementation out of arch/ directory.
These calls use information from MCFG table (PCI config space regions) and _CRS method (IO/irq resources) to initialize PCI hostbridge.
TBD: We are still not sure whether we should reassign resources after PCI bus enumeration or trust firmware to do all that work for us properly.
Signed-off-by: Tomasz Nowicki tn@semihalf.com Signed-off-by: Hanjun Guo hanjun.guo@linaro.org Signed-off-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com CC: Arnd Bergmann arnd@arndb.de CC: Catalin Marinas catalin.marinas@arm.com CC: Liviu Dudau Liviu.Dudau@arm.com CC: Lorenzo Pieralisi Lorenzo.Pieralisi@arm.com CC: Will Deacon will.deacon@arm.com Tested-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com Tested-by: Jeremy Linton jeremy.linton@arm.com
drivers/acpi/Kconfig | 5 ++ drivers/acpi/pci_root.c | 131 ++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 136 insertions(+)
diff --git a/drivers/acpi/Kconfig b/drivers/acpi/Kconfig index c3664be..e315061 100644 --- a/drivers/acpi/Kconfig +++ b/drivers/acpi/Kconfig @@ -335,6 +335,11 @@ config ACPI_PCI_SLOT i.e., segment/bus/device/function tuples, with physical slots in the system. If you are unsure, say N.
+config ACPI_PCI_HOST_GENERIC
- bool "Generic ACPI PCI host controller"
- help
Say Y here if you want to support generic ACPI PCI host
controller.
Since x86 and IA64 will not use ACPI_PCI_HOST_GENERIC in this patch, it should be defined as
config ACPI_PCI_HOST_GENERIC bool
and select for ARCHs, or will default y on x86 and IA64 too.
Right, I will fix that in v4.
Compiling the kernel on a IA64 machine, will send out the result when it's ready.
Thanks, Tomasz
Hi Tomasz,
在 2016/1/13 21:21, Tomasz Nowicki 写道:
Because of two patch series:
- Jiang Liu's common interface to support PCI host controller init
- MMCONFIG refactoring (part of this patch set)
now we can think about generic ACPI based PCI host controller init implementation out of arch/ directory.
These calls use information from MCFG table (PCI config space regions) and _CRS method (IO/irq resources) to initialize PCI hostbridge.
TBD: We are still not sure whether we should reassign resources after PCI bus enumeration or trust firmware to do all that work for us properly.
Signed-off-by: Tomasz Nowicki tn@semihalf.com Signed-off-by: Hanjun Guo hanjun.guo@linaro.org Signed-off-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com CC: Arnd Bergmann arnd@arndb.de CC: Catalin Marinas catalin.marinas@arm.com CC: Liviu Dudau Liviu.Dudau@arm.com CC: Lorenzo Pieralisi Lorenzo.Pieralisi@arm.com CC: Will Deacon will.deacon@arm.com Tested-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com Tested-by: Jeremy Linton jeremy.linton@arm.com
drivers/acpi/Kconfig | 5 ++ drivers/acpi/pci_root.c | 131 ++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 136 insertions(+)
diff --git a/drivers/acpi/Kconfig b/drivers/acpi/Kconfig index c3664be..e315061 100644 --- a/drivers/acpi/Kconfig +++ b/drivers/acpi/Kconfig @@ -335,6 +335,11 @@ config ACPI_PCI_SLOT i.e., segment/bus/device/function tuples, with physical slots in the system. If you are unsure, say N.
+config ACPI_PCI_HOST_GENERIC
- bool "Generic ACPI PCI host controller"
- help
Say Y here if you want to support generic ACPI PCI host controller.
- config X86_PM_TIMER bool "Power Management Timer Support" if EXPERT depends on X86
diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c index a65c8c2..d483e2a 100644 --- a/drivers/acpi/pci_root.c +++ b/drivers/acpi/pci_root.c @@ -24,6 +24,7 @@ #include <linux/init.h> #include <linux/types.h> #include <linux/mutex.h> +#include <linux/of_address.h> #include <linux/pm.h> #include <linux/pm_runtime.h> #include <linux/pci.h> @@ -514,6 +515,136 @@ static void negotiate_os_control(struct acpi_pci_root *root, int *no_aspm) } }
+#ifdef CONFIG_ACPI_PCI_HOST_GENERIC +static int pcibios_map_irq(struct pci_dev *dev, u8 slot, u8 pin) +{
- if (pci_dev_msi_enabled(dev))
return 0;
- acpi_pci_irq_enable(dev);
- return dev->irq;
+}
+static void pci_mcfg_release_info(struct acpi_pci_root_info *ci) +{
- pci_mmcfg_teardown_map(ci);
- kfree(ci);
+}
+static int pci_acpi_root_prepare_resources(struct acpi_pci_root_info *ci) +{
- struct list_head *list = &ci->resources;
- struct acpi_device *device = ci->bridge;
- struct resource_entry *entry, *tmp;
- unsigned long flags;
- int ret;
- flags = IORESOURCE_IO | IORESOURCE_MEM;
- ret = acpi_dev_get_resources(device, list,
acpi_dev_filter_resource_type_cb,
(void *)flags);
- if (ret < 0) {
dev_warn(&device->dev,
"failed to parse _CRS method, error code %d\n", ret);
return ret;
- } else if (ret == 0)
dev_dbg(&device->dev,
"no IO and memory resources present in _CRS\n");
- resource_list_for_each_entry_safe(entry, tmp, &ci->resources) {
resource_size_t cpu_addr, length;
struct resource *res = entry->res;
if (entry->res->flags & IORESOURCE_DISABLED)
resource_list_destroy_entry(entry);
else
res->name = ci->name;
/* PCI -> CPU space translation */
cpu_addr = res->start + entry->offset;
length = res->end - res->start + 1;
I see here is different from V2 patch, in V2 patch, 0x0000022004000000 is cpu addr. but in V3 patch, 0x0000022004000000 is pci addr. which one is right ?
0x0000022004000000 is the value of AddressMinimum. AddressMinimum evaluates to a 64-bit integer that specifies the lowest possible base address of the Memory range QWordMemory ( // 64-bit BAR Windows ResourceProducer, PosDecode, MinFixed, MaxFixed, Cacheable, ReadWrite, 0x000000000000000, // Granularity 0x0000022004000000, // Min Base Address 0x000002200fffffff, // Max Base Address 0x0000021ff9000000, // Translate 0x000000000c000000 // Length )
Thanks Dongdong
if (res->flags & IORESOURCE_MEM) {
res->start = cpu_addr;
res->end = cpu_addr + length - 1;
} else if (res->flags & IORESOURCE_IO) {
resource_size_t pci_addr = res->start;
unsigned long port;
if (pci_register_io_range(cpu_addr, length)) {
resource_list_destroy_entry(entry);
continue;
}
port = pci_address_to_pio(cpu_addr);
if (port == (unsigned long)-1) {
resource_list_destroy_entry(entry);
continue;
}
res->start = port;
res->end = port + length - 1;
entry->offset = port - pci_addr;
if (pci_remap_iospace(res, cpu_addr) < 0)
resource_list_destroy_entry(entry);
}
- }
- return ret;
+}
+static struct acpi_pci_root_ops acpi_pci_root_ops = {
- .init_info = pci_mmcfg_setup_map,
- .release_info = pci_mcfg_release_info,
- .prepare_resources = pci_acpi_root_prepare_resources,
+};
+/* Root bridge scanning */ +struct pci_bus *pci_acpi_scan_root(struct acpi_pci_root *root) +{
- int node = acpi_get_node(root->device->handle);
- int domain = root->segment;
- int busnum = root->secondary.start;
- struct acpi_pci_root_info *info;
- struct pci_host_bridge *bridge;
- struct pci_bus *bus, *child;
- if (domain && !pci_domains_supported) {
pr_warn("PCI %04x:%02x: multiple domains not supported.\n",
domain, busnum);
return NULL;
- }
- info = kzalloc_node(sizeof(*info), GFP_KERNEL, node);
- if (!info) {
dev_err(&root->device->dev,
"pci_bus %04x:%02x: ignored (out of memory)\n",
domain, busnum);
return NULL;
- }
- acpi_pci_root_ops.pci_ops = pci_mcfg_get_ops(root);
- bus = acpi_pci_root_create(root, &acpi_pci_root_ops, info, root);
- if (!bus)
return NULL;
- bridge = pci_find_host_bridge(bus);
- bridge->map_irq = pcibios_map_irq;
- pci_bus_size_bridges(bus);
- pci_bus_assign_resources(bus);
- /*
* After the PCI-E bus has been walked and all devices discovered,
* configure any settings of the fabric that might be necessary.
*/
- list_for_each_entry(child, &bus->children, node)
pcie_bus_configure_settings(child);
- return bus;
+} +#endif /* CONFIG_ARCH_PCI_HOST_GENERIC_ACPI */
- static int acpi_pci_root_add(struct acpi_device *device, const struct acpi_device_id *not_used) {
On 18.01.2016 10:25, liudongdong (C) wrote:
I see here is different from V2 patch, in V2 patch, 0x0000022004000000 is cpu addr. but in V3 patch, 0x0000022004000000 is pci addr. which one is right ?
You are right, I look back at v2 and see bug there. But v3 fixes it, for MEM & IO we always need to calculate this way: CPU address defined as: cpu_addr = res->start + entry->offset; PCI address: pci_addr = res->start;
0x0000022004000000 is the value of AddressMinimum. AddressMinimum evaluates to a 64-bit integer that specifies the lowest possible base address of the Memory range QWordMemory ( // 64-bit BAR Windows ResourceProducer, PosDecode, MinFixed, MaxFixed, Cacheable, ReadWrite, 0x000000000000000, // Granularity 0x0000022004000000, // Min Base Address 0x000002200fffffff, // Max Base Address 0x0000021ff9000000, // Translate 0x000000000c000000 // Length )
For your case: cpu_addr = 0x0000022004000000 + 0x0000021ff9000000: pci_addr = 0x0000022004000000;
Regards, Tomasz
On Wed, Jan 13, 2016 at 02:21:05PM +0100, Tomasz Nowicki wrote:
Because of two patch series:
- Jiang Liu's common interface to support PCI host controller init
- MMCONFIG refactoring (part of this patch set)
now we can think about generic ACPI based PCI host controller init implementation out of arch/ directory.
These calls use information from MCFG table (PCI config space regions) and _CRS method (IO/irq resources) to initialize PCI hostbridge.
TBD: We are still not sure whether we should reassign resources after PCI bus enumeration or trust firmware to do all that work for us properly.
We should claim resources and assign unassigned ones. I put together a patch for resource claiming instead of reinventing the wheel, waiting for feedback:
https://patchwork.ozlabs.org/patch/545669/
If we merge the code with no resources claiming, we may end up in situations where claiming can trigger regressions so we won't be able to do it anymore.
Signed-off-by: Tomasz Nowicki tn@semihalf.com Signed-off-by: Hanjun Guo hanjun.guo@linaro.org Signed-off-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com CC: Arnd Bergmann arnd@arndb.de CC: Catalin Marinas catalin.marinas@arm.com CC: Liviu Dudau Liviu.Dudau@arm.com CC: Lorenzo Pieralisi Lorenzo.Pieralisi@arm.com CC: Will Deacon will.deacon@arm.com Tested-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com Tested-by: Jeremy Linton jeremy.linton@arm.com
drivers/acpi/Kconfig | 5 ++ drivers/acpi/pci_root.c | 131 ++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 136 insertions(+)
diff --git a/drivers/acpi/Kconfig b/drivers/acpi/Kconfig index c3664be..e315061 100644 --- a/drivers/acpi/Kconfig +++ b/drivers/acpi/Kconfig @@ -335,6 +335,11 @@ config ACPI_PCI_SLOT i.e., segment/bus/device/function tuples, with physical slots in the system. If you are unsure, say N. +config ACPI_PCI_HOST_GENERIC
- bool "Generic ACPI PCI host controller"
- help
Say Y here if you want to support generic ACPI PCI host controller.
You should add a proper description here.
config X86_PM_TIMER bool "Power Management Timer Support" if EXPERT depends on X86 diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c index a65c8c2..d483e2a 100644 --- a/drivers/acpi/pci_root.c +++ b/drivers/acpi/pci_root.c @@ -24,6 +24,7 @@ #include <linux/init.h> #include <linux/types.h> #include <linux/mutex.h> +#include <linux/of_address.h>
We should move the IO space management to PCI core instead of having it in OF core, code carrying out PIO mapping does not depend on OF as far as I can see.
#include <linux/pm.h> #include <linux/pm_runtime.h> #include <linux/pci.h> @@ -514,6 +515,136 @@ static void negotiate_os_control(struct acpi_pci_root *root, int *no_aspm) } } +#ifdef CONFIG_ACPI_PCI_HOST_GENERIC +static int pcibios_map_irq(struct pci_dev *dev, u8 slot, u8 pin) +{
- if (pci_dev_msi_enabled(dev))
return 0;
- acpi_pci_irq_enable(dev);
- return dev->irq;
+}
+static void pci_mcfg_release_info(struct acpi_pci_root_info *ci) +{
- pci_mmcfg_teardown_map(ci);
- kfree(ci);
+}
+static int pci_acpi_root_prepare_resources(struct acpi_pci_root_info *ci) +{
- struct list_head *list = &ci->resources;
- struct acpi_device *device = ci->bridge;
- struct resource_entry *entry, *tmp;
- unsigned long flags;
- int ret;
- flags = IORESOURCE_IO | IORESOURCE_MEM;
- ret = acpi_dev_get_resources(device, list,
acpi_dev_filter_resource_type_cb,
(void *)flags);
- if (ret < 0) {
dev_warn(&device->dev,
"failed to parse _CRS method, error code %d\n", ret);
return ret;
- } else if (ret == 0)
dev_dbg(&device->dev,
"no IO and memory resources present in _CRS\n");
^^^^ what's the point in carrying on then ?
- resource_list_for_each_entry_safe(entry, tmp, &ci->resources) {
resource_size_t cpu_addr, length;
struct resource *res = entry->res;
if (entry->res->flags & IORESOURCE_DISABLED)
resource_list_destroy_entry(entry);
else
res->name = ci->name;
/* PCI -> CPU space translation */
cpu_addr = res->start + entry->offset;
length = res->end - res->start + 1;
if (res->flags & IORESOURCE_MEM) {
res->start = cpu_addr;
res->end = cpu_addr + length - 1;
} else if (res->flags & IORESOURCE_IO) {
resource_size_t pci_addr = res->start;
unsigned long port;
if (pci_register_io_range(cpu_addr, length)) {
resource_list_destroy_entry(entry);
continue;
}
port = pci_address_to_pio(cpu_addr);
if (port == (unsigned long)-1) {
resource_list_destroy_entry(entry);
continue;
}
res->start = port;
res->end = port + length - 1;
entry->offset = port - pci_addr;
if (pci_remap_iospace(res, cpu_addr) < 0)
resource_list_destroy_entry(entry);
}
- }
- return ret;
+}
+static struct acpi_pci_root_ops acpi_pci_root_ops = {
- .init_info = pci_mmcfg_setup_map,
- .release_info = pci_mcfg_release_info,
- .prepare_resources = pci_acpi_root_prepare_resources,
+};
+/* Root bridge scanning */ +struct pci_bus *pci_acpi_scan_root(struct acpi_pci_root *root) +{
- int node = acpi_get_node(root->device->handle);
- int domain = root->segment;
- int busnum = root->secondary.start;
- struct acpi_pci_root_info *info;
- struct pci_host_bridge *bridge;
- struct pci_bus *bus, *child;
- if (domain && !pci_domains_supported) {
pr_warn("PCI %04x:%02x: multiple domains not supported.\n",
domain, busnum);
return NULL;
- }
- info = kzalloc_node(sizeof(*info), GFP_KERNEL, node);
- if (!info) {
dev_err(&root->device->dev,
"pci_bus %04x:%02x: ignored (out of memory)\n",
domain, busnum);
return NULL;
- }
- acpi_pci_root_ops.pci_ops = pci_mcfg_get_ops(root);
- bus = acpi_pci_root_create(root, &acpi_pci_root_ops, info, root);
- if (!bus)
return NULL;
^^^ Leaking memory here.
- bridge = pci_find_host_bridge(bus);
- bridge->map_irq = pcibios_map_irq;
It would be nice to use map_irq for that, but Matthew's series seems stuck in review mode, either we take that series on and make some progress on it or you should add the irq mapping code to arm64 arch code, *temporarily* :)
Also, we should claim resources here.
- pci_bus_size_bridges(bus);
- pci_bus_assign_resources(bus);
- /*
* After the PCI-E bus has been walked and all devices discovered,
* configure any settings of the fabric that might be necessary.
*/
- list_for_each_entry(child, &bus->children, node)
pcie_bus_configure_settings(child);
- return bus;
+} +#endif /* CONFIG_ARCH_PCI_HOST_GENERIC_ACPI */
^^^ does not match the #ifdef
Lorenzo
On 19.01.2016 12:58, Lorenzo Pieralisi wrote:
On Wed, Jan 13, 2016 at 02:21:05PM +0100, Tomasz Nowicki wrote:
Because of two patch series:
- Jiang Liu's common interface to support PCI host controller init
- MMCONFIG refactoring (part of this patch set)
now we can think about generic ACPI based PCI host controller init implementation out of arch/ directory.
These calls use information from MCFG table (PCI config space regions) and _CRS method (IO/irq resources) to initialize PCI hostbridge.
TBD: We are still not sure whether we should reassign resources after PCI bus enumeration or trust firmware to do all that work for us properly.
We should claim resources and assign unassigned ones. I put together a patch for resource claiming instead of reinventing the wheel, waiting for feedback:
https://patchwork.ozlabs.org/patch/545669/
If we merge the code with no resources claiming, we may end up in situations where claiming can trigger regressions so we won't be able to do it anymore.
Signed-off-by: Tomasz Nowicki tn@semihalf.com Signed-off-by: Hanjun Guo hanjun.guo@linaro.org Signed-off-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com CC: Arnd Bergmann arnd@arndb.de CC: Catalin Marinas catalin.marinas@arm.com CC: Liviu Dudau Liviu.Dudau@arm.com CC: Lorenzo Pieralisi Lorenzo.Pieralisi@arm.com CC: Will Deacon will.deacon@arm.com Tested-by: Suravee Suthikulpanit Suravee.Suthikulpanit@amd.com Tested-by: Jeremy Linton jeremy.linton@arm.com
drivers/acpi/Kconfig | 5 ++ drivers/acpi/pci_root.c | 131 ++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 136 insertions(+)
diff --git a/drivers/acpi/Kconfig b/drivers/acpi/Kconfig index c3664be..e315061 100644 --- a/drivers/acpi/Kconfig +++ b/drivers/acpi/Kconfig @@ -335,6 +335,11 @@ config ACPI_PCI_SLOT i.e., segment/bus/device/function tuples, with physical slots in the system. If you are unsure, say N.
+config ACPI_PCI_HOST_GENERIC
- bool "Generic ACPI PCI host controller"
- help
Say Y here if you want to support generic ACPI PCI host controller.
You should add a proper description here.
Will do.
- config X86_PM_TIMER bool "Power Management Timer Support" if EXPERT depends on X86
diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c index a65c8c2..d483e2a 100644 --- a/drivers/acpi/pci_root.c +++ b/drivers/acpi/pci_root.c @@ -24,6 +24,7 @@ #include <linux/init.h> #include <linux/types.h> #include <linux/mutex.h> +#include <linux/of_address.h>
We should move the IO space management to PCI core instead of having it in OF core, code carrying out PIO mapping does not depend on OF as far as I can see.
Yes, this should be cleaned up, I will add another patch in the next series.
#include <linux/pm.h> #include <linux/pm_runtime.h> #include <linux/pci.h> @@ -514,6 +515,136 @@ static void negotiate_os_control(struct acpi_pci_root *root, int *no_aspm) } }
+#ifdef CONFIG_ACPI_PCI_HOST_GENERIC +static int pcibios_map_irq(struct pci_dev *dev, u8 slot, u8 pin) +{
- if (pci_dev_msi_enabled(dev))
return 0;
- acpi_pci_irq_enable(dev);
- return dev->irq;
+}
+static void pci_mcfg_release_info(struct acpi_pci_root_info *ci) +{
- pci_mmcfg_teardown_map(ci);
- kfree(ci);
+}
+static int pci_acpi_root_prepare_resources(struct acpi_pci_root_info *ci) +{
- struct list_head *list = &ci->resources;
- struct acpi_device *device = ci->bridge;
- struct resource_entry *entry, *tmp;
- unsigned long flags;
- int ret;
- flags = IORESOURCE_IO | IORESOURCE_MEM;
- ret = acpi_dev_get_resources(device, list,
acpi_dev_filter_resource_type_cb,
(void *)flags);
- if (ret < 0) {
dev_warn(&device->dev,
"failed to parse _CRS method, error code %d\n", ret);
return ret;
- } else if (ret == 0)
dev_dbg(&device->dev,
"no IO and memory resources present in _CRS\n");
^^^^ what's the point in carrying on then ?
There is no point, hence resource_list_for_each_entry_safe will not iterate &ci->resources and simply return. If you like I can add here return to make code more readable.
- resource_list_for_each_entry_safe(entry, tmp, &ci->resources) {
resource_size_t cpu_addr, length;
struct resource *res = entry->res;
if (entry->res->flags & IORESOURCE_DISABLED)
resource_list_destroy_entry(entry);
else
res->name = ci->name;
/* PCI -> CPU space translation */
cpu_addr = res->start + entry->offset;
length = res->end - res->start + 1;
if (res->flags & IORESOURCE_MEM) {
res->start = cpu_addr;
res->end = cpu_addr + length - 1;
} else if (res->flags & IORESOURCE_IO) {
resource_size_t pci_addr = res->start;
unsigned long port;
if (pci_register_io_range(cpu_addr, length)) {
resource_list_destroy_entry(entry);
continue;
}
port = pci_address_to_pio(cpu_addr);
if (port == (unsigned long)-1) {
resource_list_destroy_entry(entry);
continue;
}
res->start = port;
res->end = port + length - 1;
entry->offset = port - pci_addr;
if (pci_remap_iospace(res, cpu_addr) < 0)
resource_list_destroy_entry(entry);
}
- }
- return ret;
+}
+static struct acpi_pci_root_ops acpi_pci_root_ops = {
- .init_info = pci_mmcfg_setup_map,
- .release_info = pci_mcfg_release_info,
- .prepare_resources = pci_acpi_root_prepare_resources,
+};
+/* Root bridge scanning */ +struct pci_bus *pci_acpi_scan_root(struct acpi_pci_root *root) +{
- int node = acpi_get_node(root->device->handle);
- int domain = root->segment;
- int busnum = root->secondary.start;
- struct acpi_pci_root_info *info;
- struct pci_host_bridge *bridge;
- struct pci_bus *bus, *child;
- if (domain && !pci_domains_supported) {
pr_warn("PCI %04x:%02x: multiple domains not supported.\n",
domain, busnum);
return NULL;
- }
- info = kzalloc_node(sizeof(*info), GFP_KERNEL, node);
- if (!info) {
dev_err(&root->device->dev,
"pci_bus %04x:%02x: ignored (out of memory)\n",
domain, busnum);
return NULL;
- }
- acpi_pci_root_ops.pci_ops = pci_mcfg_get_ops(root);
- bus = acpi_pci_root_create(root, &acpi_pci_root_ops, info, root);
- if (!bus)
return NULL;
^^^ Leaking memory here.
No leak here. See acpi_pci_root_create->__acpi_pci_root_release_info
- bridge = pci_find_host_bridge(bus);
- bridge->map_irq = pcibios_map_irq;
It would be nice to use map_irq for that, but Matthew's series seems stuck in review mode, either we take that series on and make some progress on it or you should add the irq mapping code to arm64 arch code, *temporarily* :)
Right, I decided to decouple this and Matthew's series.
Also, we should claim resources here.
Let me investigate it more and get back to you.
- pci_bus_size_bridges(bus);
- pci_bus_assign_resources(bus);
- /*
* After the PCI-E bus has been walked and all devices discovered,
* configure any settings of the fabric that might be necessary.
*/
- list_for_each_entry(child, &bus->children, node)
pcie_bus_configure_settings(child);
- return bus;
+} +#endif /* CONFIG_ARCH_PCI_HOST_GENERIC_ACPI */
^^^ does not match the #ifdef
Fixed.
Thanks, Tomasz
Some platforms may not be fully compliant with generic set of PCI config accessors. For these cases we implement the way to overwrite accessors set prior to PCI buses enumeration. Algorithm traverses available quirk list, matches against <platform ID (DMI), domain, bus number> tuple and returns corresponding accessors. All quirks can be defined using: DECLARE_ACPI_MCFG_FIXUP() and keep self contained. Example,
static const struct dmi_system_id foo_dmi[] = { { .ident = "<Platform ident string>", .callback = <handler>, .matches = { DMI_MATCH(DMI_SYS_VENDOR, "<system vendor>"), DMI_MATCH(DMI_PRODUCT_NAME, "<product name>"), DMI_MATCH(DMI_PRODUCT_VERSION, "product version"), }, }, { } };
static struct pci_ops foo_ecam_pci_ops = { .map_bus = pci_mcfg_dev_base, .read = foo_ecam_config_read, .write = foo_ecam_config_write, }; DECLARE_ACPI_MCFG_FIXUP(foo_dmi, NULL, &foo_ecam_pci_ops, <domain_nr>, <bus_nr>);
More custom (non-DMI) matching can be done via an alternative call. Note that there is possibility to assign quirk related private data to root->sysdata which will be available along read/wriate accessor, example:
static int boo_match(struct pci_mcfg_fixup *fixup, struct acpi_pci_root *root) { return [condition] ? 1 : 0; }
int boo_ecam_config_read(struct pci_bus *bus, unsigned int devfn, int where, int size, u32 *val) { struct acpi_pci_root *root = bus->sysdata; struct boo_priv_data *boo_data = root->sysdata;
[..] }
static struct pci_ops boo_ecam_pci_ops = { .map_bus = pci_mcfg_dev_base, .read = boo_ecam_config_read, .write = boo_ecam_config_write, }; DECLARE_ACPI_MCFG_FIXUP(NULL, boo_match, &boo_ecam_pci_ops, <domain_nr>, <bus_nr>);
Signed-off-by: Tomasz Nowicki tn@semihalf.com --- drivers/acpi/mcfg.c | 33 +++++++++++++++++++++++++++++++-- include/acpi/acpi_bus.h | 1 + include/asm-generic/vmlinux.lds.h | 7 +++++++ include/linux/ecam.h | 18 ++++++++++++++++++ 4 files changed, 57 insertions(+), 2 deletions(-)
diff --git a/drivers/acpi/mcfg.c b/drivers/acpi/mcfg.c index dfc2d14..ec5fe7b 100644 --- a/drivers/acpi/mcfg.c +++ b/drivers/acpi/mcfg.c @@ -8,6 +8,7 @@ */
#include <linux/acpi.h> +#include <linux/dmi.h> #include <linux/ecam.h> #include <linux/pci.h> #include <linux/pci-acpi.h> @@ -34,6 +35,29 @@ int __weak raw_pci_write(unsigned int domain, unsigned int bus, return PCIBIOS_DEVICE_NOT_FOUND; }
+extern struct pci_mcfg_fixup __start_acpi_mcfg_fixups[]; +extern struct pci_mcfg_fixup __end_acpi_mcfg_fixups[]; + +static struct pci_ops *pci_mcfg_check_quirks(struct acpi_pci_root *root) +{ + struct pci_mcfg_fixup *f; + int bus_num = root->secondary.start; + int domain = root->segment; + + /* + * First match against PCI topology domain:bus then use DMI or + * custom match handler. + */ + for (f = __start_acpi_mcfg_fixups; f < __end_acpi_mcfg_fixups; f++) { + if ((f->domain == domain || f->domain == PCI_MCFG_DOMAIN_ANY) && + (f->bus_num == bus_num || f->bus_num == PCI_MCFG_BUS_ANY) && + (f->system ? dmi_check_system(f->system) : 0 || + f->match ? f->match(f, root) : 0)) + return f->ops; + } + return NULL; +} + void __iomem * pci_mcfg_dev_base(struct pci_bus *bus, unsigned int devfn, int offset) { @@ -56,10 +80,15 @@ static struct pci_ops default_pci_mcfg_ops = {
struct pci_ops *pci_mcfg_get_ops(struct acpi_pci_root *root) { + struct pci_ops *pci_mcfg_ops_quirk; + /* - * TODO: Match against platform specific quirks and return - * corresponding PCI config space accessor set. + * Match against platform specific quirks and return corresponding + * PCI config space accessor set. */ + pci_mcfg_ops_quirk = pci_mcfg_check_quirks(root); + if (pci_mcfg_ops_quirk) + return pci_mcfg_ops_quirk;
return &default_pci_mcfg_ops; } diff --git a/include/acpi/acpi_bus.h b/include/acpi/acpi_bus.h index ad0a5ff..ea4d2b5 100644 --- a/include/acpi/acpi_bus.h +++ b/include/acpi/acpi_bus.h @@ -554,6 +554,7 @@ struct acpi_pci_root { struct pci_bus *bus; u16 segment; struct resource secondary; /* downstream bus range */ + void *sysdata;
u32 osc_support_set; /* _OSC state of support bits */ u32 osc_control_set; /* _OSC state of control bits */ diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h index c4bd0e2..c93fc97 100644 --- a/include/asm-generic/vmlinux.lds.h +++ b/include/asm-generic/vmlinux.lds.h @@ -298,6 +298,13 @@ VMLINUX_SYMBOL(__end_pci_fixups_suspend_late) = .; \ } \ \ + /* ACPI MCFG quirks */ \ + .acpi_fixup : AT(ADDR(.acpi_fixup) - LOAD_OFFSET) { \ + VMLINUX_SYMBOL(__start_acpi_mcfg_fixups) = .; \ + *(.acpi_fixup_mcfg) \ + VMLINUX_SYMBOL(__end_acpi_mcfg_fixups) = .; \ + } \ + \ /* Built-in firmware blobs */ \ .builtin_fw : AT(ADDR(.builtin_fw) - LOAD_OFFSET) { \ VMLINUX_SYMBOL(__start_builtin_fw) = .; \ diff --git a/include/linux/ecam.h b/include/linux/ecam.h index e0f322e..21215be 100644 --- a/include/linux/ecam.h +++ b/include/linux/ecam.h @@ -20,6 +20,24 @@ struct pci_mmcfg_region { bool hot_added; };
+struct pci_mcfg_fixup { + const struct dmi_system_id *system; + int (*match)(struct pci_mcfg_fixup *, struct acpi_pci_root *); + struct pci_ops *ops; + int domain; + int bus_num; +}; + +#define PCI_MCFG_DOMAIN_ANY -1 +#define PCI_MCFG_BUS_ANY -1 + +/* Designate a routine to fix up buggy MCFG */ +#define DECLARE_ACPI_MCFG_FIXUP(system, match, ops, dom, bus) \ + static const struct pci_mcfg_fixup __mcfg_fixup_##system##dom##bus\ + __used __attribute__((__section__(".acpi_fixup_mcfg"), \ + aligned((sizeof(void *))))) = \ + { system, match, ops, dom, bus }; + struct pci_mmcfg_region *pci_mmconfig_lookup(int segment, int bus); struct pci_mmcfg_region *pci_mmconfig_alloc(int segment, int start, int end, u64 addr);
On Wed, 2016-01-13 at 14:21 +0100, Tomasz Nowicki wrote:
Some platforms may not be fully compliant with generic set of PCI config accessors. For these cases we implement the way to overwrite accessors set prior to PCI buses enumeration. Algorithm traverses available quirk list, matches against <platform ID (DMI), domain, bus number> tuple and returns corresponding accessors. All quirks can be defined using: DECLARE_ACPI_MCFG_FIXUP() and keep self contained. Example,
static const struct dmi_system_id foo_dmi[] = { { .ident = "<Platform ident string>", .callback = <handler>, .matches = { DMI_MATCH(DMI_SYS_VENDOR, "<system vendor>"), DMI_MATCH(DMI_PRODUCT_NAME, "<product name>"), DMI_MATCH(DMI_PRODUCT_VERSION, "product version"), }, }, { } };
static struct pci_ops foo_ecam_pci_ops = { .map_bus = pci_mcfg_dev_base, .read = foo_ecam_config_read, .write = foo_ecam_config_write, }; DECLARE_ACPI_MCFG_FIXUP(foo_dmi, NULL, &foo_ecam_pci_ops, <domain_nr>, <bus_nr>);
More custom (non-DMI) matching can be done via an alternative call. Note that there is possibility to assign quirk related private data to root->sysdata which will be available along read/wriate accessor, example:
static int boo_match(struct pci_mcfg_fixup *fixup, struct acpi_pci_root *root) { return [condition] ? 1 : 0; }
int boo_ecam_config_read(struct pci_bus *bus, unsigned int devfn, int where, int size, u32 *val) { struct acpi_pci_root *root = bus->sysdata; struct boo_priv_data *boo_data = root->sysdata;
[..] }
static struct pci_ops boo_ecam_pci_ops = { .map_bus = pci_mcfg_dev_base, .read = boo_ecam_config_read, .write = boo_ecam_config_write, }; DECLARE_ACPI_MCFG_FIXUP(NULL, boo_match, &boo_ecam_pci_ops, <domain_nr>, <bus_nr>);
Signed-off-by: Tomasz Nowicki tn@semihalf.com
drivers/acpi/mcfg.c | 33 +++++++++++++++++++++++++++++++-- include/acpi/acpi_bus.h | 1 + include/asm-generic/vmlinux.lds.h | 7 +++++++ include/linux/ecam.h | 18 ++++++++++++++++++ 4 files changed, 57 insertions(+), 2 deletions(-)
diff --git a/drivers/acpi/mcfg.c b/drivers/acpi/mcfg.c index dfc2d14..ec5fe7b 100644 --- a/drivers/acpi/mcfg.c +++ b/drivers/acpi/mcfg.c @@ -8,6 +8,7 @@ */ #include <linux/acpi.h> +#include <linux/dmi.h> #include <linux/ecam.h> #include <linux/pci.h> #include <linux/pci-acpi.h> @@ -34,6 +35,29 @@ int __weak raw_pci_write(unsigned int domain, unsigned int bus, return PCIBIOS_DEVICE_NOT_FOUND; } +extern struct pci_mcfg_fixup __start_acpi_mcfg_fixups[]; +extern struct pci_mcfg_fixup __end_acpi_mcfg_fixups[];
+static struct pci_ops *pci_mcfg_check_quirks(struct acpi_pci_root *root) +{
- struct pci_mcfg_fixup *f;
- int bus_num = root->secondary.start;
- int domain = root->segment;
- /*
- * First match against PCI topology domain:bus then use DMI or
- * custom match handler.
- */
- for (f = __start_acpi_mcfg_fixups; f < __end_acpi_mcfg_fixups; f++) {
if ((f->domain == domain || f->domain == PCI_MCFG_DOMAIN_ANY) &&
(f->bus_num == bus_num || f->bus_num == PCI_MCFG_BUS_ANY) &&
(f->system ? dmi_check_system(f->system) : 0 ||
f->match ? f->match(f, root) : 0))
return f->ops;
I think this would be better as:
(f->system ? dmi_check_system(f->system) : 1 && f->match ? f->match(f, root) : 1)) return f->ops;
Otherwise, one has to call dmi_check_system() from f->match() if access to root is needed.
- }
- return NULL;
+}
void __iomem * pci_mcfg_dev_base(struct pci_bus *bus, unsigned int devfn, int offset) { @@ -56,10 +80,15 @@ static struct pci_ops default_pci_mcfg_ops = { struct pci_ops *pci_mcfg_get_ops(struct acpi_pci_root *root) {
- struct pci_ops *pci_mcfg_ops_quirk;
/*
- * TODO: Match against platform specific quirks and return
- * corresponding PCI config space accessor set.
- * Match against platform specific quirks and return corresponding
- * PCI config space accessor set.
*/
- pci_mcfg_ops_quirk = pci_mcfg_check_quirks(root);
- if (pci_mcfg_ops_quirk)
return pci_mcfg_ops_quirk;
return &default_pci_mcfg_ops; } diff --git a/include/acpi/acpi_bus.h b/include/acpi/acpi_bus.h index ad0a5ff..ea4d2b5 100644 --- a/include/acpi/acpi_bus.h +++ b/include/acpi/acpi_bus.h @@ -554,6 +554,7 @@ struct acpi_pci_root { struct pci_bus *bus; u16 segment; struct resource secondary; /* downstream bus range */
- void *sysdata;
u32 osc_support_set; /* _OSC state of support bits */ u32 osc_control_set; /* _OSC state of control bits */ diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h index c4bd0e2..c93fc97 100644 --- a/include/asm-generic/vmlinux.lds.h +++ b/include/asm-generic/vmlinux.lds.h @@ -298,6 +298,13 @@ VMLINUX_SYMBOL(__end_pci_fixups_suspend_late) = .; \ } \ \
- /* ACPI MCFG quirks */ \
- .acpi_fixup : AT(ADDR(.acpi_fixup) - LOAD_OFFSET) { \
VMLINUX_SYMBOL(__start_acpi_mcfg_fixups) = .; \
*(.acpi_fixup_mcfg) \
VMLINUX_SYMBOL(__end_acpi_mcfg_fixups) = .; \
- } \
\
/* Built-in firmware blobs */ \ .builtin_fw : AT(ADDR(.builtin_fw) - LOAD_OFFSET) { \ VMLINUX_SYMBOL(__start_builtin_fw) = .; \ diff --git a/include/linux/ecam.h b/include/linux/ecam.h index e0f322e..21215be 100644 --- a/include/linux/ecam.h +++ b/include/linux/ecam.h @@ -20,6 +20,24 @@ struct pci_mmcfg_region { bool hot_added; }; +struct pci_mcfg_fixup {
- const struct dmi_system_id *system;
- int (*match)(struct pci_mcfg_fixup *, struct acpi_pci_root *);
- struct pci_ops *ops;
- int domain;
- int bus_num;
+};
+#define PCI_MCFG_DOMAIN_ANY -1 +#define PCI_MCFG_BUS_ANY -1
+/* Designate a routine to fix up buggy MCFG */ +#define DECLARE_ACPI_MCFG_FIXUP(system, match, ops, dom, bus) \
- static const struct pci_mcfg_fixup __mcfg_fixup_##system##dom##bus\
- __used __attribute__((__section__(".acpi_fixup_mcfg"), \
aligned((sizeof(void *))))) = \
- { system, match, ops, dom, bus };
struct pci_mmcfg_region *pci_mmconfig_lookup(int segment, int bus); struct pci_mmcfg_region *pci_mmconfig_alloc(int segment, int start, int end, u64 addr);
On 14.01.2016 16:36, Mark Salter wrote:
+extern struct pci_mcfg_fixup __start_acpi_mcfg_fixups[];
+extern struct pci_mcfg_fixup __end_acpi_mcfg_fixups[];
+static struct pci_ops *pci_mcfg_check_quirks(struct acpi_pci_root *root) +{
- struct pci_mcfg_fixup *f;
- int bus_num = root->secondary.start;
- int domain = root->segment;
- /*
* First match against PCI topology <domain:bus> then use DMI or
* custom match handler.
*/
- for (f = __start_acpi_mcfg_fixups; f < __end_acpi_mcfg_fixups; f++) {
if ((f->domain == domain || f->domain == PCI_MCFG_DOMAIN_ANY) &&
(f->bus_num == bus_num || f->bus_num == PCI_MCFG_BUS_ANY) &&
(f->system ? dmi_check_system(f->system) : 0 ||
f->match ? f->match(f, root) : 0))
return f->ops;
I think this would be better as:
(f->system ? dmi_check_system(f->system) : 1 && f->match ? f->match(f, root) : 1)) return f->ops;
Otherwise, one has to call dmi_check_system() from f->match() if access to root is needed.
Makes a lot of sense to me, I will modify as you suggested.
Tomasz
Hi Tomasz, Mark
在 2016/1/18 20:41, Tomasz Nowicki 写道:
On 14.01.2016 16:36, Mark Salter wrote:
+extern struct pci_mcfg_fixup __start_acpi_mcfg_fixups[];
+extern struct pci_mcfg_fixup __end_acpi_mcfg_fixups[];
+static struct pci_ops *pci_mcfg_check_quirks(struct acpi_pci_root *root) +{
- struct pci_mcfg_fixup *f;
- int bus_num = root->secondary.start;
- int domain = root->segment;
- /*
* First match against PCI topology <domain:bus> then use DMI or
* custom match handler.
*/
- for (f = __start_acpi_mcfg_fixups; f < __end_acpi_mcfg_fixups; f++) {
if ((f->domain == domain || f->domain == PCI_MCFG_DOMAIN_ANY) &&
(f->bus_num == bus_num || f->bus_num == PCI_MCFG_BUS_ANY) &&
(f->system ? dmi_check_system(f->system) : 0 ||
f->match ? f->match(f, root) : 0))
return f->ops;
I think this would be better as:
(f->system ? dmi_check_system(f->system) : 1 && f->match ? f->match(f, root) : 1)) return f->ops;
Otherwise, one has to call dmi_check_system() from f->match() if access to root is needed.
Non-DMI, we need not to call dmi_check_system() from f->match(), we can use _HID to decide to hook the pci_ops or not.
Device (PCI1) { Name (_HID, "HISI0080") // PCI Express Root Bridge Name (_CID, "PNP0A03") // Compatible PCI Root Bridge ... }
static const struct acpi_device_id hisi_pcie_ids[] = { {"HISI0080", 0}, {"", 0}, };
static int hisi_pcie_match(struct pci_mcfg_fixup *fixup, struct acpi_pci_root *root) { int ret; struct acpi_device *device;
device = root->device; ret = acpi_match_device_ids(device, hisi_pcie_ids); if (ret) return 0; ...... return 1; }
static struct pci_ops hisi_ecam_pci_ops = { .map_bus = pci_mcfg_dev_base, .read = hisi_pci_read, .write = hisi_pci_write, };
DECLARE_ACPI_MCFG_FIXUP(NULL, hisi_pcie_match, &hisi_ecam_pci_ops, PCI_MCFG_DOMAIN_ANY, PCI_MCFG_BUS_ANY);
Thanks Dongdong
Makes a lot of sense to me, I will modify as you suggested.
Tomasz
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
.
On 19.01.2016 02:49, liudongdong (C) wrote:
Hi Tomasz, Mark
在 2016/1/18 20:41, Tomasz Nowicki 写道:
On 14.01.2016 16:36, Mark Salter wrote:
+extern struct pci_mcfg_fixup __start_acpi_mcfg_fixups[];
+extern struct pci_mcfg_fixup __end_acpi_mcfg_fixups[];
+static struct pci_ops *pci_mcfg_check_quirks(struct acpi_pci_root
*root)
+{
- struct pci_mcfg_fixup *f;
- int bus_num = root->secondary.start;
- int domain = root->segment;
- /*
* First match against PCI topology <domain:bus> then use DMI or
* custom match handler.
*/
- for (f = __start_acpi_mcfg_fixups; f < __end_acpi_mcfg_fixups;
f++) {
if ((f->domain == domain || f->domain ==
PCI_MCFG_DOMAIN_ANY) &&
(f->bus_num == bus_num || f->bus_num ==
PCI_MCFG_BUS_ANY) &&
(f->system ? dmi_check_system(f->system) : 0 ||
f->match ? f->match(f, root) : 0))
return f->ops;
I think this would be better as:
(f->system ? dmi_check_system(f->system) : 1 && f->match ? f->match(f, root) : 1)) return f->ops;
Otherwise, one has to call dmi_check_system() from f->match() if access to root is needed.
Non-DMI, we need not to call dmi_check_system() from f->match(), we can use _HID to decide to hook the pci_ops or not.
Sorry, but I dont understand your point. Can you elaborate?
With Mark modification, you can use the following cases to identify platform: 1. DMI only 2. f->match() only (_HID can be used there) 3. DMI and f->match()
DMI used to be very convenient way to recognise platform, sometimes it is not enough, hence f->match() alternative.
Tomasz
在 2016/1/19 15:55, Tomasz Nowicki 写道:
On 19.01.2016 02:49, liudongdong (C) wrote:
Hi Tomasz, Mark
在 2016/1/18 20:41, Tomasz Nowicki 写道:
On 14.01.2016 16:36, Mark Salter wrote:
+extern struct pci_mcfg_fixup __start_acpi_mcfg_fixups[];
+extern struct pci_mcfg_fixup __end_acpi_mcfg_fixups[];
+static struct pci_ops *pci_mcfg_check_quirks(struct acpi_pci_root
*root)
+{
- struct pci_mcfg_fixup *f;
- int bus_num = root->secondary.start;
- int domain = root->segment;
- /*
* First match against PCI topology <domain:bus> then use DMI or
* custom match handler.
*/
- for (f = __start_acpi_mcfg_fixups; f < __end_acpi_mcfg_fixups;
f++) {
if ((f->domain == domain || f->domain ==
PCI_MCFG_DOMAIN_ANY) &&
(f->bus_num == bus_num || f->bus_num ==
PCI_MCFG_BUS_ANY) &&
(f->system ? dmi_check_system(f->system) : 0 ||
f->match ? f->match(f, root) : 0))
return f->ops;
I think this would be better as:
(f->system ? dmi_check_system(f->system) : 1 && f->match ? f->match(f, root) : 1)) return f->ops;
Otherwise, one has to call dmi_check_system() from f->match() if access to root is needed.
Non-DMI, we need not to call dmi_check_system() from f->match(), we can use _HID to decide to hook the pci_ops or not.
Sorry, but I dont understand your point. Can you elaborate?
With Mark modification, you can use the following cases to identify platform:
- DMI only
- f->match() only (_HID can be used there)
- DMI and f->match()
DMI used to be very convenient way to recognise platform, sometimes it is not enough, hence f->match() alternative.
Yes, you are right, I was wrong. In my case, I can use the second point. 2. f->match() only (_HID can be used there)
Thanks Dongdong
Tomasz
.
Lets get rid of empty PCI init stub, related ACPI header and go with full-blown PCI host bridge driver.
Signed-off-by: Tomasz Nowicki tn@semihalf.com CC: Arnd Bergmann arnd@arndb.de CC: Catalin Marinas catalin.marinas@arm.com CC: Liviu Dudau Liviu.Dudau@arm.com CC: Lorenzo Pieralisi Lorenzo.Pieralisi@arm.com CC: Will Deacon will.deacon@arm.com --- arch/arm64/Kconfig | 1 + arch/arm64/kernel/pci.c | 10 ---------- 2 files changed, 1 insertion(+), 10 deletions(-)
diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index d65d315..573bebc 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -2,6 +2,7 @@ config ARM64 def_bool y select ACPI_CCA_REQUIRED if ACPI select ACPI_GENERIC_GSI if ACPI + select ACPI_PCI_HOST_GENERIC if ACPI select ACPI_REDUCED_HARDWARE_ONLY if ACPI select ARCH_HAS_ATOMIC64_DEC_IF_POSITIVE select ARCH_HAS_ELF_RANDOMIZE diff --git a/arch/arm64/kernel/pci.c b/arch/arm64/kernel/pci.c index 023b983..a0929a6 100644 --- a/arch/arm64/kernel/pci.c +++ b/arch/arm64/kernel/pci.c @@ -10,7 +10,6 @@ * */
-#include <linux/acpi.h> #include <linux/init.h> #include <linux/io.h> #include <linux/kernel.h> @@ -60,12 +59,3 @@ int pcibios_add_device(struct pci_dev *dev)
return 0; } - -#ifdef CONFIG_ACPI -/* Root bridge scanning */ -struct pci_bus *pci_acpi_scan_root(struct acpi_pci_root *root) -{ - /* TODO: Should be revisited when implementing PCI on ACPI */ - return NULL; -} -#endif
On 1/13/2016 8:20 AM, Tomasz Nowicki wrote:
From the functionality point of view this series might be split into the following logic parts:
- Make MMCONFIG code arch-agnostic which allows all architectures to collect PCI config regions and used when necessary.
- Move non-arch specific bits to the core code.
- Use MMCONFIG code and implement generic ACPI based PCI host controller driver.
- Enable above driver on ARM64
Patches has been built on top of 4.4 and can be found here: git@github.com:semihalf-nowicki-tomasz/linux.git (pci-acpi-v3)
NOTE, this patch set depends on Matthew's patches: http://www.spinics.net/lists/linux-pci/msg45950.html https://github.com/Vality/linux/tree/pci-fixes
This has been tested on Cavium ThunderX server and QEMU. Any help in reviewing and testing is very appreciated.
v2 -> v3
- fix legacy IRQ assigning and IO ports registration
- remove reference to arch specific companion device for ia64
- move ACPI PCI host controller driver to pci_root.c
- drop generic domain assignment for x86 and ia64 as I am not able to run all necessary test variants
- drop patch which cleaned legacy IRQ assignment since it belongs to Mathew's series: https://patchwork.ozlabs.org/patch/557504/
- extend MCFG quirk code
- rebased to 4.4
v1 -> v2
- moved non-arch specific piece of code to dirver/acpi/ directory
- fixed IO resource handling
- introduced PCI config accessors quirks matching
- moved ACPI_COMPANION_SET to generic code
Liu Jiang (1): ACPI, PCI: Refine the way to handle translation_offset for ACPI resources
Tomasz Nowicki (20): x86, pci: Reorder logic of pci_mmconfig_insert() function x86, pci, acpi: Move arch-agnostic MMCONFIG (aka ECAM) and ACPI code out of arch/x86/ directory pci, acpi, mcfg: Provide generic implementation of MCFG code initialization. x86, pci: mmconfig_{32,64}.c code refactoring - remove code duplication. x86, pci, ecam: mmconfig_64.c becomes default implementation for ECAM driver. XEN / PCI: Remove the dependence on arch x86 when PCI_MMCONFIG=y pci, acpi, mcfg: Provide default RAW ACPI PCI config space accessors. arm64, acpi: Use empty PCI config space accessors from mcfg.c file. pci, acpi, ecam: Add flag to indicate whether ECAM region was hot added or not. x86, pci: Cleanup platform specific MCFG data using previously added ECAM hot_added flag. pci, acpi: Move ACPI host bridge device companion assignment to core code. x86, ia64, pci: Remove ACPI companion device from platform specific data. pci, acpi: Provide generic way to assign bus domain number. x86, ia64: Include acpi_pci_{add|remove}_bus to the default pcibios_{add|remove}_bus implementation. acpi, mcfg: Implement two calls that might be used to inject/remove MCFG region. x86, acpi, pci: Use equivalent function introduced in previous patch. acpi, mcfg: Add default PCI config accessors implementation and initial support for related quirks. pci, acpi: Support for ACPI based generic PCI host controller init pci, acpi: Match PCI config space accessors against platfrom specific quirks. arm64, pci, acpi: Start using ACPI based PCI host bridge driver for ARM64.
arch/arm64/Kconfig | 7 ++ arch/arm64/kernel/pci.c | 25 ---- arch/ia64/hp/common/sba_iommu.c | 2 +- arch/ia64/include/asm/pci.h | 1 - arch/ia64/pci/pci.c | 52 +++----- arch/ia64/sn/kernel/io_acpi_init.c | 4 +- arch/x86/Kconfig | 4 + arch/x86/include/asm/pci.h | 3 - arch/x86/include/asm/pci_x86.h | 28 +---- arch/x86/pci/acpi.c | 41 ++---- arch/x86/pci/common.c | 10 -- arch/x86/pci/irq.c | 10 -- arch/x86/pci/mmconfig-shared.c | 250 ++++++------------------------------- arch/x86/pci/mmconfig_32.c | 11 +- arch/x86/pci/mmconfig_64.c | 67 +--------- arch/x86/pci/numachip.c | 1 + drivers/acpi/Kconfig | 5 + drivers/acpi/Makefile | 1 + drivers/acpi/mcfg.c | 201 +++++++++++++++++++++++++++++ drivers/acpi/pci_root.c | 136 +++++++++++++++++++- drivers/acpi/resource.c | 12 +- drivers/pci/Kconfig | 10 ++ drivers/pci/Makefile | 5 + drivers/pci/ecam.c | 234 ++++++++++++++++++++++++++++++++++ drivers/pci/pci.c | 29 ++++- drivers/pci/probe.c | 5 + drivers/xen/pci.c | 7 +- include/acpi/acpi_bus.h | 1 + include/asm-generic/vmlinux.lds.h | 7 ++ include/linux/acpi.h | 2 + include/linux/ecam.h | 62 +++++++++ include/linux/pci-acpi.h | 17 +++ 32 files changed, 802 insertions(+), 448 deletions(-) create mode 100644 drivers/acpi/mcfg.c create mode 100644 drivers/pci/ecam.c create mode 100644 include/linux/ecam.h
Tested-by: Sinan Kaya okaya@codeaurora.org
on an ARM64 server with some nits.
I started seeing these on an ACPI system. I didn't have these before when I tested Tomasz's branch yesterday. It must have been introduced after a rebase or a recent change.
_#_dmesg_|_grep_of_irq [ 3.130774] pci 0002:00:00.0: of_irq_parse_pci() failed with rc=-22 [ 4.182526] pci 0002:00:00.0: of_irq_parse_pci() failed with rc=-22 [ 5.187764] pci 0002:00:00.0: of_irq_parse_pci() failed with rc=-22 [ 5.868518] pci 0003:00:00.0: of_irq_parse_pci() failed with rc=-22 [ 5.910629] pci 0003:00:00.0: of_irq_parse_pci() failed with rc=-22 [ 6.732817] pci 0006:00:00.0: of_irq_parse_pci() failed with rc=-22 [ 7.777698] pci 0006:00:00.0: of_irq_parse_pci() failed with rc=-22 [ 8.782882] pci 0006:00:00.0: of_irq_parse_pci() failed with rc=-22 / #
I'm also seeing an IO resource free problem with ACPI hotplug. The mappings created by pci_remap_iospace are not being freed in ACPI root port free function. Second call to pci_remap_iospace fails after a hotplug insertion as the kernel mapping already exists. I had a chat with Tomasz about this. We'll discuss the issue later.
On 13.01.2016 16:24, Sinan Kaya wrote:
on an ARM64 server with some nits.
I started seeing these on an ACPI system. I didn't have these before when I tested Tomasz's branch yesterday. It must have been introduced after a rebase or a recent change.
_#_dmesg_|_grep_of_irq [ 3.130774] pci 0002:00:00.0: of_irq_parse_pci() failed with rc=-22 [ 4.182526] pci 0002:00:00.0: of_irq_parse_pci() failed with rc=-22 [ 5.187764] pci 0002:00:00.0: of_irq_parse_pci() failed with rc=-22 [ 5.868518] pci 0003:00:00.0: of_irq_parse_pci() failed with rc=-22 [ 5.910629] pci 0003:00:00.0: of_irq_parse_pci() failed with rc=-22 [ 6.732817] pci 0006:00:00.0: of_irq_parse_pci() failed with rc=-22 [ 7.777698] pci 0006:00:00.0: of_irq_parse_pci() failed with rc=-22 [ 8.782882] pci 0006:00:00.0: of_irq_parse_pci() failed with rc=-22 / #
You got this because I had dropped patch from the previous series: https://patchwork.ozlabs.org/patch/557504/ and this is expected.
Tomasz
On Wed, Jan 13, 2016 at 02:20:46PM +0100, Tomasz Nowicki wrote:
From the functionality point of view this series might be split into the following logic parts:
- Make MMCONFIG code arch-agnostic which allows all architectures to collect PCI config regions and used when necessary.
- Move non-arch specific bits to the core code.
- Use MMCONFIG code and implement generic ACPI based PCI host controller driver.
- Enable above driver on ARM64
Patches has been built on top of 4.4 and can be found here: git@github.com:semihalf-nowicki-tomasz/linux.git (pci-acpi-v3)
NOTE, this patch set depends on Matthew's patches: http://www.spinics.net/lists/linux-pci/msg45950.html https://github.com/Vality/linux/tree/pci-fixes
This has been tested on Cavium ThunderX server and QEMU. Any help in reviewing and testing is very appreciated.
I have tested this on my AMD Overdrive so
Tested-by: Graeme Gregory graeme.gregory@linaro.org
But to actually get my r8169 network card working I also need the following patch.
diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 2fbf840..40e24e2 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -1717,7 +1717,7 @@ void pci_device_add(struct pci_dev *dev, struct pci_bus *bus) set_dev_node(&dev->dev, pcibus_to_node(bus)); dev->dev.dma_mask = &dev->dma_mask; dev->dev.dma_parms = &dev->dma_parms; - dev->dev.coherent_dma_mask = 0xffffffffull; + dev->dev.coherent_dma_mask = DMA_BIT_MASK(64); pci_dma_configure(dev);
pci_set_dma_max_seg_size(dev, 65536);
I suspect we need to set coherent_dma_mask somewhere in the platform but I do not know where this should happen. Hopefully an ARM64 expert can help.
Graeme
On Thu, Jan 14, 2016 at 01:44:40PM +0000, Graeme Gregory wrote:
On Wed, Jan 13, 2016 at 02:20:46PM +0100, Tomasz Nowicki wrote:
From the functionality point of view this series might be split into the following logic parts:
- Make MMCONFIG code arch-agnostic which allows all architectures to collect PCI config regions and used when necessary.
- Move non-arch specific bits to the core code.
- Use MMCONFIG code and implement generic ACPI based PCI host controller driver.
- Enable above driver on ARM64
Patches has been built on top of 4.4 and can be found here: git@github.com:semihalf-nowicki-tomasz/linux.git (pci-acpi-v3)
NOTE, this patch set depends on Matthew's patches: http://www.spinics.net/lists/linux-pci/msg45950.html https://github.com/Vality/linux/tree/pci-fixes
This has been tested on Cavium ThunderX server and QEMU. Any help in reviewing and testing is very appreciated.
I have tested this on my AMD Overdrive so
Tested-by: Graeme Gregory graeme.gregory@linaro.org
But to actually get my r8169 network card working I also need the following patch.
diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 2fbf840..40e24e2 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -1717,7 +1717,7 @@ void pci_device_add(struct pci_dev *dev, struct pci_bus *bus) set_dev_node(&dev->dev, pcibus_to_node(bus)); dev->dev.dma_mask = &dev->dma_mask; dev->dev.dma_parms = &dev->dma_parms;
dev->dev.coherent_dma_mask = 0xffffffffull;
dev->dev.coherent_dma_mask = DMA_BIT_MASK(64); pci_dma_configure(dev); pci_set_dma_max_seg_size(dev, 65536);
With OF, we get the coherent_dma_mask set by of_dma_configure(). But I have no idea how you do this with ACPI.
On Thu, 2016-01-14 at 14:00 +0000, Catalin Marinas wrote:
On Thu, Jan 14, 2016 at 01:44:40PM +0000, Graeme Gregory wrote:
On Wed, Jan 13, 2016 at 02:20:46PM +0100, Tomasz Nowicki wrote:
From the functionality point of view this series might be split into the following logic parts:
- Make MMCONFIG code arch-agnostic which allows all architectures to collect
PCI config regions and used when necessary. 2. Move non-arch specific bits to the core code. 3. Use MMCONFIG code and implement generic ACPI based PCI host controller driver. 4. Enable above driver on ARM64
Patches has been built on top of 4.4 and can be found here: git@github.com:semihalf-nowicki-tomasz/linux.git (pci-acpi-v3)
NOTE, this patch set depends on Matthew's patches: http://www.spinics.net/lists/linux-pci/msg45950.html https://github.com/Vality/linux/tree/pci-fixes
This has been tested on Cavium ThunderX server and QEMU. Any help in reviewing and testing is very appreciated.
I have tested this on my AMD Overdrive so
Tested-by: Graeme Gregory graeme.gregory@linaro.org
But to actually get my r8169 network card working I also need the following patch.
diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 2fbf840..40e24e2 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -1717,7 +1717,7 @@ void pci_device_add(struct pci_dev *dev, struct pci_bus *bus) set_dev_node(&dev->dev, pcibus_to_node(bus)); dev->dev.dma_mask = &dev->dma_mask; dev->dev.dma_parms = &dev->dma_parms; - dev->dev.coherent_dma_mask = 0xffffffffull; + dev->dev.coherent_dma_mask = DMA_BIT_MASK(64); pci_dma_configure(dev);
pci_set_dma_max_seg_size(dev, 65536);
With OF, we get the coherent_dma_mask set by of_dma_configure(). But I have no idea how you do this with ACPI.
That doesn't get called for a PCI device (not listed in DT) does it?
On Thu, Jan 14, 2016 at 09:09:59AM -0500, Mark Salter wrote:
On Thu, 2016-01-14 at 14:00 +0000, Catalin Marinas wrote:
On Thu, Jan 14, 2016 at 01:44:40PM +0000, Graeme Gregory wrote:
On Wed, Jan 13, 2016 at 02:20:46PM +0100, Tomasz Nowicki wrote:
From the functionality point of view this series might be split into the following logic parts:
- Make MMCONFIG code arch-agnostic which allows all architectures to collect
PCI config regions and used when necessary. 2. Move non-arch specific bits to the core code. 3. Use MMCONFIG code and implement generic ACPI based PCI host controller driver. 4. Enable above driver on ARM64
Patches has been built on top of 4.4 and can be found here: git@github.com:semihalf-nowicki-tomasz/linux.git (pci-acpi-v3)
NOTE, this patch set depends on Matthew's patches: http://www.spinics.net/lists/linux-pci/msg45950.html https://github.com/Vality/linux/tree/pci-fixes
This has been tested on Cavium ThunderX server and QEMU. Any help in reviewing and testing is very appreciated.
I have tested this on my AMD Overdrive so
Tested-by: Graeme Gregory graeme.gregory@linaro.org
But to actually get my r8169 network card working I also need the following patch.
diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 2fbf840..40e24e2 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -1717,7 +1717,7 @@ void pci_device_add(struct pci_dev *dev, struct pci_bus *bus) set_dev_node(&dev->dev, pcibus_to_node(bus)); dev->dev.dma_mask = &dev->dma_mask; dev->dev.dma_parms = &dev->dma_parms; - dev->dev.coherent_dma_mask = 0xffffffffull; + dev->dev.coherent_dma_mask = DMA_BIT_MASK(64); pci_dma_configure(dev);
pci_set_dma_max_seg_size(dev, 65536);
With OF, we get the coherent_dma_mask set by of_dma_configure(). But I have no idea how you do this with ACPI.
That doesn't get called for a PCI device (not listed in DT) does it?
pci_device_add pci_dma_configure of_dma_configure(&dev->dev, bridge->parent->of_node);
So it gets configured based on the bridge information the device is attached to.
On Thu, 2016-01-14 at 14:50 +0000, Catalin Marinas wrote:
On Thu, Jan 14, 2016 at 09:09:59AM -0500, Mark Salter wrote:
On Thu, 2016-01-14 at 14:00 +0000, Catalin Marinas wrote:
On Thu, Jan 14, 2016 at 01:44:40PM +0000, Graeme Gregory wrote:
On Wed, Jan 13, 2016 at 02:20:46PM +0100, Tomasz Nowicki wrote:
From the functionality point of view this series might be split into the following logic parts:
- Make MMCONFIG code arch-agnostic which allows all architectures to collect
PCI config regions and used when necessary. 2. Move non-arch specific bits to the core code. 3. Use MMCONFIG code and implement generic ACPI based PCI host controller driver. 4. Enable above driver on ARM64
Patches has been built on top of 4.4 and can be found here: git@github.com:semihalf-nowicki-tomasz/linux.git (pci-acpi-v3)
NOTE, this patch set depends on Matthew's patches: http://www.spinics.net/lists/linux-pci/msg45950.html https://github.com/Vality/linux/tree/pci-fixes
This has been tested on Cavium ThunderX server and QEMU. Any help in reviewing and testing is very appreciated.
I have tested this on my AMD Overdrive so
Tested-by: Graeme Gregory graeme.gregory@linaro.org
But to actually get my r8169 network card working I also need the following patch.
diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 2fbf840..40e24e2 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -1717,7 +1717,7 @@ void pci_device_add(struct pci_dev *dev, struct pci_bus *bus) set_dev_node(&dev->dev, pcibus_to_node(bus)); dev->dev.dma_mask = &dev->dma_mask; dev->dev.dma_parms = &dev->dma_parms; - dev->dev.coherent_dma_mask = 0xffffffffull; + dev->dev.coherent_dma_mask = DMA_BIT_MASK(64); pci_dma_configure(dev);
pci_set_dma_max_seg_size(dev, 65536);
With OF, we get the coherent_dma_mask set by of_dma_configure(). But I have no idea how you do this with ACPI.
That doesn't get called for a PCI device (not listed in DT) does it?
pci_device_add pci_dma_configure of_dma_configure(&dev->dev, bridge->parent->of_node);
So it gets configured based on the bridge information the device is attached to.
Oh right. And there is an ACPI path in pci_dma_configure but it only sets the coherency. But I still don't know why Graeme is having issues. I just tried an r8169 card on Mustang and Seattle with this patch series in a 4.4 kernel and it worked fine with r8169.use_dac=1.
On Thu, 2016-01-14 at 13:44 +0000, Graeme Gregory wrote:
On Wed, Jan 13, 2016 at 02:20:46PM +0100, Tomasz Nowicki wrote:
From the functionality point of view this series might be split into the following logic parts:
- Make MMCONFIG code arch-agnostic which allows all architectures to collect
PCI config regions and used when necessary. 2. Move non-arch specific bits to the core code. 3. Use MMCONFIG code and implement generic ACPI based PCI host controller driver. 4. Enable above driver on ARM64
Patches has been built on top of 4.4 and can be found here: git@github.com:semihalf-nowicki-tomasz/linux.git (pci-acpi-v3)
NOTE, this patch set depends on Matthew's patches: http://www.spinics.net/lists/linux-pci/msg45950.html https://github.com/Vality/linux/tree/pci-fixes
This has been tested on Cavium ThunderX server and QEMU. Any help in reviewing and testing is very appreciated.
I have tested this on my AMD Overdrive so
Tested-by: Graeme Gregory graeme.gregory@linaro.org
But to actually get my r8169 network card working I also need the following patch.
diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 2fbf840..40e24e2 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -1717,7 +1717,7 @@ void pci_device_add(struct pci_dev *dev, struct pci_bus *bus) set_dev_node(&dev->dev, pcibus_to_node(bus)); dev->dev.dma_mask = &dev->dma_mask; dev->dev.dma_parms = &dev->dma_parms; - dev->dev.coherent_dma_mask = 0xffffffffull; + dev->dev.coherent_dma_mask = DMA_BIT_MASK(64); pci_dma_configure(dev);
pci_set_dma_max_seg_size(dev, 65536);
I suspect we need to set coherent_dma_mask somewhere in the platform but I do not know where this should happen. Hopefully an ARM64 expert can help.
I've run into a number of cards which don't work on some arm64 machines where there are no DMA address below 4G. For the realtek card you can use r8169.use_dac=1 on the cmdline to get it to work.
Graeme
On Thu, Jan 14, 2016 at 09:01:40AM -0500, Mark Salter wrote:
On Thu, 2016-01-14 at 13:44 +0000, Graeme Gregory wrote:
On Wed, Jan 13, 2016 at 02:20:46PM +0100, Tomasz Nowicki wrote:
From the functionality point of view this series might be split into the following logic parts:
- Make MMCONFIG code arch-agnostic which allows all architectures to collect
PCI config regions and used when necessary. 2. Move non-arch specific bits to the core code. 3. Use MMCONFIG code and implement generic ACPI based PCI host controller driver. 4. Enable above driver on ARM64
Patches has been built on top of 4.4 and can be found here: git@github.com:semihalf-nowicki-tomasz/linux.git (pci-acpi-v3)
NOTE, this patch set depends on Matthew's patches: http://www.spinics.net/lists/linux-pci/msg45950.html https://github.com/Vality/linux/tree/pci-fixes
This has been tested on Cavium ThunderX server and QEMU. Any help in reviewing and testing is very appreciated.
I have tested this on my AMD Overdrive so
Tested-by: Graeme Gregory graeme.gregory@linaro.org
But to actually get my r8169 network card working I also need the following patch.
diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 2fbf840..40e24e2 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -1717,7 +1717,7 @@ void pci_device_add(struct pci_dev *dev, struct pci_bus *bus) set_dev_node(&dev->dev, pcibus_to_node(bus)); dev->dev.dma_mask = &dev->dma_mask; dev->dev.dma_parms = &dev->dma_parms; - dev->dev.coherent_dma_mask = 0xffffffffull; + dev->dev.coherent_dma_mask = DMA_BIT_MASK(64); pci_dma_configure(dev);
pci_set_dma_max_seg_size(dev, 65536);
I suspect we need to set coherent_dma_mask somewhere in the platform but I do not know where this should happen. Hopefully an ARM64 expert can help.
I've run into a number of cards which don't work on some arm64 machines where there are no DMA address below 4G. For the realtek card you can use r8169.use_dac=1 on the cmdline to get it to work.
I need r8169.use_dac=1 and this patch to get card to work.
Graeme
On Thu, 2016-01-14 at 14:15 +0000, Graeme Gregory wrote:
On Thu, Jan 14, 2016 at 09:01:40AM -0500, Mark Salter wrote:
On Thu, 2016-01-14 at 13:44 +0000, Graeme Gregory wrote:
On Wed, Jan 13, 2016 at 02:20:46PM +0100, Tomasz Nowicki wrote:
From the functionality point of view this series might be split into the following logic parts:
- Make MMCONFIG code arch-agnostic which allows all architectures to collect
PCI config regions and used when necessary. 2. Move non-arch specific bits to the core code. 3. Use MMCONFIG code and implement generic ACPI based PCI host controller driver. 4. Enable above driver on ARM64
Patches has been built on top of 4.4 and can be found here: git@github.com:semihalf-nowicki-tomasz/linux.git (pci-acpi-v3)
NOTE, this patch set depends on Matthew's patches: http://www.spinics.net/lists/linux-pci/msg45950.html https://github.com/Vality/linux/tree/pci-fixes
This has been tested on Cavium ThunderX server and QEMU. Any help in reviewing and testing is very appreciated.
I have tested this on my AMD Overdrive so
Tested-by: Graeme Gregory graeme.gregory@linaro.org
But to actually get my r8169 network card working I also need the following patch.
diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 2fbf840..40e24e2 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -1717,7 +1717,7 @@ void pci_device_add(struct pci_dev *dev, struct pci_bus *bus) set_dev_node(&dev->dev, pcibus_to_node(bus)); dev->dev.dma_mask = &dev->dma_mask; dev->dev.dma_parms = &dev->dma_parms; - dev->dev.coherent_dma_mask = 0xffffffffull; + dev->dev.coherent_dma_mask = DMA_BIT_MASK(64); pci_dma_configure(dev);
pci_set_dma_max_seg_size(dev, 65536);
I suspect we need to set coherent_dma_mask somewhere in the platform but I do not know where this should happen. Hopefully an ARM64 expert can help.
I've run into a number of cards which don't work on some arm64 machines where there are no DMA address below 4G. For the realtek card you can use r8169.use_dac=1 on the cmdline to get it to work.
I need r8169.use_dac=1 and this patch to get card to work.
Hmm. I didn't need a patch in the past but I don't think I've used that card with this patch series. I will try that when I get a chance...
Graeme
On Thu, Jan 14, 2016 at 09:24:41AM -0500, Mark Salter wrote:
On Thu, 2016-01-14 at 14:15 +0000, Graeme Gregory wrote:
On Thu, Jan 14, 2016 at 09:01:40AM -0500, Mark Salter wrote:
On Thu, 2016-01-14 at 13:44 +0000, Graeme Gregory wrote:
On Wed, Jan 13, 2016 at 02:20:46PM +0100, Tomasz Nowicki wrote:
From the functionality point of view this series might be split into the following logic parts:
- Make MMCONFIG code arch-agnostic which allows all architectures to collect
PCI config regions and used when necessary. 2. Move non-arch specific bits to the core code. 3. Use MMCONFIG code and implement generic ACPI based PCI host controller driver. 4. Enable above driver on ARM64
Patches has been built on top of 4.4 and can be found here: git@github.com:semihalf-nowicki-tomasz/linux.git (pci-acpi-v3)
NOTE, this patch set depends on Matthew's patches: http://www.spinics.net/lists/linux-pci/msg45950.html https://github.com/Vality/linux/tree/pci-fixes
This has been tested on Cavium ThunderX server and QEMU. Any help in reviewing and testing is very appreciated.
I have tested this on my AMD Overdrive so
Tested-by: Graeme Gregory graeme.gregory@linaro.org
But to actually get my r8169 network card working I also need the following patch.
diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 2fbf840..40e24e2 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -1717,7 +1717,7 @@ void pci_device_add(struct pci_dev *dev, struct pci_bus *bus) set_dev_node(&dev->dev, pcibus_to_node(bus)); dev->dev.dma_mask = &dev->dma_mask; dev->dev.dma_parms = &dev->dma_parms; - dev->dev.coherent_dma_mask = 0xffffffffull; + dev->dev.coherent_dma_mask = DMA_BIT_MASK(64); pci_dma_configure(dev);
pci_set_dma_max_seg_size(dev, 65536);
I suspect we need to set coherent_dma_mask somewhere in the platform but I do not know where this should happen. Hopefully an ARM64 expert can help.
I've run into a number of cards which don't work on some arm64 machines where there are no DMA address below 4G. For the realtek card you can use r8169.use_dac=1 on the cmdline to get it to work.
I need r8169.use_dac=1 and this patch to get card to work.
Hmm. I didn't need a patch in the past but I don't think I've used that card with this patch series. I will try that when I get a chance...
From my previous debugging on this issue I was getting an address bigger
than 32bits. As r8169 is a whole range of chips are guess they are not all equal?
I can stick debug stuff in if someone tells me where to put them.
Graeme
On Fri, Jan 15, 2016 at 12:12:31PM +0000, Graeme Gregory wrote:
On Thu, Jan 14, 2016 at 09:24:41AM -0500, Mark Salter wrote:
On Thu, 2016-01-14 at 14:15 +0000, Graeme Gregory wrote:
On Thu, Jan 14, 2016 at 09:01:40AM -0500, Mark Salter wrote:
On Thu, 2016-01-14 at 13:44 +0000, Graeme Gregory wrote:
On Wed, Jan 13, 2016 at 02:20:46PM +0100, Tomasz Nowicki wrote:
From the functionality point of view this series might be split into the following logic parts:
- Make MMCONFIG code arch-agnostic which allows all architectures to collect
PCI config regions and used when necessary. 2. Move non-arch specific bits to the core code. 3. Use MMCONFIG code and implement generic ACPI based PCI host controller driver. 4. Enable above driver on ARM64
Patches has been built on top of 4.4 and can be found here: git@github.com:semihalf-nowicki-tomasz/linux.git (pci-acpi-v3)
NOTE, this patch set depends on Matthew's patches: http://www.spinics.net/lists/linux-pci/msg45950.html https://github.com/Vality/linux/tree/pci-fixes
This has been tested on Cavium ThunderX server and QEMU. Any help in reviewing and testing is very appreciated.
I have tested this on my AMD Overdrive so
Tested-by: Graeme Gregory graeme.gregory@linaro.org
But to actually get my r8169 network card working I also need the following patch.
diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 2fbf840..40e24e2 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -1717,7 +1717,7 @@ void pci_device_add(struct pci_dev *dev, struct pci_bus *bus) set_dev_node(&dev->dev, pcibus_to_node(bus)); dev->dev.dma_mask = &dev->dma_mask; dev->dev.dma_parms = &dev->dma_parms; - dev->dev.coherent_dma_mask = 0xffffffffull; + dev->dev.coherent_dma_mask = DMA_BIT_MASK(64); pci_dma_configure(dev);
pci_set_dma_max_seg_size(dev, 65536);
I suspect we need to set coherent_dma_mask somewhere in the platform but I do not know where this should happen. Hopefully an ARM64 expert can help.
I've run into a number of cards which don't work on some arm64 machines where there are no DMA address below 4G. For the realtek card you can use r8169.use_dac=1 on the cmdline to get it to work.
I need r8169.use_dac=1 and this patch to get card to work.
Hmm. I didn't need a patch in the past but I don't think I've used that card with this patch series. I will try that when I get a chance...
From my previous debugging on this issue I was getting an address bigger than 32bits. As r8169 is a whole range of chips are guess they are not all equal?
I can stick debug stuff in if someone tells me where to put them.
After some private debugging with Mark it turned out that the difference between our configurations was I did not have.
CONFIG_DMA_CMA=y
With this enabled then the card works without coherent mask hack.
Thanks
Graeme
On Mon, Jan 18, 2016 at 02:04:05PM +0000, Graeme Gregory wrote:
On Fri, Jan 15, 2016 at 12:12:31PM +0000, Graeme Gregory wrote:
On Thu, Jan 14, 2016 at 09:24:41AM -0500, Mark Salter wrote:
On Thu, 2016-01-14 at 14:15 +0000, Graeme Gregory wrote:
On Thu, Jan 14, 2016 at 09:01:40AM -0500, Mark Salter wrote:
On Thu, 2016-01-14 at 13:44 +0000, Graeme Gregory wrote:
On Wed, Jan 13, 2016 at 02:20:46PM +0100, Tomasz Nowicki wrote: > From the functionality point of view this series might be split into the > following logic parts: > 1. Make MMCONFIG code arch-agnostic which allows all architectures to collect > PCI config regions and used when necessary. > 2. Move non-arch specific bits to the core code. > 3. Use MMCONFIG code and implement generic ACPI based PCI host controller driver. > 4. Enable above driver on ARM64 > > Patches has been built on top of 4.4 and can be found here: > git@github.com:semihalf-nowicki-tomasz/linux.git (pci-acpi-v3) > > NOTE, this patch set depends on Matthew's patches: > http://www.spinics.net/lists/linux-pci/msg45950.html > https://github.com/Vality/linux/tree/pci-fixes > > This has been tested on Cavium ThunderX server and QEMU. > Any help in reviewing and testing is very appreciated.
I have tested this on my AMD Overdrive so
Tested-by: Graeme Gregory graeme.gregory@linaro.org
But to actually get my r8169 network card working I also need the following patch.
diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 2fbf840..40e24e2 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -1717,7 +1717,7 @@ void pci_device_add(struct pci_dev *dev, struct pci_bus *bus) set_dev_node(&dev->dev, pcibus_to_node(bus)); dev->dev.dma_mask = &dev->dma_mask; dev->dev.dma_parms = &dev->dma_parms; - dev->dev.coherent_dma_mask = 0xffffffffull; + dev->dev.coherent_dma_mask = DMA_BIT_MASK(64); pci_dma_configure(dev);
pci_set_dma_max_seg_size(dev, 65536);
I suspect we need to set coherent_dma_mask somewhere in the platform but I do not know where this should happen. Hopefully an ARM64 expert can help.
I've run into a number of cards which don't work on some arm64 machines where there are no DMA address below 4G. For the realtek card you can use r8169.use_dac=1 on the cmdline to get it to work.
I need r8169.use_dac=1 and this patch to get card to work.
Hmm. I didn't need a patch in the past but I don't think I've used that card with this patch series. I will try that when I get a chance...
From my previous debugging on this issue I was getting an address bigger than 32bits. As r8169 is a whole range of chips are guess they are not all equal?
I can stick debug stuff in if someone tells me where to put them.
After some private debugging with Mark it turned out that the difference between our configurations was I did not have.
CONFIG_DMA_CMA=y
With this enabled then the card works without coherent mask hack.
I didn't follow the whole discussion here, but is this a case where the driver could fail more gracefully than it did? Can we do anything to make this easier for the next person who trips over the same problem?
Bjorn
On Tue, Jan 19, 2016 at 02:25:13PM -0600, Bjorn Helgaas wrote:
On Mon, Jan 18, 2016 at 02:04:05PM +0000, Graeme Gregory wrote:
After some private debugging with Mark it turned out that the difference between our configurations was I did not have.
CONFIG_DMA_CMA=y
With this enabled then the card works without coherent mask hack.
I didn't follow the whole discussion here, but is this a case where the driver could fail more gracefully than it did? Can we do anything to make this easier for the next person who trips over the same problem?
I've not followed the discussion at all, but reading what was in the quoted parts of the mail makes me somewhat suspicious.
The way PCI drivers (or in fact any driver) are supposed to work is:
- the bus code sets up a default mask (32-bit DMA in the case of PCI) - the driver calls dma_set_mask() or similar function with the mask the _driver_ wants to use. Arch DMA code decides whether the mask can be supported, and if it can, it re-sets the DMA mask. If the mask can't be supported, it returns an error.
The driver is doing things correctly:
if ((sizeof(dma_addr_t) > 4) && !pci_set_dma_mask(pdev, DMA_BIT_MASK(64)) && use_dac) { tp->cp_cmd |= PCIDAC; dev->features |= NETIF_F_HIGHDMA; } else { rc = pci_set_dma_mask(pdev, DMA_BIT_MASK(32));
which says: if the size of a DMA address supports 64-bit, and we can set a 64-bit DMA address mask (iow, the platform allows it), and we're permitted to use DAC, use DAC, otherwise try to set a 32-bit DMA mask.
It shouldn't matter one bit what the mask is before that point.
However, what the driver fails to do is to deal with the coherent mask - it appears to be missing a call to pci_set_consistent_dma_mask(). That may be where the issue is.
It may be worth doing what I did with the DMA API, and introducing pci_set_mask_and_coherent() which sets both masks together - which will probably allow some PCI driver code to be simplified.
On Tue, 2016-01-19 at 20:40 +0000, Russell King - ARM Linux wrote:
On Tue, Jan 19, 2016 at 02:25:13PM -0600, Bjorn Helgaas wrote:
On Mon, Jan 18, 2016 at 02:04:05PM +0000, Graeme Gregory wrote:
After some private debugging with Mark it turned out that the difference between our configurations was I did not have.
CONFIG_DMA_CMA=y
With this enabled then the card works without coherent mask hack.
I didn't follow the whole discussion here, but is this a case where the driver could fail more gracefully than it did? Can we do anything to make this easier for the next person who trips over the same problem?
I've not followed the discussion at all, but reading what was in the quoted parts of the mail makes me somewhat suspicious.
The way PCI drivers (or in fact any driver) are supposed to work is:
- the bus code sets up a default mask (32-bit DMA in the case of PCI)
- the driver calls dma_set_mask() or similar function with the mask
the _driver_ wants to use. Arch DMA code decides whether the mask can be supported, and if it can, it re-sets the DMA mask. If the mask can't be supported, it returns an error.
The driver is doing things correctly:
if ((sizeof(dma_addr_t) > 4) && !pci_set_dma_mask(pdev, DMA_BIT_MASK(64)) && use_dac) { tp->cp_cmd |= PCIDAC; dev->features |= NETIF_F_HIGHDMA; } else { rc = pci_set_dma_mask(pdev, DMA_BIT_MASK(32));
which says: if the size of a DMA address supports 64-bit, and we can set a 64-bit DMA address mask (iow, the platform allows it), and we're permitted to use DAC, use DAC, otherwise try to set a 32-bit DMA mask.
It shouldn't matter one bit what the mask is before that point.
However, what the driver fails to do is to deal with the coherent mask - it appears to be missing a call to pci_set_consistent_dma_mask(). That may be where the issue is.
It is.
The driver works with devicetree because of_dma_configure() sets up dma and coherent mask for PCI devices based on root bridge info in the DT. So the driver gets by without setting coherent mask explicitly. In the ACPI case, the coherent mask doesn't get set because there isn't DMA info in the ACPI tables (I think IORT table/IOMMU support will add this in the future). Using CONFIG_DMA_CMA avoids the issue by allowing CMA memory to be used for DMA while avoiding a check against coherent mask.
It may be worth doing what I did with the DMA API, and introducing pci_set_mask_and_coherent() which sets both masks together - which will probably allow some PCI driver code to be simplified.
On Wed, 2016-01-13 at 14:20 +0100, Tomasz Nowicki wrote:
From the functionality point of view this series might be split into the following logic parts:
- Make MMCONFIG code arch-agnostic which allows all architectures to collect
PCI config regions and used when necessary. 2. Move non-arch specific bits to the core code. 3. Use MMCONFIG code and implement generic ACPI based PCI host controller driver. 4. Enable above driver on ARM64
Patches has been built on top of 4.4 and can be found here: git@github.com:semihalf-nowicki-tomasz/linux.git (pci-acpi-v3)
NOTE, this patch set depends on Matthew's patches: http://www.spinics.net/lists/linux-pci/msg45950.html https://github.com/Vality/linux/tree/pci-fixes
This has been tested on Cavium ThunderX server and QEMU. Any help in reviewing and testing is very appreciated.
v2 -> v3
- fix legacy IRQ assigning and IO ports registration
- remove reference to arch specific companion device for ia64
- move ACPI PCI host controller driver to pci_root.c
- drop generic domain assignment for x86 and ia64 as I am not
able to run all necessary test variants
- drop patch which cleaned legacy IRQ assignment since it belongs to
Mathew's series: https://patchwork.ozlabs.org/patch/557504/
- extend MCFG quirk code
- rebased to 4.4
v1 -> v2
- moved non-arch specific piece of code to dirver/acpi/ directory
- fixed IO resource handling
- introduced PCI config accessors quirks matching
- moved ACPI_COMPANION_SET to generic code
IO resources on Mustang get disabled unless I do:
@@ -126,9 +126,10 @@ static void acpi_dev_ioresource_flags(struct resource *res, u64 len, if (!acpi_dev_resource_len_valid(res->start, res->end, len, true)) res->flags |= IORESOURCE_DISABLED | IORESOURCE_UNSET; +#if 0 if (res->end >= 0x10003) res->flags |= IORESOURCE_DISABLED | IORESOURCE_UNSET; - +#endif
res->end is way beyond 0x10003 on Mustang:
pci_bus 0000:00: root bus resource [io 0x0000-0xffff window] (bus address [0x10000000-0x1000ffff])
On 1/14/2016 10:29 AM, Mark Salter wrote: On Wed, 2016-01-13 at 14:20 +0100, Tomasz Nowicki wrote:
IO resources on Mustang get disabled unless I do:
@@ -126,9 +126,10 @@ static void acpi_dev_ioresource_flags(struct resource *res, u64 len, if (!acpi_dev_resource_len_valid(res->start, res->end, len, true)) res->flags |= IORESOURCE_DISABLED | IORESOURCE_UNSET;
+#if 0 if (res->end >= 0x10003) res->flags |= IORESOURCE_DISABLED | IORESOURCE_UNSET; - +#endif
res->end is way beyond 0x10003 on Mustang:
pci_bus 0000:00: root bus resource [io 0x0000-0xffff window] (bus address [0x10000000-0x1000ffff])
Join the club. I complained about this and I got the message that we just do what Intel does. See Arnd Bergmann's reply.
[PATCH V2 00/23] MMCONFIG refactoring and support for ARM64 PCI hostbridge init based on ACPI
1/12/2016 9:30 AM
It is an artificial limit coming from the x86 world introduced into common code.
-- 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
On Thu, Jan 14, 2016 at 10:38:19AM -0500, Sinan Kaya wrote:
On 1/14/2016 10:29 AM, Mark Salter wrote: On Wed, 2016-01-13 at 14:20 +0100, Tomasz Nowicki wrote:
IO resources on Mustang get disabled unless I do:
@@ -126,9 +126,10 @@ static void acpi_dev_ioresource_flags(struct resource *res, u64 len, if (!acpi_dev_resource_len_valid(res->start, res->end, len, true)) res->flags |= IORESOURCE_DISABLED | IORESOURCE_UNSET;
+#if 0 if (res->end >= 0x10003) res->flags |= IORESOURCE_DISABLED | IORESOURCE_UNSET; - +#endif
res->end is way beyond 0x10003 on Mustang:
pci_bus 0000:00: root bus resource [io 0x0000-0xffff window] (bus address [0x10000000-0x1000ffff])
Join the club. I complained about this and I got the message that we just do what Intel does. See Arnd Bergmann's reply.
[PATCH V2 00/23] MMCONFIG refactoring and support for ARM64 PCI hostbridge init based on ACPI
1/12/2016 9:30 AM
It is an artificial limit coming from the x86 world introduced into common code.
Guys, I think you are mixing things up here, we discussed this to death, read the archives please.
The ACPI IO descriptors AddressMinimum/AddressMaximum describe the IO space PCI bus addresses. The AddressTranslation field provides the PCI IO space -> CPU physical address translation, or put it differently, the secondary to primary bus translation in ACPI jargon, that's how ACPI tables must be written for IO space, at least that's what IA64 does (and on ia64 IO space is memory mapped, as on arm64).
I bet APM IO descriptors specify the *CPU* physical address in the AddressMinimum field, and that's where the problem lies.
Jiang's patch:
https://lkml.org/lkml/2015/12/16/249
parses the IO descriptors and stores the AddressMinimum, AddressMaximum in the IO resource (with AddressTranslation as offset which must be the *CPU* physical address mapping IO), from the log above it seems to me in AddressMinimum APM specifies the *CPU* physical address generating IO cycles.
All in all, I was right to fear this would happen, and I already raised the point within the ACPI spec working group, ACPI IO descriptors specification is ambiguous and we must agree on how they have to be specified once for all.
Lorenzo
On Thu, 2016-01-14 at 16:12 +0000, Lorenzo Pieralisi wrote:
On Thu, Jan 14, 2016 at 10:38:19AM -0500, Sinan Kaya wrote:
On 1/14/2016 10:29 AM, Mark Salter wrote: On Wed, 2016-01-13 at 14:20 +0100, Tomasz Nowicki wrote:
IO resources on Mustang get disabled unless I do:
@@ -126,9 +126,10 @@ static void acpi_dev_ioresource_flags(struct resource *res, u64 len, if (!acpi_dev_resource_len_valid(res->start, res->end, len, true)) res->flags |= IORESOURCE_DISABLED | IORESOURCE_UNSET;
+#if 0 if (res->end >= 0x10003) res->flags |= IORESOURCE_DISABLED | IORESOURCE_UNSET; - +#endif
res->end is way beyond 0x10003 on Mustang:
pci_bus 0000:00: root bus resource [io 0x0000-0xffff window] (bus address [0x10000000-0x1000ffff])
Join the club. I complained about this and I got the message that we just do what Intel does. See Arnd Bergmann's reply.
[PATCH V2 00/23] MMCONFIG refactoring and support for ARM64 PCI hostbridge init based on ACPI
1/12/2016 9:30 AM
It is an artificial limit coming from the x86 world introduced into common code.
Guys, I think you are mixing things up here, we discussed this to death, read the archives please.
The ACPI IO descriptors AddressMinimum/AddressMaximum describe the IO space PCI bus addresses. The AddressTranslation field provides the PCI IO space -> CPU physical address translation, or put it differently, the secondary to primary bus translation in ACPI jargon, that's how ACPI tables must be written for IO space, at least that's what IA64 does (and on ia64 IO space is memory mapped, as on arm64).
I bet APM IO descriptors specify the *CPU* physical address in the AddressMinimum field, and that's where the problem lies.
You would lose that bet. AddressMinimum/Maximum describe the PCI bus addresses.
QWordIO (ResourceProducer, MinFixed, MaxFixed, PosDecode, EntireRange, 0x0000000000000000, // Granularity 0x0000000010000000, // Range Minimum 0x000000001000FFFF, // Range Maximum 0x000000E000000000, // Translation Offset 0x0000000000010000, // Length ,, , TypeStatic)
Jiang's patch:
https://lkml.org/lkml/2015/12/16/249
parses the IO descriptors and stores the AddressMinimum, AddressMaximum in the IO resource (with AddressTranslation as offset which must be the *CPU* physical address mapping IO), from the log above it seems to me in AddressMinimum APM specifies the *CPU* physical address generating IO cycles.
All in all, I was right to fear this would happen, and I already raised the point within the ACPI spec working group, ACPI IO descriptors specification is ambiguous and we must agree on how they have to be specified once for all.
Lorenzo
On Thu, Jan 14, 2016 at 11:38:44AM -0500, Mark Salter wrote:
[...]
You would lose that bet. AddressMinimum/Maximum describe the PCI bus addresses.
In the mainline DT (APM Mustang), the CPU physical address corresponding to IO space is 0xe010000000, PCI bus address is 0x0.
QWordIO (ResourceProducer, MinFixed, MaxFixed, PosDecode, EntireRange, 0x0000000000000000, // Granularity 0x0000000010000000, // Range Minimum
0x000000001000FFFF, // Range Maximum 0x000000E000000000, // Translation Offset
See above, I will get the APM specifications to countercheck.
I agree with you we have to verify if this IO space limitation is real or it is just an x86ism, in which case we remove that check.
Lorenzo
0x0000000000010000, // Length ,, , TypeStatic)
Jiang's patch:
https://lkml.org/lkml/2015/12/16/249
parses the IO descriptors and stores the AddressMinimum, AddressMaximum in the IO resource (with AddressTranslation as offset which must be the *CPU* physical address mapping IO), from the log above it seems to me in AddressMinimum APM specifies the *CPU* physical address generating IO cycles.
All in all, I was right to fear this would happen, and I already raised the point within the ACPI spec working group, ACPI IO descriptors specification is ambiguous and we must agree on how they have to be specified once for all.
Lorenzo
On Thu, 2016-01-14 at 17:07 +0000, Lorenzo Pieralisi wrote:
On Thu, Jan 14, 2016 at 11:38:44AM -0500, Mark Salter wrote:
[...]
You would lose that bet. AddressMinimum/Maximum describe the PCI bus addresses.
In the mainline DT (APM Mustang), the CPU physical address corresponding to IO space is 0xe010000000, PCI bus address is 0x0.
QWordIO (ResourceProducer, MinFixed, MaxFixed, PosDecode, EntireRange, 0x0000000000000000, // Granularity 0x0000000010000000, // Range Minimum
0x000000001000FFFF, // Range Maximum 0x000000E000000000, // Translation Offset
See above, I will get the APM specifications to countercheck.
The spec won't help other than to verify that the PCIe bridge supports a 32-bit IO address space. The firmware sets the PCI bus base @ 0x10000000 with a CPU base address for that window @ 0xe010000000. The pci-xgene.c driver sets the PCI bus IO base address to whatever DT tells it too. For ACPI, we have to use whatever the firmware set it to and described it in the ACPI table.
When I looked at this a while back, neither ACPI nor PCI had anything which disallowed 32-bit IO space on the PCI bus. The 16-bit limit is an x86 limit in the instruction set.
I agree with you we have to verify if this IO space limitation is real or it is just an x86ism, in which case we remove that check.
Lorenzo
0x0000000000010000, // Length ,, , TypeStatic)
Jiang's patch:
https://lkml.org/lkml/2015/12/16/249
parses the IO descriptors and stores the AddressMinimum, AddressMaximum in the IO resource (with AddressTranslation as offset which must be the *CPU* physical address mapping IO), from the log above it seems to me in AddressMinimum APM specifies the *CPU* physical address generating IO cycles.
All in all, I was right to fear this would happen, and I already raised the point within the ACPI spec working group, ACPI IO descriptors specification is ambiguous and we must agree on how they have to be specified once for all.
Lorenzo
On Thu, Jan 14, 2016 at 12:32:03PM -0500, Mark Salter wrote:
On Thu, 2016-01-14 at 17:07 +0000, Lorenzo Pieralisi wrote:
On Thu, Jan 14, 2016 at 11:38:44AM -0500, Mark Salter wrote:
[...]
You would lose that bet. AddressMinimum/Maximum describe the PCI bus addresses.
In the mainline DT (APM Mustang), the CPU physical address corresponding to IO space is 0xe010000000, PCI bus address is 0x0.
QWordIO (ResourceProducer, MinFixed, MaxFixed, PosDecode, EntireRange, 0x0000000000000000, // Granularity 0x0000000010000000, // Range Minimum
0x000000001000FFFF, // Range Maximum 0x000000E000000000, // Translation Offset
See above, I will get the APM specifications to countercheck.
The spec won't help other than to verify that the PCIe bridge supports a 32-bit IO address space. The firmware sets the PCI bus base @ 0x10000000 with a CPU base address for that window @ 0xe010000000. The pci-xgene.c driver sets the PCI bus IO base address to whatever DT tells it too. For ACPI, we have to use whatever the firmware set it to and described it in the ACPI table.
It makes sense, thank you for clarifying (and sorry for jumping to conclusions, I am a bit worried about the ACPI IO space descriptors specification and usage on arm64).
When I looked at this a while back, neither ACPI nor PCI had anything which disallowed 32-bit IO space on the PCI bus. The 16-bit limit is an x86 limit in the instruction set.
We should ask Jiang to remove that check or to make it x86 only (does current mainline - where the offset is added to the resource start/end - work on ia64 ?)
Thanks, Lorenzo
I agree with you we have to verify if this IO space limitation is real or it is just an x86ism, in which case we remove that check.
Lorenzo
0x0000000000010000, // Length ,, , TypeStatic)
Jiang's patch:
https://lkml.org/lkml/2015/12/16/249
parses the IO descriptors and stores the AddressMinimum, AddressMaximum in the IO resource (with AddressTranslation as offset which must be the *CPU* physical address mapping IO), from the log above it seems to me in AddressMinimum APM specifies the *CPU* physical address generating IO cycles.
All in all, I was right to fear this would happen, and I already raised the point within the ACPI spec working group, ACPI IO descriptors specification is ambiguous and we must agree on how they have to be specified once for all.
Lorenzo
On Thu, 2016-01-14 at 17:59 +0000, Lorenzo Pieralisi wrote:
On Thu, Jan 14, 2016 at 12:32:03PM -0500, Mark Salter wrote:
On Thu, 2016-01-14 at 17:07 +0000, Lorenzo Pieralisi wrote:
On Thu, Jan 14, 2016 at 11:38:44AM -0500, Mark Salter wrote:
[...]
You would lose that bet. AddressMinimum/Maximum describe the PCI bus addresses.
In the mainline DT (APM Mustang), the CPU physical address corresponding to IO space is 0xe010000000, PCI bus address is 0x0.
QWordIO (ResourceProducer, MinFixed, MaxFixed, PosDecode, EntireRange, 0x0000000000000000, // Granularity 0x0000000010000000, // Range Minimum
0x000000001000FFFF, // Range Maximum 0x000000E000000000, // Translation Offset
See above, I will get the APM specifications to countercheck.
The spec won't help other than to verify that the PCIe bridge supports a 32-bit IO address space. The firmware sets the PCI bus base @ 0x10000000 with a CPU base address for that window @ 0xe010000000. The pci-xgene.c driver sets the PCI bus IO base address to whatever DT tells it too. For ACPI, we have to use whatever the firmware set it to and described it in the ACPI table.
It makes sense, thank you for clarifying (and sorry for jumping to conclusions, I am a bit worried about the ACPI IO space descriptors specification and usage on arm64).
When I looked at this a while back, neither ACPI nor PCI had anything which disallowed 32-bit IO space on the PCI bus. The 16-bit limit is an x86 limit in the instruction set.
We should ask Jiang to remove that check or to make it x86 only (does current mainline - where the offset is added to the resource start/end - work on ia64 ?)
I think current mainline ia64 would work only for IO spaces which don't exceed 64KiB in size. Otherwise it would trip on the same check disabling Mustang IO space. ia64-specific code supports multiple IO spaces but they all are expected to start at PCI bus address zero. At least my reading of that code leads me to believe it is broken otherwise. In any case, those IO spaces on the PCI bus could exceed 64K in size and any of those which do would get caught by the 64K check in generic code and disabled.
Thanks, Lorenzo
I agree with you we have to verify if this IO space limitation is real or it is just an x86ism, in which case we remove that check.
Lorenzo
0x0000000000010000, // Length ,, , TypeStatic)
Jiang's patch:
https://lkml.org/lkml/2015/12/16/249
parses the IO descriptors and stores the AddressMinimum, AddressMaximum in the IO resource (with AddressTranslation as offset which must be the *CPU* physical address mapping IO), from the log above it seems to me in AddressMinimum APM specifies the *CPU* physical address generating IO cycles.
All in all, I was right to fear this would happen, and I already raised the point within the ACPI spec working group, ACPI IO descriptors specification is ambiguous and we must agree on how they have to be specified once for all.
Lorenzo
On 01/14/2016 09:29 AM, Mark Salter wrote:
IO resources on Mustang get disabled unless I do:
@@ -126,9 +126,10 @@ static void acpi_dev_ioresource_flags(struct resource *res, u64 len, if (!acpi_dev_resource_len_valid(res->start, res->end, len, true)) res->flags |= IORESOURCE_DISABLED | IORESOURCE_UNSET;
+#if 0 if (res->end >= 0x10003) res->flags |= IORESOURCE_DISABLED | IORESOURCE_UNSET;
+#endif
res->end is way beyond 0x10003 on Mustang:
pci_bus 0000:00: root bus resource [io 0x0000-0xffff window] (bus address [0x10000000-0x1000ffff])
Juno is throwing "No Space" errors too with the following acpi definition.
DWordIo ( // IO window ResourceProducer, MinFixed, MaxFixed, PosDecode, EntireRange, 0x00000000,// Granularity 0x00000000,// Min Base Address 0x007fffff,// Max Base Address 0x5f800000, // Translate 0x00800000, // Length , , , TypeTranslation )
Although, my reading of the ACPI dword section is suspect...
On 01/13/2016 07:20 AM, Tomasz Nowicki wrote:
From the functionality point of view this series might be split into the following logic parts:
- Make MMCONFIG code arch-agnostic which allows all architectures to collect PCI config regions and used when necessary.
- Move non-arch specific bits to the core code.
- Use MMCONFIG code and implement generic ACPI based PCI host controller driver.
- Enable above driver on ARM64
Patches has been built on top of 4.4 and can be found here: git@github.com:semihalf-nowicki-tomasz/linux.git (pci-acpi-v3)
I boot tested this on JunoR2 with a recent build of Tianocore EDK2.
It works enough to boot the machine with PCIe attached sata, and ethernet, but is throwing messages like:
[ 1.094643] pci 0000:00:00.0: BAR 13: no space for [io size 0x2000] [ 1.101022] pci 0000:00:00.0: BAR 13: failed to assign [io size 0x2000]
Its also throwing:
[ 1.029958] pci 0000:00:00.0: of_irq_parse_pci() failed with rc=-22 [ 1.052784] pci 0000:00:00.0: of_irq_parse_pci() failed with rc=-22
(see other response for the DwordIo resource i'm using).
On 2016/1/13 21:20, Tomasz Nowicki wrote:
From the functionality point of view this series might be split into the following logic parts:
- Make MMCONFIG code arch-agnostic which allows all architectures to collect PCI config regions and used when necessary.
- Move non-arch specific bits to the core code.
- Use MMCONFIG code and implement generic ACPI based PCI host controller driver.
- Enable above driver on ARM64
Patches has been built on top of 4.4 and can be found here: git@github.com:semihalf-nowicki-tomasz/linux.git (pci-acpi-v3)
NOTE, this patch set depends on Matthew's patches: http://www.spinics.net/lists/linux-pci/msg45950.html https://github.com/Vality/linux/tree/pci-fixes
This has been tested on Cavium ThunderX server and QEMU. Any help in reviewing and testing is very appreciated.
v2 -> v3
- fix legacy IRQ assigning and IO ports registration
- remove reference to arch specific companion device for ia64
- move ACPI PCI host controller driver to pci_root.c
- drop generic domain assignment for x86 and ia64 as I am not able to run all necessary test variants
- drop patch which cleaned legacy IRQ assignment since it belongs to Mathew's series: https://patchwork.ozlabs.org/patch/557504/
- extend MCFG quirk code
- rebased to 4.4
Tested this patch set (and Matthew's) on a HP RX2660 IA64 machine, with the CONFIG_ACPI_PCI_HOST_GENERIC disabled on IA64 ( it's a minor comment needs to be addressed in my previous email), it can boot the system ok with NIC working properly, with the /pro/ioports [1] and the boot log [2] if anyone needs refer to it.
For this patch set, on IA64 machine,
Tested-by: Hanjun Guo hanjun.guo@linaro.org
Thanks Hanjun
[1] node11:~ # cat /proc/ioports 000003c0-000003df : vga+ 01000000-0100ffff : PCI Bus 0000:00 01001000-010010ff : 0000:00:03.0 02000000-0200ffff : PCI Bus 0000:01 02001000-020010ff : 0000:01:01.0 03000000-0300ffff : PCI Bus 0000:02 03001000-0300103f : 0000:02:01.1 03001040-0300107f : 0000:02:01.0 04000000-0400ffff : PCI Bus 0000:03 04001000-040010ff : 0000:03:01.1 04001100-040011ff : 0000:03:01.0 05000000-0500ffff : PCI Bus 0000:04 05001000-0500103f : 0000:04:01.1 05001040-0500107f : 0000:04:01.0
[2] boot log Initializing cgroup subsys cpuset Initializing cgroup subsys cpu Initializing cgroup subsys cpuacct Linux version 4.4.0.1-0.7-default (root@node11) (gcc version 4.3.4 [gcc-4_3-branch revision 152973] (SUSE Linux) ) #60 SMP Fri Jan 15 17:19:02 EST 2016 EFI v2.00 by HP: efi: SALsystab=0x3ee7a000 ACPI 2.0=0x3fde6000 SMBIOS=0x3ee7c000 HCDP=0x3fde4000 booting generic kernel on platform hpzx1 PCDP: v3 at 0x3fde4000 PCDP: VGA console ACPI: Early table checksum verification disabled ACPI: RSDP 0x000000003FDE6000 000028 (v02 HP ) ACPI: XSDT 0x000000003FDE602C 0000BC (v01 HP rx2660 00000000 HP 00000000) ACPI: FACP 0x000000003FDF7268 0000F4 (v03 HP rx2660 00000000 HP 00000000) ACPI BIOS Warning (bug): 32/64X length mismatch in FADT/Gpe0Block: 32/16 (20150930/tbfadt-623) ACPI BIOS Warning (bug): 32/64X length mismatch in FADT/Gpe1Block: 32/16 (20150930/tbfadt-623) ACPI: DSDT 0x000000003FDE61C8 00E566 (v01 HP rx2660 00000007 INTL 20050309) ACPI: FACS 0x000000003FDF7360 000040 ACPI: SPCR 0x000000003FDF73A0 000050 (v01 HP 00000000 HP 00000000) ACPI: DBGP 0x000000003FDF73F0 000034 (v01 HP rx2660 00000000 HP 00000000) ACPI: APIC 0x000000003FDF7810 0000E8 (v01 HP rx2660 00000000 HP 00000000) ACPI: SPMI 0x000000003FDF7428 000050 (v04 HP rx2660 00000000 HP 00000000) ACPI: CPEP 0x000000003FDF76E0 000034 (v01 HP rx2660 00000000 HP 00000000) ACPI: SSDT 0x000000003FDF4738 0004B3 (v01 HP rx2660 00000006 INTL 20050309) ACPI: SSDT 0x000000003FDF4BF8 000456 (v01 HP rx2660 00000006 INTL 20050309) ACPI: SSDT 0x000000003FDF5058 000866 (v01 HP rx2660 00000006 INTL 20050309) ACPI: SSDT 0x000000003FDF58C8 000866 (v01 HP rx2660 00000006 INTL 20050309) ACPI: SSDT 0x000000003FDF6138 000866 (v01 HP rx2660 00000006 INTL 20050309) ACPI: SSDT 0x000000003FDF69A8 000138 (v01 HP rx2660 00000006 INTL 20050309) ACPI: SSDT 0x000000003FDF6AE8 0000E0 (v01 HP rx2660 00000006 INTL 20050309) ACPI: SSDT 0x000000003FDF6BD8 00013C (v01 HP rx2660 00000006 INTL 20050309) ACPI: SSDT 0x000000003FDF6D18 0000E0 (v01 HP rx2660 00000006 INTL 20050309) ACPI: SSDT 0x000000003FDF6E08 00013C (v01 HP rx2660 00000006 INTL 20050309) ACPI: SSDT 0x000000003FDF6F48 0000E0 (v01 HP rx2660 00000006 INTL 20050309) ACPI: SSDT 0x000000003FDF7038 00013C (v01 HP rx2660 00000006 INTL 20050309) ACPI: SSDT 0x000000003FDF7178 0000E0 (v01 HP rx2660 00000006 INTL 20050309) ACPI: Local APIC address c0000000fee00000 4 CPUs available, 8 CPUs total SMP: Allowing 8 CPUs, 4 hotplug CPUs Reserving 512MB of memory at 128MB for crashkernel (System RAM: 16347MB) Initial ramdisk at: 0xe0000100fe2af000 (13610859 bytes) SAL 3.20: HP version 4.11 SAL Platform features: None SAL: AP wakeup using external interrupt vector 0xff MCA related initialization done Virtual mem_map starts at 0xa07fffffc7c80000 Zone ranges: DMA [mem 0x0000000001000000-0x00000000ffffffff] Normal [mem 0x0000000100000000-0x00000100ffffffff] Movable zone start for each node Early memory node ranges node 0: [mem 0x0000000001000000-0x000000003e67ffff] node 0: [mem 0x000000003eaf0000-0x000000003ee6ffff] node 0: [mem 0x000000003fc00000-0x000000003fd6ffff] node 0: [mem 0x0000000100000000-0x00000003ffffffff] node 0: [mem 0x0000010040000000-0x00000100ff1effff] node 0: [mem 0x00000100ff200000-0x00000100ffffffff] Initmem setup node 0 [mem 0x0000000001000000-0x00000100ffffffff] On node 0 totalpages: 261558 free_area_init_node: node 0, pgdat e000000001a00000, node_mem_map a07fffffc7c83800 DMA zone: 56 pages used for memmap DMA zone: 0 pages reserved DMA zone: 15799 pages, LIFO batch:1 Normal zone: 14336 pages used for memmap Normal zone: 245759 pages, LIFO batch:1 pcpu-alloc: s33560 r8192 d220392 u262144 alloc=4*65536 pcpu-alloc: [0] 0 [0] 1 [0] 2 [0] 3 [0] 4 [0] 5 [0] 6 [0] 7 Built 1 zonelists in Node order, mobility grouping on. Total pages: 247166 Policy zone: Normal Kernel command line: BOOT_IMAGE=dev000:\efi\SuSE\vmlinuz-4.4.0.1-0.7-default root=/dev/disk/by-id/scsi-35000cca00a9965f4-part2 splash=silent crashkernel=1024M-:512M PID hash table entries: 4096 (order: -1, 32768 bytes) Sorting __ex_table... Memory: 16146944K/16739712K available (10928K kernel code, 5820K rwdata, 1828K rodata, 1280K init, 8346K bss, 592768K reserved, 0K cma-reserved) Hierarchical RCU implementation. Build-time adjustment of leaf fanout to 64. RCU restricting CPUs from NR_CPUS=4096 to nr_cpu_ids=8. RCU: Adjusting geometry for rcu_fanout_leaf=64, nr_cpu_ids=8 NR_IRQS:1024 ACPI: Local APIC address c0000000fee00000 GSI 25 (level, low) -> CPU 0 (0x0000) vector 49 CPU 0: base freq=266.116MHz, ITC ratio=6/4, ITC freq=399.174MHz+/-150ppm clocksource: itc: mask: 0xffffffffffffffff max_cycles: 0x5c0fd820f3, max_idle_ns: 440795212316 ns Console: colour VGA+ 80x25 console [tty0] enabled Calibrating delay loop... 2834.43 BogoMIPS (lpj=5668864) pid_max: default: 32768 minimum: 301 ACPI: Core revision 20150930 ACPI: 14 ACPI AML tables successfully acquired and loaded Security Framework initialized SELinux: Disabled at boot. AppArmor: AppArmor disabled by boot time parameter Dentry cache hash table entries: 2097152 (order: 8, 16777216 bytes) Inode-cache hash table entries: 1048576 (order: 7, 8388608 bytes) Mount-cache hash table entries: 32768 (order: 2, 262144 bytes) Mountpoint-cache hash table entries: 32768 (order: 2, 262144 bytes) Initializing cgroup subsys devices Initializing cgroup subsys freezer Initializing cgroup subsys net_cls Boot processor id 0x0/0x0 Fixed BSP b0 value from CPU 1 CPU 1: synchronized ITC with CPU 0 (last diff -6 cycles, maxerr 172 cycles) CPU 1: base freq=266.116MHz, ITC ratio=6/4, ITC freq=399.174MHz+/-150ppm CPU 2: synchronized ITC with CPU 0 (last diff -6 cycles, maxerr 173 cycles) CPU 2: base freq=266.116MHz, ITC ratio=6/4, ITC freq=399.174MHz+/-150ppm CPU 3: synchronized ITC with CPU 0 (last diff -6 cycles, maxerr 173 cycles) CPU 3: base freq=266.116MHz, ITC ratio=6/4, ITC freq=399.174MHz+/-150ppm Brought up 4 CPUs Total of 4 processors activated (11337.72 BogoMIPS). devtmpfs: initialized SMBIOS 2.4 present. DMI: hp server rx2660 , BIOS 04.11 10/20/2008 clocksource: jiffies: mask: 0xffffffff max_cycles: 0xffffffff, max_idle_ns: 7645041785100000 ns NET: Registered protocol family 16 ACPI: bus type PCI registered ACPI: Added _OSI(Module Device) ACPI: Added _OSI(Processor Device) ACPI: Added _OSI(3.0 _SCP Extensions) ACPI: Added _OSI(Processor Aggregator Device) _SB_:_OSC invalid UUID _OSC request data:1 e ACPI: Interpreter enabled ACPI: (supports S0 S5) ACPI: Using IOSAPIC for interrupt routing ACPI: PCI Root Bridge [L000] (domain 0000 [bus 00]) acpi HWP0002:00: _OSC: OS supports [ExtendedConfig ASPM ClockPM Segments MSI] acpi HWP0002:00: _OSC failed (AE_NOT_FOUND); disabling ASPM debug: resource flag 0x200201, name PCI Bus 0000:00 with resouce [0xff5e2000 - 0xff5e2007] debug: resource flag 0x200105, name PCI Bus 0000:00 with resouce [0x1000000 - 0x100ffff]debug: resource flag 0x200201, name PCI Bus 0000:00 with resouce [0xa0000 - 0xfffff] debug: resource flag 0x200201, name PCI Bus 0000:00 with resouce [0x80000000 - 0x8fffffff]debug: resource flag 0x200201, name PCI Bus 0000:00 with resouce [0x80004000000 - 0x800ffffffff] acpi HWP0002:00: PCI host bridge to bus 0000:00 pci_bus 0000:00: root bus resource [io 0x1000000-0x100ffff window] (bus address [0x0000-0xffff]) pci_bus 0000:00: root bus resource [mem 0x000a0000-0x000fffff window] pci_bus 0000:00: root bus resource [mem 0x80000000-0x8fffffff window] pci_bus 0000:00: root bus resource [mem 0x80004000000-0x800ffffffff window] pci_bus 0000:00: root bus resource [bus 00] pci 0000:00:01.0: [103c:1303] type 00 class 0xff0000 pci 0000:00:01.1: [103c:1302] type 00 class 0x078000 pci 0000:00:01.1: reg 0x14: [mem 0x88034000-0x88034fff 64bit] pci 0000:00:01.1: reg 0x1c: [mem 0x80080000000-0x8008001ffff 64bit pref] pci 0000:00:01.2: [103c:1048] type 00 class 0x070002 pci 0000:00:01.2: reg 0x14: [mem 0x88033000-0x88033fff 64bit] pci 0000:00:02.0: [1033:0035] type 00 class 0x0c0310 pci 0000:00:02.0: reg 0x10: [mem 0x88032000-0x88032fff] pci 0000:00:02.0: supports D1 D2 pci 0000:00:02.0: PME# supported from D0 D1 D2 D3hot pci 0000:00:02.1: [1033:0035] type 00 class 0x0c0310 pci 0000:00:02.1: reg 0x10: [mem 0x88031000-0x88031fff] pci 0000:00:02.1: supports D1 D2 pci 0000:00:02.1: PME# supported from D0 D1 D2 D3hot pci 0000:00:02.2: [1033:00e0] type 00 class 0x0c0320 pci 0000:00:02.2: reg 0x10: [mem 0x88030000-0x880300ff] pci 0000:00:02.2: supports D1 D2 pci 0000:00:02.2: PME# supported from D0 D1 D2 D3hot pci 0000:00:03.0: [1002:515e] type 00 class 0x030000 pci 0000:00:03.0: reg 0x10: [mem 0x80000000-0x87ffffff pref] pci 0000:00:03.0: reg 0x14: [io 0x1001000-0x10010ff] pci 0000:00:03.0: reg 0x18: [mem 0x88020000-0x8802ffff] pci 0000:00:03.0: reg 0x30: [mem 0x88000000-0x8801ffff pref] pci 0000:00:03.0: supports D1 D2 ACPI: PCI Root Bridge [L002] (domain 0000 [bus 01]) acpi HWP0002:01: _OSC: OS supports [ExtendedConfig ASPM ClockPM Segments MSI] acpi HWP0002:01: _OSC failed (AE_NOT_FOUND); disabling ASPM debug: resource flag 0x200105, name PCI Bus 0000:01 with resouce [0x2000000 - 0x200ffff] debug: resource flag 0x200201, name PCI Bus 0000:01 with resouce [0xa0000000 - 0xafffffff]debug: resource flag 0x200201, name PCI Bus 0000:01 with resouce [0x80204000000 - 0x802ffffffff] acpi HWP0002:01: PCI host bridge to bus 0000:01 pci_bus 0000:01: root bus resource [io 0x2000000-0x200ffff window] (bus address [0x0000-0xffff]) pci_bus 0000:01: root bus resource [mem 0xa0000000-0xafffffff window] pci_bus 0000:01: root bus resource [mem 0x80204000000-0x802ffffffff window] pci_bus 0000:01: root bus resource [bus 01] pci 0000:01:01.0: [1000:0054] type 00 class 0x010000 pci 0000:01:01.0: reg 0x10: [io 0x2001000-0x20010ff] pci 0000:01:01.0: reg 0x14: [mem 0xa0470000-0xa0473fff 64bit] pci 0000:01:01.0: reg 0x1c: [mem 0xa0460000-0xa046ffff 64bit] pci 0000:01:01.0: reg 0x30: [mem 0xa0000000-0xa03fffff pref] pci 0000:01:01.0: supports D1 D2 pci 0000:01:02.0: [14e4:1648] type 00 class 0x020000 pci 0000:01:02.0: reg 0x10: [mem 0xa0450000-0xa045ffff 64bit] pci 0000:01:02.0: reg 0x30: [mem 0xa0420000-0xa043ffff pref] pci 0000:01:02.0: PME# supported from D3hot D3cold pci 0000:01:02.1: [14e4:1648] type 00 class 0x020000 pci 0000:01:02.1: reg 0x10: [mem 0xa0440000-0xa044ffff 64bit] pci 0000:01:02.1: reg 0x30: [mem 0xa0400000-0xa041ffff pref] pci 0000:01:02.1: PME# supported from D3hot D3cold ACPI: PCI Root Bridge [L003] (domain 0000 [bus 02]) acpi HWP0002:02: _OSC: OS supports [ExtendedConfig ASPM ClockPM Segments MSI] acpi HWP0002:02: _OSC failed (AE_NOT_FOUND); disabling ASPM debug: resource flag 0x200105, name PCI Bus 0000:02 with resouce [0x3000000 - 0x300ffff] debug: resource flag 0x200201, name PCI Bus 0000:02 with resouce [0xb0000000 - 0xbfffffff]debug: resource flag 0x200201, name PCI Bus 0000:02 with resouce [0x80304000000 - 0x803ffffffff] acpi HWP0002:02: PCI host bridge to bus 0000:02 pci_bus 0000:02: root bus resource [io 0x3000000-0x300ffff window] (bus address [0x0000-0xffff]) pci_bus 0000:02: root bus resource [mem 0xb0000000-0xbfffffff window] pci_bus 0000:02: root bus resource [mem 0x80304000000-0x803ffffffff window] pci_bus 0000:02: root bus resource [bus 02] pci 0000:02:01.0: [8086:1079] type 00 class 0x020000 pci 0000:02:01.0: reg 0x10: [mem 0xb0120000-0xb013ffff 64bit] pci 0000:02:01.0: reg 0x18: [mem 0xb00c0000-0xb00fffff 64bit] pci 0000:02:01.0: reg 0x20: [io 0x3001040-0x300107f] pci 0000:02:01.0: reg 0x30: [mem 0xb0080000-0xb00bffff pref] pci 0000:02:01.0: PME# supported from D0 D3hot pci 0000:02:01.1: [8086:1079] type 00 class 0x020000 pci 0000:02:01.1: reg 0x10: [mem 0xb0100000-0xb011ffff 64bit] pci 0000:02:01.1: reg 0x18: [mem 0xb0040000-0xb007ffff 64bit] pci 0000:02:01.1: reg 0x20: [io 0x3001000-0x300103f] pci 0000:02:01.1: reg 0x30: [mem 0xb0000000-0xb003ffff pref] pci 0000:02:01.1: PME# supported from D0 D3hot ACPI: PCI Root Bridge [L006] (domain 0000 [bus 03]) acpi HWP0002:03: _OSC: OS supports [ExtendedConfig ASPM ClockPM Segments MSI] acpi HWP0002:03: _OSC failed (AE_NOT_FOUND); disabling ASPM debug: resource flag 0x200105, name PCI Bus 0000:03 with resouce [0x4000000 - 0x400ffff] debug: resource flag 0x200201, name PCI Bus 0000:03 with resouce [0xe0000000 - 0xefffffff]debug: resource flag 0x200201, name PCI Bus 0000:03 with resouce [0x80604000000 - 0x806ffffffff] acpi HWP0002:03: PCI host bridge to bus 0000:03 pci_bus 0000:03: root bus resource [io 0x4000000-0x400ffff window] (bus address [0x0000-0xffff]) pci_bus 0000:03: root bus resource [mem 0xe0000000-0xefffffff window] pci_bus 0000:03: root bus resource [mem 0x80604000000-0x806ffffffff window] pci_bus 0000:03: root bus resource [bus 03] pci 0000:03:01.0: [1000:0030] type 00 class 0x010000 pci 0000:03:01.0: reg 0x10: [io 0x4001100-0x40011ff] pci 0000:03:01.0: reg 0x14: [mem 0xe0260000-0xe027ffff 64bit] pci 0000:03:01.0: reg 0x1c: [mem 0xe0240000-0xe025ffff 64bit] pci 0000:03:01.0: reg 0x30: [mem 0xe0100000-0xe01fffff pref] pci 0000:03:01.0: supports D1 D2 pci 0000:03:01.1: [1000:0030] type 00 class 0x010000 pci 0000:03:01.1: reg 0x10: [io 0x4001000-0x40010ff] pci 0000:03:01.1: reg 0x14: [mem 0xe0220000-0xe023ffff 64bit] pci 0000:03:01.1: reg 0x1c: [mem 0xe0200000-0xe021ffff 64bit] pci 0000:03:01.1: reg 0x30: [mem 0xe0000000-0xe00fffff pref] pci 0000:03:01.1: supports D1 D2 ACPI: PCI Root Bridge [L007] (domain 0000 [bus 04]) acpi HWP0002:04: _OSC: OS supports [ExtendedConfig ASPM ClockPM Segments MSI] acpi HWP0002:04: _OSC failed (AE_NOT_FOUND); disabling ASPM debug: resource flag 0x200105, name PCI Bus 0000:04 with resouce [0x5000000 - 0x500ffff] debug: resource flag 0x200201, name PCI Bus 0000:04 with resouce [0xf0000000 - 0xfdffffff]debug: resource flag 0x200201, name PCI Bus 0000:04 with resouce [0x80704000000 - 0x807ffffffff] acpi HWP0002:04: PCI host bridge to bus 0000:04 pci_bus 0000:04: root bus resource [io 0x5000000-0x500ffff window] (bus address [0x0000-0xffff]) pci_bus 0000:04: root bus resource [mem 0xf0000000-0xfdffffff window] pci_bus 0000:04: root bus resource [mem 0x80704000000-0x807ffffffff window] pci_bus 0000:04: root bus resource [bus 04] pci 0000:04:01.0: [8086:1079] type 00 class 0x020000 pci 0000:04:01.0: reg 0x10: [mem 0xf0120000-0xf013ffff 64bit] pci 0000:04:01.0: reg 0x18: [mem 0xf00c0000-0xf00fffff 64bit] pci 0000:04:01.0: reg 0x20: [io 0x5001040-0x500107f] pci 0000:04:01.0: reg 0x30: [mem 0xf0080000-0xf00bffff pref] pci 0000:04:01.0: PME# supported from D0 D3hot pci 0000:04:01.1: [8086:1079] type 00 class 0x020000 pci 0000:04:01.1: reg 0x10: [mem 0xf0100000-0xf011ffff 64bit] pci 0000:04:01.1: reg 0x18: [mem 0xf0040000-0xf007ffff 64bit] pci 0000:04:01.1: reg 0x20: [io 0x5001000-0x500103f] pci 0000:04:01.1: reg 0x30: [mem 0xf0000000-0xf003ffff pref] pci 0000:04:01.1: PME# supported from D0 D3hot ACPI: Enabled 2 GPEs in block 10 to 1F vgaarb: setting as boot device: PCI:0000:00:03.0 vgaarb: device added: PCI:0000:00:03.0,decodes=io+mem,owns=io+mem,locks=none vgaarb: loaded vgaarb: bridge control possible 0000:00:03.0 IOC: zx2 2.0 HPA 0xfed01000 IOVA space 1024Mb at 0x40000000 NetLabel: Initializing NetLabel: domain hash size = 128 NetLabel: protocols = UNLABELED CIPSOv4 NetLabel: unlabeled traffic allowed by default clocksource: Switched to clocksource itc pnp: PnP ACPI init GSI 24 (level, low) -> CPU 1 (0x0200) vector 50 pnp 00:00: Plug and Play ACPI device, IDs PNP0501 (active) pnp: PnP ACPI: found 1 devices NET: Registered protocol family 2 TCP established hash table entries: 131072 (order: 4, 1048576 bytes) TCP bind hash table entries: 65536 (order: 4, 1048576 bytes) TCP: Hash tables configured (established 131072 bind 65536) UDP hash table entries: 8192 (order: 2, 262144 bytes) UDP-Lite hash table entries: 8192 (order: 2, 262144 bytes) NET: Registered protocol family 1 GSI 17 (level, low) -> CPU 2 (0x0400) vector 51 GSI 17 (level, low) -> CPU 2 (0x0400) vector 51 unregistered GSI 18 (level, low) -> CPU 3 (0x0600) vector 51 GSI 18 (level, low) -> CPU 3 (0x0600) vector 51 unregistered GSI 19 (level, low) -> CPU 0 (0x0000) vector 51 GSI 19 (level, low) -> CPU 0 (0x0000) vector 51 unregistered pci 0000:00:03.0: Video device with shadowed ROM PCI: CLS 128 bytes, default 128 Unpacking initramfs... Freeing initrd memory: 13248kB freed perfmon: version 2.0 IRQ 238 perfmon: Montecito PMU detected, 27 PMCs, 35 PMDs, 12 counters (47 bits) PAL Information Facility v0.5 perfmon: added sampling format default_format perfmon_default_smpl: default_format v2.0 registered futex hash table entries: 2048 (order: 2, 262144 bytes) audit: initializing netlink subsys (disabled) audit: type=2000 audit(1452897622.516:1): initialized HugeTLB registered 256 MB page size, pre-allocated 0 pages VFS: Disk quotas dquot_6.6.0 VFS: Dquot-cache hash table entries: 8192 (order 0, 65536 bytes) Block layer SCSI generic (bsg) driver version 0.4 loaded (major 253) io scheduler noop registered io scheduler deadline registered io scheduler cfq registered (default) Serial: 8250/16550 driver, 4 ports, IRQ sharing enabled 00:00: ttyS0 at MMIO 0xff5e2000 (irq = 50, base_baud = 115200) is a 16550A GSI 16 (level, low) -> CPU 1 (0x0200) vector 51 0000:00:01.2: ttyS1 at MMIO 0x88033000 (irq = 54, base_baud = 115200) is a 16550A EFI Time Services Driver v0.4 libphy: Fixed MDIO Bus: probed mousedev: PS/2 mouse device common for all mice EFI Variables Facility v0.08 2004-May-17 registered taskstats version 1 Freeing unused kernel memory: 1280K (e000000004d80000 - e000000004ec0000) thermal LNXTHERM:00: registered as thermal_zone0 ACPI: Thermal Zone [THM0] (27 C) SCSI subsystem initialized Fusion MPT base driver 3.04.20 Copyright (c) 1999-2008 LSI Corporation Fusion MPT SPI Host driver 3.04.20 GSI 49 (level, low) -> CPU 2 (0x0400) vector 52 mptbase: ioc0: Initiating bringup ioc0: LSI53C1030 C0: Capabilities={Initiator,Target} scsi host0: ioc0: LSI53C1030 C0, FwRev=01032346h, Ports=1, MaxQ=255, IRQ=55 GSI 50 (level, low) -> CPU 3 (0x0600) vector 53 mptbase: ioc1: Initiating bringup ioc1: LSI53C1030 C0: Capabilities={Initiator,Target} scsi host1: ioc1: LSI53C1030 C0, FwRev=01032346h, Ports=1, MaxQ=255, IRQ=56 random: nonblocking pool is initialized Fusion MPT SAS Host driver 3.04.20 GSI 27 (level, low) -> CPU 0 (0x0000) vector 54 mptbase: ioc2: Initiating bringup ioc2: LSISAS1068 B0: Capabilities={Initiator} scsi host2: ioc2: LSISAS1068 B0, FwRev=01172a00h, Ports=1, MaxQ=163, IRQ=57 mptsas: ioc2: attaching ssp device: fw_channel 0, fw_id 8, phy 0, sas_addr 0x5000cca00a9965f5 scsi 2:0:0:0: Direct-Access HP DG0146FARVU HPD6 PQ: 0 ANSI: 5 mptsas: ioc2: attaching ssp device: fw_channel 0, fw_id 7, phy 1, sas_addr 0x5000cca00a99675d scsi 2:0:1:0: Direct-Access HP DG0146FARVU HPD6 PQ: 0 ANSI: 5 udevd (181): /proc/181/oom_adj is deprecated, please use /proc/181/oom_score_adj instead. udevd version 128 started sd 2:0:0:0: [sda] 286749488 512-byte logical blocks: (146 GB/136 GiB) sd 2:0:1:0: [sdb] 286749488 512-byte logical blocks: (146 GB/136 GiB) sd 2:0:0:0: [sda] Write Protect is off sd 2:0:0:0: [sda] Mode Sense: eb 00 10 08 sd 2:0:1:0: [sdb] Write Protect is off sd 2:0:1:0: [sdb] Mode Sense: eb 00 10 08 sd 2:0:0:0: [sda] Write cache: disabled, read cache: enabled, supports DPO and FUA sd 2:0:1:0: [sdb] Write cache: disabled, read cache: enabled, supports DPO and FUA sda: sda1 sda2 sdb: sdb1 sdb2 sdb3 sdb4 sdb5 sdb6 sdb7 sd 2:0:0:0: [sda] Attached SCSI disk sd 2:0:1:0: [sdb] Attached SCSI disk EXT4-fs (sda2): mounting ext3 file system using the ext4 subsystem EXT4-fs (sda2): mounted filesystem with ordered data mode. Opts: acl,user_xattr EXT4-fs (sda2): re-mounted. Opts: acl,user_xattr udevd version 128 started ipmi message handler version 39.2 e1000: Intel(R) PRO/1000 Network Driver - version 7.3.21-k8-NAPI e1000: Copyright (c) 1999-2006 Intel Corporation. GSI 38 (level, low) -> CPU 1 (0x0200) vector 55 input: Power Button as /devices/LNXSYSTM:00/LNXPWRBN:00/input/input0 ACPI: Power Button [PWRF] input: Sleep Button as /devices/LNXSYSTM:00/LNXSLPBN:00/input/input1 ACPI: Sleep Button [SLPF] e1000 0000:02:01.0 eth0: (PCI-X:133MHz:64-bit) 00:24:81:77:73:a4 e1000 0000:02:01.0 eth0: Intel(R) PRO/1000 Network Connection GSI 39 (level, low) -> CPU 2 (0x0400) vector 56 ipmi_si IPI0001:00: ipmi_si: probing via ACPI ipmi_si IPI0001:00: [mem 0xff5bc000-0xff5bc003] regsize 1 spacing 1 irq 0 ipmi_si: Adding ACPI-specified kcs state machine IPMI System Interface driver. ipmi_si: probing via SPMI ipmi_si: SPMI: mem 0xff5bc000 regsize 1 spacing 1 irq 0 ipmi_si: Adding SPMI-specified kcs state machine duplicate interface ipmi_si: Trying ACPI-specified kcs state machine at mem address 0xff5bc000, slave address 0x0, irq 0 e1000 0000:02:01.1 eth1: (PCI-X:133MHz:64-bit) 00:24:81:77:73:a5 e1000 0000:02:01.1 eth1: Intel(R) PRO/1000 Network Connection pps_core: LinuxPPS API ver. 1 registered pps_core: Software ver. 5.3.6 - Copyright 2005-2007 Rodolfo Giometti giometti@linux.it Fusion MPT misc device (ioctl) driver 3.04.20 GSI 60 (level, low) -> CPU 3 (0x0600) vector 57 mptctl: Registered with Fusion MPT base driver mptctl: /dev/mptctl @ (major,minor=10,220) sd 2:0:0:0: Attached scsi generic sg0 type 0 sd 2:0:1:0: Attached scsi generic sg1 type 0 PTP clock support registered ipmi_si IPI0001:00: Found new BMC (man_id: 0x00000b, prod_id: 0x8201, dev_id: 0x32) ipmi_si IPI0001:00: IPMI kcs interface initialized ACPI: bus type USB registered usbcore: registered new interface driver usbfs usbcore: registered new interface driver hub usbcore: registered new device driver usb tg3.c:v3.137 (May 11, 2014) GSI 29 (level, low) -> CPU 0 (0x0000) vector 58 ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver e1000 0000:04:01.0 eth2: (PCI-X:133MHz:64-bit) 00:24:81:77:75:70 e1000 0000:04:01.0 eth2: Intel(R) PRO/1000 Network Connection GSI 61 (level, low) -> CPU 1 (0x0200) vector 59 ohci_hcd: USB 1.1 'Open' Host Controller (OHCI) Driver ehci-pci: EHCI PCI platform driver GSI 19 (level, low) -> CPU 2 (0x0400) vector 60 ehci-pci 0000:00:02.2: EHCI Host Controller ehci-pci 0000:00:02.2: new USB bus registered, assigned bus number 1 ehci-pci 0000:00:02.2: irq 53, io mem 0x88030000 ehci-pci 0000:00:02.2: USB 2.0 started, EHCI 1.00 usb usb1: New USB device found, idVendor=1d6b, idProduct=0002 usb usb1: New USB device strings: Mfr=3, Product=2, SerialNumber=1 usb usb1: Product: EHCI Host Controller usb usb1: Manufacturer: Linux 4.4.0.1-0.7-default ehci_hcd usb usb1: SerialNumber: 0000:00:02.2 hub 1-0:1.0: USB hub found hub 1-0:1.0: 5 ports detected ohci-pci: OHCI PCI platform driver GSI 17 (level, low) -> CPU 3 (0x0600) vector 61 ohci-pci 0000:00:02.0: OHCI PCI host controller ohci-pci 0000:00:02.0: new USB bus registered, assigned bus number 2 ohci-pci 0000:00:02.0: irq 51, io mem 0x88032000 e1000 0000:04:01.1 eth3: (PCI-X:133MHz:64-bit) 00:24:81:77:75:71 e1000 0000:04:01.1 eth3: Intel(R) PRO/1000 Network Connection e1000 0000:02:01.0 eth0_rename: renamed from eth0 usb usb2: New USB device found, idVendor=1d6b, idProduct=0001 usb usb2: New USB device strings: Mfr=3, Product=2, SerialNumber=1 usb usb2: Product: OHCI PCI host controller usb usb2: Manufacturer: Linux 4.4.0.1-0.7-default ohci_hcd usb usb2: SerialNumber: 0000:00:02.0 hub 2-0:1.0: USB hub found hub 2-0:1.0: 3 ports detected GSI 18 (level, low) -> CPU 0 (0x0000) vector 62 ohci-pci 0000:00:02.1: OHCI PCI host controller ohci-pci 0000:00:02.1: new USB bus registered, assigned bus number 3 ohci-pci 0000:00:02.1: irq 52, io mem 0x88031000 e1000 0000:02:01.1 eth1_rename: renamed from eth1 e1000 0000:04:01.1 eth5: renamed from eth3 e1000 0000:04:01.0 eth4: renamed from eth2 tg3 0000:01:02.0 eth0: Tigon3 [partno(BCM95700A6) rev 2100] (PCIX:66MHz:64-bit) MAC address 00:22:64:94:09:3f tg3 0000:01:02.0 eth0: attached PHY is 5704 (10/100/1000Base-T Ethernet) (WireSpeed[1], EEE[0]) tg3 0000:01:02.0 eth0: RXcsums[1] LinkChgREG[0] MIirq[0] ASF[0] TSOcap[1] tg3 0000:01:02.0 eth0: dma_rwctrl[769f0000] dma_mask[64-bit] e1000 0000:02:01.1 eth3: renamed from eth1_rename GSI 30 (level, low) -> CPU 1 (0x0200) vector 63 usb usb3: New USB device found, idVendor=1d6b, idProduct=0001 usb usb3: New USB device strings: Mfr=3, Product=2, SerialNumber=1 usb usb3: Product: OHCI PCI host controller usb usb3: Manufacturer: Linux 4.4.0.1-0.7-default ohci_hcd usb usb3: SerialNumber: 0000:00:02.1 hub 3-0:1.0: USB hub found hub 3-0:1.0: 2 ports detected e1000 0000:02:01.0 eth2: renamed from eth0_rename tg3 0000:01:02.1 eth1: Tigon3 [partno(BCM95700A6) rev 2100] (PCIX:66MHz:64-bit) MAC address 00:22:64:94:09:3e tg3 0000:01:02.1 eth1: attached PHY is 5704 (10/100/1000Base-T Ethernet) (WireSpeed[1], EEE[0]) tg3 0000:01:02.1 eth1: RXcsums[1] LinkChgREG[0] MIirq[0] ASF[0] TSOcap[1] tg3 0000:01:02.1 eth1: dma_rwctrl[769f0000] dma_mask[64-bit] usb 1-2: new high-speed USB device number 3 using ehci-pci usb 2-1: new full-speed USB device number 2 using ohci-pci Adding 2104448k swap on /dev/sdb2. Priority:-1 extents:1 across:2104448k usb 1-2: New USB device found, idVendor=03f0, idProduct=0126 usb 1-2: New USB device strings: Mfr=51, Product=80, SerialNumber=67 usb 1-2: Product: INTEGRITY SERVER usb 1-2: Manufacturer: HEWLETT PACKARD usb 1-2: SerialNumber: A60020000001 usb 2-1: New USB device found, idVendor=03f0, idProduct=1126 usb 2-1: New USB device strings: Mfr=1, Product=2, SerialNumber=63 usb 2-1: Product: Virtual Management Device usb 2-1: Manufacturer: HP usb 2-1: SerialNumber: 0022649407fa hidraw: raw HID events driver (C) Jiri Kosina usb-storage 1-2:1.0: USB Mass Storage device detected scsi host3: usb-storage 1-2:1.0 usbcore: registered new interface driver usb-storage usbcore: registered new interface driver usbhid usbhid: USB HID core driver usb 3-2: new low-speed USB device number 2 using ohci-pci input: HP Virtual Management Device as /devices/LNXSYSTM:00/LNXSYBUS:00/HWP0001:00/HWP0002:00/pci0000:00/0000:00:02.0/usb2/2-1/2-1:1.0/0003:03F0:1126.0001/input/input2 hid-generic 0003:03F0:1126.0001: input,hidraw0: USB HID v1.11 Keyboard [HP Virtual Management Device] on usb-0000:00:02.0-1/input0 input: HP Virtual Management Device as /devices/LNXSYSTM:00/LNXSYBUS:00/HWP0001:00/HWP0002:00/pci0000:00/0000:00:02.0/usb2/2-1/2-1:1.1/0003:03F0:1126.0002/input/input3 hid-generic 0003:03F0:1126.0002: input,hidraw1: USB HID v1.01 Mouse [HP Virtual Management Device] on usb-0000:00:02.0-1/input1 usb 3-2: New USB device found, idVendor=0557, idProduct=2221 usb 3-2: New USB device strings: Mfr=1, Product=2, SerialNumber=3 usb 3-2: Product: UC-10KM V1.3.121 usb 3-2: Manufacturer: ATEN usb 3-2: SerialNumber: USB Composite device input: ATEN UC-10KM V1.3.121 as /devices/LNXSYSTM:00/LNXSYBUS:00/HWP0001:00/HWP0002:00/pci0000:00/0000:00:02.1/usb3/3-2/3-2:1.0/0003:0557:2221.0003/input/input4 hid-generic 0003:0557:2221.0003: input,hidraw2: USB HID v1.10 Keyboard [ATEN UC-10KM V1.3.121] on usb-0000:00:02.1-2/input0 input: ATEN UC-10KM V1.3.121 as /devices/LNXSYSTM:00/LNXSYBUS:00/HWP0001:00/HWP0002:00/pci0000:00/0000:00:02.1/usb3/3-2/3-2:1.1/0003:0557:2221.0004/input/input5 hid-generic 0003:0557:2221.0004: input,hidraw3: USB HID v1.10 Mouse [ATEN UC-10KM V1.3.121] on usb-0000:00:02.1-2/input1 device-mapper: uevent: version 1.0.3 device-mapper: ioctl: 4.34.0-ioctl (2015-10-28) initialised: dm-devel@redhat.com loop: module loaded scsi 3:0:0:0: CD-ROM TEAC DVD-ROM DW-224EV K.DA PQ: 0 ANSI: 0 scsi 3:0:0:0: Attached scsi generic sg2 type 5 sr 3:0:0:0: [sr0] scsi3-mmc drive: 24x/24x cd/rw xa/form2 cdda tray cdrom: Uniform CD-ROM driver Revision: 3.20 sr 3:0:0:0: Attached scsi CD-ROM sr0 fuse init (API version 7.23) ip_local_port_range: prefer different parity for start/end values. hangcheck_timer: disagrees about version of symbol module_layout tg3 0000:01:02.0 eth0: Link is up at 1000 Mbps, full duplex tg3 0000:01:02.0 eth0: Flow control is off for TX and off for RX NET: Registered protocol family 10 IPv6: ADDRCONF(NETDEV_UP): eth1: link is not ready tg3 0000:01:02.1 eth1: Link is up at 1000 Mbps, full duplex tg3 0000:01:02.1 eth1: Flow control is off for TX and off for RX IPv6: ADDRCONF(NETDEV_CHANGE): eth1: link becomes ready RPC: Registered named UNIX socket transport module. RPC: Registered udp transport module. RPC: Registered tcp transport module. RPC: Registered tcp NFSv4.1 backchannel transport module. Installing knfsd (copyright (C) 1996 okir@monad.swb.de). NFSD: Using /var/lib/nfs/v4recovery as the NFSv4 state recovery directory NFSD: starting 90-second grace period (net a0000001014af100) GSI 20 (level, low) -> CPU 2 (0x0400) vector 64 gnome-settings-(4170): floating-point assist fault at ip 2000000006b7bdc2, isr 0000020000001001 gnome-settings-(4170): floating-point assist fault at ip 2000000006b7bdc2, isr 0000020000001001 gnome-settings-(4170): floating-point assist fault at ip 2000000006b7bdc2, isr 0000020000001001 gnome-settings-(4170): floating-point assist fault at ip 2000000006b7bdc2, isr 0000020000001001 gnome-settings-(4170): floating-point assist fault at ip 2000000006b7bdc2, isr 0000020000001001 gdm-simple-gree(4179): unaligned access to 0x600fffffffadf34c, ip=0x20000000012216a0
On 2016/1/15 19:00, Hanjun Guo wrote:
On 2016/1/13 21:20, Tomasz Nowicki wrote:
From the functionality point of view this series might be split into the following logic parts:
- Make MMCONFIG code arch-agnostic which allows all architectures to
collect PCI config regions and used when necessary. 2. Move non-arch specific bits to the core code. 3. Use MMCONFIG code and implement generic ACPI based PCI host controller driver. 4. Enable above driver on ARM64
Patches has been built on top of 4.4 and can be found here: git@github.com:semihalf-nowicki-tomasz/linux.git (pci-acpi-v3)
NOTE, this patch set depends on Matthew's patches: http://www.spinics.net/lists/linux-pci/msg45950.html https://github.com/Vality/linux/tree/pci-fixes
This has been tested on Cavium ThunderX server and QEMU. Any help in reviewing and testing is very appreciated.
v2 -> v3
- fix legacy IRQ assigning and IO ports registration
- remove reference to arch specific companion device for ia64
- move ACPI PCI host controller driver to pci_root.c
- drop generic domain assignment for x86 and ia64 as I am not able to run all necessary test variants
- drop patch which cleaned legacy IRQ assignment since it belongs to Mathew's series: https://patchwork.ozlabs.org/patch/557504/
- extend MCFG quirk code
- rebased to 4.4
Tested this patch set (and Matthew's) on a HP RX2660 IA64 machine, with the CONFIG_ACPI_PCI_HOST_GENERIC disabled on IA64 ( it's a minor comment needs to be addressed in my previous email), it can boot the system ok with NIC working properly, with the /pro/ioports [1] and the boot log [2] if anyone needs refer to it.
For this patch set, on IA64 machine,
Tested-by: Hanjun Guo hanjun.guo@linaro.org
I tested this patch set on x86 server with CONFIG_ACPI_PCI_HOST_GENERIC disabled today, also booted ok with NIC working properly.
Thanks Hanjun
在 2016/1/13 21:20, Tomasz Nowicki 写道:
From the functionality point of view this series might be split into the
following logic parts:
- Make MMCONFIG code arch-agnostic which allows all architectures to collect PCI config regions and used when necessary.
- Move non-arch specific bits to the core code.
- Use MMCONFIG code and implement generic ACPI based PCI host controller driver.
- Enable above driver on ARM64
Patches has been built on top of 4.4 and can be found here: git@github.com:semihalf-nowicki-tomasz/linux.git (pci-acpi-v3)
NOTE, this patch set depends on Matthew's patches: http://www.spinics.net/lists/linux-pci/msg45950.html https://github.com/Vality/linux/tree/pci-fixes
This has been tested on Cavium ThunderX server and QEMU. Any help in reviewing and testing is very appreciated.
v2 -> v3
- fix legacy IRQ assigning and IO ports registration
- remove reference to arch specific companion device for ia64
- move ACPI PCI host controller driver to pci_root.c
- drop generic domain assignment for x86 and ia64 as I am not able to run all necessary test variants
- drop patch which cleaned legacy IRQ assignment since it belongs to Mathew's series: https://patchwork.ozlabs.org/patch/557504/
- extend MCFG quirk code
- rebased to 4.4
Based on the patchset and add the Hip05 ACPI specific quirks. Tested on the HiSilicon ARM64 D02 board. I made a test based on Intel 82599 networking card,it can work ok. this is the bootup log which contains PCIe host and Intel 82599 networking card part.
Tested-by: Dongdong Liu liudongdong3@huawei.com
EFI stub: Booting Linux Kernel... EFI stub: Using DTB from command line EFI stub: Exiting boot services and installing virtual address map... GMAC ExitBootServicesEvent SMMU ExitBootServicesEvent Booting Linux on physical CPU 0x20000 Initializing cgroup subsys cpu Linux version 4.4.0-rc2+ (l00290354@linux-ioko) (gcc version 4.9.3 20150211 (prerelease) (20150316) ) #194 SMP PREEMPT Thu Jan 28 09:51:15 CST 2016 Boot CPU: AArch64 Processor [411fd071] earlycon: Early serial console at MMIO32 0x80300000 (options '') bootconsole [uart0] enabled efi: Getting EFI parameters from FDT: EFI v2.50 by ARM Versatile Express EFI Jan 13 2016 20:43:10 efi: SMBIOS=0x7aae0000 SMBIOS 3.0=0x7aac0000 ACPI=0x7ab20000 ACPI 2.0=0x7ab20014 cma: Reserved 16 MiB at 0x000000007e800000 ACPI: Early table checksum verification disabled ACPI: RSDP 0x000000007AB20014 000024 (v02 HISI ) ACPI: XSDT 0x000000007AB100E8 00006C (v01 HISI HISI-D02 20140727 01000013) ACPI: FACP 0x000000007AA80000 00010C (v05 HISI HISI-D02 20140727 HISI 00000099) ACPI: Override [DSDT-HISI-D02], this is unsafe: tainting kernel Disabling lock debugging due to kernel taint ACPI: DSDT 0x000000007AA10000 Logical table override, new table: 0xFFFFFFC0009CB8B8 ACPI: DSDT 0xFFFFFFC0009CB8B8 0015B5 (v01 HISI HISI-D02 20140727 INTL 20150619) ACPI: DBG2 0x000000007AAA0000 00005A (v00 HISI HISI-D02 20140727 HISI 00000099) ACPI: GTDT 0x000000007AA70000 000060 (v02 HISI HISI-D02 20140727 HISI 00000099) ACPI: APIC 0x000000007AA60000 000564 (v01 HISI HISI-D02 20140727 HISI 00000099) ACPI: MCFG 0x000000007AA50000 00004C (v01 HISI HISI-D02 20140727 HISI 00000099) ACPI: SLIT 0x000000007AA40000 0001BC (v01 HISI HISI-D02 20140727 HISI 00000099) ACPI: SPCR 0x000000007AA30000 000050 (v02 HISI HISI-D02 20140727 HISI 00000099) ACPI: SRAT 0x000000007AA20000 0001B0 (v03 HISI HISI-D02 20140727 HISI 00000099) ACPI: IORT 0x000000007AA00000 0001FC (v00 INTEL TEMPLATE 00000000 INTL 20150619) psci: probing for conduit method from ACPI. NOTICE: [psci_smc_handler]:[349L] PSCI_VERSION CALL NOTICE: [psci_version]:[101L] PSCI_MAJOR_VER: 10000: PSCI_MINOR_VER: 0
0008;<54 psci: PSCIv1.0 detected in firmware. psci: Using standard PSCI v0.2 function IDs
0008;<54 psci: MIGRATE_INFO_TYPE not supported.
0008;<54
0008;<54 PERCPU: Embedded 15 pages/cpu @ffffffd1ffedb000 s24576 r8192 d28672 u61440 Detected PIPT I-cache on CPU0 CPU features: enabling workaround for ARM erratum 832075 Built 1 zonelists in Zone order, mobility grouping on. Total pages: 2063376 Kernel command line: dtb=hip05-d02.dtb console=ttyS0,115200 earlycon=uart8250,mmio32,0x80300000 initrd=filesystem.cpio.gz acpi=force log_buf_len individual max cpu contribution: 4096 bytes log_buf_len total cpu_extra contributions: 61440 bytes log_buf_len min size: 16384 bytes log_buf_len: 131072 bytes early log buf free: 12732(77%) PID hash table entries: 4096 (order: 3, 32768 bytes) Dentry cache hash table entries: 1048576 (order: 11, 8388608 bytes) Inode-cache hash table entries: 524288 (order: 10, 4194304 bytes) software IO TLB [mem 0x76a00000-0x7aa00000] (64MB) mapped at [ffffffc076a00000-ffffffc07a9fffff] Memory: 8052312K/8384512K available (6074K kernel code, 519K rwdata, 2556K rodata, 572K init, 212K bss, 315816K reserved, 16384K cma-reserved) Virtual kernel memory layout: vmalloc : 0xffffff8000000000 - 0xffffffbdbfff0000 ( 246 GB) vmemmap : 0xffffffbdc0000000 - 0xffffffbfc0000000 ( 8 GB maximum) 0xffffffbdc0000000 - 0xffffffbe08000000 ( 1152 MB actual) fixed : 0xffffffbffa7fd000 - 0xffffffbffac00000 ( 4108 KB) PCI I/O : 0xffffffbffae00000 - 0xffffffbffbe00000 ( 16 MB) modules : 0xffffffbffc000000 - 0xffffffc000000000 ( 64 MB) memory : 0xffffffc000000000 - 0xffffffd200000000 ( 73728 MB) .init : 0xffffffc0008f0000 - 0xffffffc00097f000 ( 572 KB) .text : 0xffffffc000080000 - 0xffffffc0008efa34 ( 8639 KB) .data : 0xffffffc000994000 - 0xffffffc000a15c00 ( 519 KB) SLUB: HWalign=64, Order=0-3, MinObjects=0, CPUs=16, Nodes=1 Preemptible hierarchical RCU implementation. Build-time adjustment of leaf fanout to 64. RCU restricting CPUs from NR_CPUS=64 to nr_cpu_ids=16. RCU: Adjusting geometry for rcu_fanout_leaf=64, nr_cpu_ids=16 NR_IRQS:64 nr_irqs:64 0 GIC: Using split EOI/Deactivate mode ITS@0x8c000000 ITS: allocated 65536 Devices @11f6c80000 (psz 4K, shr 1) ITS: allocated 512 Virtual CPUs @11f6c0d000 (psz 4K, shr 1) ITS: allocated 512 Interrupt Collections @11f6c0e000 (psz 4K, shr 1) ITS@0xc6000000 ITS: allocated 65536 Devices @11f6d00000 (psz 4K, shr 1) ITS: allocated 512 Virtual CPUs @11f6c0f000 (psz 4K, shr 1) ITS: allocated 512 Interrupt Collections @11f6c30000 (psz 4K, shr 1) ITS@0xa3000000 ITS: allocated 65536 Devices @11f6d80000 (psz 4K, shr 1) ITS: allocated 512 Virtual CPUs @11f6c33000 (psz 4K, shr 1) ITS: allocated 512 Interrupt Collections @11f6c34000 (psz 4K, shr 1) ITS@0xb7000000 ITS: allocated 65536 Devices @11f6e00000 (psz 4K, shr 1) ITS: allocated 512 Virtual CPUs @11f6c35000 (psz 4K, shr 1) ITS: allocated 512 Interrupt Collections @11f6c36000 (psz 4K, shr 1) GIC: using LPI property table @0x00000011f6c60000 ITS: Allocated 1792 chunks for LPIs CPU0: found redistributor 20000 region 0:0x000000008d100000 CPU0: using LPI pending table @0x00000011f6c70000 Architected cp15 timer(s) running at 50.00MHz (phys). clocksource: arch_sys_counter: mask: 0xffffffffffffff max_cycles: 0xb8812736b, max_idle_ns: 440795202655 ns sched_clock: 56 bits at 50MHz, resolution 20ns, wraps every 4398046511100ns Console: colour dummy device 80x25 Calibrating delay loop (skipped), value calculated using timer frequency.. 100.00 BogoMIPS (lpj=200000) pid_max: default: 32768 minimum: 301 ACPI: Core revision 20150930 ACPI: 1 ACPI AML tables successfully acquired and loaded Security Framework initialized Mount-cache hash table entries: 16384 (order: 5, 131072 bytes) Mountpoint-cache hash table entries: 16384 (order: 5, 131072 bytes) Initializing cgroup subsys memory Initializing cgroup subsys hugetlb Remapping and enabling EFI services. EFI remap 0x000000007aab0000 => 0000000040000000 EFI remap 0x000000007aaf0000 => 0000000040040000 EFI remap 0x000000007ab00000 => 0000000040050000 EFI remap 0x000000007ab30000 => 0000000040060000 EFI remap 0x000000007ab70000 => 00000000400a0000 EFI remap 0x000000007aba0000 => 00000000400c0000 EFI remap 0x000000007fb90000 => 00000000400d0000 EFI remap 0x000000007fbb0000 => 00000000400e0000 EFI remap 0x0000000080300000 => 0000000040110000 EFI remap 0x00000000a00f0000 => 0000000040120000 ASID allocator initialised with 65536 entries PCI/MSI: ITS@0x8c000000 domain created PCI/MSI: ITS@0xc6000000 domain created PCI/MSI: ITS@0xa3000000 domain created PCI/MSI: ITS@0xb7000000 domain created Platform MSI: irqchip@000000008c000000 domain created Platform MSI: irqchip@00000000c6000000 domain created Platform MSI: irqchip@00000000a3000000 domain created Platform MSI: irqchip@00000000b7000000 domain created NOTICE: [psci_smc_handler]:[410L] PSCI_CPU_ON_AARCH64 CALL NOTICE: [psci_smc_handler]:[411L] x1=0x20001 x2=0x827d0 x3=0x0 NOTICE: [scpi_set_css_power_state]:[89L] domain_cluster=0x1
NOTICE: [scpi_set_css_power_state]:[97L] domain_cluster=0x1
0008;<AB54 NOTICE: [psci_afflvl_power_on_finish]:[508L] NOTICE: [cm_prepare_el3_exit]:[266L] read_tpidr_el3 = 7fc3c080 NOTICE: [cm_prepare_el3_exit]:[321L] ctx add = 7fc3d190 NOTICE: [psci_afflvl_power_on_finish]:[566L]
000827=0 NOTICE: [psci_smc_handler]:[410L] PSCI_CPU_ON_AARCH64 CALL NOTICE: [psci_smc_handler]:[411L] x1=0x20002 x2=0x827d0 x3=0x0 NOTICE: [scpi_set_css_power_state]:[89L] domain_cluster=0x1
NOTICE: [scpi_set_css_power_state]:[97L] domain_cluster=0x1
0008;<AB54 NOTICE: [psci_afflvl_power_on_finish]:[508L] NOTICE: [cm_prepare_el3_exit]:[266L] read_tpidr_el3 = 7fc3c100 NOTICE: [cm_prepare_el3_exit]:[321L] ctx add = 7fc3d3a0 NOTICE: [psci_afflvl_power_on_finish]:[566L]
000827=0 NOTICE: [psci_smc_handler]:[410L] PSCI_CPU_ON_AARCH64 CALL NOTICE: [psci_smc_handler]:[411L] x1=0x20003 x2=0x827d0 x3=0x0 NOTICE: [scpi_set_css_power_state]:[89L] domain_cluster=0x1
NOTICE: [scpi_set_css_power_state]:[97L] domain_cluster=0x1
0008;<AB54 NOTICE: [psci_afflvl_power_on_finish]:[508L] NOTICE: [cm_prepare_el3_exit]:[266L] read_tpidr_el3 = 7fc3c180 NOTICE: [cm_prepare_el3_exit]:[321L] ctx add = 7fc3d5b0 NOTICE: [psci_afflvl_power_on_finish]:[566L]
000827=0 NOTICE: [psci_smc_handler]:[410L] PSCI_CPU_ON_AARCH64 CALL NOTICE: [psci_smc_handler]:[411L] x1=0x20100 x2=0x827d0 x3=0x0 NOTICE: [scpi_set_css_power_state]:[89L] domain_cluster=0x1
NOTICE: [scpi_set_css_power_state]:[97L] domain_cluster=0x3
0008;<AB54 NOTICE: [psci_afflvl_power_on_finish]:[508L] NOTICE: [cm_prepare_el3_exit]:[266L] read_tpidr_el3 = 7fc3c200 NOTICE: [cm_prepare_el3_exit]:[321L] ctx add = 7fc3d7c0 NOTICE: [psci_afflvl_power_on_finish]:[566L]
000827=0 NOTICE: [psci_smc_handler]:[410L] PSCI_CPU_ON_AARCH64 CALL NOTICE: [psci_smc_handler]:[411L] x1=0x20101 x2=0x827d0 x3=0x0 NOTICE: [scpi_set_css_power_state]:[89L] domain_cluster=0x3
NOTICE: [scpi_set_css_power_state]:[97L] domain_cluster=0x3
0008;<AB54 NOTICE: [psci_afflvl_power_on_finish]:[508L] NOTICE: [cm_prepare_el3_exit]:[266L] read_tpidr_el3 = 7fc3c280 NOTICE: [cm_prepare_el3_exit]:[321L] ctx add = 7fc3d9d0 NOTICE: [psci_afflvl_power_on_finish]:[566L]
000827=0 NOTICE: [psci_smc_handler]:[410L] PSCI_CPU_ON_AARCH64 CALL NOTICE: [psci_smc_handler]:[411L] x1=0x20102 x2=0x827d0 x3=0x0 NOTICE: [scpi_set_css_power_state]:[89L] domain_cluster=0x3
NOTICE: [scpi_set_css_power_state]:[97L] domain_cluster=0x3
0008;<AB54 NOTICE: [psci_afflvl_power_on_finish]:[508L] NOTICE: [cm_prepare_el3_exit]:[266L] read_tpidr_el3 = 7fc3c300 NOTICE: [cm_prepare_el3_exit]:[321L] ctx add = 7fc3dbe0 NOTICE: [psci_afflvl_power_on_finish]:[566L]
000827=0 NOTICE: [psci_smc_handler]:[410L] PSCI_CPU_ON_AARCH64 CALL NOTICE: [psci_smc_handler]:[411L] x1=0x20103 x2=0x827d0 x3=0x0 NOTICE: [scpi_set_css_power_state]:[89L] domain_cluster=0x3
NOTICE: [scpi_set_css_power_state]:[97L] domain_cluster=0x3
0008;<AB54 NOTICE: [psci_afflvl_power_on_finish]:[508L] NOTICE: [cm_prepare_el3_exit]:[266L] read_tpidr_el3 = 7fc3c380 NOTICE: [cm_prepare_el3_exit]:[321L] ctx add = 7fc3ddf0 NOTICE: [psci_afflvl_power_on_finish]:[566L]
000827=0 NOTICE: [psci_smc_handler]:[410L] PSCI_CPU_ON_AARCH64 CALL NOTICE: [psci_smc_handler]:[411L] x1=0x20200 x2=0x827d0 x3=0x0 NOTICE: [scpi_set_css_power_state]:[89L] domain_cluster=0x3
NOTICE: [scpi_set_css_power_state]:[97L] domain_cluster=0x7
0008;<AB54 NOTICE: [psci_afflvl_power_on_finish]:[508L] NOTICE: [cm_prepare_el3_exit]:[266L] read_tpidr_el3 = 7fc3c400 NOTICE: [cm_prepare_el3_exit]:[321L] ctx add = 7fc3e000 NOTICE: [psci_afflvl_power_on_finish]:[566L]
000827=0 NOTICE: [psci_smc_handler]:[410L] PSCI_CPU_ON_AARCH64 CALL NOTICE: [psci_smc_handler]:[411L] x1=0x20201 x2=0x827d0 x3=0x0 NOTICE: [scpi_set_css_power_state]:[89L] domain_cluster=0x7
NOTICE: [scpi_set_css_power_state]:[97L] domain_cluster=0x7
0008;<AB54 NOTICE: [psci_afflvl_power_on_finish]:[508L] NOTICE: [cm_prepare_el3_exit]:[266L] read_tpidr_el3 = 7fc3c480 NOTICE: [cm_prepare_el3_exit]:[321L] ctx add = 7fc3e210 NOTICE: [psci_afflvl_power_on_finish]:[566L]
000827=0 NOTICE: [psci_smc_handler]:[410L] PSCI_CPU_ON_AARCH64 CALL NOTICE: [psci_smc_handler]:[411L] x1=0x20202 x2=0x827d0 x3=0x0 NOTICE: [scpi_set_css_power_state]:[89L] domain_cluster=0x7
NOTICE: [scpi_set_css_power_state]:[97L] domain_cluster=0x7
0008;<AB54 NOTICE: [psci_afflvl_power_on_finish]:[508L] NOTICE: [cm_prepare_el3_exit]:[266L] read_tpidr_el3 = 7fc3c500 NOTICE: [cm_prepare_el3_exit]:[321L] ctx add = 7fc3e420 NOTICE: [psci_afflvl_power_on_finish]:[566L]
000827=0 NOTICE: [psci_smc_handler]:[410L] PSCI_CPU_ON_AARCH64 CALL NOTICE: [psci_smc_handler]:[411L] x1=0x20203 x2=0x827d0 x3=0x0 NOTICE: [scpi_set_css_power_state]:[89L] domain_cluster=0x7
NOTICE: [scpi_set_css_power_state]:[97L] domain_cluster=0x7
0008;<AB54 NOTICE: [psci_afflvl_power_on_finish]:[508L] NOTICE: [cm_prepare_el3_exit]:[266L] read_tpidr_el3 = 7fc3c580 NOTICE: [cm_prepare_el3_exit]:[321L] ctx add = 7fc3e630 NOTICE: [psci_afflvl_power_on_finish]:[566L]
000827=0 NOTICE: [psci_smc_handler]:[410L] PSCI_CPU_ON_AARCH64 CALL NOTICE: [psci_smc_handler]:[411L] x1=0x20300 x2=0x827d0 x3=0x0 NOTICE: [scpi_set_css_power_state]:[89L] domain_cluster=0x7
NOTICE: [scpi_set_css_power_state]:[97L] domain_cluster=0xf
0008;<AB54 NOTICE: [psci_afflvl_power_on_finish]:[508L] NOTICE: [cm_prepare_el3_exit]:[266L] read_tpidr_el3 = 7fc3c600 NOTICE: [cm_prepare_el3_exit]:[321L] ctx add = 7fc3e840 NOTICE: [psci_afflvl_power_on_finish]:[566L]
000827=0 NOTICE: [psci_smc_handler]:[410L] PSCI_CPU_ON_AARCH64 CALL NOTICE: [psci_smc_handler]:[411L] x1=0x20301 x2=0x827d0 x3=0x0 NOTICE: [scpi_set_css_power_state]:[89L] domain_cluster=0xf
NOTICE: [scpi_set_css_power_state]:[97L] domain_cluster=0xf
0008;<AB54 NOTICE: [psci_afflvl_power_on_finish]:[508L] NOTICE: [cm_prepare_el3_exit]:[266L] read_tpidr_el3 = 7fc3c680 NOTICE: [cm_prepare_el3_exit]:[321L] ctx add = 7fc3ea50 NOTICE: [psci_afflvl_power_on_finish]:[566L]
000827=0 NOTICE: [psci_smc_handler]:[410L] PSCI_CPU_ON_AARCH64 CALL NOTICE: [psci_smc_handler]:[411L] x1=0x20302 x2=0x827d0 x3=0x0 NOTICE: [scpi_set_css_power_state]:[89L] domain_cluster=0xf
NOTICE: [scpi_set_css_power_state]:[97L] domain_cluster=0xf
0008;<AB54 NOTICE: [psci_afflvl_power_on_finish]:[508L] NOTICE: [cm_prepare_el3_exit]:[266L] read_tpidr_el3 = 7fc3c700 NOTICE: [cm_prepare_el3_exit]:[321L] ctx add = 7fc3ec60 NOTICE: [psci_afflvl_power_on_finish]:[566L]
000827=0 NOTICE: [psci_smc_handler]:[410L] PSCI_CPU_ON_AARCH64 CALL NOTICE: [psci_smc_handler]:[411L] x1=0x20303 x2=0x827d0 x3=0x0 NOTICE: [scpi_set_css_power_state]:[89L] domain_cluster=0xf
NOTICE: [scpi_set_css_power_state]:[97L] domain_cluster=0xf
0008;<AB54 NOTICE: [psci_afflvl_power_on_finish]:[508L] NOTICE: [cm_prepare_el3_exit]:[266L] read_tpidr_el3 = 7fc3c780 NOTICE: [cm_prepare_el3_exit]:[321L] ctx add = 7fc3ee70 NOTICE: [psci_afflvl_power_on_finish]:[566L]
000827=0 Detected PIPT I-cache on CPU1 CPU1: found redistributor 20001 region 0:0x000000008d130000 CPU1: using LPI pending table @0x00000011f6400000 CPU1: Booted secondary processor [411fd071] Detected PIPT I-cache on CPU2 CPU2: found redistributor 20002 region 0:0x000000008d160000 CPU2: using LPI pending table @0x00000011f6430000 CPU2: Booted secondary processor [411fd071] Detected PIPT I-cache on CPU3 CPU3: found redistributor 20003 region 0:0x000000008d190000 CPU3: using LPI pending table @0x00000011f6460000 CPU3: Booted secondary processor [411fd071] Detected PIPT I-cache on CPU4 CPU4: found redistributor 20100 region 0:0x000000008d1c0000 CPU4: using LPI pending table @0x00000011f64a0000 CPU4: Booted secondary processor [411fd071] Detected PIPT I-cache on CPU5 CPU5: found redistributor 20101 region 0:0x000000008d1f0000 CPU5: using LPI pending table @0x00000011f64d0000 CPU5: Booted secondary processor [411fd071] Detected PIPT I-cache on CPU6 CPU6: found redistributor 20102 region 0:0x000000008d220000 CPU6: using LPI pending table @0x00000011f6500000 CPU6: Booted secondary processor [411fd071] Detected PIPT I-cache on CPU7 CPU7: found redistributor 20103 region 0:0x000000008d250000 CPU7: using LPI pending table @0x00000011f6540000 CPU7: Booted secondary processor [411fd071] Detected PIPT I-cache on CPU8 CPU8: found redistributor 20200 region 0:0x000000008d280000 CPU8: using LPI pending table @0x00000011f6560000 CPU8: Booted secondary processor [411fd071] Detected PIPT I-cache on CPU9 CPU9: found redistributor 20201 region 0:0x000000008d2b0000 CPU9: using LPI pending table @0x00000011f65a0000 CPU9: Booted secondary processor [411fd071] Detected PIPT I-cache on CPU10 CPU10: found redistributor 20202 region 0:0x000000008d2e0000 CPU10: using LPI pending table @0x00000011f65d0000 CPU10: Booted secondary processor [411fd071] Detected PIPT I-cache on CPU11 CPU11: found redistributor 20203 region 0:0x000000008d310000 CPU11: using LPI pending table @0x00000011f6600000 CPU11: Booted secondary processor [411fd071] Detected PIPT I-cache on CPU12 CPU12: found redistributor 20300 region 0:0x000000008d340000 CPU12: using LPI pending table @0x00000011f6630000 CPU12: Booted secondary processor [411fd071] Detected PIPT I-cache on CPU13 CPU13: found redistributor 20301 region 0:0x000000008d370000 CPU13: using LPI pending table @0x00000011f6670000 CPU13: Booted secondary processor [411fd071] Detected PIPT I-cache on CPU14 CPU14: found redistributor 20302 region 0:0x000000008d3a0000 CPU14: using LPI pending table @0x00000011f66a0000 CPU14: Booted secondary processor [411fd071] Detected PIPT I-cache on CPU15 CPU15: found redistributor 20303 region 0:0x000000008d3d0000 CPU15: using LPI pending table @0x00000011f66d0000 CPU15: Booted secondary processor [411fd071] Brought up 16 CPUs SMP: Total of 16 processors activated. CPU features: detected feature: GIC system register CPU interface CPU: All CPU(s) started at EL2 alternatives: patching kernel code devtmpfs: initialized SMBIOS 3.0.0 present. clocksource: jiffies: mask: 0xffffffff max_cycles: 0xffffffff, max_idle_ns: 7645041785100000 ns pinctrl core: initialized pinctrl subsystem NET: Registered protocol family 16 cpuidle: using governor ladder cpuidle: using governor menu vdso: 2 pages (1 code @ ffffffc000999000, 1 data @ ffffffc000998000) hw-breakpoint: found 6 breakpoint and 4 watchpoint registers. DMA: preallocated 256 KiB pool for atomic allocations ACPI: bus type PCI registered Serial: AMBA PL011 UART driver ACPI: Added _OSI(Module Device) ACPI: Added _OSI(Processor Device) ACPI: Added _OSI(3.0 _SCP Extensions) ACPI: Added _OSI(Processor Aggregator Device) ACPI: Interpreter enabled ACPI: Using GIC for interrupt routing PCI: MMCONFIG for domain 0001 [bus 00-3f] at [mem 0x22000000000-0x22003ffffff] (base 0x22000000000) PCI: MMCONFIG for domain 0002 [bus 40-7f] at [mem 0x24004000000-0x24007ffffff] (base 0x24000000000) ACPI: IORT: can't find node related to (null) device ACPI: IORT: can't find node related to (null) device ACPI: IORT: can't find node related to (null) device ACPI: IORT: can't find node related to (null) device ACPI: IORT: can't find node related to (null) device ACPI: IORT: can't find node related to (null) device ACPI: IORT: can't find node related to (null) device ACPI: IORT: can't find node related to (null) device ACPI: IORT: can't find node related to (null) device ACPI: IORT: can't find node related to (null) device ACPI: IORT: can't find node related to (null) device ACPI: IORT: can't find node related to (null) device ACPI: IORT: can't find node related to (null) device ACPI: PCI Root Bridge [PCI1] (domain 0001 [bus 00-3f]) acpi HISI0080:00: _OSC: OS supports [ExtendedConfig ASPM ClockPM Segments MSI] acpi HISI0080:00: _OSC failed (AE_NOT_FOUND); disabling ASPM acpi HISI0080:00: PCI host bridge to bus 0001:00 pci_bus 0001:00: root bus resource [mem 0x22004000000-0x2200ffeffff window] (bus address [0xb0000000-0xbbfeffff]) pci_bus 0001:00: root bus resource [io 0x0000-0xffff window] pci_bus 0001:00: root bus resource [bus 00-3f] pci 0001:00:00.0: of_irq_parse_pci() failed with rc=-22 pci 0001:00:00.0: of_irq_parse_pci() failed with rc=-22 pci 0001:00:00.0: BAR 8: assigned [mem 0x22004000000-0x220041fffff] pci 0001:00:00.0: BAR 7: assigned [io 0x1000-0x1fff] pci 0001:01:00.0: BAR 6: assigned [mem 0x22004000000-0x220040fffff pref] pci 0001:01:00.0: BAR 1: assigned [mem 0x22004100000-0x2200410ffff 64bit] pci 0001:01:00.0: BAR 0: assigned [io 0x1000-0x10ff] pci 0001:00:00.0: PCI bridge to [bus 01] pci 0001:00:00.0: bridge window [io 0x1000-0x1fff] pci 0001:00:00.0: bridge window [mem 0x22004000000-0x220041fffff] ACPI: PCI Root Bridge [PCI2] (domain 0002 [bus 40-7f]) acpi HISI0080:01: _OSC: OS supports [ExtendedConfig ASPM ClockPM Segments MSI] acpi HISI0080:01: _OSC failed (AE_NOT_FOUND); disabling ASPM acpi HISI0080:01: PCI host bridge to bus 0002:40 pci_bus 0002:40: root bus resource [mem 0x24008000000-0x2400ffeffff window] (bus address [0xc0000000-0xc7feffff]) pci_bus 0002:40: root bus resource [io 0x10000-0x1ffff window] (bus address [0x0000-0xffff]) pci_bus 0002:40: root bus resource [bus 40-7f] pci 0002:40:00.0: of_irq_parse_pci() failed with rc=-22 pci 0002:40:00.0: Primary bus is hard wired to 0 pci 0002:40:00.0: bridge configuration invalid ([bus 01-01]), reconfiguring pci 0002:40:00.0: of_irq_parse_pci() failed with rc=-22 pci 0002:40:00.0: of_irq_parse_pci() failed with rc=-22 pci 0002:40:00.0: BAR 8: assigned [mem 0x24008000000-0x240087fffff] pci 0002:40:00.0: BAR 9: assigned [mem 0x24008800000-0x240091fffff 64bit pref] pci 0002:40:00.0: BAR 7: assigned [io 0x10000-0x10fff] pci 0002:41:00.0: BAR 0: assigned [mem 0x24008800000-0x24008bfffff 64bit pref] pci 0002:41:00.0: BAR 6: assigned [mem 0x24008000000-0x240083fffff pref] pci 0002:41:00.1: BAR 0: assigned [mem 0x24008c00000-0x24008ffffff 64bit pref] pci 0002:41:00.1: BAR 6: assigned [mem 0x24008400000-0x240087fffff pref] pci 0002:41:00.0: BAR 4: assigned [mem 0x24009000000-0x24009003fff 64bit pref] pci 0002:41:00.1: BAR 4: assigned [mem 0x24009004000-0x24009007fff 64bit pref] pci 0002:41:00.0: BAR 2: assigned [io 0x10000-0x1001f] pci 0002:41:00.1: BAR 2: assigned [io 0x10020-0x1003f] pci 0002:40:00.0: PCI bridge to [bus 41] pci 0002:40:00.0: bridge window [io 0x10000-0x10fff] pci 0002:40:00.0: bridge window [mem 0x24008000000-0x240087fffff] pci 0002:40:00.0: bridge window [mem 0x24008800000-0x240091fffff 64bit pref] vgaarb: loaded SCSI subsystem initialized ACPI: bus type USB registered usbcore: registered new interface driver usbfs usbcore: registered new interface driver hub usbcore: registered new device driver usb pps_core: LinuxPPS API ver. 1 registered pps_core: Software ver. 5.3.6 - Copyright 2005-2007 Rodolfo Giometti giometti@linux.it PTP clock support registered clocksource: Switched to clocksource arch_sys_counter pnp: PnP ACPI init pnp: PnP ACPI: found 0 devices NET: Registered protocol family 2 TCP established hash table entries: 65536 (order: 7, 524288 bytes) TCP bind hash table entries: 65536 (order: 8, 1048576 bytes) TCP: Hash tables configured (established 65536 bind 65536) UDP hash table entries: 4096 (order: 5, 131072 bytes) UDP-Lite hash table entries: 4096 (order: 5, 131072 bytes) NET: Registered protocol family 1 RPC: Registered named UNIX socket transport module. RPC: Registered udp transport module. RPC: Registered tcp transport module. RPC: Registered tcp NFSv4.1 backchannel transport module. Unpacking initramfs... Freeing initrd memory: 27492K (ffffffc01e520000 - ffffffc01fff9000) kvm [1]: error: no compatible GIC node found kvm [1]: error initializing Hyp mode: -19 ACPI: IORT: can't find node related to (null) device futex hash table entries: 4096 (order: 7, 524288 bytes) audit: initializing netlink subsys (disabled) audit: type=2000 audit(1.720:1): initialized HugeTLB registered 2 MB page size, pre-allocated 0 pages VFS: Disk quotas dquot_6.6.0 VFS: Dquot-cache hash table entries: 512 (order 0, 4096 bytes) NFS: Registering the id_resolver key type Key type id_resolver registered Key type id_legacy registered fuse init (API version 7.23) 9p: Installing v9fs 9p2000 file system support Block layer SCSI generic (bsg) driver version 0.4 loaded (major 251) io scheduler noop registered io scheduler cfq registered (default) pcieport 0001:00:00.0: can't derive routing for PCI INT A pcieport 0001:00:00.0: PCI INT A: no GSI pcieport 0002:40:00.0: can't derive routing for PCI INT A pcieport 0002:40:00.0: PCI INT A: no GSI Serial: 8250/16550 driver, 4 ports, IRQ sharing disabled console [ttyS0] disabled APMC0D08:00: ttyS0 at MMIO 0x80300000 (irq = 5, base_baud = 12500000) is a 16550A console [ttyS0] enabled console [ttyS0] enabled bootconsole [uart0] disabled bootconsole [uart0] disabled msm_serial: driver initialized Failed to find cpu0 device node Unable to detect cache hierarchy from DT for CPU 0 loop: module loaded hisi_sas: driver version v1.0 scsi host0: hisi_sas scsi host1: hisi_sas tun: Universal TUN/TAP device driver, 1.6 tun: (C) 1999-2004 Max Krasnyansky maxk@qualcomm.com libphy: Hisilicon MII Bus: probed hisi_sas_v1_hw HISI0161:01: phyup: phy7 link_rate=10 scsi 1:0:0:0: Direct-Access SEAGATE ST9500620SS AS02 PQ: 0 ANSI: 5 sd 1:0:0:0: [sda] 976773168 512-byte logical blocks: (500 GB/465 GiB) sd 1:0:0:0: [sda] Write Protect is off sd 1:0:0:0: [sda] Write cache: disabled, read cache: enabled, supports DPO and FUA sd 1:0:0:0: [sda] Attached SCSI disk hns-nic HISI00C1:00: No valid mac, use random mac da:3a:ad:09:a3:14 netif_napi_add() called with weight 256 on device eth%d hns-nic HISI00C1:01: No valid mac, use random mac ea:58:90:0b:08:bc hns-nic HISI00C1:02: No valid mac, use random mac 2a:ef:14:d5:0f:41 hns-nic HISI00C1:03: No valid mac, use random mac 7a:77:63:4e:57:89 ixgbe: Intel(R) 10 Gigabit PCI Express Network Driver - version 4.2.1-k ixgbe: Copyright (c) 1999-2015 Intel Corporation. pcieport 0002:40:00.0: can't derive routing for PCI INT A ixgbe 0002:41:00.0: PCI INT A: no GSI ixgbe 0002:41:00.0: enabling device (0000 -> 0002) ixgbe 0002:41:00.0: Multiqueue Enabled: Rx Queue count = 16, Tx Queue count = 16 ixgbe 0002:41:00.0: PCI Express bandwidth of 32GT/s available ixgbe 0002:41:00.0: (Speed:5.0GT/s, Width: x8, Encoding Loss:20%) ixgbe 0002:41:00.0: MAC: 2, PHY: 17, SFP+: 5, PBA No: FFFFFF-0FF ixgbe 0002:41:00.0: 68:a8:28:2e:c9:10 ixgbe 0002:41:00.0: Intel(R) 10 Gigabit Network Connection pcieport 0002:40:00.0: can't derive routing for PCI INT B ixgbe 0002:41:00.1: PCI INT B: no GSI ixgbe 0002:41:00.1: enabling device (0000 -> 0002) ixgbe 0002:41:00.1: Multiqueue Enabled: Rx Queue count = 16, Tx Queue count = 16 ixgbe 0002:41:00.1: PCI Express bandwidth of 32GT/s available ixgbe 0002:41:00.1: (Speed:5.0GT/s, Width: x8, Encoding Loss:20%) ixgbe 0002:41:00.1: MAC: 2, PHY: 1, PBA No: FFFFFF-0FF ixgbe 0002:41:00.1: 68:a8:28:2e:c9:11 ixgbe 0002:41:00.1: Intel(R) 10 Gigabit Network Connection sky2: driver version 1.30 ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver ehci-pci: EHCI PCI platform driver ehci-platform: EHCI generic platform driver ehci-platform PNP0D20:00: EHCI Host Controller ehci-platform PNP0D20:00: new USB bus registered, assigned bus number 1 ehci-platform PNP0D20:00: irq 6, io mem 0xa1000000 ehci-platform PNP0D20:00: USB 2.0 started, EHCI 1.00 hub 1-0:1.0: USB hub found hub 1-0:1.0: 1 port detected ohci_hcd: USB 1.1 'Open' Host Controller (OHCI) Driver ohci-pci: OHCI PCI platform driver ohci-platform: OHCI generic platform driver usbcore: registered new interface driver usb-storage mousedev: PS/2 mouse device common for all mice rtc-efi rtc-efi: rtc core: registered rtc-efi as rtc0 sdhci: Secure Digital Host Controller Interface driver sdhci: Copyright(c) Pierre Ossman Synopsys Designware Multimedia Card Interface Driver sdhci-pltfm: SDHCI platform and OF driver helper ledtrig-cpu: registered to indicate activity on CPUs usbcore: registered new interface driver usbhid usbhid: USB HID core driver NET: Registered protocol family 17 9pnet: Installing 9P2000 support Key type dns_resolver registered registered taskstats version 1 rtc-efi rtc-efi: setting system clock to 2100-01-01 00:02:52 UTC (4102444972) ttyS0 - failed to request DMA Freeing unused kernel memory: 572K (ffffffc0008f0000 - ffffffc00097f000) Freeing alternatives memory: 60K (ffffffc00097f000 - ffffffc00098e000) root@(none)$ ifconfig eth4 192.168.20.188 ixgbe 0002:41:00.0: registered PHC device on eth4 root@(none)$ ixgbe 0002:41:00.0 eth4: detected SFP+: 5 ixgbe 0002:41:00.0 eth4: NIC Link is Up 10 Gbps, Flow Control: RX/TX
root@(none)$ ping 192.168.20.4 PING 192.168.20.4 (192.168.20.4): 56 data bytes 64 bytes from 192.168.20.4: seq=21 ttl=128 time=0.368 ms 64 bytes from 192.168.20.4: seq=22 ttl=128 time=0.257 ms 64 bytes from 192.168.20.4: seq=23 ttl=128 time=0.316 ms 64 bytes from 192.168.20.4: seq=24 ttl=128 time=0.355 ms 64 bytes from 192.168.20.4: seq=25 ttl=128 time=0.257 ms 64 bytes from 192.168.20.4: seq=26 ttl=128 time=0.305 ms 64 bytes from 192.168.20.4: seq=27 ttl=128 time=0.290 ms 64 bytes from 192.168.20.4: seq=28 ttl=128 time=0.227 ms 64 bytes from 192.168.20.4: seq=29 ttl=128 time=3.865 ms 64 bytes from 192.168.20.4: seq=30 ttl=128 time=0.297 ms
On Wed, Jan 13, 2016 at 5:20 AM, Tomasz Nowicki tn@semihalf.com wrote:
From the functionality point of view this series might be split into the following logic parts:
- Make MMCONFIG code arch-agnostic which allows all architectures to collect PCI config regions and used when necessary.
- Move non-arch specific bits to the core code.
- Use MMCONFIG code and implement generic ACPI based PCI host controller driver.
- Enable above driver on ARM64
Patches has been built on top of 4.4 and can be found here: git@github.com:semihalf-nowicki-tomasz/linux.git (pci-acpi-v3)
NOTE, this patch set depends on Matthew's patches: http://www.spinics.net/lists/linux-pci/msg45950.html https://github.com/Vality/linux/tree/pci-fixes
This has been tested on Cavium ThunderX server and QEMU. Any help in reviewing and testing is very appreciated.
Hi Tomasz,
I tested your tree with APM X-Gene v1 and X-Gene v2 platforms and PCIe came up fine on both platforms (with X-Gene ECAM fixup patch that I will post separately). Thanks for your work and please feel free to add my Tested-by.
Regards, Duc Dang.
v2 -> v3
- fix legacy IRQ assigning and IO ports registration
- remove reference to arch specific companion device for ia64
- move ACPI PCI host controller driver to pci_root.c
- drop generic domain assignment for x86 and ia64 as I am not able to run all necessary test variants
- drop patch which cleaned legacy IRQ assignment since it belongs to Mathew's series: https://patchwork.ozlabs.org/patch/557504/
- extend MCFG quirk code
- rebased to 4.4
v1 -> v2
- moved non-arch specific piece of code to dirver/acpi/ directory
- fixed IO resource handling
- introduced PCI config accessors quirks matching
- moved ACPI_COMPANION_SET to generic code
Liu Jiang (1): ACPI, PCI: Refine the way to handle translation_offset for ACPI resources
Tomasz Nowicki (20): x86, pci: Reorder logic of pci_mmconfig_insert() function x86, pci, acpi: Move arch-agnostic MMCONFIG (aka ECAM) and ACPI code out of arch/x86/ directory pci, acpi, mcfg: Provide generic implementation of MCFG code initialization. x86, pci: mmconfig_{32,64}.c code refactoring - remove code duplication. x86, pci, ecam: mmconfig_64.c becomes default implementation for ECAM driver. XEN / PCI: Remove the dependence on arch x86 when PCI_MMCONFIG=y pci, acpi, mcfg: Provide default RAW ACPI PCI config space accessors. arm64, acpi: Use empty PCI config space accessors from mcfg.c file. pci, acpi, ecam: Add flag to indicate whether ECAM region was hot added or not. x86, pci: Cleanup platform specific MCFG data using previously added ECAM hot_added flag. pci, acpi: Move ACPI host bridge device companion assignment to core code. x86, ia64, pci: Remove ACPI companion device from platform specific data. pci, acpi: Provide generic way to assign bus domain number. x86, ia64: Include acpi_pci_{add|remove}_bus to the default pcibios_{add|remove}_bus implementation. acpi, mcfg: Implement two calls that might be used to inject/remove MCFG region. x86, acpi, pci: Use equivalent function introduced in previous patch. acpi, mcfg: Add default PCI config accessors implementation and initial support for related quirks. pci, acpi: Support for ACPI based generic PCI host controller init pci, acpi: Match PCI config space accessors against platfrom specific quirks. arm64, pci, acpi: Start using ACPI based PCI host bridge driver for ARM64.
arch/arm64/Kconfig | 7 ++ arch/arm64/kernel/pci.c | 25 ---- arch/ia64/hp/common/sba_iommu.c | 2 +- arch/ia64/include/asm/pci.h | 1 - arch/ia64/pci/pci.c | 52 +++----- arch/ia64/sn/kernel/io_acpi_init.c | 4 +- arch/x86/Kconfig | 4 + arch/x86/include/asm/pci.h | 3 - arch/x86/include/asm/pci_x86.h | 28 +---- arch/x86/pci/acpi.c | 41 ++---- arch/x86/pci/common.c | 10 -- arch/x86/pci/irq.c | 10 -- arch/x86/pci/mmconfig-shared.c | 250 ++++++------------------------------- arch/x86/pci/mmconfig_32.c | 11 +- arch/x86/pci/mmconfig_64.c | 67 +--------- arch/x86/pci/numachip.c | 1 + drivers/acpi/Kconfig | 5 + drivers/acpi/Makefile | 1 + drivers/acpi/mcfg.c | 201 +++++++++++++++++++++++++++++ drivers/acpi/pci_root.c | 136 +++++++++++++++++++- drivers/acpi/resource.c | 12 +- drivers/pci/Kconfig | 10 ++ drivers/pci/Makefile | 5 + drivers/pci/ecam.c | 234 ++++++++++++++++++++++++++++++++++ drivers/pci/pci.c | 29 ++++- drivers/pci/probe.c | 5 + drivers/xen/pci.c | 7 +- include/acpi/acpi_bus.h | 1 + include/asm-generic/vmlinux.lds.h | 7 ++ include/linux/acpi.h | 2 + include/linux/ecam.h | 62 +++++++++ include/linux/pci-acpi.h | 17 +++ 32 files changed, 802 insertions(+), 448 deletions(-) create mode 100644 drivers/acpi/mcfg.c create mode 100644 drivers/pci/ecam.c create mode 100644 include/linux/ecam.h
-- 1.9.1
-- To unsubscribe from this list: send the line "unsubscribe linux-acpi" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html