From 99935a7a59eaca0292c1a5880e10bae03f4a5e3d Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Sun, 4 Oct 2009 21:54:24 -0700 Subject: [PATCH 001/109] x86/PCI: read root resources from IOH on Intel For intel systems with multi IOH, we should read peer root resources directly from PCI config space, and don't trust _CRS. Signed-off-by: Yinghai Lu Signed-off-by: Jesse Barnes --- arch/x86/pci/Makefile | 1 + arch/x86/pci/amd_bus.c | 45 ++++++++------------ arch/x86/pci/bus_numa.h | 26 ++++++++++++ arch/x86/pci/intel_bus.c | 90 ++++++++++++++++++++++++++++++++++++++++ 4 files changed, 135 insertions(+), 27 deletions(-) create mode 100644 arch/x86/pci/bus_numa.h create mode 100644 arch/x86/pci/intel_bus.c diff --git a/arch/x86/pci/Makefile b/arch/x86/pci/Makefile index d49202e740ea..56d917b556c6 100644 --- a/arch/x86/pci/Makefile +++ b/arch/x86/pci/Makefile @@ -15,3 +15,4 @@ obj-$(CONFIG_X86_NUMAQ) += numaq_32.o obj-y += common.o early.o obj-y += amd_bus.o +obj-$(CONFIG_X86_64) += intel_bus.o diff --git a/arch/x86/pci/amd_bus.c b/arch/x86/pci/amd_bus.c index 572ee9782f2a..995f36096a42 100644 --- a/arch/x86/pci/amd_bus.c +++ b/arch/x86/pci/amd_bus.c @@ -10,6 +10,8 @@ #include #endif +#include "bus_numa.h" + /* * This discovers the pcibus <-> node mapping on AMD K8. * also get peer root bus resource for io,mmio @@ -17,25 +19,9 @@ #ifdef CONFIG_X86_64 -/* - * sub bus (transparent) will use entres from 3 to store extra from root, - * so need to make sure have enought slot there, increase PCI_BUS_NUM_RESOURCES? - */ -#define RES_NUM 16 -struct pci_root_info { - char name[12]; - unsigned int res_num; - struct resource res[RES_NUM]; - int bus_min; - int bus_max; - int node; - int link; -}; - -/* 4 at this time, it may become to 32 */ -#define PCI_ROOT_NR 4 -static int pci_root_num; -static struct pci_root_info pci_root_info[PCI_ROOT_NR]; +int pci_root_num; +struct pci_root_info pci_root_info[PCI_ROOT_NR]; +static int found_all_numa_early; void x86_pci_root_bus_res_quirks(struct pci_bus *b) { @@ -48,8 +34,11 @@ void x86_pci_root_bus_res_quirks(struct pci_bus *b) b->resource[1] != &iomem_resource) return; - /* if only one root bus, don't need to anything */ - if (pci_root_num < 2) + if (!pci_root_num) + return; + + /* for amd, if only one root bus, don't need to do anything */ + if (pci_root_num < 2 && found_all_numa_early) return; for (i = 0; i < pci_root_num; i++) { @@ -130,12 +119,15 @@ static void __init update_range(struct res_range *range, size_t start, } } -static void __init update_res(struct pci_root_info *info, size_t start, +void __init update_res(struct pci_root_info *info, size_t start, size_t end, unsigned long flags, int merge) { int i; struct resource *res; + if (start > end) + return; + if (!merge) goto addit; @@ -230,7 +222,6 @@ static int __init early_fill_mp_bus_info(void) int j; unsigned bus; unsigned slot; - int found; int node; int link; int def_node; @@ -247,7 +238,7 @@ static int __init early_fill_mp_bus_info(void) if (!early_pci_allowed()) return -1; - found = 0; + found_all_numa_early = 0; for (i = 0; i < ARRAY_SIZE(pci_probes); i++) { u32 id; u16 device; @@ -261,12 +252,12 @@ static int __init early_fill_mp_bus_info(void) device = (id>>16) & 0xffff; if (pci_probes[i].vendor == vendor && pci_probes[i].device == device) { - found = 1; + found_all_numa_early = 1; break; } } - if (!found) + if (!found_all_numa_early) return 0; pci_root_num = 0; @@ -488,7 +479,7 @@ static int __init early_fill_mp_bus_info(void) info = &pci_root_info[i]; res_num = info->res_num; busnum = info->bus_min; - printk(KERN_DEBUG "bus: [%02x,%02x] on node %x link %x\n", + printk(KERN_DEBUG "bus: [%02x, %02x] on node %x link %x\n", info->bus_min, info->bus_max, info->node, info->link); for (j = 0; j < res_num; j++) { res = &info->res[j]; diff --git a/arch/x86/pci/bus_numa.h b/arch/x86/pci/bus_numa.h new file mode 100644 index 000000000000..4ff126a3e887 --- /dev/null +++ b/arch/x86/pci/bus_numa.h @@ -0,0 +1,26 @@ +#ifdef CONFIG_X86_64 + +/* + * sub bus (transparent) will use entres from 3 to store extra from + * root, so need to make sure we have enought slot there, Should we + * increase PCI_BUS_NUM_RESOURCES? + */ +#define RES_NUM 16 +struct pci_root_info { + char name[12]; + unsigned int res_num; + struct resource res[RES_NUM]; + int bus_min; + int bus_max; + int node; + int link; +}; + +/* 4 at this time, it may become to 32 */ +#define PCI_ROOT_NR 4 +extern int pci_root_num; +extern struct pci_root_info pci_root_info[PCI_ROOT_NR]; + +extern void update_res(struct pci_root_info *info, size_t start, + size_t end, unsigned long flags, int merge); +#endif diff --git a/arch/x86/pci/intel_bus.c b/arch/x86/pci/intel_bus.c new file mode 100644 index 000000000000..b7a55dc55d13 --- /dev/null +++ b/arch/x86/pci/intel_bus.c @@ -0,0 +1,90 @@ +/* + * to read io range from IOH pci conf, need to do it after mmconfig is there + */ + +#include +#include +#include +#include +#include + +#include "bus_numa.h" + +static inline void print_ioh_resources(struct pci_root_info *info) +{ + int res_num; + int busnum; + int i; + + printk(KERN_DEBUG "IOH bus: [%02x, %02x]\n", + info->bus_min, info->bus_max); + res_num = info->res_num; + busnum = info->bus_min; + for (i = 0; i < res_num; i++) { + struct resource *res; + + res = &info->res[i]; + printk(KERN_DEBUG "IOH bus: %02x index %x %s: [%llx, %llx]\n", + busnum, i, + (res->flags & IORESOURCE_IO) ? "io port" : + "mmio", + res->start, res->end); + } +} + +#define IOH_LIO 0x108 +#define IOH_LMMIOL 0x10c +#define IOH_LMMIOH 0x110 +#define IOH_LMMIOH_BASEU 0x114 +#define IOH_LMMIOH_LIMITU 0x118 +#define IOH_LCFGBUS 0x11c + +static void __devinit pci_root_bus_res(struct pci_dev *dev) +{ + u16 word; + u32 dword; + struct pci_root_info *info; + u16 io_base, io_end; + u32 mmiol_base, mmiol_end; + u64 mmioh_base, mmioh_end; + int bus_base, bus_end; + + if (pci_root_num >= PCI_ROOT_NR) { + printk(KERN_DEBUG "intel_bus.c: PCI_ROOT_NR is too small\n"); + return; + } + + info = &pci_root_info[pci_root_num]; + pci_root_num++; + + pci_read_config_word(dev, IOH_LCFGBUS, &word); + bus_base = (word & 0xff); + bus_end = (word & 0xff00) >> 8; + sprintf(info->name, "PCI Bus #%02x", bus_base); + info->bus_min = bus_base; + info->bus_max = bus_end; + + pci_read_config_word(dev, IOH_LIO, &word); + io_base = (word & 0xf0) << (12 - 4); + io_end = (word & 0xf000) | 0xfff; + update_res(info, io_base, io_end, IORESOURCE_IO, 0); + + pci_read_config_dword(dev, IOH_LMMIOL, &dword); + mmiol_base = (dword & 0xff00) << (24 - 8); + mmiol_end = (dword & 0xff000000) | 0xffffff; + update_res(info, mmiol_base, mmiol_end, IORESOURCE_MEM, 0); + + pci_read_config_dword(dev, IOH_LMMIOH, &dword); + mmioh_base = ((u64)(dword & 0xfc00)) << (26 - 10); + mmioh_end = ((u64)(dword & 0xfc000000) | 0x3ffffff); + pci_read_config_dword(dev, IOH_LMMIOH_BASEU, &dword); + mmioh_base |= ((u64)(dword & 0x7ffff)) << 32; + pci_read_config_dword(dev, IOH_LMMIOH_LIMITU, &dword); + mmioh_end |= ((u64)(dword & 0x7ffff)) << 32; + update_res(info, mmioh_base, mmioh_end, IORESOURCE_MEM, 0); + + print_ioh_resources(info); +} + +/* intel IOH */ +DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x342e, pci_root_bus_res); From ac1aa47b131416a6ff37eb1005a0a1d2541aad6c Mon Sep 17 00:00:00 2001 From: Jesse Barnes Date: Mon, 26 Oct 2009 13:20:44 -0700 Subject: [PATCH 002/109] PCI: determine CLS more intelligently Till now, CLS has been determined either by arch code or as L1_CACHE_BYTES. Only x86 and ia64 set CLS explicitly and x86 doesn't always get it right. On most configurations, the chance is that firmware configures the correct value during boot. This patch makes pci_init() determine CLS by looking at what firmware has configured. It scans all devices and if all non-zero values agree, the value is used. If none is configured or there is a disagreement, pci_dfl_cache_line_size is used. arch can set the dfl value (via PCI_CACHE_LINE_BYTES or pci_dfl_cache_line_size) or override the actual one. ia64, x86 and sparc64 updated to set the default cls instead of the actual one. While at it, declare pci_cache_line_size and pci_dfl_cache_line_size in pci.h and drop private declarations from arch code. Signed-off-by: Tejun Heo Acked-by: David Miller Acked-by: Greg KH Cc: Ingo Molnar Cc: Thomas Gleixner Cc: Tony Luck Signed-off-by: Jesse Barnes --- arch/ia64/pci/pci.c | 9 +++------ arch/x86/pci/common.c | 8 +++----- drivers/pci/pci.c | 21 +++++++++++++-------- drivers/pci/quirks.c | 28 ++++++++++++++++++++++++++++ include/linux/pci.h | 2 ++ 5 files changed, 49 insertions(+), 19 deletions(-) diff --git a/arch/ia64/pci/pci.c b/arch/ia64/pci/pci.c index c0fca2c1c858..d60e7195b7dd 100644 --- a/arch/ia64/pci/pci.c +++ b/arch/ia64/pci/pci.c @@ -720,9 +720,6 @@ int ia64_pci_legacy_write(struct pci_bus *bus, u16 port, u32 val, u8 size) return ret; } -/* It's defined in drivers/pci/pci.c */ -extern u8 pci_cache_line_size; - /** * set_pci_cacheline_size - determine cacheline size for PCI devices * @@ -731,7 +728,7 @@ extern u8 pci_cache_line_size; * * Code mostly taken from arch/ia64/kernel/palinfo.c:cache_info(). */ -static void __init set_pci_cacheline_size(void) +static void __init set_pci_dfl_cacheline_size(void) { unsigned long levels, unique_caches; long status; @@ -751,7 +748,7 @@ static void __init set_pci_cacheline_size(void) "(status=%ld)\n", __func__, status); return; } - pci_cache_line_size = (1 << cci.pcci_line_size) / 4; + pci_dfl_cache_line_size = (1 << cci.pcci_line_size) / 4; } u64 ia64_dma_get_required_mask(struct device *dev) @@ -782,7 +779,7 @@ EXPORT_SYMBOL_GPL(dma_get_required_mask); static int __init pcibios_init(void) { - set_pci_cacheline_size(); + set_pci_dfl_cacheline_size(); return 0; } diff --git a/arch/x86/pci/common.c b/arch/x86/pci/common.c index 1331fcf26143..fbeec31316cf 100644 --- a/arch/x86/pci/common.c +++ b/arch/x86/pci/common.c @@ -410,8 +410,6 @@ struct pci_bus * __devinit pcibios_scan_root(int busnum) return bus; } -extern u8 pci_cache_line_size; - int __init pcibios_init(void) { struct cpuinfo_x86 *c = &boot_cpu_data; @@ -426,11 +424,11 @@ int __init pcibios_init(void) * and P4. It's also good for 386/486s (which actually have 16) * as quite a few PCI devices do not support smaller values. */ - pci_cache_line_size = 32 >> 2; + pci_dfl_cache_line_size = 32 >> 2; if (c->x86 >= 6 && c->x86_vendor == X86_VENDOR_AMD) - pci_cache_line_size = 64 >> 2; /* K7 & K8 */ + pci_dfl_cache_line_size = 64 >> 2; /* K7 & K8 */ else if (c->x86 > 6 && c->x86_vendor == X86_VENDOR_INTEL) - pci_cache_line_size = 128 >> 2; /* P4 */ + pci_dfl_cache_line_size = 128 >> 2; /* P4 */ pcibios_resource_survey(); diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index 4e4c295a049f..1f9a7a03847b 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -47,6 +47,19 @@ unsigned long pci_cardbus_mem_size = DEFAULT_CARDBUS_MEM_SIZE; unsigned long pci_hotplug_io_size = DEFAULT_HOTPLUG_IO_SIZE; unsigned long pci_hotplug_mem_size = DEFAULT_HOTPLUG_MEM_SIZE; +#ifndef PCI_CACHE_LINE_BYTES +#define PCI_CACHE_LINE_BYTES L1_CACHE_BYTES +#endif + +/* + * The default CLS is used if arch didn't set CLS explicitly and not + * all pci devices agree on the same value. Arch can override either + * the dfl or actual value as it sees fit. Don't forget this is + * measured in 32-bit words, not bytes. + */ +u8 pci_dfl_cache_line_size __initdata = PCI_CACHE_LINE_BYTES >> 2; +u8 pci_cache_line_size; + /** * pci_bus_max_busnr - returns maximum PCI bus number of given bus' children * @bus: pointer to PCI bus structure to search @@ -1883,14 +1896,6 @@ void pci_clear_mwi(struct pci_dev *dev) #else -#ifndef PCI_CACHE_LINE_BYTES -#define PCI_CACHE_LINE_BYTES L1_CACHE_BYTES -#endif - -/* This can be overridden by arch code. */ -/* Don't forget this is measured in 32-bit words, not bytes */ -u8 pci_cache_line_size = PCI_CACHE_LINE_BYTES / 4; - /** * pci_set_cacheline_size - ensure the CACHE_LINE_SIZE register is programmed * @dev: the PCI device for which MWI is to be enabled diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c index 245d2cdb4765..1812ae7698de 100644 --- a/drivers/pci/quirks.c +++ b/drivers/pci/quirks.c @@ -2595,9 +2595,37 @@ void pci_fixup_device(enum pci_fixup_pass pass, struct pci_dev *dev) static int __init pci_apply_final_quirks(void) { struct pci_dev *dev = NULL; + u8 cls = 0; + u8 tmp; + + if (pci_cache_line_size) + printk(KERN_DEBUG "PCI: CLS %u bytes\n", + pci_cache_line_size << 2); while ((dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, dev)) != NULL) { pci_fixup_device(pci_fixup_final, dev); + /* + * If arch hasn't set it explicitly yet, use the CLS + * value shared by all PCI devices. If there's a + * mismatch, fall back to the default value. + */ + if (!pci_cache_line_size) { + pci_read_config_byte(dev, PCI_CACHE_LINE_SIZE, &tmp); + if (!cls) + cls = tmp; + if (!tmp || cls == tmp) + continue; + + printk(KERN_DEBUG "PCI: CLS mismatch (%u != %u), " + "using %u bytes\n", cls << 2, tmp << 2, + pci_dfl_cache_line_size << 2); + pci_cache_line_size = pci_dfl_cache_line_size; + } + } + if (!pci_cache_line_size) { + printk(KERN_DEBUG "PCI: CLS %u bytes, default %u\n", + cls << 2, pci_dfl_cache_line_size << 2); + pci_cache_line_size = cls; } return 0; diff --git a/include/linux/pci.h b/include/linux/pci.h index f5c7cd343e56..b849861d78e6 100644 --- a/include/linux/pci.h +++ b/include/linux/pci.h @@ -1246,6 +1246,8 @@ extern int pci_pci_problems; extern unsigned long pci_cardbus_io_size; extern unsigned long pci_cardbus_mem_size; +extern u8 pci_dfl_cache_line_size; +extern u8 pci_cache_line_size; extern unsigned long pci_hotplug_io_size; extern unsigned long pci_hotplug_mem_size; From 4c0eec7a86303ce6e3edf7825d0ef1d414e76767 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 22 Sep 2009 17:34:17 +0900 Subject: [PATCH 003/109] sparc64/PCI: drop PCI_CACHE_LINE_BYTES sparc64 is now the only user of PCI_CACHE_LINE_BYTES. Drop it and set pci_dfl_cache_line_size from pcibios_init() instead and drop PCI_CACHE_LINE_BYTES handling from generic pci code. Orignally-From: David Miller Signed-off-by: Tejun Heo Signed-off-by: Jesse Barnes --- arch/sparc/include/asm/pci_64.h | 2 -- arch/sparc/kernel/pci.c | 7 +++++++ drivers/pci/pci.c | 6 +----- 3 files changed, 8 insertions(+), 7 deletions(-) diff --git a/arch/sparc/include/asm/pci_64.h b/arch/sparc/include/asm/pci_64.h index b63e51c3c3ee..b0576df6ec83 100644 --- a/arch/sparc/include/asm/pci_64.h +++ b/arch/sparc/include/asm/pci_64.h @@ -16,8 +16,6 @@ #define PCI_IRQ_NONE 0xffffffff -#define PCI_CACHE_LINE_BYTES 64 - static inline void pcibios_set_master(struct pci_dev *dev) { /* No special bus mastering setup handling */ diff --git a/arch/sparc/kernel/pci.c b/arch/sparc/kernel/pci.c index c68648662802..b85374f7cf94 100644 --- a/arch/sparc/kernel/pci.c +++ b/arch/sparc/kernel/pci.c @@ -1081,3 +1081,10 @@ void pci_resource_to_user(const struct pci_dev *pdev, int bar, *start = rp->start - offset; *end = rp->end - offset; } + +static int __init pcibios_init(void) +{ + pci_dfl_cache_line_size = 64 >> 2; + return 0; +} +subsys_initcall(pcibios_init); diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index 1f9a7a03847b..01337b7a215f 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -47,17 +47,13 @@ unsigned long pci_cardbus_mem_size = DEFAULT_CARDBUS_MEM_SIZE; unsigned long pci_hotplug_io_size = DEFAULT_HOTPLUG_IO_SIZE; unsigned long pci_hotplug_mem_size = DEFAULT_HOTPLUG_MEM_SIZE; -#ifndef PCI_CACHE_LINE_BYTES -#define PCI_CACHE_LINE_BYTES L1_CACHE_BYTES -#endif - /* * The default CLS is used if arch didn't set CLS explicitly and not * all pci devices agree on the same value. Arch can override either * the dfl or actual value as it sees fit. Don't forget this is * measured in 32-bit words, not bytes. */ -u8 pci_dfl_cache_line_size __initdata = PCI_CACHE_LINE_BYTES >> 2; +u8 pci_dfl_cache_line_size __initdata = L1_CACHE_BYTES >> 2; u8 pci_cache_line_size; /** From 15ea76d407d560f985224b65fe59c9db01692a0d Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 22 Sep 2009 17:34:48 +0900 Subject: [PATCH 004/109] pccard: configure CLS on attach For non hotplug PCI devices, the system firmware usually configures CLS correctly. For pccard devices system firmware can't do it and Linux PCI layer doesn't do it either. Unfortunately this leads to poor performance for certain devices (sata_sil). Unless MWI, which requires separate configuration, is to be used, CLS doesn't affect correctness, so the configuration should be harmless. This patch makes pci_set_cacheline_size() always built and export it and make pccard call it during attach. Please note that some other PCI hotplug drivers (shpchp and pciehp) also configure CLS on hotplug. Signed-off-by: Tejun Heo Cc: Daniel Ritz Cc: Dominik Brodowski Cc: Greg KH Cc: Kenji Kaneshige Cc: Axel Birndt Signed-off-by: Jesse Barnes --- drivers/pci/pci.c | 40 ++++++++++++++++++++-------------------- drivers/pcmcia/cardbus.c | 23 +++++++++++++++-------- include/linux/pci.h | 1 + 3 files changed, 36 insertions(+), 28 deletions(-) diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index 01337b7a215f..d1afbae5b1fb 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -1875,23 +1875,6 @@ void pci_clear_master(struct pci_dev *dev) __pci_set_master(dev, false); } -#ifdef PCI_DISABLE_MWI -int pci_set_mwi(struct pci_dev *dev) -{ - return 0; -} - -int pci_try_set_mwi(struct pci_dev *dev) -{ - return 0; -} - -void pci_clear_mwi(struct pci_dev *dev) -{ -} - -#else - /** * pci_set_cacheline_size - ensure the CACHE_LINE_SIZE register is programmed * @dev: the PCI device for which MWI is to be enabled @@ -1902,13 +1885,12 @@ void pci_clear_mwi(struct pci_dev *dev) * * RETURNS: An appropriate -ERRNO error value on error, or zero for success. */ -static int -pci_set_cacheline_size(struct pci_dev *dev) +int pci_set_cacheline_size(struct pci_dev *dev) { u8 cacheline_size; if (!pci_cache_line_size) - return -EINVAL; /* The system doesn't support MWI. */ + return -EINVAL; /* Validate current setting: the PCI_CACHE_LINE_SIZE must be equal to or multiple of the right value. */ @@ -1929,6 +1911,24 @@ pci_set_cacheline_size(struct pci_dev *dev) return -EINVAL; } +EXPORT_SYMBOL_GPL(pci_set_cacheline_size); + +#ifdef PCI_DISABLE_MWI +int pci_set_mwi(struct pci_dev *dev) +{ + return 0; +} + +int pci_try_set_mwi(struct pci_dev *dev) +{ + return 0; +} + +void pci_clear_mwi(struct pci_dev *dev) +{ +} + +#else /** * pci_set_mwi - enables memory-write-invalidate PCI transaction diff --git a/drivers/pcmcia/cardbus.c b/drivers/pcmcia/cardbus.c index db77e1f3309a..98789c031a7c 100644 --- a/drivers/pcmcia/cardbus.c +++ b/drivers/pcmcia/cardbus.c @@ -184,26 +184,33 @@ fail: =====================================================================*/ -/* - * Since there is only one interrupt available to CardBus - * devices, all devices downstream of this device must - * be using this IRQ. - */ -static void cardbus_assign_irqs(struct pci_bus *bus, int irq) +static void cardbus_config_irq_and_cls(struct pci_bus *bus, int irq) { struct pci_dev *dev; list_for_each_entry(dev, &bus->devices, bus_list) { u8 irq_pin; + /* + * Since there is only one interrupt available to + * CardBus devices, all devices downstream of this + * device must be using this IRQ. + */ pci_read_config_byte(dev, PCI_INTERRUPT_PIN, &irq_pin); if (irq_pin) { dev->irq = irq; pci_write_config_byte(dev, PCI_INTERRUPT_LINE, dev->irq); } + /* + * Some controllers transfer very slowly with 0 CLS. + * Configure it. This may fail as CLS configuration + * is mandatory only for MWI. + */ + pci_set_cacheline_size(dev); + if (dev->subordinate) - cardbus_assign_irqs(dev->subordinate, irq); + cardbus_config_irq_and_cls(dev->subordinate, irq); } } @@ -228,7 +235,7 @@ int __ref cb_alloc(struct pcmcia_socket * s) */ pci_bus_size_bridges(bus); pci_bus_assign_resources(bus); - cardbus_assign_irqs(bus, s->pci_irq); + cardbus_config_irq_and_cls(bus, s->pci_irq); /* socket specific tune function */ if (s->tune_bridge) diff --git a/include/linux/pci.h b/include/linux/pci.h index b849861d78e6..da4128f6e916 100644 --- a/include/linux/pci.h +++ b/include/linux/pci.h @@ -701,6 +701,7 @@ void pci_disable_device(struct pci_dev *dev); void pci_set_master(struct pci_dev *dev); void pci_clear_master(struct pci_dev *dev); int pci_set_pcie_reset_state(struct pci_dev *dev, enum pcie_reset_state state); +int pci_set_cacheline_size(struct pci_dev *dev); #define HAVE_PCI_SET_MWI int __must_check pci_set_mwi(struct pci_dev *dev); int pci_try_set_mwi(struct pci_dev *dev); From 98e724c791924c0dfc5b1dcf053ed3841cc89c78 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 8 Oct 2009 18:59:53 +0900 Subject: [PATCH 005/109] PCI: pci_dfl_cache_line_size is __devinitdata pci_dfl_cache_line_size is marked as __initdata but referenced by pci_init() which is __devinit. Make it __devinitdata instead of __initdata. Signed-off-by: Tejun Heo Reported-by: Stephen Rothwell Signed-off-by: Jesse Barnes --- drivers/pci/pci.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index d1afbae5b1fb..3f8d971ac36e 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -53,7 +53,7 @@ unsigned long pci_hotplug_mem_size = DEFAULT_HOTPLUG_MEM_SIZE; * the dfl or actual value as it sees fit. Don't forget this is * measured in 32-bit words, not bytes. */ -u8 pci_dfl_cache_line_size __initdata = L1_CACHE_BYTES >> 2; +u8 pci_dfl_cache_line_size __devinitdata = L1_CACHE_BYTES >> 2; u8 pci_cache_line_size; /** From 76b1a87b217927f905f4b01c586452b2a1d33913 Mon Sep 17 00:00:00 2001 From: Dave Jones Date: Wed, 14 Oct 2009 16:31:39 -0400 Subject: [PATCH 006/109] x86/PCI: Use generic cacheline sizing instead of per-vendor tests. Instead of the PCI code needing to have code to determine the cacheline size of each processor, use the data the cpu identification code should have already determined during early boot. (The vendor checks are also incomplete, and don't take into account modern CPUs) I've been carrying a variant of this code in Fedora for a while, that prints debug information. There are a number of cases where we are currently setting the PCI cacheline size to 32 bytes, when the CPU cacheline size is 64 bytes. With this patch, we set them both the same. Signed-off-by: Dave Jones Signed-off-by: Jesse Barnes --- arch/x86/pci/common.c | 18 +++++++++++------- 1 file changed, 11 insertions(+), 7 deletions(-) diff --git a/arch/x86/pci/common.c b/arch/x86/pci/common.c index fbeec31316cf..d2552c68e94d 100644 --- a/arch/x86/pci/common.c +++ b/arch/x86/pci/common.c @@ -420,15 +420,19 @@ int __init pcibios_init(void) } /* - * Assume PCI cacheline size of 32 bytes for all x86s except K7/K8 - * and P4. It's also good for 386/486s (which actually have 16) + * Set PCI cacheline size to that of the CPU if the CPU has reported it. + * (For older CPUs that don't support cpuid, we se it to 32 bytes + * It's also good for 386/486s (which actually have 16) * as quite a few PCI devices do not support smaller values. */ - pci_dfl_cache_line_size = 32 >> 2; - if (c->x86 >= 6 && c->x86_vendor == X86_VENDOR_AMD) - pci_dfl_cache_line_size = 64 >> 2; /* K7 & K8 */ - else if (c->x86 > 6 && c->x86_vendor == X86_VENDOR_INTEL) - pci_dfl_cache_line_size = 128 >> 2; /* P4 */ + if (c->x86_clflush_size > 0) { + pci_dfl_cache_line_size = c->x86_clflush_size >> 2; + printk(KERN_DEBUG "PCI: pci_cache_line_size set to %d bytes\n", + pci_dfl_cache_line_size << 2); + } else { + pci_dfl_cache_line_size = 32 >> 2; + printk(KERN_DEBUG "PCI: Unknown cacheline size. Setting to 32 bytes\n"); + } pcibios_resource_survey(); From 17d67152793c43344930bda9b723c80186598aad Mon Sep 17 00:00:00 2001 From: Stefan Assmann Date: Mon, 26 Oct 2009 14:44:46 +0100 Subject: [PATCH 007/109] PCI hotplug: change PCI nomenclature Change PCI nomenclature according to http://www.pcisig.com/developers/procedures/logos/Trademark_and_Logo_Usage_Guidelines_updated_112206.pdf. Signed-off-by: Stefan Assmann Signed-off-by: Jesse Barnes --- drivers/pci/hotplug/pci_hotplug_core.c | 22 +++++++++++----------- 1 file changed, 11 insertions(+), 11 deletions(-) diff --git a/drivers/pci/hotplug/pci_hotplug_core.c b/drivers/pci/hotplug/pci_hotplug_core.c index 0325d989bb46..38183a534b65 100644 --- a/drivers/pci/hotplug/pci_hotplug_core.c +++ b/drivers/pci/hotplug/pci_hotplug_core.c @@ -68,26 +68,26 @@ static DEFINE_MUTEX(pci_hp_mutex); static char *pci_bus_speed_strings[] = { "33 MHz PCI", /* 0x00 */ "66 MHz PCI", /* 0x01 */ - "66 MHz PCIX", /* 0x02 */ - "100 MHz PCIX", /* 0x03 */ - "133 MHz PCIX", /* 0x04 */ + "66 MHz PCI-X", /* 0x02 */ + "100 MHz PCI-X", /* 0x03 */ + "133 MHz PCI-X", /* 0x04 */ NULL, /* 0x05 */ NULL, /* 0x06 */ NULL, /* 0x07 */ NULL, /* 0x08 */ - "66 MHz PCIX 266", /* 0x09 */ - "100 MHz PCIX 266", /* 0x0a */ - "133 MHz PCIX 266", /* 0x0b */ + "66 MHz PCI-X 266", /* 0x09 */ + "100 MHz PCI-X 266", /* 0x0a */ + "133 MHz PCI-X 266", /* 0x0b */ NULL, /* 0x0c */ NULL, /* 0x0d */ NULL, /* 0x0e */ NULL, /* 0x0f */ NULL, /* 0x10 */ - "66 MHz PCIX 533", /* 0x11 */ - "100 MHz PCIX 533", /* 0x12 */ - "133 MHz PCIX 533", /* 0x13 */ - "2.5 GT/s PCI-E", /* 0x14 */ - "5.0 GT/s PCI-E", /* 0x15 */ + "66 MHz PCI-X 533", /* 0x11 */ + "100 MHz PCI-X 533", /* 0x12 */ + "133 MHz PCI-X 533", /* 0x13 */ + "2.5 GT/s PCIe", /* 0x14 */ + "5.0 GT/s PCIe", /* 0x15 */ }; #ifdef CONFIG_HOTPLUG_PCI_CPCI From 3368dd29586c6460b629ac5b4f6b86a6fd3dd421 Mon Sep 17 00:00:00 2001 From: Matthew Garrett Date: Mon, 26 Oct 2009 13:18:22 -0400 Subject: [PATCH 008/109] PCI hotplug: acpiphp should be linked after vendor drivers As a followup to 71a082efc9fdc12068a3cee6cebb1330b00ebeee, it's conceivable that some vendors may expose PCI hotplug functionality through both vendor mechanisms and ACPI. The native mechanism will generally be a superset of any functionality provided via ACPI, so the acpiphp driver should always be initialised after any others. Change the link order such that acpiphp will not be initialised until any other statically linked drivers have had an opportunity to claim the hardware. Signed-off-by: Matthew Garrett Signed-off-by: Jesse Barnes --- drivers/pci/hotplug/Makefile | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) diff --git a/drivers/pci/hotplug/Makefile b/drivers/pci/hotplug/Makefile index 3625b094bf7e..6cd9f3c9887d 100644 --- a/drivers/pci/hotplug/Makefile +++ b/drivers/pci/hotplug/Makefile @@ -6,18 +6,22 @@ obj-$(CONFIG_HOTPLUG_PCI) += pci_hotplug.o obj-$(CONFIG_HOTPLUG_PCI_COMPAQ) += cpqphp.o obj-$(CONFIG_HOTPLUG_PCI_IBM) += ibmphp.o -# pciehp should be linked before acpiphp in order to allow the native driver -# to attempt to bind first. We can then fall back to generic support. +# native drivers should be linked before acpiphp in order to allow the +# native driver to attempt to bind first. We can then fall back to +# generic support. obj-$(CONFIG_HOTPLUG_PCI_PCIE) += pciehp.o -obj-$(CONFIG_HOTPLUG_PCI_ACPI) += acpiphp.o -obj-$(CONFIG_HOTPLUG_PCI_ACPI_IBM) += acpiphp_ibm.o obj-$(CONFIG_HOTPLUG_PCI_CPCI_ZT5550) += cpcihp_zt5550.o obj-$(CONFIG_HOTPLUG_PCI_CPCI_GENERIC) += cpcihp_generic.o obj-$(CONFIG_HOTPLUG_PCI_SHPC) += shpchp.o obj-$(CONFIG_HOTPLUG_PCI_RPA) += rpaphp.o obj-$(CONFIG_HOTPLUG_PCI_RPA_DLPAR) += rpadlpar_io.o obj-$(CONFIG_HOTPLUG_PCI_SGI) += sgi_hotplug.o +obj-$(CONFIG_HOTPLUG_PCI_ACPI) += acpiphp.o + +# acpiphp_ibm extends acpiphp, so should be linked afterwards. + +obj-$(CONFIG_HOTPLUG_PCI_ACPI_IBM) += acpiphp_ibm.o # Link this last so it doesn't claim devices that have a real hotplug driver obj-$(CONFIG_HOTPLUG_PCI_FAKE) += fakephp.o From 2840537228fba95e05cab1a6b5719c61982db279 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Tue, 6 Oct 2009 15:33:29 -0600 Subject: [PATCH 009/109] vsprintf: fix io/mem resource width The leading "0x" consumes field width, so leave space for it in addition to the 4 or 8 hex digits. This means we'll print "0x0000-0x01df" rather than "0x00-0x1df", for example. Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- lib/vsprintf.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/lib/vsprintf.c b/lib/vsprintf.c index 33bed5e67a21..7830576018c0 100644 --- a/lib/vsprintf.c +++ b/lib/vsprintf.c @@ -598,11 +598,11 @@ static char *resource_string(char *buf, char *end, struct resource *res, struct printf_spec spec) { #ifndef IO_RSRC_PRINTK_SIZE -#define IO_RSRC_PRINTK_SIZE 4 +#define IO_RSRC_PRINTK_SIZE 6 #endif #ifndef MEM_RSRC_PRINTK_SIZE -#define MEM_RSRC_PRINTK_SIZE 8 +#define MEM_RSRC_PRINTK_SIZE 10 #endif struct printf_spec num_spec = { .base = 16, From c91d3376e5f4277173a22f0ef9989125c318bacb Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Tue, 6 Oct 2009 15:33:34 -0600 Subject: [PATCH 010/109] vsprintf: add %pR support for IRQ and DMA resources Print addresses (IO port numbers and memory addresses) in hex, but print others (IRQs and DMA channels) in decimal. Only print the end if it's different from the start. Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- lib/vsprintf.c | 31 +++++++++++++++++++++---------- 1 file changed, 21 insertions(+), 10 deletions(-) diff --git a/lib/vsprintf.c b/lib/vsprintf.c index 7830576018c0..1b60aedab44d 100644 --- a/lib/vsprintf.c +++ b/lib/vsprintf.c @@ -604,26 +604,37 @@ static char *resource_string(char *buf, char *end, struct resource *res, #ifndef MEM_RSRC_PRINTK_SIZE #define MEM_RSRC_PRINTK_SIZE 10 #endif - struct printf_spec num_spec = { + struct printf_spec hex_spec = { .base = 16, .precision = -1, .flags = SPECIAL | SMALL | ZEROPAD, }; - /* room for the actual numbers, the two "0x", -, [, ] and the final zero */ - char sym[4*sizeof(resource_size_t) + 8]; + struct printf_spec dec_spec = { + .base = 10, + .precision = -1, + .flags = 0, + }; + /* room for two actual numbers (decimal or hex), the two "0x", -, [, ] + * and the final zero */ + char sym[2*3*sizeof(resource_size_t) + 8]; char *p = sym, *pend = sym + sizeof(sym); - int size = -1; + int size = -1, addr = 0; - if (res->flags & IORESOURCE_IO) + if (res->flags & IORESOURCE_IO) { size = IO_RSRC_PRINTK_SIZE; - else if (res->flags & IORESOURCE_MEM) + addr = 1; + } else if (res->flags & IORESOURCE_MEM) { size = MEM_RSRC_PRINTK_SIZE; + addr = 1; + } *p++ = '['; - num_spec.field_width = size; - p = number(p, pend, res->start, num_spec); - *p++ = '-'; - p = number(p, pend, res->end, num_spec); + hex_spec.field_width = size; + p = number(p, pend, res->start, addr ? hex_spec : dec_spec); + if (res->start != res->end) { + *p++ = '-'; + p = number(p, pend, res->end, addr ? hex_spec : dec_spec); + } *p++ = ']'; *p = 0; From fd95541e23e2c9acb1e38cd41fc0c7cc37fceb53 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Tue, 6 Oct 2009 15:33:39 -0600 Subject: [PATCH 011/109] vsprintf: add %pRt, %pRf to print struct resource details This adds support for printing struct resource type and flag information. For example, "%pRt" looks like "[mem 0x80080000000-0x8008001ffff 64bit pref]", and "%pRf" looks like "[mem 0xff5e2000-0xff5e2007 pref flags 0x1]". Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- lib/vsprintf.c | 51 +++++++++++++++++++++++++++++++++++++++++++------- 1 file changed, 44 insertions(+), 7 deletions(-) diff --git a/lib/vsprintf.c b/lib/vsprintf.c index 1b60aedab44d..a6e195163eb3 100644 --- a/lib/vsprintf.c +++ b/lib/vsprintf.c @@ -595,7 +595,7 @@ static char *symbol_string(char *buf, char *end, void *ptr, } static char *resource_string(char *buf, char *end, struct resource *res, - struct printf_spec spec) + struct printf_spec spec, const char *fmt) { #ifndef IO_RSRC_PRINTK_SIZE #define IO_RSRC_PRINTK_SIZE 6 @@ -614,9 +614,21 @@ static char *resource_string(char *buf, char *end, struct resource *res, .precision = -1, .flags = 0, }; - /* room for two actual numbers (decimal or hex), the two "0x", -, [, ] - * and the final zero */ - char sym[2*3*sizeof(resource_size_t) + 8]; + struct printf_spec str_spec = { + .field_width = -1, + .precision = 10, + .flags = LEFT, + }; + struct printf_spec flag_spec = { + .base = 16, + .precision = -1, + .flags = SPECIAL | SMALL, + }; + /* + * room for three actual numbers (decimal or hex), plus + * "[mem 0x-0x 64bit pref disabled flags 0x]\0" + */ + char sym[3*3*sizeof(resource_size_t) + 41]; char *p = sym, *pend = sym + sizeof(sym); int size = -1, addr = 0; @@ -629,12 +641,35 @@ static char *resource_string(char *buf, char *end, struct resource *res, } *p++ = '['; + if (fmt[1] == 't' || fmt[1] == 'f') { + if (res->flags & IORESOURCE_IO) + p = string(p, pend, "io ", str_spec); + else if (res->flags & IORESOURCE_MEM) + p = string(p, pend, "mem ", str_spec); + else if (res->flags & IORESOURCE_IRQ) + p = string(p, pend, "irq ", str_spec); + else if (res->flags & IORESOURCE_DMA) + p = string(p, pend, "dma ", str_spec); + } hex_spec.field_width = size; p = number(p, pend, res->start, addr ? hex_spec : dec_spec); if (res->start != res->end) { *p++ = '-'; p = number(p, pend, res->end, addr ? hex_spec : dec_spec); } + if (fmt[1] == 't' || fmt[1] == 'f') { + if (res->flags & IORESOURCE_MEM_64) + p = string(p, pend, " 64bit", str_spec); + if (res->flags & IORESOURCE_PREFETCH) + p = string(p, pend, " pref", str_spec); + if (res->flags & IORESOURCE_DISABLED) + p = string(p, pend, " disabled", str_spec); + if (fmt[1] == 'f') { + p = string(p, pend, " flags ", str_spec); + p = number(p, pend, res->flags & ~IORESOURCE_TYPE_BITS, + flag_spec); + } + } *p++ = ']'; *p = 0; @@ -812,8 +847,10 @@ static char *ip4_addr_string(char *buf, char *end, const u8 *addr, * - 'f' For simple symbolic function names without offset * - 'S' For symbolic direct pointers with offset * - 's' For symbolic direct pointers without offset - * - 'R' For a struct resource pointer, it prints the range of - * addresses (not the name nor the flags) + * - 'R' For a struct resource pointer, print: + * R address range only ([0x0-0x1f]) + * Rt type and range ([mem 0x0-0x1f 64bit pref]) + * Rf type, range, and flags ([mem 0x0-0x1f 64bit pref flags 0x1]) * - 'M' For a 6-byte MAC address, it prints the address in the * usual colon-separated hex notation * - 'm' For a 6-byte MAC address, it prints the hex address without colons @@ -844,7 +881,7 @@ static char *pointer(const char *fmt, char *buf, char *end, void *ptr, case 'S': return symbol_string(buf, end, ptr, spec, *fmt); case 'R': - return resource_string(buf, end, ptr, spec); + return resource_string(buf, end, ptr, spec, fmt); case 'M': /* Colon separated: 00:01:02:03:04:05 */ case 'm': /* Contiguous: 000102030405 */ return mac_address_string(buf, end, ptr, spec, fmt); From a369c791e881503a6253dafc0d0ad5e41e5557e5 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Tue, 6 Oct 2009 15:33:44 -0600 Subject: [PATCH 012/109] PCI: print resources consistently with %pRt This uses %pRt to print additional resource information (type, size, prefetchability, etc.) consistently. Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- drivers/pci/pci.c | 4 +--- drivers/pci/probe.c | 26 ++++++++------------------ drivers/pci/setup-bus.c | 9 +++------ drivers/pci/setup-res.c | 28 ++++++++++++---------------- 4 files changed, 24 insertions(+), 43 deletions(-) diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index 3f8d971ac36e..4859669f0ab5 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -1678,9 +1678,7 @@ static int __pci_request_region(struct pci_dev *pdev, int bar, const char *res_n return 0; err_out: - dev_warn(&pdev->dev, "BAR %d: can't reserve %s region %pR\n", - bar, - pci_resource_flags(pdev, bar) & IORESOURCE_IO ? "I/O" : "mem", + dev_warn(&pdev->dev, "BAR %d: can't reserve %pRt\n", bar, &pdev->resource[bar]); return -EBUSY; } diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 8105e32117f6..2adb47574d86 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -222,6 +222,8 @@ int __pci_read_base(struct pci_dev *dev, enum pci_bar_type type, if (!sz64) goto fail; + res->flags |= IORESOURCE_MEM_64; + if ((sizeof(resource_size_t) < 8) && (sz64 > 0x100000000ULL)) { dev_err(&dev->dev, "can't handle 64-bit BAR\n"); goto fail; @@ -234,14 +236,9 @@ int __pci_read_base(struct pci_dev *dev, enum pci_bar_type type, } else { res->start = l64; res->end = l64 + sz64; - dev_printk(KERN_DEBUG, &dev->dev, - "reg %x %s: %pR\n", pos, - (res->flags & IORESOURCE_PREFETCH) ? - "64bit mmio pref" : "64bit mmio", - res); + dev_printk(KERN_DEBUG, &dev->dev, "reg %x: %pRt\n", + pos, res); } - - res->flags |= IORESOURCE_MEM_64; } else { sz = pci_size(l, sz, mask); @@ -251,11 +248,7 @@ int __pci_read_base(struct pci_dev *dev, enum pci_bar_type type, res->start = l; res->end = l + sz; - dev_printk(KERN_DEBUG, &dev->dev, "reg %x %s: %pR\n", pos, - (res->flags & IORESOURCE_IO) ? "io port" : - ((res->flags & IORESOURCE_PREFETCH) ? - "32bit mmio pref" : "32bit mmio"), - res); + dev_printk(KERN_DEBUG, &dev->dev, "reg %x: %pRt\n", pos, res); } out: @@ -323,7 +316,7 @@ void __devinit pci_read_bridge_bases(struct pci_bus *child) res->start = base; if (!res->end) res->end = limit + 0xfff; - dev_printk(KERN_DEBUG, &dev->dev, "bridge io port: %pR\n", res); + dev_printk(KERN_DEBUG, &dev->dev, "bridge window: %pRt\n", res); } res = child->resource[1]; @@ -335,8 +328,7 @@ void __devinit pci_read_bridge_bases(struct pci_bus *child) res->flags = (mem_base_lo & PCI_MEMORY_RANGE_TYPE_MASK) | IORESOURCE_MEM; res->start = base; res->end = limit + 0xfffff; - dev_printk(KERN_DEBUG, &dev->dev, "bridge 32bit mmio: %pR\n", - res); + dev_printk(KERN_DEBUG, &dev->dev, "bridge window: %pRt\n", res); } res = child->resource[2]; @@ -375,9 +367,7 @@ void __devinit pci_read_bridge_bases(struct pci_bus *child) res->flags |= IORESOURCE_MEM_64; res->start = base; res->end = limit + 0xfffff; - dev_printk(KERN_DEBUG, &dev->dev, "bridge %sbit mmio pref: %pR\n", - (res->flags & PCI_PREF_RANGE_TYPE_64) ? "64" : "32", - res); + dev_printk(KERN_DEBUG, &dev->dev, "bridge window: %pRt\n", res); } } diff --git a/drivers/pci/setup-bus.c b/drivers/pci/setup-bus.c index cb1a027eb552..ceb75333862b 100644 --- a/drivers/pci/setup-bus.c +++ b/drivers/pci/setup-bus.c @@ -390,8 +390,8 @@ static int pbus_size_mem(struct pci_bus *bus, unsigned long mask, align = pci_resource_alignment(dev, r); order = __ffs(align) - 20; if (order > 11) { - dev_warn(&dev->dev, "BAR %d bad alignment %llx: " - "%pR\n", i, (unsigned long long)align, r); + dev_warn(&dev->dev, "BAR %d: bad alignment %llx: " + "%pRt\n", i, (unsigned long long)align, r); r->flags = 0; continue; } @@ -582,10 +582,7 @@ static void pci_bus_dump_res(struct pci_bus *bus) if (!res || !res->end) continue; - dev_printk(KERN_DEBUG, &bus->dev, "resource %d %s %pR\n", i, - (res->flags & IORESOURCE_IO) ? "io: " : - ((res->flags & IORESOURCE_PREFETCH)? "pref mem":"mem:"), - res); + dev_printk(KERN_DEBUG, &bus->dev, "resource %d %pRt\n", i, res); } } diff --git a/drivers/pci/setup-res.c b/drivers/pci/setup-res.c index c54526b206b5..5e78f2096ce8 100644 --- a/drivers/pci/setup-res.c +++ b/drivers/pci/setup-res.c @@ -51,11 +51,9 @@ void pci_update_resource(struct pci_dev *dev, int resno) pcibios_resource_to_bus(dev, ®ion, res); - dev_dbg(&dev->dev, "BAR %d: got res %pR bus [%#llx-%#llx] " - "flags %#lx\n", resno, res, - (unsigned long long)region.start, - (unsigned long long)region.end, - (unsigned long)res->flags); + dev_dbg(&dev->dev, "BAR %d: got %pRf (bus addr [%#llx-%#llx])\n", + resno, res, (unsigned long long)region.start, + (unsigned long long)region.end); new = region.start | (res->flags & PCI_REGION_FLAG_MASK); if (res->flags & IORESOURCE_IO) @@ -91,9 +89,9 @@ void pci_update_resource(struct pci_dev *dev, int resno) } } res->flags &= ~IORESOURCE_UNSET; - dev_dbg(&dev->dev, "BAR %d: moved to bus [%#llx-%#llx] flags %#lx\n", + dev_dbg(&dev->dev, "BAR %d: moved to bus addr [%#llx-%#llx]\n", resno, (unsigned long long)region.start, - (unsigned long long)region.end, res->flags); + (unsigned long long)region.end); } int pci_claim_resource(struct pci_dev *dev, int resource) @@ -110,7 +108,7 @@ int pci_claim_resource(struct pci_dev *dev, int resource) if (err) { const char *dtype = resource < PCI_BRIDGE_RESOURCES ? "device" : "bridge"; - dev_err(&dev->dev, "BAR %d: %s of %s %pR\n", + dev_err(&dev->dev, "BAR %d: %s %s %pRt\n", resource, root ? "address space collision on" : "no parent found for", @@ -181,9 +179,8 @@ int pci_assign_resource(struct pci_dev *dev, int resno) align = pci_resource_alignment(dev, res); if (!align) { - dev_info(&dev->dev, "BAR %d: can't allocate resource (bogus " - "alignment) %pR flags %#lx\n", - resno, res, res->flags); + dev_info(&dev->dev, "BAR %d: can't allocate %pRf " + "(bogus alignment)\n", resno, res); return -EINVAL; } @@ -199,8 +196,8 @@ int pci_assign_resource(struct pci_dev *dev, int resno) } if (ret) - dev_info(&dev->dev, "BAR %d: can't allocate %s resource %pR\n", - resno, res->flags & IORESOURCE_IO ? "I/O" : "mem", res); + dev_info(&dev->dev, "BAR %d: can't allocate %pRt\n", + resno, res); return ret; } @@ -225,9 +222,8 @@ void pdev_sort_resources(struct pci_dev *dev, struct resource_list *head) r_align = pci_resource_alignment(dev, r); if (!r_align) { - dev_warn(&dev->dev, "BAR %d: bogus alignment " - "%pR flags %#lx\n", - i, r, r->flags); + dev_warn(&dev->dev, "BAR %d: bogus alignment %pRf\n", + i, r); continue; } for (list = head; ; list = list->next) { From 42887b29ced263ec3b8bd26ef157a324789b89d9 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Tue, 6 Oct 2009 15:33:49 -0600 Subject: [PATCH 013/109] x86/PCI: print resources consistently with %pRt This uses %pRt to print additional resource information (type, size, prefetchability, etc.) consistently. Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/pci/acpi.c | 14 +++++++++++--- arch/x86/pci/i386.c | 12 +++++------- 2 files changed, 16 insertions(+), 10 deletions(-) diff --git a/arch/x86/pci/acpi.c b/arch/x86/pci/acpi.c index 1014eb4bfc37..9b3daf976732 100644 --- a/arch/x86/pci/acpi.c +++ b/arch/x86/pci/acpi.c @@ -7,6 +7,7 @@ #include struct pci_root_info { + struct acpi_device *bridge; char *name; unsigned int res_num; struct resource *res; @@ -107,12 +108,18 @@ setup_resource(struct acpi_resource *acpi_res, void *data) res->child = NULL; if (insert_resource(root, res)) { - printk(KERN_ERR "PCI: Failed to allocate 0x%lx-0x%lx " - "from %s for %s\n", (unsigned long) res->start, - (unsigned long) res->end, root->name, info->name); + dev_err(&info->bridge->dev, "can't allocate %pRt\n", res); } else { info->bus->resource[info->res_num] = res; info->res_num++; + if (addr.translation_offset) + dev_info(&info->bridge->dev, "host bridge window: %pRt " + "(PCI address [%#llx-%#llx])\n", + res, res->start - addr.translation_offset, + res->end - addr.translation_offset); + else + dev_info(&info->bridge->dev, + "host bridge window: %pRt\n", res); } return AE_OK; } @@ -124,6 +131,7 @@ get_current_resources(struct acpi_device *device, int busnum, struct pci_root_info info; size_t size; + info.bridge = device; info.bus = bus; info.res_num = 0; acpi_walk_resources(device->handle, METHOD_NAME__CRS, count_resource, diff --git a/arch/x86/pci/i386.c b/arch/x86/pci/i386.c index b22d13b0c71d..a70a85de5e84 100644 --- a/arch/x86/pci/i386.c +++ b/arch/x86/pci/i386.c @@ -129,7 +129,7 @@ static void __init pcibios_allocate_bus_resources(struct list_head *bus_list) continue; if (!r->start || pci_claim_resource(dev, idx) < 0) { - dev_info(&dev->dev, "BAR %d: can't allocate resource\n", idx); + dev_info(&dev->dev, "BAR %d: can't allocate %pRt\n", idx, r); /* * Something is wrong with the region. * Invalidate the resource to prevent @@ -164,12 +164,10 @@ static void __init pcibios_allocate_resources(int pass) else disabled = !(command & PCI_COMMAND_MEMORY); if (pass == disabled) { - dev_dbg(&dev->dev, "resource %#08llx-%#08llx (f=%lx, d=%d, p=%d)\n", - (unsigned long long) r->start, - (unsigned long long) r->end, - r->flags, disabled, pass); + dev_dbg(&dev->dev, "%pRf (d=%d, p=%d)\n", r, + disabled, pass); if (pci_claim_resource(dev, idx) < 0) { - dev_info(&dev->dev, "BAR %d: can't allocate resource\n", idx); + dev_info(&dev->dev, "BAR %d: can't allocate %pRt\n", idx, r); /* We'll assign a new address later */ r->end -= r->start; r->start = 0; @@ -182,7 +180,7 @@ static void __init pcibios_allocate_resources(int pass) /* Turn the ROM off, leave the resource region, * but keep it unregistered. */ u32 reg; - dev_dbg(&dev->dev, "disabling ROM\n"); + dev_dbg(&dev->dev, "disabling ROM %pRt\n", r); r->flags &= ~IORESOURCE_ROM_ENABLE; pci_read_config_dword(dev, dev->rom_base_reg, ®); From 637b363e86f71effe74cba5d509f18bd3199a65b Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Tue, 6 Oct 2009 15:33:54 -0600 Subject: [PATCH 014/109] ia64/PCI: print resources consistently with %pRt This uses %pRt to print additional resource information (type, size, prefetchability, etc.) consistently. Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/ia64/pci/pci.c | 21 ++++++++++++++++----- 1 file changed, 16 insertions(+), 5 deletions(-) diff --git a/arch/ia64/pci/pci.c b/arch/ia64/pci/pci.c index d60e7195b7dd..06413b827e97 100644 --- a/arch/ia64/pci/pci.c +++ b/arch/ia64/pci/pci.c @@ -131,6 +131,7 @@ alloc_pci_controller (int seg) } struct pci_root_info { + struct acpi_device *bridge; struct pci_controller *controller; char *name; }; @@ -297,9 +298,19 @@ static __devinit acpi_status add_window(struct acpi_resource *res, void *data) window->offset = offset; if (insert_resource(root, &window->resource)) { - printk(KERN_ERR "alloc 0x%llx-0x%llx from %s for %s failed\n", - window->resource.start, window->resource.end, - root->name, info->name); + dev_err(&info->bridge->dev, "can't allocate %pRt\n", + &window->resource); + } else { + if (offset) + dev_info(&info->bridge->dev, "host bridge window: %pRt " + "(PCI address [%#llx-%#llx])\n", + &window->resource, + window->resource.start - offset, + window->resource.end - offset); + else + dev_info(&info->bridge->dev, + "host bridge window: %pRt\n", + &window->resource); } return AE_OK; @@ -319,8 +330,7 @@ pcibios_setup_root_windows(struct pci_bus *bus, struct pci_controller *ctrl) (res->end - res->start < 16)) continue; if (j >= PCI_BUS_NUM_RESOURCES) { - printk("Ignoring range [%#llx-%#llx] (%lx)\n", - res->start, res->end, res->flags); + dev_warn(&bus->dev, "ignoring %pRf (no space)\n", res); continue; } bus->resource[j++] = res; @@ -364,6 +374,7 @@ pci_acpi_scan_root(struct acpi_device *device, int domain, int bus) goto out3; sprintf(name, "PCI Bus %04x:%02x", domain, bus); + info.bridge = device; info.controller = controller; info.name = name; acpi_walk_resources(device->handle, METHOD_NAME__CRS, From 9a007b3791cdba3601d835ea10e68c14115b9afb Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Tue, 6 Oct 2009 15:34:00 -0600 Subject: [PATCH 015/109] PNP: print resources consistently with %pRt This uses %pRt and %pRf to print additional resource information (type, size, prefetchability, etc.) consistently. Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- drivers/pnp/quirks.c | 12 +++--------- drivers/pnp/resource.c | 10 ++++------ drivers/pnp/support.c | 43 +++++------------------------------------- drivers/pnp/system.c | 14 ++++++-------- 4 files changed, 18 insertions(+), 61 deletions(-) diff --git a/drivers/pnp/quirks.c b/drivers/pnp/quirks.c index 8473fe5ed7ff..3a2031b25c3c 100644 --- a/drivers/pnp/quirks.c +++ b/drivers/pnp/quirks.c @@ -285,15 +285,9 @@ static void quirk_system_pci_resources(struct pnp_dev *dev) * the PCI region, and that might prevent a PCI * driver from requesting its resources. */ - dev_warn(&dev->dev, "%s resource " - "(0x%llx-0x%llx) overlaps %s BAR %d " - "(0x%llx-0x%llx), disabling\n", - pnp_resource_type_name(res), - (unsigned long long) pnp_start, - (unsigned long long) pnp_end, - pci_name(pdev), i, - (unsigned long long) pci_start, - (unsigned long long) pci_end); + dev_warn(&dev->dev, "resource %pRt overlaps %s " + "BAR %d %pRt, disabling\n", res, + pci_name(pdev), i, &pdev->resource[i]); res->flags |= IORESOURCE_DISABLED; } } diff --git a/drivers/pnp/resource.c b/drivers/pnp/resource.c index ba9765427886..18557d7bb0b9 100644 --- a/drivers/pnp/resource.c +++ b/drivers/pnp/resource.c @@ -517,7 +517,7 @@ struct pnp_resource *pnp_add_irq_resource(struct pnp_dev *dev, int irq, res->start = irq; res->end = irq; - pnp_dbg(&dev->dev, " add irq %d flags %#x\n", irq, flags); + pnp_dbg(&dev->dev, " add %pRf\n", res); return pnp_res; } @@ -538,7 +538,7 @@ struct pnp_resource *pnp_add_dma_resource(struct pnp_dev *dev, int dma, res->start = dma; res->end = dma; - pnp_dbg(&dev->dev, " add dma %d flags %#x\n", dma, flags); + pnp_dbg(&dev->dev, " add %pRf\n", res); return pnp_res; } @@ -562,8 +562,7 @@ struct pnp_resource *pnp_add_io_resource(struct pnp_dev *dev, res->start = start; res->end = end; - pnp_dbg(&dev->dev, " add io %#llx-%#llx flags %#x\n", - (unsigned long long) start, (unsigned long long) end, flags); + pnp_dbg(&dev->dev, " add %pRf\n", res); return pnp_res; } @@ -587,8 +586,7 @@ struct pnp_resource *pnp_add_mem_resource(struct pnp_dev *dev, res->start = start; res->end = end; - pnp_dbg(&dev->dev, " add mem %#llx-%#llx flags %#x\n", - (unsigned long long) start, (unsigned long long) end, flags); + pnp_dbg(&dev->dev, " add %pRf\n", res); return pnp_res; } diff --git a/drivers/pnp/support.c b/drivers/pnp/support.c index 63087d5ce609..1f8a33b0abac 100644 --- a/drivers/pnp/support.c +++ b/drivers/pnp/support.c @@ -75,47 +75,14 @@ char *pnp_resource_type_name(struct resource *res) void dbg_pnp_show_resources(struct pnp_dev *dev, char *desc) { - char buf[128]; - int len; struct pnp_resource *pnp_res; - struct resource *res; - if (list_empty(&dev->resources)) { + if (list_empty(&dev->resources)) pnp_dbg(&dev->dev, "%s: no current resources\n", desc); - return; - } - - pnp_dbg(&dev->dev, "%s: current resources:\n", desc); - list_for_each_entry(pnp_res, &dev->resources, list) { - res = &pnp_res->res; - len = 0; - - len += scnprintf(buf + len, sizeof(buf) - len, " %-3s ", - pnp_resource_type_name(res)); - - if (res->flags & IORESOURCE_DISABLED) { - pnp_dbg(&dev->dev, "%sdisabled\n", buf); - continue; - } - - switch (pnp_resource_type(res)) { - case IORESOURCE_IO: - case IORESOURCE_MEM: - len += scnprintf(buf + len, sizeof(buf) - len, - "%#llx-%#llx flags %#lx", - (unsigned long long) res->start, - (unsigned long long) res->end, - res->flags); - break; - case IORESOURCE_IRQ: - case IORESOURCE_DMA: - len += scnprintf(buf + len, sizeof(buf) - len, - "%lld flags %#lx", - (unsigned long long) res->start, - res->flags); - break; - } - pnp_dbg(&dev->dev, "%s\n", buf); + else { + pnp_dbg(&dev->dev, "%s: current resources:\n", desc); + list_for_each_entry(pnp_res, &dev->resources, list) + pnp_dbg(&dev->dev, "%pRf\n", &pnp_res->res); } } diff --git a/drivers/pnp/system.c b/drivers/pnp/system.c index 59b90922da8c..242d3a872011 100644 --- a/drivers/pnp/system.c +++ b/drivers/pnp/system.c @@ -22,11 +22,11 @@ static const struct pnp_device_id pnp_dev_table[] = { {"", 0} }; -static void reserve_range(struct pnp_dev *dev, resource_size_t start, - resource_size_t end, int port) +static void reserve_range(struct pnp_dev *dev, struct resource *r, int port) { char *regionid; const char *pnpid = dev_name(&dev->dev); + resource_size_t start = r->start, end = r->end; struct resource *res; regionid = kmalloc(16, GFP_KERNEL); @@ -48,10 +48,8 @@ static void reserve_range(struct pnp_dev *dev, resource_size_t start, * example do reserve stuff they know about too, so we may well * have double reservations. */ - dev_info(&dev->dev, "%s range 0x%llx-0x%llx %s reserved\n", - port ? "ioport" : "iomem", - (unsigned long long) start, (unsigned long long) end, - res ? "has been" : "could not be"); + dev_info(&dev->dev, "%pRt %s reserved\n", r, + res ? "has been" : "could not be"); } static void reserve_resources_of_dev(struct pnp_dev *dev) @@ -77,14 +75,14 @@ static void reserve_resources_of_dev(struct pnp_dev *dev) if (res->end < res->start) continue; /* invalid */ - reserve_range(dev, res->start, res->end, 1); + reserve_range(dev, res, 1); } for (i = 0; (res = pnp_get_resource(dev, IORESOURCE_MEM, i)); i++) { if (res->flags & IORESOURCE_DISABLED) continue; - reserve_range(dev, res->start, res->end, 0); + reserve_range(dev, res, 0); } } From 2992e545ea006992ec9dc91c4fa996ce1e15f921 Mon Sep 17 00:00:00 2001 From: Suresh Siddha Date: Mon, 26 Oct 2009 13:21:32 -0800 Subject: [PATCH 016/109] x86/PCI/PAT: return EINVAL for pci mmap WC request for !pat_enabled Thomas Schlichter reported: > X.org uses libpciaccess which tries to mmap with write combining enabled via > /sys/bus/pci/devices/*/resource0_wc. Currently, when PAT is not enabled, the > kernel does fall back to uncached mmap. Then libpciaccess thinks it succeeded > mapping with write combining enabled and does not set up suited MTRR entries. > ;-( Instead of silently mapping pci mmap region as UC minus in the case of !pat_enabled and wc request, we can return error. Eric Anholt mentioned that caller (like X) typically follows up with UC minus pci mmap request and if there is a free mtrr slot, caller will manage adding WC mtrr. Jesse Barnes says: > Older versions of libpciaccess will behave better if we do it that way > (iirc it only allocates an MTRR if the resource_wc file doesn't exist or > fails to get mapped). Reported-by: Thomas Schlichter Signed-off-by: Thomas Schlichter Signed-off-by: Suresh Siddha Acked-by: Eric Anholt Acked-by: Jesse Barnes Signed-off-by: Jesse Barnes --- arch/x86/pci/i386.c | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/arch/x86/pci/i386.c b/arch/x86/pci/i386.c index a70a85de5e84..52e656f17781 100644 --- a/arch/x86/pci/i386.c +++ b/arch/x86/pci/i386.c @@ -280,6 +280,15 @@ int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma, return -EINVAL; prot = pgprot_val(vma->vm_page_prot); + + /* + * Return error if pat is not enabled and write_combine is requested. + * Caller can followup with UC MINUS request and add a WC mtrr if there + * is a free mtrr slot. + */ + if (!pat_enabled && write_combine) + return -EINVAL; + if (pat_enabled && write_combine) prot |= _PAGE_CACHE_WC; else if (pat_enabled || boot_cpu_data.x86 > 3) From 9a08f7d3506019e3833cd4394ca0d7da0ae3689f Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Fri, 23 Oct 2009 15:20:33 -0600 Subject: [PATCH 017/109] x86/PCI: allow MMCONFIG above 4GB The current whitelist requires a kernel change for every machine that has MMCONFIG regions above 4GB, even if BIOS provides a correct MCFG table. This patch expands the whitelist to include machines with a rev 1 or newer MCFG table and a DMI_BIOS_DATE of 2010 or later. That way, we only need kernel changes for new machines that provide incorrect MCFG tables. Signed-off-by: Bjorn Helgaas CC: Matthew Wilcox CC: John Keller CC: Yinghai Lu CC: Kenji Kaneshige CC: Andi Kleen Acked-by: Ingo Molnar Signed-off-by: Jesse Barnes --- arch/x86/pci/mmconfig-shared.c | 33 +++++++++++++++++++++------------ 1 file changed, 21 insertions(+), 12 deletions(-) diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index 602c172d3bd5..02642773c29d 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -15,6 +15,7 @@ #include #include #include +#include #include #include #include @@ -527,18 +528,31 @@ reject: static int __initdata known_bridge; -static int acpi_mcfg_64bit_base_addr __initdata = FALSE; - /* The physical address of the MMCONFIG aperture. Set from ACPI tables. */ struct acpi_mcfg_allocation *pci_mmcfg_config; int pci_mmcfg_config_num; -static int __init acpi_mcfg_oem_check(struct acpi_table_mcfg *mcfg) +static int __init acpi_mcfg_check_entry(struct acpi_table_mcfg *mcfg, + struct acpi_mcfg_allocation *cfg) { - if (!strcmp(mcfg->header.oem_id, "SGI")) - acpi_mcfg_64bit_base_addr = TRUE; + int year; - return 0; + if (cfg->address < 0xFFFFFFFF) + return 0; + + if (!strcmp(mcfg->header.oem_id, "SGI")) + return 0; + + if (mcfg->header.revision >= 1) { + if (dmi_get_date(DMI_BIOS_DATE, &year, NULL, NULL) && + year >= 2010) + return 0; + } + + printk(KERN_ERR PREFIX "MCFG region for %04x:%02x-%02x at %#llx " + "is above 4GB, ignored\n", cfg->pci_segment, + cfg->start_bus_number, cfg->end_bus_number, cfg->address); + return -EINVAL; } static int __init pci_parse_mcfg(struct acpi_table_header *header) @@ -574,13 +588,8 @@ static int __init pci_parse_mcfg(struct acpi_table_header *header) memcpy(pci_mmcfg_config, &mcfg[1], config_size); - acpi_mcfg_oem_check(mcfg); - for (i = 0; i < pci_mmcfg_config_num; ++i) { - if ((pci_mmcfg_config[i].address > 0xFFFFFFFF) && - !acpi_mcfg_64bit_base_addr) { - printk(KERN_ERR PREFIX - "MMCONFIG not in low 4GB of memory\n"); + if (acpi_mcfg_check_entry(mcfg, &pci_mmcfg_config[i])) { kfree(pci_mmcfg_config); pci_mmcfg_config_num = 0; return -ENODEV; From 1ccbf5344c3daef046d2323190cc6807c44f1917 Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Tue, 6 Oct 2009 15:11:14 -0700 Subject: [PATCH 018/109] xen: move Xen-testing predicates to common header Move xen_domain and related tests out of asm-x86 to xen/xen.h so they can be included whenever they are necessary. Signed-off-by: Jeremy Fitzhardinge Signed-off-by: Jesse Barnes --- arch/x86/include/asm/xen/hypervisor.h | 27 ---------------------- arch/x86/xen/enlighten.c | 1 + drivers/block/xen-blkfront.c | 1 + drivers/char/hvc_xen.c | 2 ++ drivers/input/xen-kbdfront.c | 3 +++ drivers/net/xen-netfront.c | 1 + drivers/video/xen-fbfront.c | 3 +++ drivers/xen/balloon.c | 2 ++ drivers/xen/cpu_hotplug.c | 1 + drivers/xen/evtchn.c | 2 ++ drivers/xen/grant-table.c | 1 + drivers/xen/sys-hypervisor.c | 1 + drivers/xen/xenbus/xenbus_probe.c | 2 ++ drivers/xen/xenfs/super.c | 2 ++ include/xen/xen.h | 32 +++++++++++++++++++++++++++ 15 files changed, 54 insertions(+), 27 deletions(-) create mode 100644 include/xen/xen.h diff --git a/arch/x86/include/asm/xen/hypervisor.h b/arch/x86/include/asm/xen/hypervisor.h index d5b7e90c0edf..396ff4cc8ed4 100644 --- a/arch/x86/include/asm/xen/hypervisor.h +++ b/arch/x86/include/asm/xen/hypervisor.h @@ -37,31 +37,4 @@ extern struct shared_info *HYPERVISOR_shared_info; extern struct start_info *xen_start_info; -enum xen_domain_type { - XEN_NATIVE, /* running on bare hardware */ - XEN_PV_DOMAIN, /* running in a PV domain */ - XEN_HVM_DOMAIN, /* running in a Xen hvm domain */ -}; - -#ifdef CONFIG_XEN -extern enum xen_domain_type xen_domain_type; -#else -#define xen_domain_type XEN_NATIVE -#endif - -#define xen_domain() (xen_domain_type != XEN_NATIVE) -#define xen_pv_domain() (xen_domain() && \ - xen_domain_type == XEN_PV_DOMAIN) -#define xen_hvm_domain() (xen_domain() && \ - xen_domain_type == XEN_HVM_DOMAIN) - -#ifdef CONFIG_XEN_DOM0 -#include - -#define xen_initial_domain() (xen_pv_domain() && \ - xen_start_info->flags & SIF_INITDOMAIN) -#else /* !CONFIG_XEN_DOM0 */ -#define xen_initial_domain() (0) -#endif /* CONFIG_XEN_DOM0 */ - #endif /* _ASM_X86_XEN_HYPERVISOR_H */ diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index 23a4d80fb39e..5bccd706232c 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -28,6 +28,7 @@ #include #include +#include #include #include #include diff --git a/drivers/block/xen-blkfront.c b/drivers/block/xen-blkfront.c index b8578bb3f4c9..05a31e55d278 100644 --- a/drivers/block/xen-blkfront.c +++ b/drivers/block/xen-blkfront.c @@ -42,6 +42,7 @@ #include #include +#include #include #include #include diff --git a/drivers/char/hvc_xen.c b/drivers/char/hvc_xen.c index eba999f8598d..93d33816c9d9 100644 --- a/drivers/char/hvc_xen.c +++ b/drivers/char/hvc_xen.c @@ -25,6 +25,8 @@ #include #include + +#include #include #include #include diff --git a/drivers/input/xen-kbdfront.c b/drivers/input/xen-kbdfront.c index b115726dc088..c721c0a23eb8 100644 --- a/drivers/input/xen-kbdfront.c +++ b/drivers/input/xen-kbdfront.c @@ -21,7 +21,10 @@ #include #include #include + #include + +#include #include #include #include diff --git a/drivers/net/xen-netfront.c b/drivers/net/xen-netfront.c index baa051d5bfbe..a869b45d3d37 100644 --- a/drivers/net/xen-netfront.c +++ b/drivers/net/xen-netfront.c @@ -42,6 +42,7 @@ #include #include +#include #include #include #include diff --git a/drivers/video/xen-fbfront.c b/drivers/video/xen-fbfront.c index 54cd91610174..966b226c858c 100644 --- a/drivers/video/xen-fbfront.c +++ b/drivers/video/xen-fbfront.c @@ -25,7 +25,10 @@ #include #include #include + #include + +#include #include #include #include diff --git a/drivers/xen/balloon.c b/drivers/xen/balloon.c index d31505b6f7a4..826dda414166 100644 --- a/drivers/xen/balloon.c +++ b/drivers/xen/balloon.c @@ -52,6 +52,8 @@ #include #include + +#include #include #include #include diff --git a/drivers/xen/cpu_hotplug.c b/drivers/xen/cpu_hotplug.c index bdfd584ad853..6625ffe1a6fd 100644 --- a/drivers/xen/cpu_hotplug.c +++ b/drivers/xen/cpu_hotplug.c @@ -1,5 +1,6 @@ #include +#include #include #include diff --git a/drivers/xen/evtchn.c b/drivers/xen/evtchn.c index 79bedba44fee..f70a4f4698c5 100644 --- a/drivers/xen/evtchn.c +++ b/drivers/xen/evtchn.c @@ -48,6 +48,8 @@ #include #include #include + +#include #include #include #include diff --git a/drivers/xen/grant-table.c b/drivers/xen/grant-table.c index 7d8f531fb8e8..4c6c0bd636a8 100644 --- a/drivers/xen/grant-table.c +++ b/drivers/xen/grant-table.c @@ -37,6 +37,7 @@ #include #include +#include #include #include #include diff --git a/drivers/xen/sys-hypervisor.c b/drivers/xen/sys-hypervisor.c index 88a60e03ccf0..ae5cb05a1a1c 100644 --- a/drivers/xen/sys-hypervisor.c +++ b/drivers/xen/sys-hypervisor.c @@ -14,6 +14,7 @@ #include #include +#include #include #include #include diff --git a/drivers/xen/xenbus/xenbus_probe.c b/drivers/xen/xenbus/xenbus_probe.c index d42e25d5968d..08638adec9fc 100644 --- a/drivers/xen/xenbus/xenbus_probe.c +++ b/drivers/xen/xenbus/xenbus_probe.c @@ -49,6 +49,8 @@ #include #include #include + +#include #include #include #include diff --git a/drivers/xen/xenfs/super.c b/drivers/xen/xenfs/super.c index 6559e0c752ce..8924d93136f1 100644 --- a/drivers/xen/xenfs/super.c +++ b/drivers/xen/xenfs/super.c @@ -13,6 +13,8 @@ #include #include +#include + #include "xenfs.h" #include diff --git a/include/xen/xen.h b/include/xen/xen.h new file mode 100644 index 000000000000..a16402418d31 --- /dev/null +++ b/include/xen/xen.h @@ -0,0 +1,32 @@ +#ifndef _XEN_XEN_H +#define _XEN_XEN_H + +enum xen_domain_type { + XEN_NATIVE, /* running on bare hardware */ + XEN_PV_DOMAIN, /* running in a PV domain */ + XEN_HVM_DOMAIN, /* running in a Xen hvm domain */ +}; + +#ifdef CONFIG_XEN +extern enum xen_domain_type xen_domain_type; +#else +#define xen_domain_type XEN_NATIVE +#endif + +#define xen_domain() (xen_domain_type != XEN_NATIVE) +#define xen_pv_domain() (xen_domain() && \ + xen_domain_type == XEN_PV_DOMAIN) +#define xen_hvm_domain() (xen_domain() && \ + xen_domain_type == XEN_HVM_DOMAIN) + +#ifdef CONFIG_XEN_DOM0 +#include +#include + +#define xen_initial_domain() (xen_pv_domain() && \ + xen_start_info->flags & SIF_INITDOMAIN) +#else /* !CONFIG_XEN_DOM0 */ +#define xen_initial_domain() (0) +#endif /* CONFIG_XEN_DOM0 */ + +#endif /* _XEN_XEN_H */ From ae21ee65e8bc228416bbcc8a1da01c56a847a60c Mon Sep 17 00:00:00 2001 From: Allen Kay Date: Wed, 7 Oct 2009 10:27:17 -0700 Subject: [PATCH 019/109] PCI: acs p2p upsteram forwarding enabling Note: dom0 checking in v4 has been separated out into 2/2. This patch enables P2P upstream forwarding in ACS capable PCIe switches. It solves two potential problems in virtualization environment where a PCIe device is assigned to a guest domain using a HW iommu such as VT-d: 1) Unintentional failure caused by guest physical address programmed into the device's DMA that happens to match the memory address range of other downstream ports in the same PCIe switch. This causes the PCI transaction to go to the matching downstream port instead of go to the root complex to get translated by VT-d as it should be. 2) Malicious guest software intentionally attacks another downstream PCIe device by programming the DMA address into the assigned device that matches memory address range of the downstream PCIe port. We are in process of implementing device filtering software in KVM/XEN management software to allow device assignment of PCIe devices behind a PCIe switch only if it has ACS capability and with the P2P upstream forwarding bits enabled. This patch is intended to work for both KVM and Xen environments. Signed-off-by: Allen Kay Reviewed-by: Mathew Wilcox Reviewed-by: Chris Wright Signed-off-by: Jesse Barnes --- drivers/pci/pci.c | 35 +++++++++++++++++++++++++++++++++++ drivers/pci/pci.h | 2 ++ drivers/pci/probe.c | 5 +++++ include/linux/pci_regs.h | 13 +++++++++++++ 4 files changed, 55 insertions(+) diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index 4859669f0ab5..557218222826 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -1545,6 +1545,41 @@ void pci_enable_ari(struct pci_dev *dev) bridge->ari_enabled = 1; } +/** + * pci_enable_acs - enable ACS if hardware support it + * @dev: the PCI device + */ +void pci_enable_acs(struct pci_dev *dev) +{ + int pos; + u16 cap; + u16 ctrl; + + if (!dev->is_pcie) + return; + + pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ACS); + if (!pos) + return; + + pci_read_config_word(dev, pos + PCI_ACS_CAP, &cap); + pci_read_config_word(dev, pos + PCI_ACS_CTRL, &ctrl); + + /* Source Validation */ + ctrl |= (cap & PCI_ACS_SV); + + /* P2P Request Redirect */ + ctrl |= (cap & PCI_ACS_RR); + + /* P2P Completion Redirect */ + ctrl |= (cap & PCI_ACS_CR); + + /* Upstream Forwarding */ + ctrl |= (cap & PCI_ACS_UF); + + pci_write_config_word(dev, pos + PCI_ACS_CTRL, ctrl); +} + /** * pci_swizzle_interrupt_pin - swizzle INTx for device behind bridge * @dev: the PCI device diff --git a/drivers/pci/pci.h b/drivers/pci/pci.h index d92d1954a2fb..33ed8e0aba1e 100644 --- a/drivers/pci/pci.h +++ b/drivers/pci/pci.h @@ -311,4 +311,6 @@ static inline int pci_resource_alignment(struct pci_dev *dev, return resource_alignment(res); } +extern void pci_enable_acs(struct pci_dev *dev); + #endif /* DRIVERS_PCI_H */ diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 2adb47574d86..aac5b156a5c5 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -10,6 +10,7 @@ #include #include #include +#include #include "pci.h" #define CARDBUS_LATENCY_TIMER 176 /* secondary latency timer */ @@ -1004,6 +1005,10 @@ static void pci_init_capabilities(struct pci_dev *dev) /* Single Root I/O Virtualization */ pci_iov_init(dev); + + /* Enable ACS P2P upstream forwarding */ + if (iommu_found()) + pci_enable_acs(dev); } void pci_device_add(struct pci_dev *dev, struct pci_bus *bus) diff --git a/include/linux/pci_regs.h b/include/linux/pci_regs.h index dd0bed4f1cf0..d798770f08cd 100644 --- a/include/linux/pci_regs.h +++ b/include/linux/pci_regs.h @@ -502,6 +502,7 @@ #define PCI_EXT_CAP_ID_VC 2 #define PCI_EXT_CAP_ID_DSN 3 #define PCI_EXT_CAP_ID_PWR 4 +#define PCI_EXT_CAP_ID_ACS 13 #define PCI_EXT_CAP_ID_ARI 14 #define PCI_EXT_CAP_ID_ATS 15 #define PCI_EXT_CAP_ID_SRIOV 16 @@ -662,4 +663,16 @@ #define PCI_SRIOV_VFM_MO 0x2 /* Active.MigrateOut */ #define PCI_SRIOV_VFM_AV 0x3 /* Active.Available */ +/* Access Control Service */ +#define PCI_ACS_CAP 0x04 /* ACS Capability Register */ +#define PCI_ACS_SV 0x01 /* Source Validation */ +#define PCI_ACS_TB 0x02 /* Translation Blocking */ +#define PCI_ACS_RR 0x04 /* P2P Request Redirect */ +#define PCI_ACS_CR 0x08 /* P2P Completion Redirect */ +#define PCI_ACS_UF 0x10 /* Upstream Forwarding */ +#define PCI_ACS_EC 0x20 /* P2P Egress Control */ +#define PCI_ACS_DT 0x40 /* Direct Translated P2P */ +#define PCI_ACS_CTRL 0x06 /* ACS Control Register */ +#define PCI_ACS_EGRESS_CTL_V 0x08 /* ACS Egress Control Vector */ + #endif /* LINUX_PCI_REGS_H */ From df0e97c6f1f2fdca686036998fe816cefd8e27d7 Mon Sep 17 00:00:00 2001 From: Allen Kay Date: Wed, 7 Oct 2009 10:27:51 -0700 Subject: [PATCH 020/109] PCI: add xen dom0 checking before ACS initialization This patch is predicated on Jeremy's patch in include/xen/xen.h. It'll prevent ACS init unless the platform has both an IOMMU and we're running as dom0. Signed-off-by: Allen Kay Signed-off-by: Jesse Barnes --- drivers/pci/probe.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index aac5b156a5c5..bb2cc39b64ff 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -11,6 +11,7 @@ #include #include #include +#include #include "pci.h" #define CARDBUS_LATENCY_TIMER 176 /* secondary latency timer */ @@ -1007,7 +1008,7 @@ static void pci_init_capabilities(struct pci_dev *dev) pci_iov_init(dev); /* Enable ACS P2P upstream forwarding */ - if (iommu_found()) + if (iommu_found() || xen_initial_domain()) pci_enable_acs(dev); } From af5a8ee05404112f38fb2904747c688bdc31a746 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Wed, 14 Oct 2009 10:27:42 -0600 Subject: [PATCH 021/109] x86/PCI: use -DDEBUG when CONFIG_PCI_DEBUG set We use dev_dbg() in arch/x86/pci, but there's no easy way to turn it on. Add -DDEBUG when CONFIG_PCI_DEBUG=y, just like we do in drivers/pci. Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/pci/Makefile | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/arch/x86/pci/Makefile b/arch/x86/pci/Makefile index 56d917b556c6..d8a0a6279a4d 100644 --- a/arch/x86/pci/Makefile +++ b/arch/x86/pci/Makefile @@ -16,3 +16,7 @@ obj-$(CONFIG_X86_NUMAQ) += numaq_32.o obj-y += common.o early.o obj-y += amd_bus.o obj-$(CONFIG_X86_64) += intel_bus.o + +ifeq ($(CONFIG_PCI_DEBUG),y) +EXTRA_CFLAGS += -DDEBUG +endif From 1ed6743918abbec69c0f0b725fa56e3c3248bbab Mon Sep 17 00:00:00 2001 From: "Michael S. Tsirkin" Date: Thu, 29 Oct 2009 17:24:59 +0200 Subject: [PATCH 022/109] PCI: fix nit in ROM BAR size probing When probing for ROM BAR size, we should not change bits 1:10 in this BAR, because these bits are marked as "reserved for future use" in PCI spec, so changing them might have side effects. No such issue for I/O or memory, as there is an implementation note in PCI spec which explicitly allows writing 0xfffffffff there. Signed-off-by: Michael S. Tsirkin Signed-off-by: Jesse Barnes --- drivers/pci/probe.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index bb2cc39b64ff..9cefc54a0125 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -165,12 +165,12 @@ int __pci_read_base(struct pci_dev *dev, enum pci_bar_type type, { u32 l, sz, mask; - mask = type ? ~PCI_ROM_ADDRESS_ENABLE : ~0; + mask = type ? PCI_ROM_ADDRESS_MASK : ~0; res->name = pci_name(dev); pci_read_config_dword(dev, pos, &l); - pci_write_config_dword(dev, pos, mask); + pci_write_config_dword(dev, pos, l | mask); pci_read_config_dword(dev, pos, &sz); pci_write_config_dword(dev, pos, l); From f22daf1fb9970f1565f224a0951ba58b5d044605 Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Mon, 5 Oct 2009 17:40:02 +0900 Subject: [PATCH 023/109] PCI: pciehp: disable DLL state changed event notification Current pciehp doesn't handle Data Link Layer State Changed Event notification. So it needs to be disabled at initialization time, otherwise other event notifications are not generated. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/hotplug/pciehp_hpc.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/pci/hotplug/pciehp_hpc.c b/drivers/pci/hotplug/pciehp_hpc.c index 9ef4605c1ef6..36f8545a2813 100644 --- a/drivers/pci/hotplug/pciehp_hpc.c +++ b/drivers/pci/hotplug/pciehp_hpc.c @@ -866,7 +866,8 @@ static void pcie_disable_notification(struct controller *ctrl) u16 mask; mask = (PCI_EXP_SLTCTL_PDCE | PCI_EXP_SLTCTL_ABPE | PCI_EXP_SLTCTL_MRLSCE | PCI_EXP_SLTCTL_PFDE | - PCI_EXP_SLTCTL_HPIE | PCI_EXP_SLTCTL_CCIE); + PCI_EXP_SLTCTL_HPIE | PCI_EXP_SLTCTL_CCIE | + PCI_EXP_SLTCTL_DLLSCE); if (pcie_write_cmd(ctrl, 0, mask)) ctrl_warn(ctrl, "Cannot disable software notification\n"); } From 3c3a1b1759616e6603027108f8abcbec54271e62 Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Mon, 5 Oct 2009 17:40:48 +0900 Subject: [PATCH 024/109] PCI: pciehp: remove wrong workaround for bad DLLP Remove wrong workaround for BAD DLLP error, which confused surprise down error with DLL errors. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/hotplug/pciehp_hpc.c | 53 ++------------------------------ 1 file changed, 3 insertions(+), 50 deletions(-) diff --git a/drivers/pci/hotplug/pciehp_hpc.c b/drivers/pci/hotplug/pciehp_hpc.c index 36f8545a2813..88b654ec5afc 100644 --- a/drivers/pci/hotplug/pciehp_hpc.c +++ b/drivers/pci/hotplug/pciehp_hpc.c @@ -535,54 +535,12 @@ int pciehp_power_on_slot(struct slot * slot) return retval; } -static inline int pcie_mask_bad_dllp(struct controller *ctrl) -{ - struct pci_dev *dev = ctrl->pcie->port; - int pos; - u32 reg; - - pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ERR); - if (!pos) - return 0; - pci_read_config_dword(dev, pos + PCI_ERR_COR_MASK, ®); - if (reg & PCI_ERR_COR_BAD_DLLP) - return 0; - reg |= PCI_ERR_COR_BAD_DLLP; - pci_write_config_dword(dev, pos + PCI_ERR_COR_MASK, reg); - return 1; -} - -static inline void pcie_unmask_bad_dllp(struct controller *ctrl) -{ - struct pci_dev *dev = ctrl->pcie->port; - u32 reg; - int pos; - - pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ERR); - if (!pos) - return; - pci_read_config_dword(dev, pos + PCI_ERR_COR_MASK, ®); - if (!(reg & PCI_ERR_COR_BAD_DLLP)) - return; - reg &= ~PCI_ERR_COR_BAD_DLLP; - pci_write_config_dword(dev, pos + PCI_ERR_COR_MASK, reg); -} - int pciehp_power_off_slot(struct slot * slot) { struct controller *ctrl = slot->ctrl; u16 slot_cmd; u16 cmd_mask; - int retval = 0; - int changed; - - /* - * Set Bad DLLP Mask bit in Correctable Error Mask - * Register. This is the workaround against Bad DLLP error - * that sometimes happens during turning power off the slot - * which conforms to PCI Express 1.0a spec. - */ - changed = pcie_mask_bad_dllp(ctrl); + int retval; slot_cmd = POWER_OFF; cmd_mask = PCI_EXP_SLTCTL_PCC; @@ -595,16 +553,11 @@ int pciehp_power_off_slot(struct slot * slot) retval = pcie_write_cmd(ctrl, slot_cmd, cmd_mask); if (retval) { ctrl_err(ctrl, "Write command failed!\n"); - retval = -1; - goto out; + return retval; } ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n", __func__, ctrl->cap_base + PCI_EXP_SLTCTL, slot_cmd); - out: - if (changed) - pcie_unmask_bad_dllp(ctrl); - - return retval; + return 0; } static irqreturn_t pcie_isr(int irq, void *dev_id) From 586f1d6688c68a6c7fa4e6a00fa3968b16daef75 Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Mon, 5 Oct 2009 17:41:37 +0900 Subject: [PATCH 025/109] PCI: pciehp: create files only for existing capabilities Current pciehp driver creates 'attention' and 'latch' files even if the controller doesn't support them. In this case, the contents of those files are meaningless and unpredictable. Those files should be created only if the controller has the corresponding capabilities. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/hotplug/pciehp_core.c | 38 ++++++++++++++++++------------- 1 file changed, 22 insertions(+), 16 deletions(-) diff --git a/drivers/pci/hotplug/pciehp_core.c b/drivers/pci/hotplug/pciehp_core.c index bc234719b1df..77feafd39e29 100644 --- a/drivers/pci/hotplug/pciehp_core.c +++ b/drivers/pci/hotplug/pciehp_core.c @@ -72,18 +72,6 @@ static int get_adapter_status (struct hotplug_slot *slot, u8 *value); static int get_max_bus_speed (struct hotplug_slot *slot, enum pci_bus_speed *value); static int get_cur_bus_speed (struct hotplug_slot *slot, enum pci_bus_speed *value); -static struct hotplug_slot_ops pciehp_hotplug_slot_ops = { - .set_attention_status = set_attention_status, - .enable_slot = enable_slot, - .disable_slot = disable_slot, - .get_power_status = get_power_status, - .get_attention_status = get_attention_status, - .get_latch_status = get_latch_status, - .get_adapter_status = get_adapter_status, - .get_max_bus_speed = get_max_bus_speed, - .get_cur_bus_speed = get_cur_bus_speed, -}; - /** * release_slot - free up the memory used by a slot * @hotplug_slot: slot to free @@ -95,6 +83,7 @@ static void release_slot(struct hotplug_slot *hotplug_slot) ctrl_dbg(slot->ctrl, "%s: physical_slot = %s\n", __func__, hotplug_slot_name(hotplug_slot)); + kfree(hotplug_slot->ops); kfree(hotplug_slot->info); kfree(hotplug_slot); } @@ -104,6 +93,7 @@ static int init_slot(struct controller *ctrl) struct slot *slot = ctrl->slot; struct hotplug_slot *hotplug = NULL; struct hotplug_slot_info *info = NULL; + struct hotplug_slot_ops *ops = NULL; char name[SLOT_NAME_SIZE]; int retval = -ENOMEM; @@ -115,11 +105,28 @@ static int init_slot(struct controller *ctrl) if (!info) goto out; + /* Setup hotplug slot ops */ + ops = kzalloc(sizeof(*ops), GFP_KERNEL); + if (!ops) + goto out; + ops->enable_slot = enable_slot; + ops->disable_slot = disable_slot; + ops->get_power_status = get_power_status; + ops->get_adapter_status = get_adapter_status; + ops->get_max_bus_speed = get_max_bus_speed; + ops->get_cur_bus_speed = get_cur_bus_speed; + if (MRL_SENS(ctrl)) + ops->get_latch_status = get_latch_status; + if (ATTN_LED(ctrl)) { + ops->get_attention_status = get_attention_status; + ops->set_attention_status = set_attention_status; + } + /* register this slot with the hotplug pci core */ hotplug->info = info; hotplug->private = slot; hotplug->release = &release_slot; - hotplug->ops = &pciehp_hotplug_slot_ops; + hotplug->ops = ops; slot->hotplug_slot = hotplug; snprintf(name, SLOT_NAME_SIZE, "%u", PSN(ctrl)); @@ -139,6 +146,7 @@ static int init_slot(struct controller *ctrl) get_adapter_status(hotplug, &info->adapter_status); out: if (retval) { + kfree(ops); kfree(info); kfree(hotplug); } @@ -161,9 +169,7 @@ static int set_attention_status(struct hotplug_slot *hotplug_slot, u8 status) __func__, slot_name(slot)); hotplug_slot->info->attention_status = status; - - if (ATTN_LED(slot->ctrl)) - pciehp_set_attention_status(slot, status); + pciehp_set_attention_status(slot, status); return 0; } From 445f798555e218a5601222ca5849e8553ddd866a Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Mon, 5 Oct 2009 17:42:59 +0900 Subject: [PATCH 026/109] PCI: pciehp: return error on read/write failure Current pciehp returns successfully on read/write failure with dummy state values. It should return error instead. With this patch, pciehp no longer uses hotplug_slot_info data structure. So this also removes hotplug_slot_info related code. But note that it still allocates hotplug_slot_info because it is required by pci hotplug core. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/hotplug/pciehp_core.c | 55 +++++-------------------------- drivers/pci/hotplug/pciehp_ctrl.c | 30 +---------------- drivers/pci/hotplug/pciehp_hpc.c | 27 +++++++-------- 3 files changed, 21 insertions(+), 91 deletions(-) diff --git a/drivers/pci/hotplug/pciehp_core.c b/drivers/pci/hotplug/pciehp_core.c index 77feafd39e29..d19f87594df4 100644 --- a/drivers/pci/hotplug/pciehp_core.c +++ b/drivers/pci/hotplug/pciehp_core.c @@ -135,15 +135,9 @@ static int init_slot(struct controller *ctrl) ctrl->pcie->port->subordinate->number, PSN(ctrl)); retval = pci_hp_register(hotplug, ctrl->pcie->port->subordinate, 0, name); - if (retval) { + if (retval) ctrl_err(ctrl, "pci_hp_register failed with error %d\n", retval); - goto out; - } - get_power_status(hotplug, &info->power_status); - get_attention_status(hotplug, &info->attention_status); - get_latch_status(hotplug, &info->latch_status); - get_adapter_status(hotplug, &info->adapter_status); out: if (retval) { kfree(ops); @@ -168,10 +162,7 @@ static int set_attention_status(struct hotplug_slot *hotplug_slot, u8 status) ctrl_dbg(slot->ctrl, "%s: physical_slot = %s\n", __func__, slot_name(slot)); - hotplug_slot->info->attention_status = status; - pciehp_set_attention_status(slot, status); - - return 0; + return pciehp_set_attention_status(slot, status); } @@ -199,92 +190,62 @@ static int disable_slot(struct hotplug_slot *hotplug_slot) static int get_power_status(struct hotplug_slot *hotplug_slot, u8 *value) { struct slot *slot = hotplug_slot->private; - int retval; ctrl_dbg(slot->ctrl, "%s: physical_slot = %s\n", __func__, slot_name(slot)); - retval = pciehp_get_power_status(slot, value); - if (retval < 0) - *value = hotplug_slot->info->power_status; - - return 0; + return pciehp_get_power_status(slot, value); } static int get_attention_status(struct hotplug_slot *hotplug_slot, u8 *value) { struct slot *slot = hotplug_slot->private; - int retval; ctrl_dbg(slot->ctrl, "%s: physical_slot = %s\n", __func__, slot_name(slot)); - retval = pciehp_get_attention_status(slot, value); - if (retval < 0) - *value = hotplug_slot->info->attention_status; - - return 0; + return pciehp_get_attention_status(slot, value); } static int get_latch_status(struct hotplug_slot *hotplug_slot, u8 *value) { struct slot *slot = hotplug_slot->private; - int retval; ctrl_dbg(slot->ctrl, "%s: physical_slot = %s\n", __func__, slot_name(slot)); - retval = pciehp_get_latch_status(slot, value); - if (retval < 0) - *value = hotplug_slot->info->latch_status; - - return 0; + return pciehp_get_latch_status(slot, value); } static int get_adapter_status(struct hotplug_slot *hotplug_slot, u8 *value) { struct slot *slot = hotplug_slot->private; - int retval; ctrl_dbg(slot->ctrl, "%s: physical_slot = %s\n", __func__, slot_name(slot)); - retval = pciehp_get_adapter_status(slot, value); - if (retval < 0) - *value = hotplug_slot->info->adapter_status; - - return 0; + return pciehp_get_adapter_status(slot, value); } static int get_max_bus_speed(struct hotplug_slot *hotplug_slot, enum pci_bus_speed *value) { struct slot *slot = hotplug_slot->private; - int retval; ctrl_dbg(slot->ctrl, "%s: physical_slot = %s\n", __func__, slot_name(slot)); - retval = pciehp_get_max_link_speed(slot, value); - if (retval < 0) - *value = PCI_SPEED_UNKNOWN; - - return 0; + return pciehp_get_max_link_speed(slot, value); } static int get_cur_bus_speed(struct hotplug_slot *hotplug_slot, enum pci_bus_speed *value) { struct slot *slot = hotplug_slot->private; - int retval; ctrl_dbg(slot->ctrl, "%s: physical_slot = %s\n", __func__, slot_name(slot)); - retval = pciehp_get_cur_link_speed(slot, value); - if (retval < 0) - *value = PCI_SPEED_UNKNOWN; - - return 0; + return pciehp_get_cur_link_speed(slot, value); } static int pciehp_probe(struct pcie_device *dev) diff --git a/drivers/pci/hotplug/pciehp_ctrl.c b/drivers/pci/hotplug/pciehp_ctrl.c index 84487d126e4d..15ce2a3cc0f1 100644 --- a/drivers/pci/hotplug/pciehp_ctrl.c +++ b/drivers/pci/hotplug/pciehp_ctrl.c @@ -363,25 +363,6 @@ void pciehp_queue_pushbutton_work(struct work_struct *work) mutex_unlock(&p_slot->lock); } -static int update_slot_info(struct slot *slot) -{ - struct hotplug_slot_info *info; - int result; - - info = kmalloc(sizeof(*info), GFP_KERNEL); - if (!info) - return -ENOMEM; - - pciehp_get_power_status(slot, &info->power_status); - pciehp_get_attention_status(slot, &info->attention_status); - pciehp_get_latch_status(slot, &info->latch_status); - pciehp_get_adapter_status(slot, &info->adapter_status); - - result = pci_hp_change_slot_info(slot->hotplug_slot, info); - kfree (info); - return result; -} - /* * Note: This function must be called with slot->lock held */ @@ -442,7 +423,6 @@ static void handle_button_press_event(struct slot *p_slot) * to hot-add or hot-remove is undergoing */ ctrl_info(ctrl, "Button ignore on Slot(%s)\n", slot_name(p_slot)); - update_slot_info(p_slot); break; default: ctrl_warn(ctrl, "Not a valid state\n"); @@ -500,11 +480,9 @@ static void interrupt_event_handler(struct work_struct *work) if (!HP_SUPR_RM(ctrl)) break; ctrl_dbg(ctrl, "Surprise Removal\n"); - update_slot_info(p_slot); handle_surprise_event(p_slot); break; default: - update_slot_info(p_slot); break; } mutex_unlock(&p_slot->lock); @@ -547,9 +525,6 @@ int pciehp_enable_slot(struct slot *p_slot) if (rc) { pciehp_get_latch_status(p_slot, &getstatus); } - - update_slot_info(p_slot); - return rc; } @@ -590,10 +565,7 @@ int pciehp_disable_slot(struct slot *p_slot) } } - ret = remove_board(p_slot); - update_slot_info(p_slot); - - return ret; + return remove_board(p_slot); } int pciehp_sysfs_enable_slot(struct slot *p_slot) diff --git a/drivers/pci/hotplug/pciehp_hpc.c b/drivers/pci/hotplug/pciehp_hpc.c index 88b654ec5afc..7f35aff22362 100644 --- a/drivers/pci/hotplug/pciehp_hpc.c +++ b/drivers/pci/hotplug/pciehp_hpc.c @@ -427,27 +427,24 @@ int pciehp_set_attention_status(struct slot *slot, u8 value) struct controller *ctrl = slot->ctrl; u16 slot_cmd; u16 cmd_mask; - int rc; cmd_mask = PCI_EXP_SLTCTL_AIC; switch (value) { - case 0 : /* turn off */ - slot_cmd = 0x00C0; - break; - case 1: /* turn on */ - slot_cmd = 0x0040; - break; - case 2: /* turn blink */ - slot_cmd = 0x0080; - break; - default: - return -1; + case 0 : /* turn off */ + slot_cmd = 0x00C0; + break; + case 1: /* turn on */ + slot_cmd = 0x0040; + break; + case 2: /* turn blink */ + slot_cmd = 0x0080; + break; + default: + return -EINVAL; } - rc = pcie_write_cmd(ctrl, slot_cmd, cmd_mask); ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n", __func__, ctrl->cap_base + PCI_EXP_SLTCTL, slot_cmd); - - return rc; + return pcie_write_cmd(ctrl, slot_cmd, cmd_mask); } void pciehp_green_led_on(struct slot *slot) From 65b947bc5f32d8d4fe1f925a6146a4088d5466f3 Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Mon, 5 Oct 2009 17:43:29 +0900 Subject: [PATCH 027/109] PCI: pciehp: fix typo in pciehp_probe Fix typo that might cause memory leak in pciehp_probe(). Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/hotplug/pciehp_core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/pci/hotplug/pciehp_core.c b/drivers/pci/hotplug/pciehp_core.c index d19f87594df4..e1b29a59032b 100644 --- a/drivers/pci/hotplug/pciehp_core.c +++ b/drivers/pci/hotplug/pciehp_core.c @@ -285,7 +285,7 @@ static int pciehp_probe(struct pcie_device *dev) rc = pcie_init_notification(ctrl); if (rc) { ctrl_err(ctrl, "Notification initialization failed\n"); - goto err_out_release_ctlr; + goto err_out_free_ctrl_slot; } /* Check if slot is occupied */ From 8792e11f1c54bcba34412f03959e70ee217f2231 Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Mon, 5 Oct 2009 17:46:43 +0900 Subject: [PATCH 028/109] PCI: pciehp: prevent unnecessary power off Prevent unnecessary power off at initialization time. If slot power is already off, we don't need to power off the slot. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/hotplug/pciehp_core.c | 26 ++++++++++---------------- 1 file changed, 10 insertions(+), 16 deletions(-) diff --git a/drivers/pci/hotplug/pciehp_core.c b/drivers/pci/hotplug/pciehp_core.c index e1b29a59032b..5674b2075bdc 100644 --- a/drivers/pci/hotplug/pciehp_core.c +++ b/drivers/pci/hotplug/pciehp_core.c @@ -253,14 +253,13 @@ static int pciehp_probe(struct pcie_device *dev) int rc; struct controller *ctrl; struct slot *slot; - u8 value; - struct pci_dev *pdev = dev->port; + u8 occupied, poweron; if (pciehp_force) dev_info(&dev->device, "Bypassing BIOS check for pciehp use on %s\n", - pci_name(pdev)); - else if (pciehp_get_hp_hw_control_from_firmware(pdev)) + pci_name(dev->port)); + else if (pciehp_get_hp_hw_control_from_firmware(dev->port)) goto err_out_none; ctrl = pcie_init(dev); @@ -290,18 +289,13 @@ static int pciehp_probe(struct pcie_device *dev) /* Check if slot is occupied */ slot = ctrl->slot; - pciehp_get_adapter_status(slot, &value); - if (value) { - if (pciehp_force) - pciehp_enable_slot(slot); - } else { - /* Power off slot if not occupied */ - if (POWER_CTRL(ctrl)) { - rc = pciehp_power_off_slot(slot); - if (rc) - goto err_out_free_ctrl_slot; - } - } + pciehp_get_adapter_status(slot, &occupied); + pciehp_get_power_status(slot, &poweron); + if (occupied && pciehp_force) + pciehp_enable_slot(slot); + /* If empty slot's power status is on, turn power off */ + if (!occupied && poweron && POWER_CTRL(ctrl)) + pciehp_power_off_slot(slot); return 0; From 0584396157ad2d008e2cc76b4ed6254151183a25 Mon Sep 17 00:00:00 2001 From: Matt Domsch Date: Mon, 2 Nov 2009 11:51:24 -0600 Subject: [PATCH 029/109] PCI: PCIe AER: honor ACPI HEST FIRMWARE FIRST mode Feedback from Hidetoshi Seto and Kenji Kaneshige incorporated. This correctly handles PCI-X bridges, PCIe root ports and endpoints, and prints debug messages when invalid/reserved types are found in the HEST. PCI devices not in domain/segment 0 are not represented in HEST, thus will be ignored. Today, the PCIe Advanced Error Reporting (AER) driver attaches itself to every PCIe root port for which BIOS reports it should, via ACPI _OSC. However, _OSC alone is insufficient for newer BIOSes. Part of ACPI 4.0 is the new APEI (ACPI Platform Error Interfaces) which is a way for OS and BIOS to handshake over which errors for which components each will handle. One table in ACPI 4.0 is the Hardware Error Source Table (HEST), where BIOS can define that errors for certain PCIe devices (or all devices), should be handled by BIOS ("Firmware First mode"), rather than be handled by the OS. Dell PowerEdge 11G server BIOS defines Firmware First mode in HEST, so that it may manage such errors, log them to the System Event Log, and possibly take other actions. The aer driver should honor this, and not attach itself to devices noted as such. Furthermore, Kenji Kaneshige reminded us to disallow changing the AER registers when respecting Firmware First mode. Platform firmware is expected to manage these, and if changes to them are allowed, it could break that firmware's behavior. The HEST parsing code may be replaced in the future by a more feature-rich implementation. This patch provides the minimum needed to prevent breakage until that implementation is available. Reviewed-by: Kenji Kaneshige Reviewed-by: Hidetoshi Seto Signed-off-by: Matt Domsch Signed-off-by: Jesse Barnes --- drivers/acpi/Makefile | 1 + drivers/acpi/hest.c | 135 +++++++++++++++++++++++++++++ drivers/pci/pcie/aer/aerdrv_core.c | 24 ++++- drivers/pci/probe.c | 8 ++ include/acpi/acpi_hest.h | 12 +++ include/linux/pci.h | 1 + 6 files changed, 179 insertions(+), 2 deletions(-) create mode 100644 drivers/acpi/hest.c create mode 100644 include/acpi/acpi_hest.h diff --git a/drivers/acpi/Makefile b/drivers/acpi/Makefile index 7702118509a0..c7b10b4298e9 100644 --- a/drivers/acpi/Makefile +++ b/drivers/acpi/Makefile @@ -19,6 +19,7 @@ obj-y += acpi.o \ # All the builtin files are in the "acpi." module_param namespace. acpi-y += osl.o utils.o reboot.o +acpi-y += hest.o # sleep related files acpi-y += wakeup.o diff --git a/drivers/acpi/hest.c b/drivers/acpi/hest.c new file mode 100644 index 000000000000..4bb18c980ac6 --- /dev/null +++ b/drivers/acpi/hest.c @@ -0,0 +1,135 @@ +#include +#include + +#define PREFIX "ACPI: " + +static inline unsigned long parse_acpi_hest_ia_machine_check(struct acpi_hest_ia_machine_check *p) +{ + return sizeof(*p) + + (sizeof(struct acpi_hest_ia_error_bank) * p->num_hardware_banks); +} + +static inline unsigned long parse_acpi_hest_ia_corrected(struct acpi_hest_ia_corrected *p) +{ + return sizeof(*p) + + (sizeof(struct acpi_hest_ia_error_bank) * p->num_hardware_banks); +} + +static inline unsigned long parse_acpi_hest_ia_nmi(struct acpi_hest_ia_nmi *p) +{ + return sizeof(*p); +} + +static inline unsigned long parse_acpi_hest_generic(struct acpi_hest_generic *p) +{ + return sizeof(*p); +} + +static inline unsigned int hest_match_pci(struct acpi_hest_aer_common *p, struct pci_dev *pci) +{ + return (0 == pci_domain_nr(pci->bus) && + p->bus == pci->bus->number && + p->device == PCI_SLOT(pci->devfn) && + p->function == PCI_FUNC(pci->devfn)); +} + +static unsigned long parse_acpi_hest_aer(void *hdr, int type, struct pci_dev *pci, int *firmware_first) +{ + struct acpi_hest_aer_common *p = hdr + sizeof(struct acpi_hest_header); + unsigned long rc=0; + u8 pcie_type = 0; + u8 bridge = 0; + switch (type) { + case ACPI_HEST_TYPE_AER_ROOT_PORT: + rc = sizeof(struct acpi_hest_aer_root); + pcie_type = PCI_EXP_TYPE_ROOT_PORT; + break; + case ACPI_HEST_TYPE_AER_ENDPOINT: + rc = sizeof(struct acpi_hest_aer); + pcie_type = PCI_EXP_TYPE_ENDPOINT; + break; + case ACPI_HEST_TYPE_AER_BRIDGE: + rc = sizeof(struct acpi_hest_aer_bridge); + if ((pci->class >> 16) == PCI_BASE_CLASS_BRIDGE) + bridge = 1; + break; + } + + if (p->flags & ACPI_HEST_GLOBAL) { + if ((pci->is_pcie && (pci->pcie_type == pcie_type)) || bridge) + *firmware_first = !!(p->flags & ACPI_HEST_FIRMWARE_FIRST); + } + else + if (hest_match_pci(p, pci)) + *firmware_first = !!(p->flags & ACPI_HEST_FIRMWARE_FIRST); + return rc; +} + +static int acpi_hest_firmware_first(struct acpi_table_header *stdheader, struct pci_dev *pci) +{ + struct acpi_table_hest *hest = (struct acpi_table_hest *)stdheader; + void *p = (void *)hest + sizeof(*hest); /* defined by the ACPI 4.0 spec */ + struct acpi_hest_header *hdr = p; + + int i; + int firmware_first = 0; + static unsigned char printed_unused = 0; + static unsigned char printed_reserved = 0; + + for (i=0, hdr=p; p < (((void *)hest) + hest->header.length) && i < hest->error_source_count; i++) { + switch (hdr->type) { + case ACPI_HEST_TYPE_IA32_CHECK: + p += parse_acpi_hest_ia_machine_check(p); + break; + case ACPI_HEST_TYPE_IA32_CORRECTED_CHECK: + p += parse_acpi_hest_ia_corrected(p); + break; + case ACPI_HEST_TYPE_IA32_NMI: + p += parse_acpi_hest_ia_nmi(p); + break; + /* These three should never appear */ + case ACPI_HEST_TYPE_NOT_USED3: + case ACPI_HEST_TYPE_NOT_USED4: + case ACPI_HEST_TYPE_NOT_USED5: + if (!printed_unused) { + printk(KERN_DEBUG PREFIX + "HEST Error Source list contains an obsolete type (%d).\n", hdr->type); + printed_unused = 1; + } + break; + case ACPI_HEST_TYPE_AER_ROOT_PORT: + case ACPI_HEST_TYPE_AER_ENDPOINT: + case ACPI_HEST_TYPE_AER_BRIDGE: + p += parse_acpi_hest_aer(p, hdr->type, pci, &firmware_first); + break; + case ACPI_HEST_TYPE_GENERIC_ERROR: + p += parse_acpi_hest_generic(p); + break; + /* These should never appear either */ + case ACPI_HEST_TYPE_RESERVED: + default: + if (!printed_reserved) { + printk(KERN_DEBUG PREFIX + "HEST Error Source list contains a reserved type (%d).\n", hdr->type); + printed_reserved = 1; + } + break; + } + } + return firmware_first; +} + +int acpi_hest_firmware_first_pci(struct pci_dev *pci) +{ + acpi_status status = AE_NOT_FOUND; + struct acpi_table_header *hest = NULL; + status = acpi_get_table(ACPI_SIG_HEST, 1, &hest); + + if (ACPI_SUCCESS(status)) { + if (acpi_hest_firmware_first(hest, pci)) { + return 1; + } + } + return 0; +} +EXPORT_SYMBOL_GPL(acpi_hest_firmware_first_pci); diff --git a/drivers/pci/pcie/aer/aerdrv_core.c b/drivers/pci/pcie/aer/aerdrv_core.c index 9f5ccbeb4fa5..f4512feac12b 100644 --- a/drivers/pci/pcie/aer/aerdrv_core.c +++ b/drivers/pci/pcie/aer/aerdrv_core.c @@ -35,6 +35,9 @@ int pci_enable_pcie_error_reporting(struct pci_dev *dev) u16 reg16 = 0; int pos; + if (dev->aer_firmware_first) + return -EIO; + pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ERR); if (!pos) return -EIO; @@ -60,6 +63,9 @@ int pci_disable_pcie_error_reporting(struct pci_dev *dev) u16 reg16 = 0; int pos; + if (dev->aer_firmware_first) + return -EIO; + pos = pci_find_capability(dev, PCI_CAP_ID_EXP); if (!pos) return -EIO; @@ -874,8 +880,22 @@ void aer_delete_rootport(struct aer_rpc *rpc) */ int aer_init(struct pcie_device *dev) { - if (aer_osc_setup(dev) && !forceload) - return -ENXIO; + if (dev->port->aer_firmware_first) { + dev_printk(KERN_DEBUG, &dev->device, + "PCIe errors handled by platform firmware.\n"); + goto out; + } + + if (aer_osc_setup(dev)) + goto out; return 0; +out: + if (forceload) { + dev_printk(KERN_DEBUG, &dev->device, + "aerdrv forceload requested.\n"); + dev->port->aer_firmware_first = 0; + return 0; + } + return -ENXIO; } diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 9cefc54a0125..118463befef0 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -11,6 +11,7 @@ #include #include #include +#include #include #include "pci.h" @@ -706,6 +707,12 @@ static void set_pcie_hotplug_bridge(struct pci_dev *pdev) pdev->is_hotplug_bridge = 1; } +static void set_pci_aer_firmware_first(struct pci_dev *pdev) +{ + if (acpi_hest_firmware_first_pci(pdev)) + pdev->aer_firmware_first = 1; +} + #define LEGACY_IO_RESOURCE (IORESOURCE_IO | IORESOURCE_PCI_FIXED) /** @@ -734,6 +741,7 @@ int pci_setup_device(struct pci_dev *dev) dev->multifunction = !!(hdr_type & 0x80); dev->error_state = pci_channel_io_normal; set_pcie_port_type(dev); + set_pci_aer_firmware_first(dev); list_for_each_entry(slot, &dev->bus->slots, list) if (PCI_SLOT(dev->devfn) == slot->number) diff --git a/include/acpi/acpi_hest.h b/include/acpi/acpi_hest.h new file mode 100644 index 000000000000..63194d03cb2d --- /dev/null +++ b/include/acpi/acpi_hest.h @@ -0,0 +1,12 @@ +#ifndef __ACPI_HEST_H +#define __ACPI_HEST_H + +#include + +#ifdef CONFIG_ACPI +extern int acpi_hest_firmware_first_pci(struct pci_dev *pci); +#else +static inline int acpi_hest_firmware_first_pci(struct pci_dev *pci) { return 0; } +#endif + +#endif diff --git a/include/linux/pci.h b/include/linux/pci.h index da4128f6e916..9d646e60cae0 100644 --- a/include/linux/pci.h +++ b/include/linux/pci.h @@ -280,6 +280,7 @@ struct pci_dev { unsigned int is_virtfn:1; unsigned int reset_fn:1; unsigned int is_hotplug_bridge:1; + unsigned int aer_firmware_first:1; pci_dev_flags_t dev_flags; atomic_t enable_cnt; /* pci_enable_device has been called */ From bc577d2bb98cc44371287fce3e892d26ad4050a8 Mon Sep 17 00:00:00 2001 From: Gabe Black Date: Tue, 6 Oct 2009 10:45:19 -0500 Subject: [PATCH 030/109] PCI: populate subsystem vendor and device IDs for PCI bridges Change to populate the subsystem vendor and subsytem device IDs for PCI-PCI bridges that implement the PCI Subsystem Vendor ID capability. Previously bridges left subsystem vendor IDs unpopulated. Signed-off-by: Gabe Black Signed-off-by: Jesse Barnes --- drivers/pci/probe.c | 6 ++++++ include/linux/pci_regs.h | 5 +++++ 2 files changed, 11 insertions(+) diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 118463befef0..4842b09b7f3c 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -730,6 +730,7 @@ int pci_setup_device(struct pci_dev *dev) u32 class; u8 hdr_type; struct pci_slot *slot; + int pos = 0; if (pci_read_config_byte(dev, PCI_HEADER_TYPE, &hdr_type)) return -EIO; @@ -822,6 +823,11 @@ int pci_setup_device(struct pci_dev *dev) dev->transparent = ((dev->class & 0xff) == 1); pci_read_bases(dev, 2, PCI_ROM_ADDRESS1); set_pcie_hotplug_bridge(dev); + pos = pci_find_capability(dev, PCI_CAP_ID_SSVID); + if (pos) { + pci_read_config_word(dev, pos + PCI_SSVID_VENDOR_ID, &dev->subsystem_vendor); + pci_read_config_word(dev, pos + PCI_SSVID_DEVICE_ID, &dev->subsystem_device); + } break; case PCI_HEADER_TYPE_CARDBUS: /* CardBus bridge header */ diff --git a/include/linux/pci_regs.h b/include/linux/pci_regs.h index d798770f08cd..9f2ad0aa3c39 100644 --- a/include/linux/pci_regs.h +++ b/include/linux/pci_regs.h @@ -365,6 +365,11 @@ #define PCI_X_STATUS_266MHZ 0x40000000 /* 266 MHz capable */ #define PCI_X_STATUS_533MHZ 0x80000000 /* 533 MHz capable */ +/* PCI Bridge Subsystem ID registers */ + +#define PCI_SSVID_VENDOR_ID 4 /* PCI-Bridge subsystem vendor id register */ +#define PCI_SSVID_DEVICE_ID 6 /* PCI-Bridge subsystem device id register */ + /* PCI Express capability registers */ #define PCI_EXP_FLAGS 2 /* Capabilities register */ From 3c299dc22635e500214707aa28be119ff2b3901c Mon Sep 17 00:00:00 2001 From: Andrew Patterson Date: Mon, 12 Oct 2009 13:14:00 -0600 Subject: [PATCH 031/109] PCI: add pci_get_domain_bus_and_slot function Added the pci_get_domain_and_slot_function which is analogous to pci_get_bus_and_slot. It returns a pci_dev given a domain (segment) number, bus number, and devnr. Like pci_get_bus_and_slot, pci_get_domain_bus_and_slot holds a reference to the returned pci_dev. Converted pci_get_bus_and_slot to a wrapper that calls pci_get_domain_bus_and_slot with the domain hard-coded to 0. This routine was patterned off code suggested by Bjorn Helgaas. Acked-by: Huang Ying Signed-off-by: Andrew Patterson Signed-off-by: Jesse Barnes --- drivers/pci/search.c | 34 +++++++++++++++++----------------- include/linux/pci.h | 8 +++++++- 2 files changed, 24 insertions(+), 18 deletions(-) diff --git a/drivers/pci/search.c b/drivers/pci/search.c index ec415352d9ba..75826482c71a 100644 --- a/drivers/pci/search.c +++ b/drivers/pci/search.c @@ -149,32 +149,33 @@ struct pci_dev * pci_get_slot(struct pci_bus *bus, unsigned int devfn) } /** - * pci_get_bus_and_slot - locate PCI device from a given PCI bus & slot - * @bus: number of PCI bus on which desired PCI device resides - * @devfn: encodes number of PCI slot in which the desired PCI - * device resides and the logical device number within that slot - * in case of multi-function devices. + * pci_get_domain_bus_and_slot - locate PCI device for a given PCI domain (segment), bus, and slot + * @domain: PCI domain/segment on which the PCI device resides. + * @bus: PCI bus on which desired PCI device resides + * @devfn: encodes number of PCI slot in which the desired PCI device + * resides and the logical device number within that slot in case of + * multi-function devices. * - * Note: the bus/slot search is limited to PCI domain (segment) 0. - * - * Given a PCI bus and slot/function number, the desired PCI device - * is located in system global list of PCI devices. If the device - * is found, a pointer to its data structure is returned. If no - * device is found, %NULL is returned. The returned device has its - * reference count bumped by one. + * Given a PCI domain, bus, and slot/function number, the desired PCI + * device is located in the list of PCI devices. If the device is + * found, its reference count is increased and this function returns a + * pointer to its data structure. The caller must decrement the + * reference count by calling pci_dev_put(). If no device is found, + * %NULL is returned. */ - -struct pci_dev * pci_get_bus_and_slot(unsigned int bus, unsigned int devfn) +struct pci_dev *pci_get_domain_bus_and_slot(int domain, unsigned int bus, + unsigned int devfn) { struct pci_dev *dev = NULL; while ((dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, dev)) != NULL) { - if (pci_domain_nr(dev->bus) == 0 && - (dev->bus->number == bus && dev->devfn == devfn)) + if (pci_domain_nr(dev->bus) == domain && + (dev->bus->number == bus && dev->devfn == devfn)) return dev; } return NULL; } +EXPORT_SYMBOL(pci_get_domain_bus_and_slot); static int match_pci_dev_by_id(struct device *dev, void *data) { @@ -354,5 +355,4 @@ EXPORT_SYMBOL(pci_find_next_bus); EXPORT_SYMBOL(pci_get_device); EXPORT_SYMBOL(pci_get_subsys); EXPORT_SYMBOL(pci_get_slot); -EXPORT_SYMBOL(pci_get_bus_and_slot); EXPORT_SYMBOL(pci_get_class); diff --git a/include/linux/pci.h b/include/linux/pci.h index 9d646e60cae0..86c31ac454d1 100644 --- a/include/linux/pci.h +++ b/include/linux/pci.h @@ -636,7 +636,13 @@ struct pci_dev *pci_get_subsys(unsigned int vendor, unsigned int device, unsigned int ss_vendor, unsigned int ss_device, struct pci_dev *from); struct pci_dev *pci_get_slot(struct pci_bus *bus, unsigned int devfn); -struct pci_dev *pci_get_bus_and_slot(unsigned int bus, unsigned int devfn); +struct pci_dev *pci_get_domain_bus_and_slot(int domain, unsigned int bus, + unsigned int devfn); +static inline struct pci_dev *pci_get_bus_and_slot(unsigned int bus, + unsigned int devfn) +{ + return pci_get_domain_bus_and_slot(0, bus, devfn); +} struct pci_dev *pci_get_class(unsigned int class, struct pci_dev *from); int pci_dev_present(const struct pci_device_id *ids); From cc5d153a0ca794e3781ef34c76f32ad3e991b13d Mon Sep 17 00:00:00 2001 From: Andrew Patterson Date: Mon, 12 Oct 2009 13:14:05 -0600 Subject: [PATCH 032/109] PCI: add support for PCI domains to aer_inject Add support for PCI domains (segments) to aer_inject. Acked-by: Huang Ying Signed-off-by: Andrew Patterson Signed-off-by: Jesse Barnes --- drivers/pci/pcie/aer/aer_inject.c | 46 +++++++++++++++++++++++-------- 1 file changed, 34 insertions(+), 12 deletions(-) diff --git a/drivers/pci/pcie/aer/aer_inject.c b/drivers/pci/pcie/aer/aer_inject.c index 62d15f652bb6..ac0b5e7bb3f5 100644 --- a/drivers/pci/pcie/aer/aer_inject.c +++ b/drivers/pci/pcie/aer/aer_inject.c @@ -23,6 +23,7 @@ #include #include #include +#include #include "aerdrv.h" struct aer_error_inj { @@ -35,10 +36,12 @@ struct aer_error_inj { u32 header_log1; u32 header_log2; u32 header_log3; + u16 domain; }; struct aer_error { struct list_head list; + u16 domain; unsigned int bus; unsigned int devfn; int pos_cap_err; @@ -66,22 +69,27 @@ static LIST_HEAD(pci_bus_ops_list); /* Protect einjected and pci_bus_ops_list */ static DEFINE_SPINLOCK(inject_lock); -static void aer_error_init(struct aer_error *err, unsigned int bus, - unsigned int devfn, int pos_cap_err) +static void aer_error_init(struct aer_error *err, u16 domain, + unsigned int bus, unsigned int devfn, + int pos_cap_err) { INIT_LIST_HEAD(&err->list); + err->domain = domain; err->bus = bus; err->devfn = devfn; err->pos_cap_err = pos_cap_err; } /* inject_lock must be held before calling */ -static struct aer_error *__find_aer_error(unsigned int bus, unsigned int devfn) +static struct aer_error *__find_aer_error(u16 domain, unsigned int bus, + unsigned int devfn) { struct aer_error *err; list_for_each_entry(err, &einjected, list) { - if (bus == err->bus && devfn == err->devfn) + if (domain == err->domain && + bus == err->bus && + devfn == err->devfn) return err; } return NULL; @@ -90,7 +98,10 @@ static struct aer_error *__find_aer_error(unsigned int bus, unsigned int devfn) /* inject_lock must be held before calling */ static struct aer_error *__find_aer_error_by_dev(struct pci_dev *dev) { - return __find_aer_error(dev->bus->number, dev->devfn); + int domain = pci_domain_nr(dev->bus); + if (domain < 0) + return NULL; + return __find_aer_error((u16)domain, dev->bus->number, dev->devfn); } /* inject_lock must be held before calling */ @@ -172,11 +183,15 @@ static int pci_read_aer(struct pci_bus *bus, unsigned int devfn, int where, struct aer_error *err; unsigned long flags; struct pci_ops *ops; + int domain; spin_lock_irqsave(&inject_lock, flags); if (size != sizeof(u32)) goto out; - err = __find_aer_error(bus->number, devfn); + domain = pci_domain_nr(bus); + if (domain < 0) + goto out; + err = __find_aer_error((u16)domain, bus->number, devfn); if (!err) goto out; @@ -200,11 +215,15 @@ int pci_write_aer(struct pci_bus *bus, unsigned int devfn, int where, int size, unsigned long flags; int rw1cs; struct pci_ops *ops; + int domain; spin_lock_irqsave(&inject_lock, flags); if (size != sizeof(u32)) goto out; - err = __find_aer_error(bus->number, devfn); + domain = pci_domain_nr(bus); + if (domain < 0) + goto out; + err = __find_aer_error((u16)domain, bus->number, devfn); if (!err) goto out; @@ -305,7 +324,7 @@ static int aer_inject(struct aer_error_inj *einj) u32 sever; int ret = 0; - dev = pci_get_bus_and_slot(einj->bus, devfn); + dev = pci_get_domain_bus_and_slot((int)einj->domain, einj->bus, devfn); if (!dev) return -EINVAL; rpdev = pcie_find_root_port(dev); @@ -344,7 +363,8 @@ static int aer_inject(struct aer_error_inj *einj) if (!err) { err = err_alloc; err_alloc = NULL; - aer_error_init(err, einj->bus, devfn, pos_cap_err); + aer_error_init(err, einj->domain, einj->bus, devfn, + pos_cap_err); list_add(&err->list, &einjected); } err->uncor_status |= einj->uncor_status; @@ -358,7 +378,8 @@ static int aer_inject(struct aer_error_inj *einj) if (!rperr) { rperr = rperr_alloc; rperr_alloc = NULL; - aer_error_init(rperr, rpdev->bus->number, rpdev->devfn, + aer_error_init(rperr, pci_domain_nr(rpdev->bus), + rpdev->bus->number, rpdev->devfn, rp_pos_cap_err); list_add(&rperr->list, &einjected); } @@ -411,10 +432,11 @@ static ssize_t aer_inject_write(struct file *filp, const char __user *ubuf, if (!capable(CAP_SYS_ADMIN)) return -EPERM; - - if (usize != sizeof(struct aer_error_inj)) + if (usize < offsetof(struct aer_error_inj, domain) || + usize > sizeof(einj)) return -EINVAL; + memset(&einj, 0, sizeof(einj)); if (copy_from_user(&einj, ubuf, usize)) return -EFAULT; From 1d0243559497b9cab00099c49a5ba3222cd6576f Mon Sep 17 00:00:00 2001 From: Andrew Patterson Date: Mon, 12 Oct 2009 13:14:10 -0600 Subject: [PATCH 033/109] PCI: use better error return values in aer_inject Replaced some error return values in aer_inject. Use -ENODEV when we can't find a device and -ENOTTY when the device does not support PCIe AER. Acked-by: Huang Ying Signed-off-by: Andrew Patterson Signed-off-by: Jesse Barnes --- drivers/pci/pcie/aer/aer_inject.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/pci/pcie/aer/aer_inject.c b/drivers/pci/pcie/aer/aer_inject.c index ac0b5e7bb3f5..da2ad6eaee17 100644 --- a/drivers/pci/pcie/aer/aer_inject.c +++ b/drivers/pci/pcie/aer/aer_inject.c @@ -326,23 +326,23 @@ static int aer_inject(struct aer_error_inj *einj) dev = pci_get_domain_bus_and_slot((int)einj->domain, einj->bus, devfn); if (!dev) - return -EINVAL; + return -ENODEV; rpdev = pcie_find_root_port(dev); if (!rpdev) { - ret = -EINVAL; + ret = -ENOTTY; goto out_put; } pos_cap_err = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ERR); if (!pos_cap_err) { - ret = -EIO; + ret = -ENOTTY; goto out_put; } pci_read_config_dword(dev, pos_cap_err + PCI_ERR_UNCOR_SEVER, &sever); rp_pos_cap_err = pci_find_ext_capability(rpdev, PCI_EXT_CAP_ID_ERR); if (!rp_pos_cap_err) { - ret = -EIO; + ret = -ENOTTY; goto out_put; } From 476f644edf7c22b47e6a118e4a1e138112a5ef14 Mon Sep 17 00:00:00 2001 From: Andrew Patterson Date: Mon, 12 Oct 2009 13:14:15 -0600 Subject: [PATCH 034/109] PCI: fix memory leak in aer_inject Fixed probable typo in aer_inject cleanup code resulting in a memory leak. Acked-by: Huang Ying Signed-off-by: Andrew Patterson Signed-off-by: Jesse Barnes --- drivers/pci/pcie/aer/aer_inject.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/pci/pcie/aer/aer_inject.c b/drivers/pci/pcie/aer/aer_inject.c index da2ad6eaee17..2246bf7aee7e 100644 --- a/drivers/pci/pcie/aer/aer_inject.c +++ b/drivers/pci/pcie/aer/aer_inject.c @@ -474,7 +474,7 @@ static void __exit aer_inject_exit(void) } spin_lock_irqsave(&inject_lock, flags); - list_for_each_entry_safe(err, err_next, &pci_bus_ops_list, list) { + list_for_each_entry_safe(err, err_next, &einjected, list) { list_del(&err->list); kfree(err); } From 204d49a5613a06eb2fa5c3b842a29b1336cc7995 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Mon, 26 Oct 2009 11:20:47 -0600 Subject: [PATCH 035/109] PCI hotplug: move IOAPIC support from acpiphp to ioapic driver This patch moves PCI I/O APIC support from acpiphp to a separate driver. Like pciehp and shpchp, acpiphp handles PCI hotplug, i.e., addition and removal of PCI adapters. But in addition, acpiphp handles some ACPI hotplug, such as the addition of new host bridges, and the I/O APIC support was tangled up with that. I don't think the I/O APIC support needs to be in acpiphp; PCI I/O APICs usually appear as a function on a PCI host bridge, and we'll enumerate the APIC before any of the devices behind the bridge that use it. As far as I know, nobody actually uses I/O APIC hotplug. It depends on acpi_register_ioapic(), which is only implemented for ia64, and I don't think any vendors have supported I/O chassis hotplug yet. Signed-off-by: Bjorn Helgaas Reviewed-by: Kenji Kaneshige CC: Satoru Takeuchi CC: MUNEDA Takahiro Signed-off-by: Jesse Barnes --- drivers/pci/Kconfig | 7 ++ drivers/pci/Makefile | 2 + drivers/pci/hotplug/acpiphp.h | 6 - drivers/pci/hotplug/acpiphp_glue.c | 190 ----------------------------- drivers/pci/ioapic.c | 127 +++++++++++++++++++ 5 files changed, 136 insertions(+), 196 deletions(-) create mode 100644 drivers/pci/ioapic.c diff --git a/drivers/pci/Kconfig b/drivers/pci/Kconfig index fdc864f9cf23..d7d28aef8a27 100644 --- a/drivers/pci/Kconfig +++ b/drivers/pci/Kconfig @@ -69,3 +69,10 @@ config PCI_IOV physical resources. If unsure, say N. + +config PCI_IOAPIC + bool + depends on PCI + depends on ACPI + depends on HOTPLUG + default y diff --git a/drivers/pci/Makefile b/drivers/pci/Makefile index 4a7f11d8f432..4df48d58eaa6 100644 --- a/drivers/pci/Makefile +++ b/drivers/pci/Makefile @@ -14,6 +14,8 @@ CFLAGS_legacy.o += -Wno-deprecated-declarations # Build PCI Express stuff if needed obj-$(CONFIG_PCIEPORTBUS) += pcie/ +obj-$(CONFIG_PCI_IOAPIC) += ioapic.o + obj-$(CONFIG_HOTPLUG) += hotplug.o # Build the PCI Hotplug drivers if we were asked to diff --git a/drivers/pci/hotplug/acpiphp.h b/drivers/pci/hotplug/acpiphp.h index 7d938df79206..bab52047baa8 100644 --- a/drivers/pci/hotplug/acpiphp.h +++ b/drivers/pci/hotplug/acpiphp.h @@ -146,12 +146,6 @@ struct acpiphp_attention_info struct module *owner; }; -struct acpiphp_ioapic { - struct pci_dev *dev; - u32 gsi_base; - struct list_head list; -}; - /* PCI bus bridge HID */ #define ACPI_PCI_HOST_HID "PNP0A03" diff --git a/drivers/pci/hotplug/acpiphp_glue.c b/drivers/pci/hotplug/acpiphp_glue.c index 58d25a163a8b..392e4b3fa2e4 100644 --- a/drivers/pci/hotplug/acpiphp_glue.c +++ b/drivers/pci/hotplug/acpiphp_glue.c @@ -52,8 +52,6 @@ #include "acpiphp.h" static LIST_HEAD(bridge_list); -static LIST_HEAD(ioapic_list); -static DEFINE_SPINLOCK(ioapic_list_lock); #define MY_NAME "acpiphp_glue" @@ -606,190 +604,6 @@ static void remove_bridge(acpi_handle handle) handle_hotplug_event_bridge); } -static struct pci_dev * get_apic_pci_info(acpi_handle handle) -{ - struct pci_dev *dev; - - dev = acpi_get_pci_dev(handle); - if (!dev) - return NULL; - - if ((dev->class != PCI_CLASS_SYSTEM_PIC_IOAPIC) && - (dev->class != PCI_CLASS_SYSTEM_PIC_IOXAPIC)) - { - pci_dev_put(dev); - return NULL; - } - - return dev; -} - -static int get_gsi_base(acpi_handle handle, u32 *gsi_base) -{ - acpi_status status; - int result = -1; - unsigned long long gsb; - struct acpi_buffer buffer = {ACPI_ALLOCATE_BUFFER, NULL}; - union acpi_object *obj; - void *table; - - status = acpi_evaluate_integer(handle, "_GSB", NULL, &gsb); - if (ACPI_SUCCESS(status)) { - *gsi_base = (u32)gsb; - return 0; - } - - status = acpi_evaluate_object(handle, "_MAT", NULL, &buffer); - if (ACPI_FAILURE(status) || !buffer.length || !buffer.pointer) - return -1; - - obj = buffer.pointer; - if (obj->type != ACPI_TYPE_BUFFER) - goto out; - - table = obj->buffer.pointer; - switch (((struct acpi_subtable_header *)table)->type) { - case ACPI_MADT_TYPE_IO_SAPIC: - *gsi_base = ((struct acpi_madt_io_sapic *)table)->global_irq_base; - result = 0; - break; - case ACPI_MADT_TYPE_IO_APIC: - *gsi_base = ((struct acpi_madt_io_apic *)table)->global_irq_base; - result = 0; - break; - default: - break; - } - out: - kfree(buffer.pointer); - return result; -} - -static acpi_status -ioapic_add(acpi_handle handle, u32 lvl, void *context, void **rv) -{ - acpi_status status; - unsigned long long sta; - acpi_handle tmp; - struct pci_dev *pdev; - u32 gsi_base; - u64 phys_addr; - struct acpiphp_ioapic *ioapic; - - /* Evaluate _STA if present */ - status = acpi_evaluate_integer(handle, "_STA", NULL, &sta); - if (ACPI_SUCCESS(status) && sta != ACPI_STA_ALL) - return AE_CTRL_DEPTH; - - /* Scan only PCI bus scope */ - status = acpi_get_handle(handle, "_HID", &tmp); - if (ACPI_SUCCESS(status)) - return AE_CTRL_DEPTH; - - if (get_gsi_base(handle, &gsi_base)) - return AE_OK; - - ioapic = kmalloc(sizeof(*ioapic), GFP_KERNEL); - if (!ioapic) - return AE_NO_MEMORY; - - pdev = get_apic_pci_info(handle); - if (!pdev) - goto exit_kfree; - - if (pci_enable_device(pdev)) - goto exit_pci_dev_put; - - pci_set_master(pdev); - - if (pci_request_region(pdev, 0, "I/O APIC(acpiphp)")) - goto exit_pci_disable_device; - - phys_addr = pci_resource_start(pdev, 0); - if (acpi_register_ioapic(handle, phys_addr, gsi_base)) - goto exit_pci_release_region; - - ioapic->gsi_base = gsi_base; - ioapic->dev = pdev; - spin_lock(&ioapic_list_lock); - list_add_tail(&ioapic->list, &ioapic_list); - spin_unlock(&ioapic_list_lock); - - return AE_OK; - - exit_pci_release_region: - pci_release_region(pdev, 0); - exit_pci_disable_device: - pci_disable_device(pdev); - exit_pci_dev_put: - pci_dev_put(pdev); - exit_kfree: - kfree(ioapic); - - return AE_OK; -} - -static acpi_status -ioapic_remove(acpi_handle handle, u32 lvl, void *context, void **rv) -{ - acpi_status status; - unsigned long long sta; - acpi_handle tmp; - u32 gsi_base; - struct acpiphp_ioapic *pos, *n, *ioapic = NULL; - - /* Evaluate _STA if present */ - status = acpi_evaluate_integer(handle, "_STA", NULL, &sta); - if (ACPI_SUCCESS(status) && sta != ACPI_STA_ALL) - return AE_CTRL_DEPTH; - - /* Scan only PCI bus scope */ - status = acpi_get_handle(handle, "_HID", &tmp); - if (ACPI_SUCCESS(status)) - return AE_CTRL_DEPTH; - - if (get_gsi_base(handle, &gsi_base)) - return AE_OK; - - acpi_unregister_ioapic(handle, gsi_base); - - spin_lock(&ioapic_list_lock); - list_for_each_entry_safe(pos, n, &ioapic_list, list) { - if (pos->gsi_base != gsi_base) - continue; - ioapic = pos; - list_del(&ioapic->list); - break; - } - spin_unlock(&ioapic_list_lock); - - if (!ioapic) - return AE_OK; - - pci_release_region(ioapic->dev, 0); - pci_disable_device(ioapic->dev); - pci_dev_put(ioapic->dev); - kfree(ioapic); - - return AE_OK; -} - -static int acpiphp_configure_ioapics(acpi_handle handle) -{ - ioapic_add(handle, 0, NULL, NULL); - acpi_walk_namespace(ACPI_TYPE_DEVICE, handle, - ACPI_UINT32_MAX, ioapic_add, NULL, NULL); - return 0; -} - -static int acpiphp_unconfigure_ioapics(acpi_handle handle) -{ - ioapic_remove(handle, 0, NULL, NULL); - acpi_walk_namespace(ACPI_TYPE_DEVICE, handle, - ACPI_UINT32_MAX, ioapic_remove, NULL, NULL); - return 0; -} - static int power_on_slot(struct acpiphp_slot *slot) { acpi_status status; @@ -1014,8 +828,6 @@ static int __ref enable_device(struct acpiphp_slot *slot) pci_bus_assign_resources(bus); acpiphp_sanitize_bus(bus); acpiphp_set_hpp_values(bus); - list_for_each_entry(func, &slot->funcs, sibling) - acpiphp_configure_ioapics(func->handle); pci_enable_bridges(bus); pci_bus_add_devices(bus); @@ -1091,7 +903,6 @@ static int disable_device(struct acpiphp_slot *slot) } list_for_each_entry(func, &slot->funcs, sibling) { - acpiphp_unconfigure_ioapics(func->handle); acpiphp_bus_trim(func->handle); } @@ -1275,7 +1086,6 @@ static int acpiphp_configure_bridge (acpi_handle handle) acpiphp_sanitize_bus(bus); acpiphp_set_hpp_values(bus); pci_enable_bridges(bus); - acpiphp_configure_ioapics(handle); return 0; } diff --git a/drivers/pci/ioapic.c b/drivers/pci/ioapic.c new file mode 100644 index 000000000000..3e0d7b5dd1b9 --- /dev/null +++ b/drivers/pci/ioapic.c @@ -0,0 +1,127 @@ +/* + * IOAPIC/IOxAPIC/IOSAPIC driver + * + * Copyright (C) 2009 Fujitsu Limited. + * (c) Copyright 2009 Hewlett-Packard Development Company, L.P. + * + * 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. + */ + +/* + * This driver manages PCI I/O APICs added by hotplug after boot. We try to + * claim all I/O APIC PCI devices, but those present at boot were registered + * when we parsed the ACPI MADT, so we'll fail when we try to re-register + * them. + */ + +#include +#include +#include + +struct ioapic { + acpi_handle handle; + u32 gsi_base; +}; + +static int ioapic_probe(struct pci_dev *dev, const struct pci_device_id *ent) +{ + acpi_handle handle; + acpi_status status; + unsigned long long gsb; + struct ioapic *ioapic; + u64 addr; + int ret; + char *type; + + handle = DEVICE_ACPI_HANDLE(&dev->dev); + if (!handle) + return -EINVAL; + + status = acpi_evaluate_integer(handle, "_GSB", NULL, &gsb); + if (ACPI_FAILURE(status)) + return -EINVAL; + + /* + * The previous code in acpiphp evaluated _MAT if _GSB failed, but + * ACPI spec 4.0 sec 6.2.2 requires _GSB for hot-pluggable I/O APICs. + */ + + ioapic = kzalloc(sizeof(*ioapic), GFP_KERNEL); + if (!ioapic) + return -ENOMEM; + + ioapic->handle = handle; + ioapic->gsi_base = (u32) gsb; + + if (dev->class == PCI_CLASS_SYSTEM_PIC_IOAPIC) + type = "IOAPIC"; + else + type = "IOxAPIC"; + + ret = pci_enable_device(dev); + if (ret < 0) + goto exit_free; + + pci_set_master(dev); + + if (pci_request_region(dev, 0, type)) + goto exit_disable; + + addr = pci_resource_start(dev, 0); + if (acpi_register_ioapic(ioapic->handle, addr, ioapic->gsi_base)) + goto exit_release; + + pci_set_drvdata(dev, ioapic); + dev_info(&dev->dev, "%s at %#llx, GSI %u\n", type, addr, + ioapic->gsi_base); + return 0; + +exit_release: + pci_release_region(dev, 0); +exit_disable: + pci_disable_device(dev); +exit_free: + kfree(ioapic); + return -ENODEV; +} + +static void ioapic_remove(struct pci_dev *dev) +{ + struct ioapic *ioapic = pci_get_drvdata(dev); + + acpi_unregister_ioapic(ioapic->handle, ioapic->gsi_base); + pci_release_region(dev, 0); + pci_disable_device(dev); + kfree(ioapic); +} + + +static struct pci_device_id ioapic_devices[] = { + { PCI_ANY_ID, PCI_ANY_ID, PCI_ANY_ID, PCI_ANY_ID, + PCI_CLASS_SYSTEM_PIC_IOAPIC << 8, 0xffff00, }, + { PCI_ANY_ID, PCI_ANY_ID, PCI_ANY_ID, PCI_ANY_ID, + PCI_CLASS_SYSTEM_PIC_IOXAPIC << 8, 0xffff00, }, + { } +}; + +static struct pci_driver ioapic_driver = { + .name = "ioapic", + .id_table = ioapic_devices, + .probe = ioapic_probe, + .remove = __devexit_p(ioapic_remove), +}; + +static int __init ioapic_init(void) +{ + return pci_register_driver(&ioapic_driver); +} + +static void __exit ioapic_exit(void) +{ + pci_unregister_driver(&ioapic_driver); +} + +module_init(ioapic_init); +module_exit(ioapic_exit); From 58c08628c4fe664bfd5f8b7e773b4b157bb9686f Mon Sep 17 00:00:00 2001 From: Alex Chiang Date: Mon, 26 Oct 2009 21:25:27 -0600 Subject: [PATCH 036/109] PCI Hotplug: acpiphp: clean up list traversals Using list_for_each_entry instead of list_for_each allows us to enhance readability and minorly reduce some stack usage. Signed-off-by: Alex Chiang Signed-off-by: Jesse Barnes --- drivers/pci/hotplug/acpiphp_glue.c | 58 +++++++++--------------------- 1 file changed, 17 insertions(+), 41 deletions(-) diff --git a/drivers/pci/hotplug/acpiphp_glue.c b/drivers/pci/hotplug/acpiphp_glue.c index 392e4b3fa2e4..8f4a2073d83a 100644 --- a/drivers/pci/hotplug/acpiphp_glue.c +++ b/drivers/pci/hotplug/acpiphp_glue.c @@ -309,17 +309,13 @@ static void init_bridge_misc(struct acpiphp_bridge *bridge) /* find acpiphp_func from acpiphp_bridge */ static struct acpiphp_func *acpiphp_bridge_handle_to_function(acpi_handle handle) { - struct list_head *node, *l; struct acpiphp_bridge *bridge; struct acpiphp_slot *slot; struct acpiphp_func *func; - list_for_each(node, &bridge_list) { - bridge = list_entry(node, struct acpiphp_bridge, list); + list_for_each_entry(bridge, &bridge_list, list) { for (slot = bridge->slots; slot; slot = slot->next) { - list_for_each(l, &slot->funcs) { - func = list_entry(l, struct acpiphp_func, - sibling); + list_for_each_entry(func, &slot->funcs, sibling) { if (func->handle == handle) return func; } @@ -493,21 +489,19 @@ static int add_bridge(acpi_handle handle) static struct acpiphp_bridge *acpiphp_handle_to_bridge(acpi_handle handle) { - struct list_head *head; - list_for_each(head, &bridge_list) { - struct acpiphp_bridge *bridge = list_entry(head, - struct acpiphp_bridge, list); + struct acpiphp_bridge *bridge; + + list_for_each_entry(bridge, &bridge_list, list) if (bridge->handle == handle) return bridge; - } return NULL; } static void cleanup_bridge(struct acpiphp_bridge *bridge) { - struct list_head *list, *tmp; - struct acpiphp_slot *slot; + struct acpiphp_slot *slot, *next; + struct acpiphp_func *func, *tmp; acpi_status status; acpi_handle handle = bridge->handle; @@ -528,10 +522,8 @@ static void cleanup_bridge(struct acpiphp_bridge *bridge) slot = bridge->slots; while (slot) { - struct acpiphp_slot *next = slot->next; - list_for_each_safe (list, tmp, &slot->funcs) { - struct acpiphp_func *func; - func = list_entry(list, struct acpiphp_func, sibling); + next = slot->next; + list_for_each_entry_safe(func, tmp, &slot->funcs, sibling) { if (is_dock_device(func->handle)) { unregister_hotplug_dock_device(func->handle); unregister_dock_notifier(&func->nb); @@ -543,7 +535,7 @@ static void cleanup_bridge(struct acpiphp_bridge *bridge) if (ACPI_FAILURE(status)) err("failed to remove notify handler\n"); } - list_del(list); + list_del(&func->sibling); kfree(func); } acpiphp_unregister_hotplug_slot(slot); @@ -608,16 +600,13 @@ static int power_on_slot(struct acpiphp_slot *slot) { acpi_status status; struct acpiphp_func *func; - struct list_head *l; int retval = 0; /* if already enabled, just skip */ if (slot->flags & SLOT_POWEREDON) goto err_exit; - list_for_each (l, &slot->funcs) { - func = list_entry(l, struct acpiphp_func, sibling); - + list_for_each_entry(func, &slot->funcs, sibling) { if (func->flags & FUNC_HAS_PS0) { dbg("%s: executing _PS0\n", __func__); status = acpi_evaluate_object(func->handle, "_PS0", NULL, NULL); @@ -643,7 +632,6 @@ static int power_off_slot(struct acpiphp_slot *slot) { acpi_status status; struct acpiphp_func *func; - struct list_head *l; int retval = 0; @@ -651,9 +639,7 @@ static int power_off_slot(struct acpiphp_slot *slot) if ((slot->flags & SLOT_POWEREDON) == 0) goto err_exit; - list_for_each (l, &slot->funcs) { - func = list_entry(l, struct acpiphp_func, sibling); - + list_for_each_entry(func, &slot->funcs, sibling) { if (func->flags & FUNC_HAS_PS3) { status = acpi_evaluate_object(func->handle, "_PS3", NULL, NULL); if (ACPI_FAILURE(status)) { @@ -780,7 +766,6 @@ static int __ref enable_device(struct acpiphp_slot *slot) { struct pci_dev *dev; struct pci_bus *bus = slot->bridge->pci_bus; - struct list_head *l; struct acpiphp_func *func; int retval = 0; int num, max, pass; @@ -820,10 +805,8 @@ static int __ref enable_device(struct acpiphp_slot *slot) } } - list_for_each (l, &slot->funcs) { - func = list_entry(l, struct acpiphp_func, sibling); + list_for_each_entry(func, &slot->funcs, sibling) acpiphp_bus_add(func); - } pci_bus_assign_resources(bus); acpiphp_sanitize_bus(bus); @@ -831,8 +814,7 @@ static int __ref enable_device(struct acpiphp_slot *slot) pci_enable_bridges(bus); pci_bus_add_devices(bus); - list_for_each (l, &slot->funcs) { - func = list_entry(l, struct acpiphp_func, sibling); + list_for_each_entry(func, &slot->funcs, sibling) { dev = pci_get_slot(bus, PCI_DEVFN(slot->device, func->function)); if (!dev) @@ -930,12 +912,9 @@ static unsigned int get_slot_status(struct acpiphp_slot *slot) acpi_status status; unsigned long long sta = 0; u32 dvid; - struct list_head *l; struct acpiphp_func *func; - list_for_each (l, &slot->funcs) { - func = list_entry(l, struct acpiphp_func, sibling); - + list_for_each_entry(func, &slot->funcs, sibling) { if (func->flags & FUNC_HAS_STA) { status = acpi_evaluate_integer(func->handle, "_STA", NULL, &sta); if (ACPI_SUCCESS(status) && sta) @@ -963,13 +942,10 @@ int acpiphp_eject_slot(struct acpiphp_slot *slot) { acpi_status status; struct acpiphp_func *func; - struct list_head *l; struct acpi_object_list arg_list; union acpi_object arg; - list_for_each (l, &slot->funcs) { - func = list_entry(l, struct acpiphp_func, sibling); - + list_for_each_entry(func, &slot->funcs, sibling) { /* We don't want to call _EJ0 on non-existing functions. */ if ((func->flags & FUNC_HAS_EJ0)) { /* _EJ0 method take one argument */ @@ -1352,7 +1328,7 @@ int __init acpiphp_get_num_slots(void) struct acpiphp_bridge *bridge; int num_slots = 0; - list_for_each_entry (bridge, &bridge_list, list) { + list_for_each_entry(bridge, &bridge_list, list) { dbg("Bus %04x:%02x has %d slot%s\n", pci_domain_nr(bridge->pci_bus), bridge->pci_bus->number, bridge->nr_slots, From 4fd8bdc567e70c02fab7eeaaa7d2a64232add789 Mon Sep 17 00:00:00 2001 From: Stefan Assmann Date: Tue, 27 Oct 2009 08:57:42 +0100 Subject: [PATCH 037/109] PCI: avoid boot interrupt quirk for AMD 813x B1 devices AMD 813x rev. B1 (like rev. B2) devices generate no interrupts if quirk_disable_amd_813x_boot_interrupt is executed, add an exception. http://bugzilla.kernel.org/show_bug.cgi?id=14159 Patch also adds missing cases for DECLARE_PCI_FIXUP_RESUME and DECLARE_PCI_FIXUP_FINAL calls to quirk_disable_amd_813x_boot_interrupt. Signed-off-by: Stefan Assmann Tested-by: Gabriele Giorgetti Signed-off-by: Jesse Barnes --- drivers/pci/quirks.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c index 1812ae7698de..c0c4537d66da 100644 --- a/drivers/pci/quirks.c +++ b/drivers/pci/quirks.c @@ -1680,6 +1680,7 @@ DECLARE_PCI_FIXUP_RESUME(PCI_VENDOR_ID_SERVERWORKS, PCI_DEVICE_ID_SERVERWORKS_ */ #define AMD_813X_MISC 0x40 #define AMD_813X_NOIOAMODE (1<<0) +#define AMD_813X_REV_B1 0x12 #define AMD_813X_REV_B2 0x13 static void quirk_disable_amd_813x_boot_interrupt(struct pci_dev *dev) @@ -1688,7 +1689,8 @@ static void quirk_disable_amd_813x_boot_interrupt(struct pci_dev *dev) if (noioapicquirk) return; - if (dev->revision == AMD_813X_REV_B2) + if ((dev->revision == AMD_813X_REV_B1) || + (dev->revision == AMD_813X_REV_B2)) return; pci_read_config_dword(dev, AMD_813X_MISC, &pci_config_dword); @@ -1698,8 +1700,10 @@ static void quirk_disable_amd_813x_boot_interrupt(struct pci_dev *dev) dev_info(&dev->dev, "disabled boot interrupts on device [%04x:%04x]\n", dev->vendor, dev->device); } -DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_8131_BRIDGE, quirk_disable_amd_813x_boot_interrupt); -DECLARE_PCI_FIXUP_RESUME(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_8132_BRIDGE, quirk_disable_amd_813x_boot_interrupt); +DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_8131_BRIDGE, quirk_disable_amd_813x_boot_interrupt); +DECLARE_PCI_FIXUP_RESUME(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_8131_BRIDGE, quirk_disable_amd_813x_boot_interrupt); +DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_8132_BRIDGE, quirk_disable_amd_813x_boot_interrupt); +DECLARE_PCI_FIXUP_RESUME(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_8132_BRIDGE, quirk_disable_amd_813x_boot_interrupt); #define AMD_8111_PCI_IRQ_ROUTING 0x56 From c7dabef8a2c59e6a3de9d66fc35fb6a43ef7172d Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Tue, 27 Oct 2009 13:26:47 -0600 Subject: [PATCH 038/109] vsprintf: use %pR, %pr instead of %pRt, %pRf Jesse accidentally applied v1 [1] of the patchset instead of v2 [2]. This is the diff between v1 and v2. The changes in this patch are: - tidied vsprintf stack buffer to shrink and compute size more accurately - use %pR for decoding and %pr for "raw" (with type and flags) instead of adding %pRt and %pRf [1] http://lkml.org/lkml/2009/10/6/491 [2] http://lkml.org/lkml/2009/10/13/441 Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/ia64/pci/pci.c | 11 +++--- arch/x86/pci/acpi.c | 7 ++-- arch/x86/pci/i386.c | 11 +++--- drivers/pci/pci.c | 2 +- drivers/pci/probe.c | 17 ++++----- drivers/pci/quirks.c | 2 +- drivers/pci/setup-bus.c | 79 +++++++++++++++++++---------------------- drivers/pci/setup-res.c | 20 ++++++----- drivers/pnp/quirks.c | 5 +-- drivers/pnp/resource.c | 8 ++--- drivers/pnp/support.c | 2 +- drivers/pnp/system.c | 2 +- lib/vsprintf.c | 55 +++++++++++++++------------- 13 files changed, 114 insertions(+), 107 deletions(-) diff --git a/arch/ia64/pci/pci.c b/arch/ia64/pci/pci.c index 06413b827e97..df639db779f9 100644 --- a/arch/ia64/pci/pci.c +++ b/arch/ia64/pci/pci.c @@ -298,18 +298,19 @@ static __devinit acpi_status add_window(struct acpi_resource *res, void *data) window->offset = offset; if (insert_resource(root, &window->resource)) { - dev_err(&info->bridge->dev, "can't allocate %pRt\n", + dev_err(&info->bridge->dev, + "can't allocate host bridge window %pR\n", &window->resource); } else { if (offset) - dev_info(&info->bridge->dev, "host bridge window: %pRt " + dev_info(&info->bridge->dev, "host bridge window %pR " "(PCI address [%#llx-%#llx])\n", &window->resource, window->resource.start - offset, window->resource.end - offset); else dev_info(&info->bridge->dev, - "host bridge window: %pRt\n", + "host bridge window %pR\n", &window->resource); } @@ -330,7 +331,9 @@ pcibios_setup_root_windows(struct pci_bus *bus, struct pci_controller *ctrl) (res->end - res->start < 16)) continue; if (j >= PCI_BUS_NUM_RESOURCES) { - dev_warn(&bus->dev, "ignoring %pRf (no space)\n", res); + dev_warn(&bus->dev, + "ignoring host bridge window %pR (no space)\n", + res); continue; } bus->resource[j++] = res; diff --git a/arch/x86/pci/acpi.c b/arch/x86/pci/acpi.c index 9b3daf976732..6bf8091d2fd5 100644 --- a/arch/x86/pci/acpi.c +++ b/arch/x86/pci/acpi.c @@ -108,18 +108,19 @@ setup_resource(struct acpi_resource *acpi_res, void *data) res->child = NULL; if (insert_resource(root, res)) { - dev_err(&info->bridge->dev, "can't allocate %pRt\n", res); + dev_err(&info->bridge->dev, + "can't allocate host bridge window %pR\n", res); } else { info->bus->resource[info->res_num] = res; info->res_num++; if (addr.translation_offset) - dev_info(&info->bridge->dev, "host bridge window: %pRt " + dev_info(&info->bridge->dev, "host bridge window %pR " "(PCI address [%#llx-%#llx])\n", res, res->start - addr.translation_offset, res->end - addr.translation_offset); else dev_info(&info->bridge->dev, - "host bridge window: %pRt\n", res); + "host bridge window %pR\n", res); } return AE_OK; } diff --git a/arch/x86/pci/i386.c b/arch/x86/pci/i386.c index 52e656f17781..d49d17de7b3f 100644 --- a/arch/x86/pci/i386.c +++ b/arch/x86/pci/i386.c @@ -129,7 +129,7 @@ static void __init pcibios_allocate_bus_resources(struct list_head *bus_list) continue; if (!r->start || pci_claim_resource(dev, idx) < 0) { - dev_info(&dev->dev, "BAR %d: can't allocate %pRt\n", idx, r); + dev_info(&dev->dev, "BAR %d: can't allocate %pR\n", idx, r); /* * Something is wrong with the region. * Invalidate the resource to prevent @@ -164,10 +164,11 @@ static void __init pcibios_allocate_resources(int pass) else disabled = !(command & PCI_COMMAND_MEMORY); if (pass == disabled) { - dev_dbg(&dev->dev, "%pRf (d=%d, p=%d)\n", r, - disabled, pass); + dev_dbg(&dev->dev, + "BAR %d: claiming %pr (d=%d, p=%d)\n", + idx, r, disabled, pass); if (pci_claim_resource(dev, idx) < 0) { - dev_info(&dev->dev, "BAR %d: can't allocate %pRt\n", idx, r); + dev_info(&dev->dev, "BAR %d: can't claim %pR\n", idx, r); /* We'll assign a new address later */ r->end -= r->start; r->start = 0; @@ -180,7 +181,7 @@ static void __init pcibios_allocate_resources(int pass) /* Turn the ROM off, leave the resource region, * but keep it unregistered. */ u32 reg; - dev_dbg(&dev->dev, "disabling ROM %pRt\n", r); + dev_dbg(&dev->dev, "disabling ROM %pR\n", r); r->flags &= ~IORESOURCE_ROM_ENABLE; pci_read_config_dword(dev, dev->rom_base_reg, ®); diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index 557218222826..f0da1676d2be 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -1713,7 +1713,7 @@ static int __pci_request_region(struct pci_dev *pdev, int bar, const char *res_n return 0; err_out: - dev_warn(&pdev->dev, "BAR %d: can't reserve %pRt\n", bar, + dev_warn(&pdev->dev, "BAR %d: can't reserve %pR\n", bar, &pdev->resource[bar]); return -EBUSY; } diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 4842b09b7f3c..4c4aca53ae09 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -225,12 +225,13 @@ int __pci_read_base(struct pci_dev *dev, enum pci_bar_type type, if (!sz64) goto fail; - res->flags |= IORESOURCE_MEM_64; - if ((sizeof(resource_size_t) < 8) && (sz64 > 0x100000000ULL)) { dev_err(&dev->dev, "can't handle 64-bit BAR\n"); goto fail; - } else if ((sizeof(resource_size_t) < 8) && l) { + } + + res->flags |= IORESOURCE_MEM_64; + if ((sizeof(resource_size_t) < 8) && l) { /* Address above 32-bit boundary; disable the BAR */ pci_write_config_dword(dev, pos, 0); pci_write_config_dword(dev, pos + 4, 0); @@ -239,7 +240,7 @@ int __pci_read_base(struct pci_dev *dev, enum pci_bar_type type, } else { res->start = l64; res->end = l64 + sz64; - dev_printk(KERN_DEBUG, &dev->dev, "reg %x: %pRt\n", + dev_printk(KERN_DEBUG, &dev->dev, "reg %x: %pR\n", pos, res); } } else { @@ -251,7 +252,7 @@ int __pci_read_base(struct pci_dev *dev, enum pci_bar_type type, res->start = l; res->end = l + sz; - dev_printk(KERN_DEBUG, &dev->dev, "reg %x: %pRt\n", pos, res); + dev_printk(KERN_DEBUG, &dev->dev, "reg %x: %pR\n", pos, res); } out: @@ -319,7 +320,7 @@ void __devinit pci_read_bridge_bases(struct pci_bus *child) res->start = base; if (!res->end) res->end = limit + 0xfff; - dev_printk(KERN_DEBUG, &dev->dev, "bridge window: %pRt\n", res); + dev_printk(KERN_DEBUG, &dev->dev, " bridge window %pR\n", res); } res = child->resource[1]; @@ -331,7 +332,7 @@ void __devinit pci_read_bridge_bases(struct pci_bus *child) res->flags = (mem_base_lo & PCI_MEMORY_RANGE_TYPE_MASK) | IORESOURCE_MEM; res->start = base; res->end = limit + 0xfffff; - dev_printk(KERN_DEBUG, &dev->dev, "bridge window: %pRt\n", res); + dev_printk(KERN_DEBUG, &dev->dev, " bridge window %pR\n", res); } res = child->resource[2]; @@ -370,7 +371,7 @@ void __devinit pci_read_bridge_bases(struct pci_bus *child) res->flags |= IORESOURCE_MEM_64; res->start = base; res->end = limit + 0xfffff; - dev_printk(KERN_DEBUG, &dev->dev, "bridge window: %pRt\n", res); + dev_printk(KERN_DEBUG, &dev->dev, " bridge window %pR\n", res); } } diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c index c0c4537d66da..7cfa7c38d318 100644 --- a/drivers/pci/quirks.c +++ b/drivers/pci/quirks.c @@ -357,7 +357,7 @@ static void __devinit quirk_io_region(struct pci_dev *dev, unsigned region, pcibios_bus_to_resource(dev, res, &bus_region); pci_claim_resource(dev, nr); - dev_info(&dev->dev, "quirk: region %04x-%04x claimed by %s\n", region, region + size - 1, name); + dev_info(&dev->dev, "quirk: %pR claimed by %s\n", res, name); } } diff --git a/drivers/pci/setup-bus.c b/drivers/pci/setup-bus.c index ceb75333862b..ed6916bac675 100644 --- a/drivers/pci/setup-bus.c +++ b/drivers/pci/setup-bus.c @@ -71,53 +71,50 @@ static void pbus_assign_resources_sorted(const struct pci_bus *bus) void pci_setup_cardbus(struct pci_bus *bus) { struct pci_dev *bridge = bus->self; + struct resource *res; struct pci_bus_region region; dev_info(&bridge->dev, "CardBus bridge, secondary bus %04x:%02x\n", pci_domain_nr(bus), bus->number); - pcibios_resource_to_bus(bridge, ®ion, bus->resource[0]); - if (bus->resource[0]->flags & IORESOURCE_IO) { + res = bus->resource[0]; + pcibios_resource_to_bus(bridge, ®ion, res); + if (res->flags & IORESOURCE_IO) { /* * The IO resource is allocated a range twice as large as it * would normally need. This allows us to set both IO regs. */ - dev_info(&bridge->dev, " IO window: %#08lx-%#08lx\n", - (unsigned long)region.start, - (unsigned long)region.end); + dev_info(&bridge->dev, " bridge window %pR\n", res); pci_write_config_dword(bridge, PCI_CB_IO_BASE_0, region.start); pci_write_config_dword(bridge, PCI_CB_IO_LIMIT_0, region.end); } - pcibios_resource_to_bus(bridge, ®ion, bus->resource[1]); - if (bus->resource[1]->flags & IORESOURCE_IO) { - dev_info(&bridge->dev, " IO window: %#08lx-%#08lx\n", - (unsigned long)region.start, - (unsigned long)region.end); + res = bus->resource[1]; + pcibios_resource_to_bus(bridge, ®ion, res); + if (res->flags & IORESOURCE_IO) { + dev_info(&bridge->dev, " bridge window %pR\n", res); pci_write_config_dword(bridge, PCI_CB_IO_BASE_1, region.start); pci_write_config_dword(bridge, PCI_CB_IO_LIMIT_1, region.end); } - pcibios_resource_to_bus(bridge, ®ion, bus->resource[2]); - if (bus->resource[2]->flags & IORESOURCE_MEM) { - dev_info(&bridge->dev, " PREFETCH window: %#08lx-%#08lx\n", - (unsigned long)region.start, - (unsigned long)region.end); + res = bus->resource[2]; + pcibios_resource_to_bus(bridge, ®ion, res); + if (res->flags & IORESOURCE_MEM) { + dev_info(&bridge->dev, " bridge window %pR\n", res); pci_write_config_dword(bridge, PCI_CB_MEMORY_BASE_0, region.start); pci_write_config_dword(bridge, PCI_CB_MEMORY_LIMIT_0, region.end); } - pcibios_resource_to_bus(bridge, ®ion, bus->resource[3]); - if (bus->resource[3]->flags & IORESOURCE_MEM) { - dev_info(&bridge->dev, " MEM window: %#08lx-%#08lx\n", - (unsigned long)region.start, - (unsigned long)region.end); + res = bus->resource[3]; + pcibios_resource_to_bus(bridge, ®ion, res); + if (res->flags & IORESOURCE_MEM) { + dev_info(&bridge->dev, " bridge window %pR\n", res); pci_write_config_dword(bridge, PCI_CB_MEMORY_BASE_1, region.start); pci_write_config_dword(bridge, PCI_CB_MEMORY_LIMIT_1, @@ -140,6 +137,7 @@ EXPORT_SYMBOL(pci_setup_cardbus); static void pci_setup_bridge(struct pci_bus *bus) { struct pci_dev *bridge = bus->self; + struct resource *res; struct pci_bus_region region; u32 l, bu, lu, io_upper16; int pref_mem64; @@ -151,23 +149,22 @@ static void pci_setup_bridge(struct pci_bus *bus) pci_domain_nr(bus), bus->number); /* Set up the top and bottom of the PCI I/O segment for this bus. */ - pcibios_resource_to_bus(bridge, ®ion, bus->resource[0]); - if (bus->resource[0]->flags & IORESOURCE_IO) { + res = bus->resource[0]; + pcibios_resource_to_bus(bridge, ®ion, res); + if (res->flags & IORESOURCE_IO) { pci_read_config_dword(bridge, PCI_IO_BASE, &l); l &= 0xffff0000; l |= (region.start >> 8) & 0x00f0; l |= region.end & 0xf000; /* Set up upper 16 bits of I/O base/limit. */ io_upper16 = (region.end & 0xffff0000) | (region.start >> 16); - dev_info(&bridge->dev, " IO window: %#04lx-%#04lx\n", - (unsigned long)region.start, - (unsigned long)region.end); + dev_info(&bridge->dev, " bridge window %pR\n", res); } else { /* Clear upper 16 bits of I/O base/limit. */ io_upper16 = 0; l = 0x00f0; - dev_info(&bridge->dev, " IO window: disabled\n"); + dev_info(&bridge->dev, " bridge window [io disabled]\n"); } /* Temporarily disable the I/O range before updating PCI_IO_BASE. */ pci_write_config_dword(bridge, PCI_IO_BASE_UPPER16, 0x0000ffff); @@ -178,17 +175,16 @@ static void pci_setup_bridge(struct pci_bus *bus) /* Set up the top and bottom of the PCI Memory segment for this bus. */ - pcibios_resource_to_bus(bridge, ®ion, bus->resource[1]); - if (bus->resource[1]->flags & IORESOURCE_MEM) { + res = bus->resource[1]; + pcibios_resource_to_bus(bridge, ®ion, res); + if (res->flags & IORESOURCE_MEM) { l = (region.start >> 16) & 0xfff0; l |= region.end & 0xfff00000; - dev_info(&bridge->dev, " MEM window: %#08lx-%#08lx\n", - (unsigned long)region.start, - (unsigned long)region.end); + dev_info(&bridge->dev, " bridge window %pR\n", res); } else { l = 0x0000fff0; - dev_info(&bridge->dev, " MEM window: disabled\n"); + dev_info(&bridge->dev, " bridge window [mem disabled]\n"); } pci_write_config_dword(bridge, PCI_MEMORY_BASE, l); @@ -200,24 +196,21 @@ static void pci_setup_bridge(struct pci_bus *bus) /* Set up PREF base/limit. */ pref_mem64 = 0; bu = lu = 0; - pcibios_resource_to_bus(bridge, ®ion, bus->resource[2]); - if (bus->resource[2]->flags & IORESOURCE_PREFETCH) { - int width = 8; + res = bus->resource[2]; + pcibios_resource_to_bus(bridge, ®ion, res); + if (res->flags & IORESOURCE_PREFETCH) { l = (region.start >> 16) & 0xfff0; l |= region.end & 0xfff00000; - if (bus->resource[2]->flags & IORESOURCE_MEM_64) { + if (res->flags & IORESOURCE_MEM_64) { pref_mem64 = 1; bu = upper_32_bits(region.start); lu = upper_32_bits(region.end); - width = 16; } - dev_info(&bridge->dev, " PREFETCH window: %#0*llx-%#0*llx\n", - width, (unsigned long long)region.start, - width, (unsigned long long)region.end); + dev_info(&bridge->dev, " bridge window %pR\n", res); } else { l = 0x0000fff0; - dev_info(&bridge->dev, " PREFETCH window: disabled\n"); + dev_info(&bridge->dev, " bridge window [mem pref disabled]\n"); } pci_write_config_dword(bridge, PCI_PREF_MEMORY_BASE, l); @@ -391,7 +384,7 @@ static int pbus_size_mem(struct pci_bus *bus, unsigned long mask, order = __ffs(align) - 20; if (order > 11) { dev_warn(&dev->dev, "BAR %d: bad alignment %llx: " - "%pRt\n", i, (unsigned long long)align, r); + "%pR\n", i, (unsigned long long)align, r); r->flags = 0; continue; } @@ -582,7 +575,7 @@ static void pci_bus_dump_res(struct pci_bus *bus) if (!res || !res->end) continue; - dev_printk(KERN_DEBUG, &bus->dev, "resource %d %pRt\n", i, res); + dev_printk(KERN_DEBUG, &bus->dev, "resource %d %pR\n", i, res); } } diff --git a/drivers/pci/setup-res.c b/drivers/pci/setup-res.c index 5e78f2096ce8..357ca5c54607 100644 --- a/drivers/pci/setup-res.c +++ b/drivers/pci/setup-res.c @@ -51,9 +51,11 @@ void pci_update_resource(struct pci_dev *dev, int resno) pcibios_resource_to_bus(dev, ®ion, res); - dev_dbg(&dev->dev, "BAR %d: got %pRf (bus addr [%#llx-%#llx])\n", - resno, res, (unsigned long long)region.start, - (unsigned long long)region.end); + dev_dbg(&dev->dev, "BAR %d: got res %pR bus [%#llx-%#llx] " + "flags %#lx\n", resno, res, + (unsigned long long)region.start, + (unsigned long long)region.end, + (unsigned long)res->flags); new = region.start | (res->flags & PCI_REGION_FLAG_MASK); if (res->flags & IORESOURCE_IO) @@ -89,8 +91,8 @@ void pci_update_resource(struct pci_dev *dev, int resno) } } res->flags &= ~IORESOURCE_UNSET; - dev_dbg(&dev->dev, "BAR %d: moved to bus addr [%#llx-%#llx]\n", - resno, (unsigned long long)region.start, + dev_dbg(&dev->dev, "BAR %d: moved to %pR (bus addr [%#llx-%#llx])\n", + resno, res, (unsigned long long)region.start, (unsigned long long)region.end); } @@ -108,7 +110,7 @@ int pci_claim_resource(struct pci_dev *dev, int resource) if (err) { const char *dtype = resource < PCI_BRIDGE_RESOURCES ? "device" : "bridge"; - dev_err(&dev->dev, "BAR %d: %s %s %pRt\n", + dev_err(&dev->dev, "BAR %d: %s of %s %pR\n", resource, root ? "address space collision on" : "no parent found for", @@ -179,7 +181,7 @@ int pci_assign_resource(struct pci_dev *dev, int resno) align = pci_resource_alignment(dev, res); if (!align) { - dev_info(&dev->dev, "BAR %d: can't allocate %pRf " + dev_info(&dev->dev, "BAR %d: can't allocate %pR " "(bogus alignment)\n", resno, res); return -EINVAL; } @@ -196,7 +198,7 @@ int pci_assign_resource(struct pci_dev *dev, int resno) } if (ret) - dev_info(&dev->dev, "BAR %d: can't allocate %pRt\n", + dev_info(&dev->dev, "BAR %d: can't allocate %pR\n", resno, res); return ret; @@ -222,7 +224,7 @@ void pdev_sort_resources(struct pci_dev *dev, struct resource_list *head) r_align = pci_resource_alignment(dev, r); if (!r_align) { - dev_warn(&dev->dev, "BAR %d: bogus alignment %pRf\n", + dev_warn(&dev->dev, "BAR %d: %pR has bogus alignment\n", i, r); continue; } diff --git a/drivers/pnp/quirks.c b/drivers/pnp/quirks.c index 3a2031b25c3c..dfbd5a6cc58b 100644 --- a/drivers/pnp/quirks.c +++ b/drivers/pnp/quirks.c @@ -285,8 +285,9 @@ static void quirk_system_pci_resources(struct pnp_dev *dev) * the PCI region, and that might prevent a PCI * driver from requesting its resources. */ - dev_warn(&dev->dev, "resource %pRt overlaps %s " - "BAR %d %pRt, disabling\n", res, + dev_warn(&dev->dev, + "disabling %pR because it overlaps " + "%s BAR %d %pR\n", res, pci_name(pdev), i, &pdev->resource[i]); res->flags |= IORESOURCE_DISABLED; } diff --git a/drivers/pnp/resource.c b/drivers/pnp/resource.c index 18557d7bb0b9..64d0596bafb5 100644 --- a/drivers/pnp/resource.c +++ b/drivers/pnp/resource.c @@ -517,7 +517,7 @@ struct pnp_resource *pnp_add_irq_resource(struct pnp_dev *dev, int irq, res->start = irq; res->end = irq; - pnp_dbg(&dev->dev, " add %pRf\n", res); + pnp_dbg(&dev->dev, " add %pr\n", res); return pnp_res; } @@ -538,7 +538,7 @@ struct pnp_resource *pnp_add_dma_resource(struct pnp_dev *dev, int dma, res->start = dma; res->end = dma; - pnp_dbg(&dev->dev, " add %pRf\n", res); + pnp_dbg(&dev->dev, " add %pr\n", res); return pnp_res; } @@ -562,7 +562,7 @@ struct pnp_resource *pnp_add_io_resource(struct pnp_dev *dev, res->start = start; res->end = end; - pnp_dbg(&dev->dev, " add %pRf\n", res); + pnp_dbg(&dev->dev, " add %pr\n", res); return pnp_res; } @@ -586,7 +586,7 @@ struct pnp_resource *pnp_add_mem_resource(struct pnp_dev *dev, res->start = start; res->end = end; - pnp_dbg(&dev->dev, " add %pRf\n", res); + pnp_dbg(&dev->dev, " add %pr\n", res); return pnp_res; } diff --git a/drivers/pnp/support.c b/drivers/pnp/support.c index 1f8a33b0abac..9585c1c1cc36 100644 --- a/drivers/pnp/support.c +++ b/drivers/pnp/support.c @@ -82,7 +82,7 @@ void dbg_pnp_show_resources(struct pnp_dev *dev, char *desc) else { pnp_dbg(&dev->dev, "%s: current resources:\n", desc); list_for_each_entry(pnp_res, &dev->resources, list) - pnp_dbg(&dev->dev, "%pRf\n", &pnp_res->res); + pnp_dbg(&dev->dev, "%pr\n", &pnp_res->res); } } diff --git a/drivers/pnp/system.c b/drivers/pnp/system.c index 242d3a872011..49c1720df59a 100644 --- a/drivers/pnp/system.c +++ b/drivers/pnp/system.c @@ -48,7 +48,7 @@ static void reserve_range(struct pnp_dev *dev, struct resource *r, int port) * example do reserve stuff they know about too, so we may well * have double reservations. */ - dev_info(&dev->dev, "%pRt %s reserved\n", r, + dev_info(&dev->dev, "%pR %s reserved\n", r, res ? "has been" : "could not be"); } diff --git a/lib/vsprintf.c b/lib/vsprintf.c index a6e195163eb3..6438cd5599ee 100644 --- a/lib/vsprintf.c +++ b/lib/vsprintf.c @@ -624,13 +624,19 @@ static char *resource_string(char *buf, char *end, struct resource *res, .precision = -1, .flags = SPECIAL | SMALL, }; - /* - * room for three actual numbers (decimal or hex), plus - * "[mem 0x-0x 64bit pref disabled flags 0x]\0" - */ - char sym[3*3*sizeof(resource_size_t) + 41]; + + /* 32-bit res (sizeof==4): 10 chars in dec, 10 in hex ("0x" + 8) + * 64-bit res (sizeof==8): 20 chars in dec, 18 in hex ("0x" + 16) */ +#define RSRC_BUF_SIZE ((2 * sizeof(resource_size_t)) + 4) +#define FLAG_BUF_SIZE (2 * sizeof(res->flags)) +#define DECODED_BUF_SIZE sizeof("[mem - 64bit pref disabled]") +#define RAW_BUF_SIZE sizeof("[mem - flags 0x]") + char sym[max(2*RSRC_BUF_SIZE + DECODED_BUF_SIZE, + 2*RSRC_BUF_SIZE + FLAG_BUF_SIZE + RAW_BUF_SIZE)]; + char *p = sym, *pend = sym + sizeof(sym); int size = -1, addr = 0; + int decode = (fmt[0] == 'R') ? 1 : 0; if (res->flags & IORESOURCE_IO) { size = IO_RSRC_PRINTK_SIZE; @@ -641,15 +647,17 @@ static char *resource_string(char *buf, char *end, struct resource *res, } *p++ = '['; - if (fmt[1] == 't' || fmt[1] == 'f') { - if (res->flags & IORESOURCE_IO) - p = string(p, pend, "io ", str_spec); - else if (res->flags & IORESOURCE_MEM) - p = string(p, pend, "mem ", str_spec); - else if (res->flags & IORESOURCE_IRQ) - p = string(p, pend, "irq ", str_spec); - else if (res->flags & IORESOURCE_DMA) - p = string(p, pend, "dma ", str_spec); + if (res->flags & IORESOURCE_IO) + p = string(p, pend, "io ", str_spec); + else if (res->flags & IORESOURCE_MEM) + p = string(p, pend, "mem ", str_spec); + else if (res->flags & IORESOURCE_IRQ) + p = string(p, pend, "irq ", str_spec); + else if (res->flags & IORESOURCE_DMA) + p = string(p, pend, "dma ", str_spec); + else { + p = string(p, pend, "??? ", str_spec); + decode = 0; } hex_spec.field_width = size; p = number(p, pend, res->start, addr ? hex_spec : dec_spec); @@ -657,21 +665,19 @@ static char *resource_string(char *buf, char *end, struct resource *res, *p++ = '-'; p = number(p, pend, res->end, addr ? hex_spec : dec_spec); } - if (fmt[1] == 't' || fmt[1] == 'f') { + if (decode) { if (res->flags & IORESOURCE_MEM_64) p = string(p, pend, " 64bit", str_spec); if (res->flags & IORESOURCE_PREFETCH) p = string(p, pend, " pref", str_spec); if (res->flags & IORESOURCE_DISABLED) p = string(p, pend, " disabled", str_spec); - if (fmt[1] == 'f') { - p = string(p, pend, " flags ", str_spec); - p = number(p, pend, res->flags & ~IORESOURCE_TYPE_BITS, - flag_spec); - } + } else { + p = string(p, pend, " flags ", str_spec); + p = number(p, pend, res->flags, flag_spec); } *p++ = ']'; - *p = 0; + *p = '\0'; return string(buf, end, sym, spec); } @@ -847,10 +853,8 @@ static char *ip4_addr_string(char *buf, char *end, const u8 *addr, * - 'f' For simple symbolic function names without offset * - 'S' For symbolic direct pointers with offset * - 's' For symbolic direct pointers without offset - * - 'R' For a struct resource pointer, print: - * R address range only ([0x0-0x1f]) - * Rt type and range ([mem 0x0-0x1f 64bit pref]) - * Rf type, range, and flags ([mem 0x0-0x1f 64bit pref flags 0x1]) + * - 'R' For decoded struct resource, e.g., [mem 0x0-0x1f 64bit pref] + * - 'r' For raw struct resource, e.g., [mem 0x0-0x1f flags 0x201] * - 'M' For a 6-byte MAC address, it prints the address in the * usual colon-separated hex notation * - 'm' For a 6-byte MAC address, it prints the hex address without colons @@ -881,6 +885,7 @@ static char *pointer(const char *fmt, char *buf, char *end, void *ptr, case 'S': return symbol_string(buf, end, ptr, spec, *fmt); case 'R': + case 'r': return resource_string(buf, end, ptr, spec, fmt); case 'M': /* Colon separated: 00:01:02:03:04:05 */ case 'm': /* Contiguous: 000102030405 */ From 8d6cfdcdb50e94c92b3621422d909fa7cc41f866 Mon Sep 17 00:00:00 2001 From: Thadeu Lima de Souza Cascardo Date: Fri, 30 Oct 2009 17:46:48 -0200 Subject: [PATCH 039/109] PCI: remove pci_find_slot from PCI_LEGACY config description Commit 3b073eda has removed pci_find_slot, so there's no point in mentioning it in the config description as one of the deprecated APIs there are enabled by PCI_LEGACY and still used by some drivers. Signed-off-by: Thadeu Lima de Souza Cascardo Signed-off-by: Jesse Barnes --- drivers/pci/Kconfig | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/pci/Kconfig b/drivers/pci/Kconfig index d7d28aef8a27..b1ecefa2a23d 100644 --- a/drivers/pci/Kconfig +++ b/drivers/pci/Kconfig @@ -27,10 +27,10 @@ config PCI_LEGACY default y help Say Y here if you want to include support for the deprecated - pci_find_slot() and pci_find_device() APIs. Most drivers have - been converted over to using the proper hotplug APIs, so this - option serves to include/exclude only a few drivers that are - still using this API. + pci_find_device() API. Most drivers have been converted over + to using the proper hotplug APIs, so this option serves to + include/exclude only a few drivers that are still using this + API. config PCI_DEBUG bool "PCI Debugging" From 10c3d71d42f341775d96187eedd3e50eb34939d0 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Wed, 4 Nov 2009 10:32:42 -0700 Subject: [PATCH 040/109] PCI: make PME# messages KERN_DEBUG Messages about PME# being supported and enabled/disabled are probably useful for debug, but maybe don't need to be on the console. Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- drivers/pci/pci.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index f0da1676d2be..930eadf4670f 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -1211,7 +1211,7 @@ void pci_pme_active(struct pci_dev *dev, bool enable) pci_write_config_word(dev, dev->pm_cap + PCI_PM_CTRL, pmcsr); - dev_printk(KERN_INFO, &dev->dev, "PME# %s\n", + dev_printk(KERN_DEBUG, &dev->dev, "PME# %s\n", enable ? "enabled" : "disabled"); } @@ -1422,7 +1422,8 @@ void pci_pm_init(struct pci_dev *dev) pmc &= PCI_PM_CAP_PME_MASK; if (pmc) { - dev_info(&dev->dev, "PME# supported from%s%s%s%s%s\n", + dev_printk(KERN_DEBUG, &dev->dev, + "PME# supported from%s%s%s%s%s\n", (pmc & PCI_PM_CAP_PME_D0) ? " D0" : "", (pmc & PCI_PM_CAP_PME_D1) ? " D1" : "", (pmc & PCI_PM_CAP_PME_D2) ? " D2" : "", From 2a6bed8301f8b019717504575a3f9c6cce1fe271 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Wed, 4 Nov 2009 10:32:47 -0700 Subject: [PATCH 041/109] x86/PCI: print domain:bus in conventional format Use the dev_printk-like "%04x:%02x" format for printing PCI bus numbers. Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/pci/acpi.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/arch/x86/pci/acpi.c b/arch/x86/pci/acpi.c index 6bf8091d2fd5..68b89dc7d761 100644 --- a/arch/x86/pci/acpi.c +++ b/arch/x86/pci/acpi.c @@ -172,8 +172,9 @@ struct pci_bus * __devinit pci_acpi_scan_root(struct acpi_device *device, int do #endif if (domain && !pci_domains_supported) { - printk(KERN_WARNING "PCI: Multiple domains not supported " - "(dom %d, bus %d)\n", domain, busnum); + printk(KERN_WARNING "pci_bus %04x:%02x: " + "ignored (multiple domains not supported)\n", + domain, busnum); return NULL; } @@ -197,7 +198,8 @@ struct pci_bus * __devinit pci_acpi_scan_root(struct acpi_device *device, int do */ sd = kzalloc(sizeof(*sd), GFP_KERNEL); if (!sd) { - printk(KERN_ERR "PCI: OOM, not probing PCI bus %02x\n", busnum); + printk(KERN_WARNING "pci_bus %04x:%02x: " + "ignored (out of memory)\n", domain, busnum); return NULL; } From 0207c356ef0e2bae6ce4603080d42c130d7debc6 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Wed, 4 Nov 2009 10:32:52 -0700 Subject: [PATCH 042/109] PCI: replace pr_debug with dev_dbg Since we have a struct device, we might as well use dev_printk. Note that both pr_debug() and dev_dbg() are completely compiled out unless DEBUG or DYNAMIC_DEBUG is defined. Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- drivers/pci/probe.c | 15 +++++++-------- 1 file changed, 7 insertions(+), 8 deletions(-) diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 4c4aca53ae09..a7fdc4344cef 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -1121,7 +1121,7 @@ unsigned int __devinit pci_scan_child_bus(struct pci_bus *bus) unsigned int devfn, pass, max = bus->secondary; struct pci_dev *dev; - pr_debug("PCI: Scanning bus %04x:%02x\n", pci_domain_nr(bus), bus->number); + dev_dbg(&bus->dev, "scanning bus\n"); /* Go find them, Rover! */ for (devfn = 0; devfn < 0x100; devfn += 8) @@ -1135,8 +1135,7 @@ unsigned int __devinit pci_scan_child_bus(struct pci_bus *bus) * all PCI-to-PCI bridges on this bus. */ if (!bus->is_added) { - pr_debug("PCI: Fixups for bus %04x:%02x\n", - pci_domain_nr(bus), bus->number); + dev_dbg(&bus->dev, "fixups for bus\n"); pcibios_fixup_bus(bus); if (pci_is_root_bus(bus)) bus->is_added = 1; @@ -1156,8 +1155,7 @@ unsigned int __devinit pci_scan_child_bus(struct pci_bus *bus) * * Return how far we've got finding sub-buses. */ - pr_debug("PCI: Bus scan for %04x:%02x returning with max=%02x\n", - pci_domain_nr(bus), bus->number, max); + dev_dbg(&bus->dev, "bus scan returning with max=%02x\n", max); return max; } @@ -1165,7 +1163,7 @@ struct pci_bus * pci_create_bus(struct device *parent, int bus, struct pci_ops *ops, void *sysdata) { int error; - struct pci_bus *b; + struct pci_bus *b, *b2; struct device *dev; b = pci_alloc_bus(); @@ -1181,9 +1179,10 @@ struct pci_bus * pci_create_bus(struct device *parent, b->sysdata = sysdata; b->ops = ops; - if (pci_find_bus(pci_domain_nr(b), bus)) { + b2 = pci_find_bus(pci_domain_nr(b), bus); + if (b2) { /* If we already got to this bus through a different bridge, ignore it */ - pr_debug("PCI: Bus %04x:%02x already known\n", pci_domain_nr(b), bus); + dev_dbg(&b2->dev, "bus already known\n"); goto err_out; } From 865df576e8fc70daf297b53e61a4fbefc719d065 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Wed, 4 Nov 2009 10:32:57 -0700 Subject: [PATCH 043/109] PCI: improve discovery/configuration messages This makes PCI resource management messages more consistent and adds a few new messages to aid debugging. Whenever we assign resources to a device, update a BAR, or change a bridge aperture, it's worth noting it. Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/pci/i386.c | 9 ++++--- drivers/pci/pci.c | 2 +- drivers/pci/probe.c | 13 ++++++--- drivers/pci/setup-bus.c | 21 ++++++++++----- drivers/pci/setup-res.c | 60 ++++++++++++++++++++++------------------- 5 files changed, 63 insertions(+), 42 deletions(-) diff --git a/arch/x86/pci/i386.c b/arch/x86/pci/i386.c index d49d17de7b3f..b73c09f45210 100644 --- a/arch/x86/pci/i386.c +++ b/arch/x86/pci/i386.c @@ -129,7 +129,9 @@ static void __init pcibios_allocate_bus_resources(struct list_head *bus_list) continue; if (!r->start || pci_claim_resource(dev, idx) < 0) { - dev_info(&dev->dev, "BAR %d: can't allocate %pR\n", idx, r); + dev_info(&dev->dev, + "can't reserve window %pR\n", + r); /* * Something is wrong with the region. * Invalidate the resource to prevent @@ -165,10 +167,11 @@ static void __init pcibios_allocate_resources(int pass) disabled = !(command & PCI_COMMAND_MEMORY); if (pass == disabled) { dev_dbg(&dev->dev, - "BAR %d: claiming %pr (d=%d, p=%d)\n", + "BAR %d: reserving %pr (d=%d, p=%d)\n", idx, r, disabled, pass); if (pci_claim_resource(dev, idx) < 0) { - dev_info(&dev->dev, "BAR %d: can't claim %pR\n", idx, r); + dev_info(&dev->dev, + "can't reserve %pR\n", r); /* We'll assign a new address later */ r->end -= r->start; r->start = 0; diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index 930eadf4670f..f88de099ef43 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -2575,7 +2575,7 @@ int pci_resource_bar(struct pci_dev *dev, int resno, enum pci_bar_type *type) return reg; } - dev_err(&dev->dev, "BAR: invalid resource #%d\n", resno); + dev_err(&dev->dev, "BAR %d: invalid resource\n", resno); return 0; } diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index a7fdc4344cef..623086f9ba84 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -226,7 +226,8 @@ int __pci_read_base(struct pci_dev *dev, enum pci_bar_type type, goto fail; if ((sizeof(resource_size_t) < 8) && (sz64 > 0x100000000ULL)) { - dev_err(&dev->dev, "can't handle 64-bit BAR\n"); + dev_err(&dev->dev, "reg %x: can't handle 64-bit BAR\n", + pos); goto fail; } @@ -294,8 +295,11 @@ void __devinit pci_read_bridge_bases(struct pci_bus *child) if (pci_is_root_bus(child)) /* It's a host bus, nothing to read */ return; + dev_info(&dev->dev, "PCI bridge to [bus %02x-%02x]%s\n", + child->secondary, child->subordinate, + dev->transparent ? " (subtractive decode)": ""); + if (dev->transparent) { - dev_info(&dev->dev, "transparent bridge\n"); for(i = 3; i < PCI_BUS_NUM_RESOURCES; i++) child->resource[i] = child->parent->resource[i - 3]; } @@ -645,13 +649,14 @@ int __devinit pci_scan_bridge(struct pci_bus *bus, struct pci_dev *dev, int max, (child->number > bus->subordinate) || (child->number < bus->number) || (child->subordinate < bus->number)) { - pr_debug("PCI: Bus #%02x (-#%02x) is %s " - "hidden behind%s bridge #%02x (-#%02x)\n", + dev_info(&child->dev, "[bus %02x-%02x] %s " + "hidden behind%s bridge %s [bus %02x-%02x]\n", child->number, child->subordinate, (bus->number > child->subordinate && bus->subordinate < child->number) ? "wholly" : "partially", bus->self->transparent ? " transparent" : "", + dev_name(&bus->dev), bus->number, bus->subordinate); } bus = bus->parent; diff --git a/drivers/pci/setup-bus.c b/drivers/pci/setup-bus.c index ed6916bac675..502d1704c533 100644 --- a/drivers/pci/setup-bus.c +++ b/drivers/pci/setup-bus.c @@ -74,8 +74,8 @@ void pci_setup_cardbus(struct pci_bus *bus) struct resource *res; struct pci_bus_region region; - dev_info(&bridge->dev, "CardBus bridge, secondary bus %04x:%02x\n", - pci_domain_nr(bus), bus->number); + dev_info(&bridge->dev, "CardBus bridge to [bus %02x-%02x]\n", + bus->secondary, bus->subordinate); res = bus->resource[0]; pcibios_resource_to_bus(bridge, ®ion, res); @@ -145,8 +145,8 @@ static void pci_setup_bridge(struct pci_bus *bus) if (pci_is_enabled(bridge)) return; - dev_info(&bridge->dev, "PCI bridge, secondary bus %04x:%02x\n", - pci_domain_nr(bus), bus->number); + dev_info(&bridge->dev, "PCI bridge to [bus %02x-%02x]\n", + bus->secondary, bus->subordinate); /* Set up the top and bottom of the PCI I/O segment for this bus. */ res = bus->resource[0]; @@ -338,6 +338,10 @@ static void pbus_size_io(struct pci_bus *bus, resource_size_t min_size) #endif size = ALIGN(size + size1, 4096); if (!size) { + if (b_res->start || b_res->end) + dev_info(&bus->self->dev, "disabling bridge window " + "%pR to [bus %02x-%02x] (unused)\n", b_res, + bus->secondary, bus->subordinate); b_res->flags = 0; return; } @@ -383,8 +387,9 @@ static int pbus_size_mem(struct pci_bus *bus, unsigned long mask, align = pci_resource_alignment(dev, r); order = __ffs(align) - 20; if (order > 11) { - dev_warn(&dev->dev, "BAR %d: bad alignment %llx: " - "%pR\n", i, (unsigned long long)align, r); + dev_warn(&dev->dev, "disabling BAR %d: %pR " + "(bad alignment %#llx)\n", i, r, + (unsigned long long) align); r->flags = 0; continue; } @@ -418,6 +423,10 @@ static int pbus_size_mem(struct pci_bus *bus, unsigned long mask, } size = ALIGN(size, min_align); if (!size) { + if (b_res->start || b_res->end) + dev_info(&bus->self->dev, "disabling bridge window " + "%pR to [bus %02x-%02x] (unused)\n", b_res, + bus->secondary, bus->subordinate); b_res->flags = 0; return 1; } diff --git a/drivers/pci/setup-res.c b/drivers/pci/setup-res.c index 357ca5c54607..7d678bb15ffb 100644 --- a/drivers/pci/setup-res.c +++ b/drivers/pci/setup-res.c @@ -51,12 +51,6 @@ void pci_update_resource(struct pci_dev *dev, int resno) pcibios_resource_to_bus(dev, ®ion, res); - dev_dbg(&dev->dev, "BAR %d: got res %pR bus [%#llx-%#llx] " - "flags %#lx\n", resno, res, - (unsigned long long)region.start, - (unsigned long long)region.end, - (unsigned long)res->flags); - new = region.start | (res->flags & PCI_REGION_FLAG_MASK); if (res->flags & IORESOURCE_IO) mask = (u32)PCI_BASE_ADDRESS_IO_MASK; @@ -91,9 +85,9 @@ void pci_update_resource(struct pci_dev *dev, int resno) } } res->flags &= ~IORESOURCE_UNSET; - dev_dbg(&dev->dev, "BAR %d: moved to %pR (bus addr [%#llx-%#llx])\n", - resno, res, (unsigned long long)region.start, - (unsigned long long)region.end); + dev_info(&dev->dev, "BAR %d: set to %pR (PCI address [%#llx-%#llx]\n", + resno, res, (unsigned long long)region.start, + (unsigned long long)region.end); } int pci_claim_resource(struct pci_dev *dev, int resource) @@ -103,20 +97,17 @@ int pci_claim_resource(struct pci_dev *dev, int resource) int err; root = pci_find_parent_resource(dev, res); - - err = -EINVAL; - if (root != NULL) - err = request_resource(root, res); - - if (err) { - const char *dtype = resource < PCI_BRIDGE_RESOURCES ? "device" : "bridge"; - dev_err(&dev->dev, "BAR %d: %s of %s %pR\n", - resource, - root ? "address space collision on" : - "no parent found for", - dtype, res); + if (!root) { + dev_err(&dev->dev, "no compatible bridge window for %pR\n", + res); + return -EINVAL; } + err = request_resource(root, res); + if (err) + dev_err(&dev->dev, + "address space collision: %pR already in use\n", res); + return err; } EXPORT_SYMBOL(pci_claim_resource); @@ -124,7 +115,7 @@ EXPORT_SYMBOL(pci_claim_resource); #ifdef CONFIG_PCI_QUIRKS void pci_disable_bridge_window(struct pci_dev *dev) { - dev_dbg(&dev->dev, "Disabling bridge window.\n"); + dev_info(&dev->dev, "disabling bridge mem windows\n"); /* MMIO Base/Limit */ pci_write_config_dword(dev, PCI_MEMORY_BASE, 0x0000fff0); @@ -165,6 +156,7 @@ static int __pci_assign_resource(struct pci_bus *bus, struct pci_dev *dev, if (!ret) { res->flags &= ~IORESOURCE_STARTALIGN; + dev_info(&dev->dev, "BAR %d: assigned %pR\n", resno, res); if (resno < PCI_BRIDGE_RESOURCES) pci_update_resource(dev, resno); } @@ -178,10 +170,11 @@ int pci_assign_resource(struct pci_dev *dev, int resno) resource_size_t align; struct pci_bus *bus; int ret; + char *type; align = pci_resource_alignment(dev, res); if (!align) { - dev_info(&dev->dev, "BAR %d: can't allocate %pR " + dev_info(&dev->dev, "BAR %d: can't assign %pR " "(bogus alignment)\n", resno, res); return -EINVAL; } @@ -197,9 +190,20 @@ int pci_assign_resource(struct pci_dev *dev, int resno) break; } - if (ret) - dev_info(&dev->dev, "BAR %d: can't allocate %pR\n", - resno, res); + if (ret) { + if (res->flags & IORESOURCE_MEM) + if (res->flags & IORESOURCE_PREFETCH) + type = "mem pref"; + else + type = "mem"; + else if (res->flags & IORESOURCE_IO) + type = "io"; + else + type = "unknown"; + dev_info(&dev->dev, + "BAR %d: can't assign %s (size %#llx)\n", + resno, type, (unsigned long long) resource_size(res)); + } return ret; } @@ -272,8 +276,8 @@ int pci_enable_resources(struct pci_dev *dev, int mask) continue; if (!r->parent) { - dev_err(&dev->dev, "device not available because of " - "BAR %d %pR collisions\n", i, r); + dev_err(&dev->dev, "device not available " + "(can't reserve %pR)\n", r); return -EINVAL; } From f1db6fde09e201218f488d7205a7cd7bc448d496 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Wed, 4 Nov 2009 10:39:13 -0700 Subject: [PATCH 044/109] x86/PCI: for debuggability, show host bridge windows even when ignoring _CRS We have occasional problems with PCI resource allocation, and sometimes they could be avoided by paying attention to what ACPI tells us about the host bridges. This patch doesn't change the behavior, but it prints window information that should make debugging easier. Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/pci/acpi.c | 26 ++++++++++++++++++-------- 1 file changed, 18 insertions(+), 8 deletions(-) diff --git a/arch/x86/pci/acpi.c b/arch/x86/pci/acpi.c index 68b89dc7d761..54db5a04b5e1 100644 --- a/arch/x86/pci/acpi.c +++ b/arch/x86/pci/acpi.c @@ -92,11 +92,12 @@ setup_resource(struct acpi_resource *acpi_res, void *data) start = addr.minimum + addr.translation_offset; end = start + addr.address_length - 1; if (info->res_num >= max_root_bus_resources) { - printk(KERN_WARNING "PCI: Failed to allocate 0x%lx-0x%lx " - "from %s for %s due to _CRS returning more than " - "%d resource descriptors\n", (unsigned long) start, - (unsigned long) end, root->name, info->name, - max_root_bus_resources); + if (pci_probe & PCI_USE__CRS) + printk(KERN_WARNING "PCI: Failed to allocate " + "0x%lx-0x%lx from %s for %s due to _CRS " + "returning more than %d resource descriptors\n", + (unsigned long) start, (unsigned long) end, + root->name, info->name, max_root_bus_resources); return AE_OK; } @@ -107,6 +108,12 @@ setup_resource(struct acpi_resource *acpi_res, void *data) res->end = end; res->child = NULL; + if (!(pci_probe & PCI_USE__CRS)) { + dev_printk(KERN_DEBUG, &info->bridge->dev, + "host bridge window %pR (ignored)\n", res); + return AE_OK; + } + if (insert_resource(root, res)) { dev_err(&info->bridge->dev, "can't allocate host bridge window %pR\n", res); @@ -132,6 +139,11 @@ get_current_resources(struct acpi_device *device, int busnum, struct pci_root_info info; size_t size; + if (!(pci_probe & PCI_USE__CRS)) + dev_info(&device->dev, + "ignoring host bridge windows from ACPI; " + "boot with \"pci=use_crs\" to use them\n"); + info.bridge = device; info.bus = bus; info.res_num = 0; @@ -220,9 +232,7 @@ struct pci_bus * __devinit pci_acpi_scan_root(struct acpi_device *device, int do } else { bus = pci_create_bus(NULL, busnum, &pci_root_ops, sd); if (bus) { - if (pci_probe & PCI_USE__CRS) - get_current_resources(device, busnum, domain, - bus); + get_current_resources(device, busnum, domain, bus); bus->subordinate = pci_scan_child_bus(bus); } } From 03db42adfeeabe856dbb6894dd3aaff55838330a Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Wed, 4 Nov 2009 10:39:18 -0700 Subject: [PATCH 045/109] x86/PCI: fix bogus host bridge window start/end alignment from _CRS PCI device BARs are guaranteed to start and end on at least a four-byte (I/O) or a sixteen-byte (MMIO) boundary because they're aligned on their size and the low BAR bits are reserved. PCI-to-PCI bridge apertures have even larger alignment restrictions. However, some BIOSes (e.g., HP DL360 BIOS P31) report host bridge windows like "[io 0x0000-0x2cfe]". This is wrong because it excludes the last port at 0x2cff: it's impossible for a downstream device to claim 0x2cfe without also claiming 0x2cff. In fact, this BIOS configures a device behind the bridge to "[io 0x2c00-0x2cff]", so we know the window actually does include 0x2cff. This patch rounds the start and end of apertures to the appropriate boundary. I experimentally determined that Windows contains a similar workaround; details here: http://bugzilla.kernel.org/show_bug.cgi?id=14337 Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/pci/acpi.c | 25 +++++++++++++++++++++++++ 1 file changed, 25 insertions(+) diff --git a/arch/x86/pci/acpi.c b/arch/x86/pci/acpi.c index 54db5a04b5e1..8ddf4f4c7253 100644 --- a/arch/x86/pci/acpi.c +++ b/arch/x86/pci/acpi.c @@ -59,6 +59,30 @@ bus_has_transparent_bridge(struct pci_bus *bus) return false; } +static void +align_resource(struct acpi_device *bridge, struct resource *res) +{ + int align = (res->flags & IORESOURCE_MEM) ? 16 : 4; + + /* + * Host bridge windows are not BARs, but the decoders on the PCI side + * that claim this address space have starting alignment and length + * constraints, so fix any obvious BIOS goofs. + */ + if (res->start & (align - 1)) { + dev_printk(KERN_DEBUG, &bridge->dev, + "host bridge window %pR invalid; " + "aligning start to %d-byte boundary\n", res, align); + res->start &= ~(align - 1); + } + if ((res->end + 1) & (align - 1)) { + dev_printk(KERN_DEBUG, &bridge->dev, + "host bridge window %pR invalid; " + "aligning end to %d-byte boundary\n", res, align); + res->end = roundup(res->end, align) - 1; + } +} + static acpi_status setup_resource(struct acpi_resource *acpi_res, void *data) { @@ -107,6 +131,7 @@ setup_resource(struct acpi_resource *acpi_res, void *data) res->start = start; res->end = end; res->child = NULL; + align_resource(info->bridge, res); if (!(pci_probe & PCI_USE__CRS)) { dev_printk(KERN_DEBUG, &info->bridge->dev, From 1e5ad9679016275d422e36b12a98b0927d76f556 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Mon, 2 Nov 2009 10:45:36 -0700 Subject: [PATCH 046/109] resources: when allocate_resource() fails, leave resource untouched When "allocate_resource(root, new, size, ...)" fails, we currently clobber "new". This is inconvenient for the caller, who might care about the original contents of the resource. For example, when pci_bus_alloc_resource() fails, the "can't allocate mem resource %pR" message from pci_assign_resources() currently contains junk for the resource start/end. This patch delays the "new" update until we're about to return success. Acked-by: Linus Torvalds Acked-by: Yinghai Lu Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- kernel/resource.c | 26 ++++++++++++++------------ 1 file changed, 14 insertions(+), 12 deletions(-) diff --git a/kernel/resource.c b/kernel/resource.c index fb11a58b9594..dc15686b7a77 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -308,35 +308,37 @@ static int find_resource(struct resource *root, struct resource *new, void *alignf_data) { struct resource *this = root->child; + resource_size_t start, end; - new->start = root->start; + start = root->start; /* * Skip past an allocated resource that starts at 0, since the assignment * of this->start - 1 to new->end below would cause an underflow. */ if (this && this->start == 0) { - new->start = this->end + 1; + start = this->end + 1; this = this->sibling; } for(;;) { if (this) - new->end = this->start - 1; + end = this->start - 1; else - new->end = root->end; - if (new->start < min) - new->start = min; - if (new->end > max) - new->end = max; - new->start = ALIGN(new->start, align); + end = root->end; + if (start < min) + start = min; + if (end > max) + end = max; + start = ALIGN(start, align); if (alignf) alignf(alignf_data, new, size, align); - if (new->start < new->end && new->end - new->start >= size - 1) { - new->end = new->start + size - 1; + if (start < end && end - start >= size - 1) { + new->start = start; + new->end = start + size - 1; return 0; } if (!this) break; - new->start = this->end + 1; + start = this->end + 1; this = this->sibling; } return -EBUSY; From 0efea0006335a2425b1a12a2ad35efad626fe353 Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Thu, 5 Nov 2009 12:05:11 +0900 Subject: [PATCH 047/109] PCI: cache PCIe capability offset There are a lot of codes that searches PCI express capability offset in the PCI configuration space using pci_find_capability(). Caching it in the struct pci_dev will reduce unncecessary search. This patch adds an additional 'pcie_cap' fields into struct pci_dev, which is initialized at pci device scan time (in set_pcie_port_type()). Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/probe.c | 1 + include/linux/pci.h | 1 + 2 files changed, 2 insertions(+) diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 623086f9ba84..54b9f1501487 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -692,6 +692,7 @@ static void set_pcie_port_type(struct pci_dev *pdev) if (!pos) return; pdev->is_pcie = 1; + pdev->pcie_cap = pos; pci_read_config_word(pdev, pos + PCI_EXP_FLAGS, ®16); pdev->pcie_type = (reg16 & PCI_EXP_FLAGS_TYPE) >> 4; } diff --git a/include/linux/pci.h b/include/linux/pci.h index 86c31ac454d1..233b3a092035 100644 --- a/include/linux/pci.h +++ b/include/linux/pci.h @@ -218,6 +218,7 @@ struct pci_dev { unsigned int class; /* 3 bytes: (base,sub,prog-if) */ u8 revision; /* PCI revision, low byte of class word */ u8 hdr_type; /* PCI header type (`multi' flag masked out) */ + u8 pcie_cap; /* PCI-E capability offset */ u8 pcie_type; /* PCI-E device/port type */ u8 rom_base_reg; /* which config register controls the ROM */ u8 pin; /* which interrupt pin this device uses */ From ea7f1b6ee9dc96c5827b06ba21d7769d553efb7d Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Thu, 5 Nov 2009 11:17:11 -0600 Subject: [PATCH 048/109] x86/PCI: remove 64-bit division The roundup() caused a build error (undefined reference to `__udivdi3'). We're aligning to power-of-two boundaries, so it's simpler to just use ALIGN() anyway, which avoids the division. Signed-off-by: Bjorn Helgaas Acked-by: Randy Dunlap Signed-off-by: Jesse Barnes --- arch/x86/pci/acpi.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/arch/x86/pci/acpi.c b/arch/x86/pci/acpi.c index 8ddf4f4c7253..959e548a7039 100644 --- a/arch/x86/pci/acpi.c +++ b/arch/x86/pci/acpi.c @@ -69,17 +69,17 @@ align_resource(struct acpi_device *bridge, struct resource *res) * that claim this address space have starting alignment and length * constraints, so fix any obvious BIOS goofs. */ - if (res->start & (align - 1)) { + if (!IS_ALIGNED(res->start, align)) { dev_printk(KERN_DEBUG, &bridge->dev, "host bridge window %pR invalid; " "aligning start to %d-byte boundary\n", res, align); res->start &= ~(align - 1); } - if ((res->end + 1) & (align - 1)) { + if (!IS_ALIGNED(res->end + 1, align)) { dev_printk(KERN_DEBUG, &bridge->dev, "host bridge window %pR invalid; " "aligning end to %d-byte boundary\n", res, align); - res->end = roundup(res->end, align) - 1; + res->end = ALIGN(res->end, align) - 1; } } From e0cd5160346f5b88fe536f529066f102518f8acd Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Fri, 17 Apr 2009 12:01:55 +0200 Subject: [PATCH 049/109] PCI: derive nearby CPUs from device's instead of bus' NUMA information In case of AMD CPU northbridge functions this NUMA information might differ. Here is an example from a 4-socket system. Currently Linux shows root@hagen:/sys/devices/pci0000:00/0000:00:1a.4# cat numa_node 0 root@hagen:/sys/devices/pci0000:00/0000:00:1a.4# cat local_cpu* 0-3 00000000,0000000f which is not correct for northbridge functions as the local CPUs are those of the same socket. With this patch and a quirk for AMD CPU NB functions Linux can do better and correctly show root@hagen:/sys/devices/pci0000:00/0000:00:1a.4# cat numa_node 2 root@hagen:/sys/devices/pci0000:00/0000:00:1a.4# cat local_cpu* 8-11 00000000,00000f00 Signed-off-by: Andreas Herrmann Signed-off-by: Jesse Barnes --- drivers/pci/pci-sysfs.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/drivers/pci/pci-sysfs.c b/drivers/pci/pci-sysfs.c index 0f6382f090ee..0fa707e2a0f8 100644 --- a/drivers/pci/pci-sysfs.c +++ b/drivers/pci/pci-sysfs.c @@ -74,7 +74,11 @@ static ssize_t local_cpus_show(struct device *dev, const struct cpumask *mask; int len; +#ifdef CONFIG_NUMA + mask = cpumask_of_node(dev_to_node(dev)); +#else mask = cpumask_of_pcibus(to_pci_dev(dev)->bus); +#endif len = cpumask_scnprintf(buf, PAGE_SIZE-2, mask); buf[len++] = '\n'; buf[len] = '\0'; @@ -88,7 +92,11 @@ static ssize_t local_cpulist_show(struct device *dev, const struct cpumask *mask; int len; +#ifdef CONFIG_NUMA + mask = cpumask_of_node(dev_to_node(dev)); +#else mask = cpumask_of_pcibus(to_pci_dev(dev)->bus); +#endif len = cpulist_scnprintf(buf, PAGE_SIZE-2, mask); buf[len++] = '\n'; buf[len] = '\0'; From 9b536e0b6164d8875b4a5bb66cc102dcf0badeba Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 4 Nov 2009 05:59:55 +0200 Subject: [PATCH 050/109] PCI hotplug: fix oshp evaluation If firmware doesn't grant over native hotplug control through ACPI _OSC method, we must not evaluate OSHP. Acked-by: Andrew Patterson Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/hotplug/acpi_pcihp.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/pci/hotplug/acpi_pcihp.c b/drivers/pci/hotplug/acpi_pcihp.c index a73028ec52e5..6833d7bdbbff 100644 --- a/drivers/pci/hotplug/acpi_pcihp.c +++ b/drivers/pci/hotplug/acpi_pcihp.c @@ -362,6 +362,8 @@ int acpi_get_hp_hw_control_from_firmware(struct pci_dev *pdev, u32 flags) status = acpi_pci_osc_control_set(handle, flags); if (ACPI_SUCCESS(status)) goto got_one; + if (status == AE_SUPPORT) + goto no_control; kfree(string.pointer); string = (struct acpi_buffer){ ACPI_ALLOCATE_BUFFER, NULL }; } @@ -394,10 +396,9 @@ int acpi_get_hp_hw_control_from_firmware(struct pci_dev *pdev, u32 flags) if (ACPI_FAILURE(status)) break; } - +no_control: dbg("Cannot get control of hotplug hardware for pci %s\n", pci_name(pdev)); - kfree(string.pointer); return -ENODEV; got_one: From 5b5d94487d934be6b0aa966c9acbdf15b07ef627 Mon Sep 17 00:00:00 2001 From: Isaku Yamahata Date: Fri, 6 Nov 2009 11:11:43 +0900 Subject: [PATCH 051/109] ia64/xen: compilation fix This patch fixes the following compilation error introduced by a PCI related features. The change set of 5dd1af9f84c79bedd589db89e71ca733f3bf0ebd moves some xen related definitions from the arch header file (x86/include/asm/xen/hypervisor.h) to the common header file (include/xen/xen.h). So ia64/xen also follows it. In file included from linux-next/include/xen/grant_table.h:41, from linux-next/drivers/block/xen-blkfront.c:48: linux-next/arch/ia64/include/asm/xen/hypervisor.h:43: error: nested redefinition of 'enum xen_domain_type' linux-next/arch/ia64/include/asm/xen/hypervisor.h:43: error: redeclaration of 'enum xen_domain_type' linux-next/arch/ia64/include/asm/xen/hypervisor.h:44: error: redeclaration of enumerator 'XEN_NATIVE' linux-next/include/xen/xen.h:5: error: previous definition of 'XEN_NATIVE' was here linux-next/arch/ia64/include/asm/xen/hypervisor.h:45: error: redeclaration of enumerator 'XEN_PV_DOMAIN' linux-next/include/xen/xen.h:6: error: previous definition of 'XEN_PV_DOMAIN' was here linux-next/arch/ia64/include/asm/xen/hypervisor.h:46: error: redeclaration of enumerator 'XEN_HVM_DOMAIN' linux-next/include/xen/xen.h:7: error: previous definition of 'XEN_HVM_DOMAIN' was here Signed-off-by: Isaku Yamahata Signed-off-by: Jesse Barnes --- arch/ia64/include/asm/xen/hypervisor.h | 28 +------------------------- 1 file changed, 1 insertion(+), 27 deletions(-) diff --git a/arch/ia64/include/asm/xen/hypervisor.h b/arch/ia64/include/asm/xen/hypervisor.h index 88afb54501e4..67455c2ed2b1 100644 --- a/arch/ia64/include/asm/xen/hypervisor.h +++ b/arch/ia64/include/asm/xen/hypervisor.h @@ -37,35 +37,9 @@ #include #include /* to compile feature.c */ #include /* to comiple xen-netfront.c */ +#include #include -/* xen_domain_type is set before executing any C code by early_xen_setup */ -enum xen_domain_type { - XEN_NATIVE, /* running on bare hardware */ - XEN_PV_DOMAIN, /* running in a PV domain */ - XEN_HVM_DOMAIN, /* running in a Xen hvm domain*/ -}; - -#ifdef CONFIG_XEN -extern enum xen_domain_type xen_domain_type; -#else -#define xen_domain_type XEN_NATIVE -#endif - -#define xen_domain() (xen_domain_type != XEN_NATIVE) -#define xen_pv_domain() (xen_domain() && \ - xen_domain_type == XEN_PV_DOMAIN) -#define xen_hvm_domain() (xen_domain() && \ - xen_domain_type == XEN_HVM_DOMAIN) - -#ifdef CONFIG_XEN_DOM0 -#define xen_initial_domain() (xen_pv_domain() && \ - (xen_start_info->flags & SIF_INITDOMAIN)) -#else -#define xen_initial_domain() (0) -#endif - - #ifdef CONFIG_XEN extern struct shared_info *HYPERVISOR_shared_info; extern struct start_info *xen_start_info; From e9d1e4921d5b62a80ed02851639249e2548d24f1 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 6 Nov 2009 22:41:23 +0000 Subject: [PATCH 052/109] PCI: Replace old style lock initializer SPIN_LOCK_UNLOCKED is deprecated. Use DEFINE_SPINLOCK instead. Make the lock static while at it. Signed-off-by: Thomas Gleixner Signed-off-by: Jesse Barnes --- drivers/pci/pci.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index f88de099ef43..f36b46806513 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -2625,7 +2625,7 @@ int pci_set_vga_state(struct pci_dev *dev, bool decode, #define RESOURCE_ALIGNMENT_PARAM_SIZE COMMAND_LINE_SIZE static char resource_alignment_param[RESOURCE_ALIGNMENT_PARAM_SIZE] = {0}; -spinlock_t resource_alignment_lock = SPIN_LOCK_UNLOCKED; +static DEFINE_SPINLOCK(resource_alignment_lock); /** * pci_specified_resource_alignment - get resource alignment specified by user. From 8c8def26bfaa704db67d515da3eb92cf26067548 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Mon, 9 Nov 2009 12:04:32 -0800 Subject: [PATCH 053/109] PCI: allow matching of prefetchable resources to non-prefetchable windows I'm not entirely sure it needs to go into 32, but it's probably the right thing to do. Another way of explaining the patch is: - we currently pick the _first_ exactly matching bus resource entry, but the _last_ inexactly matching one. Normally first/last shouldn't matter, but bus resource entries aren't actually all created equal: in a transparent bus, the last resources will be the parent resources, which we should generally try to avoid unless we have no choice. So "first matching" is the thing we should always aim for. - the patch is a bit bigger than it needs to be, because I simplified the logic at the same time. It used to be a fairly incomprehensible if ((res->flags & IORESOURCE_PREFETCH) && !(r->flags & IORESOURCE_PREFETCH)) best = r; /* Approximating prefetchable by non-prefetchable */ and technically, all the patch did was to make that complex choice be even more complex (it basically added a "&& !best" to say that if we already gound a non-prefetchable window for the prefetchable resource, then we won't override an earlier one with that later one: remember "first matching"). - So instead of that complex one with three separate conditionals in one, I split it up a bit, and am taking advantage of the fact that we already handled the exact case, so if 'res->flags' has the PREFETCH bit, then we already know that 'r->flags' will _not_ have it. So the simplified code drops the redundant test, and does the new '!best' test separately. It also uses 'continue' as a way to ignore the bus resource we know doesn't work (ie a prefetchable bus resource is _not_ acceptable for anything but an exact match), so it turns into: /* We can't insert a non-prefetch resource inside a prefetchable parent .. */ if (r->flags & IORESOURCE_PREFETCH) continue; /* .. but we can put a prefetchable resource inside a non-prefetchable one */ if (!best) best = r; instead. With the comments, it's now six lines instead of two, but it's conceptually simpler, and I _could_ have written it as two lines: if ((res->flags & IORESOURCE_PREFETCH) && !best) best = r; /* Approximating prefetchable by non-prefetchable */ but I thought that was too damn subtle. Signed-off-by: Linus Torvalds Signed-off-by: Jesse Barnes --- drivers/pci/pci.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index f36b46806513..e3145f020271 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -382,8 +382,12 @@ pci_find_parent_resource(const struct pci_dev *dev, struct resource *res) continue; /* Wrong type */ if (!((res->flags ^ r->flags) & IORESOURCE_PREFETCH)) return r; /* Exact match */ - if ((res->flags & IORESOURCE_PREFETCH) && !(r->flags & IORESOURCE_PREFETCH)) - best = r; /* Approximating prefetchable by non-prefetchable */ + /* We can't insert a non-prefetch resource inside a prefetchable parent .. */ + if (r->flags & IORESOURCE_PREFETCH) + continue; + /* .. but we can put a prefetchable resource inside a non-prefetchable one */ + if (!best) + best = r; } return best; } From d7b7e60526d54da4c94afe5f137714cee7d05c41 Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 11 Nov 2009 14:29:54 +0900 Subject: [PATCH 054/109] PCI: introduce pci_pcie_cap() Introduce pci_pcie_cap() API that returns saved PCIe capability offset (currently it is saved in 'pcie_cap' field in the struct PCI dev). Using pci_pcie_cap() instead of pci_find_capability() avoids unnecessary search in PCI configuration space. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- include/linux/pci.h | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) diff --git a/include/linux/pci.h b/include/linux/pci.h index 233b3a092035..15f37f102dd3 100644 --- a/include/linux/pci.h +++ b/include/linux/pci.h @@ -1301,5 +1301,21 @@ extern void pci_hp_create_module_link(struct pci_slot *pci_slot); extern void pci_hp_remove_module_link(struct pci_slot *pci_slot); #endif +/** + * pci_pcie_cap - get the saved PCIe capability offset + * @dev: PCI device + * + * PCIe capability offset is calculated at PCI device initialization + * time and saved in the data structure. This function returns saved + * PCIe capability offset. Using this instead of pci_find_capability() + * reduces unnecessary search in the PCI configuration space. If you + * need to calculate PCIe capability offset from raw device for some + * reasons, please use pci_find_capability() instead. + */ +static inline int pci_pcie_cap(struct pci_dev *dev) +{ + return dev->pcie_cap; +} + #endif /* __KERNEL__ */ #endif /* LINUX_PCI_H */ From 06a1cbafb253c4c60d6a54a994887f5fbceabcc0 Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 11 Nov 2009 14:30:56 +0900 Subject: [PATCH 055/109] PCI: use pci_pcie_cap() in pci core Use pcie_cap() instead of pci_find_capability() to get PCIe capability offset in PCI core code. This avoids unnecessary search in PCI configuration space. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/pci.c | 12 ++++++------ drivers/pci/probe.c | 4 ++-- 2 files changed, 8 insertions(+), 8 deletions(-) diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index e3145f020271..bbc82f08d1c1 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -741,8 +741,8 @@ static int pci_save_pcie_state(struct pci_dev *dev) u16 *cap; u16 flags; - pos = pci_find_capability(dev, PCI_CAP_ID_EXP); - if (pos <= 0) + pos = pci_pcie_cap(dev); + if (!pos) return 0; save_state = pci_find_saved_cap(dev, PCI_CAP_ID_EXP); @@ -1535,7 +1535,7 @@ void pci_enable_ari(struct pci_dev *dev) if (!bridge || !bridge->is_pcie) return; - pos = pci_find_capability(bridge, PCI_CAP_ID_EXP); + pos = pci_pcie_cap(bridge); if (!pos) return; @@ -2140,7 +2140,7 @@ static int pcie_flr(struct pci_dev *dev, int probe) u32 cap; u16 status; - pos = pci_find_capability(dev, PCI_CAP_ID_EXP); + pos = pci_pcie_cap(dev); if (!pos) return -ENOTTY; @@ -2489,7 +2489,7 @@ int pcie_get_readrq(struct pci_dev *dev) int ret, cap; u16 ctl; - cap = pci_find_capability(dev, PCI_CAP_ID_EXP); + cap = pci_pcie_cap(dev); if (!cap) return -EINVAL; @@ -2519,7 +2519,7 @@ int pcie_set_readrq(struct pci_dev *dev, int rq) v = (ffs(rq) - 8) << 12; - cap = pci_find_capability(dev, PCI_CAP_ID_EXP); + cap = pci_pcie_cap(dev); if (!cap) goto out; diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 54b9f1501487..2fdffc02a308 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -703,7 +703,7 @@ static void set_pcie_hotplug_bridge(struct pci_dev *pdev) u16 reg16; u32 reg32; - pos = pci_find_capability(pdev, PCI_CAP_ID_EXP); + pos = pci_pcie_cap(pdev); if (!pos) return; pci_read_config_word(pdev, pos + PCI_EXP_FLAGS, ®16); @@ -920,7 +920,7 @@ int pci_cfg_space_size(struct pci_dev *dev) if (class == PCI_CLASS_BRIDGE_HOST) return pci_cfg_space_size_ext(dev); - pos = pci_find_capability(dev, PCI_CAP_ID_EXP); + pos = pci_pcie_cap(dev); if (!pos) { pos = pci_find_capability(dev, PCI_CAP_ID_PCIX); if (!pos) From 39a53062cb5b2ceca6035f3ed67317672f0bcf3b Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 11 Nov 2009 14:31:38 +0900 Subject: [PATCH 056/109] PCIe AER: use pci_pcie_cap() Use pcie_cap() instead of pci_find_capability() to get PCIe capability offset in PCIe AER driver. This avoids unnecessary search in PCI configuration space. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/pcie/aer/aerdrv.c | 2 +- drivers/pci/pcie/aer/aerdrv_core.c | 8 ++++---- 2 files changed, 5 insertions(+), 5 deletions(-) diff --git a/drivers/pci/pcie/aer/aerdrv.c b/drivers/pci/pcie/aer/aerdrv.c index 40c3cc5d1caf..6d30e795a10d 100644 --- a/drivers/pci/pcie/aer/aerdrv.c +++ b/drivers/pci/pcie/aer/aerdrv.c @@ -295,7 +295,7 @@ static void aer_error_resume(struct pci_dev *dev) u16 reg16; /* Clean up Root device status */ - pos = pci_find_capability(dev, PCI_CAP_ID_EXP); + pos = pci_pcie_cap(dev); pci_read_config_word(dev, pos + PCI_EXP_DEVSTA, ®16); pci_write_config_word(dev, pos + PCI_EXP_DEVSTA, reg16); diff --git a/drivers/pci/pcie/aer/aerdrv_core.c b/drivers/pci/pcie/aer/aerdrv_core.c index f4512feac12b..5391a9b412e5 100644 --- a/drivers/pci/pcie/aer/aerdrv_core.c +++ b/drivers/pci/pcie/aer/aerdrv_core.c @@ -42,7 +42,7 @@ int pci_enable_pcie_error_reporting(struct pci_dev *dev) if (!pos) return -EIO; - pos = pci_find_capability(dev, PCI_CAP_ID_EXP); + pos = pci_pcie_cap(dev); if (!pos) return -EIO; @@ -66,7 +66,7 @@ int pci_disable_pcie_error_reporting(struct pci_dev *dev) if (dev->aer_firmware_first) return -EIO; - pos = pci_find_capability(dev, PCI_CAP_ID_EXP); + pos = pci_pcie_cap(dev); if (!pos) return -EIO; @@ -224,7 +224,7 @@ static int find_device_iter(struct pci_dev *dev, void *data) */ if (atomic_read(&dev->enable_cnt) == 0) return 0; - pos = pci_find_capability(dev, PCI_CAP_ID_EXP); + pos = pci_pcie_cap(dev); if (!pos) return 0; /* Check if AER is enabled */ @@ -618,7 +618,7 @@ void aer_enable_rootport(struct aer_rpc *rpc) u16 reg16; u32 reg32; - pos = pci_find_capability(pdev, PCI_CAP_ID_EXP); + pos = pci_pcie_cap(pdev); /* Clear PCIE Capability's Device Status */ pci_read_config_word(pdev, pos+PCI_EXP_DEVSTA, ®16); pci_write_config_word(pdev, pos+PCI_EXP_DEVSTA, reg16); From dba90dfe48e2e00e79a15c95940730b6926ee176 Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 11 Nov 2009 14:32:42 +0900 Subject: [PATCH 057/109] PCIe port bus: use pci_pcie_cap() Use pci_pcie_cap() instead of pci_find_capability() to get PCIe capability offset in PCI Express Port Bus driver. This avoids unnecessary serarch in PCI configuration space. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/pcie/portdrv_core.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/drivers/pci/pcie/portdrv_core.c b/drivers/pci/pcie/portdrv_core.c index 52f84fca9f7d..ce99c7121372 100644 --- a/drivers/pci/pcie/portdrv_core.c +++ b/drivers/pci/pcie/portdrv_core.c @@ -108,7 +108,7 @@ static int pcie_port_enable_msix(struct pci_dev *dev, int *vectors, int mask) * the value in this field indicates which MSI-X Table entry is * used to generate the interrupt message." */ - pos = pci_find_capability(dev, PCI_CAP_ID_EXP); + pos = pci_pcie_cap(dev); pci_read_config_word(dev, pos + PCIE_CAPABILITIES_REG, ®16); entry = (reg16 >> 9) & PCIE_PORT_MSI_VECTOR_MASK; if (entry >= nr_entries) @@ -226,7 +226,7 @@ static int get_port_device_capability(struct pci_dev *dev) u16 reg16; u32 reg32; - pos = pci_find_capability(dev, PCI_CAP_ID_EXP); + pos = pci_pcie_cap(dev); pci_read_config_word(dev, pos + PCIE_CAPABILITIES_REG, ®16); /* Hot-Plug Capable */ if (reg16 & PORT_TO_SLOT_MASK) { @@ -305,7 +305,8 @@ int pcie_port_device_probe(struct pci_dev *dev) int pos, type; u16 reg; - if (!(pos = pci_find_capability(dev, PCI_CAP_ID_EXP))) + pos = pci_pcie_cap(dev); + if (!pos) return -ENODEV; pci_read_config_word(dev, pos + PCIE_CAPABILITIES_REG, ®); @@ -327,7 +328,7 @@ int pcie_port_device_probe(struct pci_dev *dev) int pcie_port_device_register(struct pci_dev *dev) { struct pcie_port_data *port_data; - int status, capabilities, irq_mode, i, nr_serv; + int status, capabilities, irq_mode, i, nr_serv, pos; int vectors[PCIE_PORT_DEVICE_MAXSERVICES]; u16 reg16; @@ -337,9 +338,8 @@ int pcie_port_device_register(struct pci_dev *dev) pci_set_drvdata(dev, port_data); /* Get port type */ - pci_read_config_word(dev, - pci_find_capability(dev, PCI_CAP_ID_EXP) + - PCIE_CAPABILITIES_REG, ®16); + pos = pci_pcie_cap(dev); + pci_read_config_word(dev, pos + PCIE_CAPABILITIES_REG, ®16); port_data->port_type = (reg16 >> 4) & PORT_TYPE_MASK; capabilities = get_port_device_capability(dev); From db9538a7495e33f3571c0e791c7678bc0c6ef50f Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 11 Nov 2009 14:33:30 +0900 Subject: [PATCH 058/109] PCIe ASPM: use pci_pcie_cap() Use pci_pcie_cap() instead of pci_find_capability() to get PCIe capability offset in PCIe ASPM driver. This avoids unnecessary search in PCI configuration space. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/pcie/aspm.c | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) diff --git a/drivers/pci/pcie/aspm.c b/drivers/pci/pcie/aspm.c index 745402e8e498..17baffcef5fc 100644 --- a/drivers/pci/pcie/aspm.c +++ b/drivers/pci/pcie/aspm.c @@ -122,7 +122,7 @@ static void pcie_set_clkpm_nocheck(struct pcie_link_state *link, int enable) struct pci_bus *linkbus = link->pdev->subordinate; list_for_each_entry(child, &linkbus->devices, bus_list) { - pos = pci_find_capability(child, PCI_CAP_ID_EXP); + pos = pci_pcie_cap(child); if (!pos) return; pci_read_config_word(child, pos + PCI_EXP_LNKCTL, ®16); @@ -156,7 +156,7 @@ static void pcie_clkpm_cap_init(struct pcie_link_state *link, int blacklist) /* All functions should have the same cap and state, take the worst */ list_for_each_entry(child, &linkbus->devices, bus_list) { - pos = pci_find_capability(child, PCI_CAP_ID_EXP); + pos = pci_pcie_cap(child); if (!pos) return; pci_read_config_dword(child, pos + PCI_EXP_LNKCAP, ®32); @@ -194,20 +194,20 @@ static void pcie_aspm_configure_common_clock(struct pcie_link_state *link) BUG_ON(!child->is_pcie); /* Check downstream component if bit Slot Clock Configuration is 1 */ - cpos = pci_find_capability(child, PCI_CAP_ID_EXP); + cpos = pci_pcie_cap(child); pci_read_config_word(child, cpos + PCI_EXP_LNKSTA, ®16); if (!(reg16 & PCI_EXP_LNKSTA_SLC)) same_clock = 0; /* Check upstream component if bit Slot Clock Configuration is 1 */ - ppos = pci_find_capability(parent, PCI_CAP_ID_EXP); + ppos = pci_pcie_cap(parent); pci_read_config_word(parent, ppos + PCI_EXP_LNKSTA, ®16); if (!(reg16 & PCI_EXP_LNKSTA_SLC)) same_clock = 0; /* Configure downstream component, all functions */ list_for_each_entry(child, &linkbus->devices, bus_list) { - cpos = pci_find_capability(child, PCI_CAP_ID_EXP); + cpos = pci_pcie_cap(child); pci_read_config_word(child, cpos + PCI_EXP_LNKCTL, ®16); child_reg[PCI_FUNC(child->devfn)] = reg16; if (same_clock) @@ -247,7 +247,7 @@ static void pcie_aspm_configure_common_clock(struct pcie_link_state *link) dev_printk(KERN_ERR, &parent->dev, "ASPM: Could not configure common clock\n"); list_for_each_entry(child, &linkbus->devices, bus_list) { - cpos = pci_find_capability(child, PCI_CAP_ID_EXP); + cpos = pci_pcie_cap(child); pci_write_config_word(child, cpos + PCI_EXP_LNKCTL, child_reg[PCI_FUNC(child->devfn)]); } @@ -300,7 +300,7 @@ static void pcie_get_aspm_reg(struct pci_dev *pdev, u16 reg16; u32 reg32; - pos = pci_find_capability(pdev, PCI_CAP_ID_EXP); + pos = pci_pcie_cap(pdev); pci_read_config_dword(pdev, pos + PCI_EXP_LNKCAP, ®32); info->support = (reg32 & PCI_EXP_LNKCAP_ASPMS) >> 10; info->latency_encoding_l0s = (reg32 & PCI_EXP_LNKCAP_L0SEL) >> 12; @@ -420,7 +420,7 @@ static void pcie_aspm_cap_init(struct pcie_link_state *link, int blacklist) child->pcie_type != PCI_EXP_TYPE_LEG_END) continue; - pos = pci_find_capability(child, PCI_CAP_ID_EXP); + pos = pci_pcie_cap(child); pci_read_config_dword(child, pos + PCI_EXP_DEVCAP, ®32); /* Calculate endpoint L0s acceptable latency */ encoding = (reg32 & PCI_EXP_DEVCAP_L0S) >> 6; @@ -436,7 +436,7 @@ static void pcie_aspm_cap_init(struct pcie_link_state *link, int blacklist) static void pcie_config_aspm_dev(struct pci_dev *pdev, u32 val) { u16 reg16; - int pos = pci_find_capability(pdev, PCI_CAP_ID_EXP); + int pos = pci_pcie_cap(pdev); pci_read_config_word(pdev, pos + PCI_EXP_LNKCTL, ®16); reg16 &= ~0x3; @@ -503,7 +503,7 @@ static int pcie_aspm_sanity_check(struct pci_dev *pdev) * very strange. Disable ASPM for the whole slot */ list_for_each_entry(child, &pdev->subordinate->devices, bus_list) { - pos = pci_find_capability(child, PCI_CAP_ID_EXP); + pos = pci_pcie_cap(child); if (!pos) return -EINVAL; /* From d3ccc4091f0d63a3e0976c739c27037a5666060d Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 11 Nov 2009 14:34:15 +0900 Subject: [PATCH 059/109] PCI hotplug: use pci_pcie_cap() Use pci_pcie_cap() instead of pci_find_capability() to get PCIe capability offset in PCI hotplug core. This avoids unnecessary search in PCI configuration space. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/hotplug/pcihp_slot.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/pci/hotplug/pcihp_slot.c b/drivers/pci/hotplug/pcihp_slot.c index cc8ec3aa41a7..7f61afed672a 100644 --- a/drivers/pci/hotplug/pcihp_slot.c +++ b/drivers/pci/hotplug/pcihp_slot.c @@ -102,7 +102,7 @@ static void program_hpp_type2(struct pci_dev *dev, struct hpp_type2 *hpp) return; /* Find PCI Express capability */ - pos = pci_find_capability(dev, PCI_CAP_ID_EXP); + pos = pci_pcie_cap(dev); if (!pos) return; From 1518c17ab736303098843bd306a0fc4f8f5faa42 Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 11 Nov 2009 14:34:52 +0900 Subject: [PATCH 060/109] pciehp: use pci_pcie_cap() Use pci_pcie_cap() instead of pci_find_capability() to get PCIe capability offset in pciehp driver. This avoids unnecessary search in PCI configuration space. This patch also removes 'cap_base' field in struct controller, that was used to hold PCIe capability offset by pciehp itself. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/hotplug/pciehp.h | 1 - drivers/pci/hotplug/pciehp_acpi.c | 3 +- drivers/pci/hotplug/pciehp_hpc.c | 46 +++++++++++++++---------------- 3 files changed, 25 insertions(+), 25 deletions(-) diff --git a/drivers/pci/hotplug/pciehp.h b/drivers/pci/hotplug/pciehp.h index 3070f77eb56a..4ed76b47b6dc 100644 --- a/drivers/pci/hotplug/pciehp.h +++ b/drivers/pci/hotplug/pciehp.h @@ -91,7 +91,6 @@ struct controller { struct slot *slot; wait_queue_head_t queue; /* sleep & wake process */ u32 slot_cap; - u8 cap_base; struct timer_list poll_timer; unsigned int cmd_busy:1; unsigned int no_cmd_complete:1; diff --git a/drivers/pci/hotplug/pciehp_acpi.c b/drivers/pci/hotplug/pciehp_acpi.c index 37c8d3d0323e..b09b083011d6 100644 --- a/drivers/pci/hotplug/pciehp_acpi.c +++ b/drivers/pci/hotplug/pciehp_acpi.c @@ -87,7 +87,8 @@ static int __init dummy_probe(struct pcie_device *dev) /* Note: pciehp_detect_mode != PCIEHP_DETECT_ACPI here */ if (pciehp_get_hp_hw_control_from_firmware(pdev)) return -ENODEV; - if (!(pos = pci_find_capability(pdev, PCI_CAP_ID_EXP))) + pos = pci_pcie_cap(pdev); + if (!pos) return -ENODEV; pci_read_config_dword(pdev, pos + PCI_EXP_SLTCAP, &slot_cap); slot = kzalloc(sizeof(*slot), GFP_KERNEL); diff --git a/drivers/pci/hotplug/pciehp_hpc.c b/drivers/pci/hotplug/pciehp_hpc.c index 7f35aff22362..90dac515b60f 100644 --- a/drivers/pci/hotplug/pciehp_hpc.c +++ b/drivers/pci/hotplug/pciehp_hpc.c @@ -45,25 +45,25 @@ static atomic_t pciehp_num_controllers = ATOMIC_INIT(0); static inline int pciehp_readw(struct controller *ctrl, int reg, u16 *value) { struct pci_dev *dev = ctrl->pcie->port; - return pci_read_config_word(dev, ctrl->cap_base + reg, value); + return pci_read_config_word(dev, pci_pcie_cap(dev) + reg, value); } static inline int pciehp_readl(struct controller *ctrl, int reg, u32 *value) { struct pci_dev *dev = ctrl->pcie->port; - return pci_read_config_dword(dev, ctrl->cap_base + reg, value); + return pci_read_config_dword(dev, pci_pcie_cap(dev) + reg, value); } static inline int pciehp_writew(struct controller *ctrl, int reg, u16 value) { struct pci_dev *dev = ctrl->pcie->port; - return pci_write_config_word(dev, ctrl->cap_base + reg, value); + return pci_write_config_word(dev, pci_pcie_cap(dev) + reg, value); } static inline int pciehp_writel(struct controller *ctrl, int reg, u32 value) { struct pci_dev *dev = ctrl->pcie->port; - return pci_write_config_dword(dev, ctrl->cap_base + reg, value); + return pci_write_config_dword(dev, pci_pcie_cap(dev) + reg, value); } /* Power Control Command */ @@ -318,8 +318,8 @@ int pciehp_get_attention_status(struct slot *slot, u8 *status) return retval; } - ctrl_dbg(ctrl, "%s: SLOTCTRL %x, value read %x\n", - __func__, ctrl->cap_base + PCI_EXP_SLTCTL, slot_ctrl); + ctrl_dbg(ctrl, "%s: SLOTCTRL %x, value read %x\n", __func__, + pci_pcie_cap(ctrl->pcie->port) + PCI_EXP_SLTCTL, slot_ctrl); atten_led_state = (slot_ctrl & PCI_EXP_SLTCTL_AIC) >> 6; @@ -356,8 +356,8 @@ int pciehp_get_power_status(struct slot *slot, u8 *status) ctrl_err(ctrl, "%s: Cannot read SLOTCTRL register\n", __func__); return retval; } - ctrl_dbg(ctrl, "%s: SLOTCTRL %x value read %x\n", - __func__, ctrl->cap_base + PCI_EXP_SLTCTL, slot_ctrl); + ctrl_dbg(ctrl, "%s: SLOTCTRL %x value read %x\n", __func__, + pci_pcie_cap(ctrl->pcie->port) + PCI_EXP_SLTCTL, slot_ctrl); pwr_state = (slot_ctrl & PCI_EXP_SLTCTL_PCC) >> 10; @@ -442,8 +442,8 @@ int pciehp_set_attention_status(struct slot *slot, u8 value) default: return -EINVAL; } - ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n", - __func__, ctrl->cap_base + PCI_EXP_SLTCTL, slot_cmd); + ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n", __func__, + pci_pcie_cap(ctrl->pcie->port) + PCI_EXP_SLTCTL, slot_cmd); return pcie_write_cmd(ctrl, slot_cmd, cmd_mask); } @@ -456,8 +456,8 @@ void pciehp_green_led_on(struct slot *slot) slot_cmd = 0x0100; cmd_mask = PCI_EXP_SLTCTL_PIC; pcie_write_cmd(ctrl, slot_cmd, cmd_mask); - ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n", - __func__, ctrl->cap_base + PCI_EXP_SLTCTL, slot_cmd); + ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n", __func__, + pci_pcie_cap(ctrl->pcie->port) + PCI_EXP_SLTCTL, slot_cmd); } void pciehp_green_led_off(struct slot *slot) @@ -469,8 +469,8 @@ void pciehp_green_led_off(struct slot *slot) slot_cmd = 0x0300; cmd_mask = PCI_EXP_SLTCTL_PIC; pcie_write_cmd(ctrl, slot_cmd, cmd_mask); - ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n", - __func__, ctrl->cap_base + PCI_EXP_SLTCTL, slot_cmd); + ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n", __func__, + pci_pcie_cap(ctrl->pcie->port) + PCI_EXP_SLTCTL, slot_cmd); } void pciehp_green_led_blink(struct slot *slot) @@ -482,8 +482,8 @@ void pciehp_green_led_blink(struct slot *slot) slot_cmd = 0x0200; cmd_mask = PCI_EXP_SLTCTL_PIC; pcie_write_cmd(ctrl, slot_cmd, cmd_mask); - ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n", - __func__, ctrl->cap_base + PCI_EXP_SLTCTL, slot_cmd); + ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n", __func__, + pci_pcie_cap(ctrl->pcie->port) + PCI_EXP_SLTCTL, slot_cmd); } int pciehp_power_on_slot(struct slot * slot) @@ -525,8 +525,8 @@ int pciehp_power_on_slot(struct slot * slot) ctrl_err(ctrl, "Write %x command failed!\n", slot_cmd); return retval; } - ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n", - __func__, ctrl->cap_base + PCI_EXP_SLTCTL, slot_cmd); + ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n", __func__, + pci_pcie_cap(ctrl->pcie->port) + PCI_EXP_SLTCTL, slot_cmd); ctrl->power_fault_detected = 0; return retval; @@ -552,8 +552,8 @@ int pciehp_power_off_slot(struct slot * slot) ctrl_err(ctrl, "Write command failed!\n"); return retval; } - ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n", - __func__, ctrl->cap_base + PCI_EXP_SLTCTL, slot_cmd); + ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n", __func__, + pci_pcie_cap(ctrl->pcie->port) + PCI_EXP_SLTCTL, slot_cmd); return 0; } @@ -885,7 +885,8 @@ static inline void dbg_ctrl(struct controller *ctrl) pdev->subsystem_device); ctrl_info(ctrl, " Subsystem Vendor ID : 0x%04x\n", pdev->subsystem_vendor); - ctrl_info(ctrl, " PCIe Cap offset : 0x%02x\n", ctrl->cap_base); + ctrl_info(ctrl, " PCIe Cap offset : 0x%02x\n", + pci_pcie_cap(pdev)); for (i = 0; i < DEVICE_COUNT_RESOURCE; i++) { if (!pci_resource_len(pdev, i)) continue; @@ -929,8 +930,7 @@ struct controller *pcie_init(struct pcie_device *dev) goto abort; } ctrl->pcie = dev; - ctrl->cap_base = pci_find_capability(pdev, PCI_CAP_ID_EXP); - if (!ctrl->cap_base) { + if (!pci_pcie_cap(pdev)) { ctrl_err(ctrl, "Cannot find PCI Express capability\n"); goto abort_ctrl; } From 7eb776c42e75d17bd8107a1359068d8c742639d1 Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 11 Nov 2009 14:35:22 +0900 Subject: [PATCH 061/109] PCI: introduce pci_is_pcie() Introduce pci_is_pcie() which returns true if the specified PCI device is PCI Express capable, false otherwise. The purpose of pci_is_pcie() is removing 'is_pcie' flag in the struct pci_dev, which is not needed because we can check it using 'pcie_cap' field. To remove 'is_pcie', we need to update user of 'is_pcie' to use pci_is_pcie() instead first. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- include/linux/pci.h | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/include/linux/pci.h b/include/linux/pci.h index 15f37f102dd3..2891c3d3e51a 100644 --- a/include/linux/pci.h +++ b/include/linux/pci.h @@ -1317,5 +1317,16 @@ static inline int pci_pcie_cap(struct pci_dev *dev) return dev->pcie_cap; } +/** + * pci_is_pcie - check if the PCI device is PCI Express capable + * @dev: PCI device + * + * Retrun true if the PCI device is PCI Express capable, false otherwise. + */ +static inline bool pci_is_pcie(struct pci_dev *dev) +{ + return !!pci_pcie_cap(dev); +} + #endif /* __KERNEL__ */ #endif /* LINUX_PCI_H */ From 5f4d91a1228ac85c75b099efd36fff1a3407335c Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 11 Nov 2009 14:36:17 +0900 Subject: [PATCH 062/109] PCI: use pci_is_pcie() in pci core Change for PCI core to use pci_is_pcie() instead of checking pci_dev->is_pcie. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/dmar.c | 2 +- drivers/pci/intel-iommu.c | 10 +++++----- drivers/pci/intr_remapping.c | 4 ++-- drivers/pci/iov.c | 2 +- drivers/pci/pci-acpi.c | 4 ++-- drivers/pci/pci.c | 6 +++--- drivers/pci/search.c | 4 ++-- 7 files changed, 16 insertions(+), 16 deletions(-) diff --git a/drivers/pci/dmar.c b/drivers/pci/dmar.c index 22b02c6df854..e01ca4d6b3e6 100644 --- a/drivers/pci/dmar.c +++ b/drivers/pci/dmar.c @@ -329,7 +329,7 @@ found: for (bus = dev->bus; bus; bus = bus->parent) { struct pci_dev *bridge = bus->self; - if (!bridge || !bridge->is_pcie || + if (!bridge || !pci_is_pcie(bridge) || bridge->pcie_type == PCI_EXP_TYPE_PCI_BRIDGE) return 0; diff --git a/drivers/pci/intel-iommu.c b/drivers/pci/intel-iommu.c index b1e97e682500..b8802ae4bcdb 100644 --- a/drivers/pci/intel-iommu.c +++ b/drivers/pci/intel-iommu.c @@ -1611,7 +1611,7 @@ domain_context_mapping(struct dmar_domain *domain, struct pci_dev *pdev, return ret; parent = parent->bus->self; } - if (tmp->is_pcie) /* this is a PCIE-to-PCI bridge */ + if (pci_is_pcie(tmp)) /* this is a PCIE-to-PCI bridge */ return domain_context_mapping_one(domain, pci_domain_nr(tmp->subordinate), tmp->subordinate->number, 0, @@ -1651,7 +1651,7 @@ static int domain_context_mapped(struct pci_dev *pdev) return ret; parent = parent->bus->self; } - if (tmp->is_pcie) + if (pci_is_pcie(tmp)) return device_context_mapped(iommu, tmp->subordinate->number, 0); else @@ -1821,7 +1821,7 @@ static struct dmar_domain *get_domain_for_dev(struct pci_dev *pdev, int gaw) dev_tmp = pci_find_upstream_pcie_bridge(pdev); if (dev_tmp) { - if (dev_tmp->is_pcie) { + if (pci_is_pcie(dev_tmp)) { bus = dev_tmp->subordinate->number; devfn = 0; } else { @@ -2182,7 +2182,7 @@ static int iommu_should_identity_map(struct pci_dev *pdev, int startup) * the 1:1 domain, just in _case_ one of their siblings turns out * not to be able to map all of memory. */ - if (!pdev->is_pcie) { + if (!pci_is_pcie(pdev)) { if (!pci_is_root_bus(pdev->bus)) return 0; if (pdev->class >> 8 == PCI_CLASS_BRIDGE_PCI) @@ -3280,7 +3280,7 @@ static void iommu_detach_dependent_devices(struct intel_iommu *iommu, parent->devfn); parent = parent->bus->self; } - if (tmp->is_pcie) /* this is a PCIE-to-PCI bridge */ + if (pci_is_pcie(tmp)) /* this is a PCIE-to-PCI bridge */ iommu_detach_dev(iommu, tmp->subordinate->number, 0); else /* this is a legacy PCI bridge */ diff --git a/drivers/pci/intr_remapping.c b/drivers/pci/intr_remapping.c index 0ed78a764ded..3e6d8d79a09d 100644 --- a/drivers/pci/intr_remapping.c +++ b/drivers/pci/intr_remapping.c @@ -478,7 +478,7 @@ int set_msi_sid(struct irte *irte, struct pci_dev *dev) return -1; /* PCIe device or Root Complex integrated PCI device */ - if (dev->is_pcie || !dev->bus->parent) { + if (pci_is_pcie(dev) || !dev->bus->parent) { set_irte_sid(irte, SVT_VERIFY_SID_SQ, SQ_ALL_16, (dev->bus->number << 8) | dev->devfn); return 0; @@ -486,7 +486,7 @@ int set_msi_sid(struct irte *irte, struct pci_dev *dev) bridge = pci_find_upstream_pcie_bridge(dev); if (bridge) { - if (bridge->is_pcie) /* this is a PCIE-to-PCI/PCIX bridge */ + if (pci_is_pcie(bridge))/* this is a PCIE-to-PCI/PCIX bridge */ set_irte_sid(irte, SVT_VERIFY_BUS, SQ_ALL_16, (bridge->bus->number << 8) | dev->bus->number); else /* this is a legacy PCI bridge */ diff --git a/drivers/pci/iov.c b/drivers/pci/iov.c index e03fe98f0619..b2a448e19fe6 100644 --- a/drivers/pci/iov.c +++ b/drivers/pci/iov.c @@ -555,7 +555,7 @@ int pci_iov_init(struct pci_dev *dev) { int pos; - if (!dev->is_pcie) + if (!pci_is_pcie(dev)) return -ENODEV; pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_SRIOV); diff --git a/drivers/pci/pci-acpi.c b/drivers/pci/pci-acpi.c index 33317df47699..cc617ddd33d0 100644 --- a/drivers/pci/pci-acpi.c +++ b/drivers/pci/pci-acpi.c @@ -116,7 +116,7 @@ static void acpi_pci_propagate_wakeup_enable(struct pci_bus *bus, bool enable) int ret; ret = acpi_pm_device_sleep_wake(&bridge->dev, enable); - if (!ret || bridge->is_pcie) + if (!ret || pci_is_pcie(bridge)) return; bus = bus->parent; } @@ -131,7 +131,7 @@ static int acpi_pci_sleep_wake(struct pci_dev *dev, bool enable) if (acpi_pci_can_wakeup(dev)) return acpi_pm_device_sleep_wake(&dev->dev, enable); - if (!dev->is_pcie) + if (!pci_is_pcie(dev)) acpi_pci_propagate_wakeup_enable(dev->bus, enable); return 0; diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index bbc82f08d1c1..453d6ba6811e 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -1524,7 +1524,7 @@ void pci_enable_ari(struct pci_dev *dev) u16 ctrl; struct pci_dev *bridge; - if (!dev->is_pcie || dev->devfn) + if (!pci_is_pcie(dev) || dev->devfn) return; pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ARI); @@ -1532,7 +1532,7 @@ void pci_enable_ari(struct pci_dev *dev) return; bridge = dev->bus->self; - if (!bridge || !bridge->is_pcie) + if (!bridge || !pci_is_pcie(bridge)) return; pos = pci_pcie_cap(bridge); @@ -1560,7 +1560,7 @@ void pci_enable_acs(struct pci_dev *dev) u16 cap; u16 ctrl; - if (!dev->is_pcie) + if (!pci_is_pcie(dev)) return; pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ACS); diff --git a/drivers/pci/search.c b/drivers/pci/search.c index 75826482c71a..6dae87143258 100644 --- a/drivers/pci/search.c +++ b/drivers/pci/search.c @@ -26,14 +26,14 @@ pci_find_upstream_pcie_bridge(struct pci_dev *pdev) { struct pci_dev *tmp = NULL; - if (pdev->is_pcie) + if (pci_is_pcie(pdev)) return NULL; while (1) { if (pci_is_root_bus(pdev->bus)) break; pdev = pdev->bus->self; /* a p2p bridge */ - if (!pdev->is_pcie) { + if (!pci_is_pcie(pdev)) { tmp = pdev; continue; } From 8b06477dc4fcdfc21442ad334d3f3e335225ea0c Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 11 Nov 2009 14:36:52 +0900 Subject: [PATCH 063/109] PCIe ASPM: use pci_is_pcie() Change for PCIe ASPM driver to use pci_is_pcie() instead of checking pci_dev->is_pcie. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/pcie/aspm.c | 21 ++++++++++++--------- 1 file changed, 12 insertions(+), 9 deletions(-) diff --git a/drivers/pci/pcie/aspm.c b/drivers/pci/pcie/aspm.c index 17baffcef5fc..05b8e25b551b 100644 --- a/drivers/pci/pcie/aspm.c +++ b/drivers/pci/pcie/aspm.c @@ -191,7 +191,7 @@ static void pcie_aspm_configure_common_clock(struct pcie_link_state *link) * Configuration, so just check one function */ child = list_entry(linkbus->devices.next, struct pci_dev, bus_list); - BUG_ON(!child->is_pcie); + BUG_ON(!pci_is_pcie(child)); /* Check downstream component if bit Slot Clock Configuration is 1 */ cpos = pci_pcie_cap(child); @@ -563,7 +563,7 @@ void pcie_aspm_init_link_state(struct pci_dev *pdev) struct pcie_link_state *link; int blacklist = !!pcie_aspm_sanity_check(pdev); - if (aspm_disabled || !pdev->is_pcie || pdev->link_state) + if (aspm_disabled || !pci_is_pcie(pdev) || pdev->link_state) return; if (pdev->pcie_type != PCI_EXP_TYPE_ROOT_PORT && pdev->pcie_type != PCI_EXP_TYPE_DOWNSTREAM) @@ -629,7 +629,8 @@ void pcie_aspm_exit_link_state(struct pci_dev *pdev) struct pci_dev *parent = pdev->bus->self; struct pcie_link_state *link, *root, *parent_link; - if (aspm_disabled || !pdev->is_pcie || !parent || !parent->link_state) + if (aspm_disabled || !pci_is_pcie(pdev) || + !parent || !parent->link_state) return; if ((parent->pcie_type != PCI_EXP_TYPE_ROOT_PORT) && (parent->pcie_type != PCI_EXP_TYPE_DOWNSTREAM)) @@ -668,7 +669,7 @@ void pcie_aspm_pm_state_change(struct pci_dev *pdev) { struct pcie_link_state *link = pdev->link_state; - if (aspm_disabled || !pdev->is_pcie || !link) + if (aspm_disabled || !pci_is_pcie(pdev) || !link) return; if ((pdev->pcie_type != PCI_EXP_TYPE_ROOT_PORT) && (pdev->pcie_type != PCI_EXP_TYPE_DOWNSTREAM)) @@ -694,7 +695,7 @@ void pci_disable_link_state(struct pci_dev *pdev, int state) struct pci_dev *parent = pdev->bus->self; struct pcie_link_state *link; - if (aspm_disabled || !pdev->is_pcie) + if (aspm_disabled || !pci_is_pcie(pdev)) return; if (pdev->pcie_type == PCI_EXP_TYPE_ROOT_PORT || pdev->pcie_type == PCI_EXP_TYPE_DOWNSTREAM) @@ -839,8 +840,9 @@ void pcie_aspm_create_sysfs_dev_files(struct pci_dev *pdev) { struct pcie_link_state *link_state = pdev->link_state; - if (!pdev->is_pcie || (pdev->pcie_type != PCI_EXP_TYPE_ROOT_PORT && - pdev->pcie_type != PCI_EXP_TYPE_DOWNSTREAM) || !link_state) + if (!pci_is_pcie(pdev) || + (pdev->pcie_type != PCI_EXP_TYPE_ROOT_PORT && + pdev->pcie_type != PCI_EXP_TYPE_DOWNSTREAM) || !link_state) return; if (link_state->aspm_support) @@ -855,8 +857,9 @@ void pcie_aspm_remove_sysfs_dev_files(struct pci_dev *pdev) { struct pcie_link_state *link_state = pdev->link_state; - if (!pdev->is_pcie || (pdev->pcie_type != PCI_EXP_TYPE_ROOT_PORT && - pdev->pcie_type != PCI_EXP_TYPE_DOWNSTREAM) || !link_state) + if (!pci_is_pcie(pdev) || + (pdev->pcie_type != PCI_EXP_TYPE_ROOT_PORT && + pdev->pcie_type != PCI_EXP_TYPE_DOWNSTREAM) || !link_state) return; if (link_state->aspm_support) From b44d7db36480a3b27e78141fc9d6597aa577744b Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 11 Nov 2009 14:37:24 +0900 Subject: [PATCH 064/109] PCIe AER: use pci_is_pcie() Changes for PCIe AER driver to use pci_is_pcie() instead of checking pci_dev->is_pcie. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/pcie/aer/aer_inject.c | 2 +- drivers/pci/pcie/aer/ecrc.c | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/pci/pcie/aer/aer_inject.c b/drivers/pci/pcie/aer/aer_inject.c index 2246bf7aee7e..7fcd5331b14c 100644 --- a/drivers/pci/pcie/aer/aer_inject.c +++ b/drivers/pci/pcie/aer/aer_inject.c @@ -281,7 +281,7 @@ out: static struct pci_dev *pcie_find_root_port(struct pci_dev *dev) { while (1) { - if (!dev->is_pcie) + if (!pci_is_pcie(dev)) break; if (dev->pcie_type == PCI_EXP_TYPE_ROOT_PORT) return dev; diff --git a/drivers/pci/pcie/aer/ecrc.c b/drivers/pci/pcie/aer/ecrc.c index a928d8ab6bda..a2747a663bc9 100644 --- a/drivers/pci/pcie/aer/ecrc.c +++ b/drivers/pci/pcie/aer/ecrc.c @@ -51,7 +51,7 @@ static int enable_ecrc_checking(struct pci_dev *dev) int pos; u32 reg32; - if (!dev->is_pcie) + if (!pci_is_pcie(dev)) return -ENODEV; pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ERR); @@ -79,7 +79,7 @@ static int disable_ecrc_checking(struct pci_dev *dev) int pos; u32 reg32; - if (!dev->is_pcie) + if (!pci_is_pcie(dev)) return -ENODEV; pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ERR); From 13598378f29c125d78047b23330eb2294b03d414 Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 11 Nov 2009 14:38:16 +0900 Subject: [PATCH 065/109] PCI hotplug: use pci_is_pcie() Change for PCI hotplug to use pci_is_pcie() instead of checking pci_dev->is_pcie. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/hotplug/pcihp_slot.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/pci/hotplug/pcihp_slot.c b/drivers/pci/hotplug/pcihp_slot.c index 7f61afed672a..80b461c98557 100644 --- a/drivers/pci/hotplug/pcihp_slot.c +++ b/drivers/pci/hotplug/pcihp_slot.c @@ -43,7 +43,7 @@ static void program_hpp_type0(struct pci_dev *dev, struct hpp_type0 *hpp) * Perhaps we *should* use default settings for PCIe, but * pciehp didn't, so we won't either. */ - if (dev->is_pcie) + if (pci_is_pcie(dev)) return; dev_info(&dev->dev, "using default PCI settings\n"); hpp = &pci_default_type0; From 5651c48cfafef1b9a7ebdc00ebeb32f2af887a89 Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Fri, 13 Nov 2009 15:14:10 +0900 Subject: [PATCH 066/109] PCI pciehp: fix power fault interrupt storm problem Enabling power fault detected event notification in current pciehp might cause power fault interrupt storm on some machines. On those machines. On those machines, power fault detected bit in the slot status register was set again immediately when it is cleared in the interrupt service routine, and next power fault detected interrupt was notified again. Therefore, disable power fault detected event notification for now. This patch also removes unnecessary handling for power fault cleared event because this event is not supported by PCIe spec. Tested-by: Jens Axboe Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/hotplug/pciehp_ctrl.c | 27 ++++++--------------------- drivers/pci/hotplug/pciehp_hpc.c | 26 +++++++++++--------------- 2 files changed, 17 insertions(+), 36 deletions(-) diff --git a/drivers/pci/hotplug/pciehp_ctrl.c b/drivers/pci/hotplug/pciehp_ctrl.c index 15ce2a3cc0f1..d6ac1b261dd9 100644 --- a/drivers/pci/hotplug/pciehp_ctrl.c +++ b/drivers/pci/hotplug/pciehp_ctrl.c @@ -142,23 +142,9 @@ u8 pciehp_handle_power_fault(struct slot *p_slot) /* power fault */ ctrl_dbg(ctrl, "Power fault interrupt received\n"); - - if (!pciehp_query_power_fault(p_slot)) { - /* - * power fault Cleared - */ - ctrl_info(ctrl, "Power fault cleared on Slot(%s)\n", - slot_name(p_slot)); - event_type = INT_POWER_FAULT_CLEAR; - } else { - /* - * power fault - */ - ctrl_info(ctrl, "Power fault on Slot(%s)\n", slot_name(p_slot)); - event_type = INT_POWER_FAULT; - ctrl_info(ctrl, "Power fault bit %x set\n", 0); - } - + ctrl_err(ctrl, "Power fault on slot %s\n", slot_name(p_slot)); + event_type = INT_POWER_FAULT; + ctrl_info(ctrl, "Power fault bit %x set\n", 0); queue_interrupt_event(p_slot, event_type); return 1; @@ -224,13 +210,12 @@ static int board_added(struct slot *p_slot) retval = pciehp_check_link_status(ctrl); if (retval) { ctrl_err(ctrl, "Failed to check link status\n"); - set_slot_off(ctrl, p_slot); - return retval; + goto err_exit; } /* Check for a power fault */ - if (pciehp_query_power_fault(p_slot)) { - ctrl_dbg(ctrl, "Power fault detected\n"); + if (ctrl->power_fault_detected || pciehp_query_power_fault(p_slot)) { + ctrl_err(ctrl, "Power fault on slot %s\n", slot_name(p_slot)); retval = -EIO; goto err_exit; } diff --git a/drivers/pci/hotplug/pciehp_hpc.c b/drivers/pci/hotplug/pciehp_hpc.c index 90dac515b60f..10040d58c8ef 100644 --- a/drivers/pci/hotplug/pciehp_hpc.c +++ b/drivers/pci/hotplug/pciehp_hpc.c @@ -511,15 +511,10 @@ int pciehp_power_on_slot(struct slot * slot) return retval; } } + ctrl->power_fault_detected = 0; slot_cmd = POWER_ON; cmd_mask = PCI_EXP_SLTCTL_PCC; - if (!pciehp_poll_mode) { - /* Enable power fault detection turned off at power off time */ - slot_cmd |= PCI_EXP_SLTCTL_PFDE; - cmd_mask |= PCI_EXP_SLTCTL_PFDE; - } - retval = pcie_write_cmd(ctrl, slot_cmd, cmd_mask); if (retval) { ctrl_err(ctrl, "Write %x command failed!\n", slot_cmd); @@ -528,7 +523,6 @@ int pciehp_power_on_slot(struct slot * slot) ctrl_dbg(ctrl, "%s: SLOTCTRL %x write cmd %x\n", __func__, pci_pcie_cap(ctrl->pcie->port) + PCI_EXP_SLTCTL, slot_cmd); - ctrl->power_fault_detected = 0; return retval; } @@ -541,12 +535,6 @@ int pciehp_power_off_slot(struct slot * slot) slot_cmd = POWER_OFF; cmd_mask = PCI_EXP_SLTCTL_PCC; - if (!pciehp_poll_mode) { - /* Disable power fault detection */ - slot_cmd &= ~PCI_EXP_SLTCTL_PFDE; - cmd_mask |= PCI_EXP_SLTCTL_PFDE; - } - retval = pcie_write_cmd(ctrl, slot_cmd, cmd_mask); if (retval) { ctrl_err(ctrl, "Write command failed!\n"); @@ -790,11 +778,19 @@ int pcie_enable_notification(struct controller *ctrl) { u16 cmd, mask; + /* + * TBD: Power fault detected software notification support. + * + * Power fault detected software notification is not enabled + * now, because it caused power fault detected interrupt storm + * on some machines. On those machines, power fault detected + * bit in the slot status register was set again immediately + * when it is cleared in the interrupt service routine, and + * next power fault detected interrupt was notified again. + */ cmd = PCI_EXP_SLTCTL_PDCE; if (ATTN_BUTTN(ctrl)) cmd |= PCI_EXP_SLTCTL_ABPE; - if (POWER_CTRL(ctrl)) - cmd |= PCI_EXP_SLTCTL_PFDE; if (MRL_SENS(ctrl)) cmd |= PCI_EXP_SLTCTL_MRLSCE; if (!pciehp_poll_mode) From 2ed7a806d864bde5903b73da1c65b0316b21efd3 Mon Sep 17 00:00:00 2001 From: Alex Chiang Date: Mon, 16 Nov 2009 14:21:13 -0700 Subject: [PATCH 067/109] x86/PCI: remove early PCI pr_debug statements commit db635adc turned -DDEBUG for x86/pci on when CONFIG_PCI_DEBUG is set. In general, I agree with that change. However, it exposes a bunch of very low level PCI debugging in the early x86 path, such as: 0 reading 2 from a: ffff 1 reading 2 from a: ffff 2 reading 2 from a: ffff 3 reading 2 from a: 300 3 reading 2 from 0: 1002 3 reading 2 from 2: 515e These statements add a lot of noise to the boot and aren't likely to be necessary even when handling random upstream bug reports. [In contrast, statements such as these: pci 0000:02:04.0: found [14e4:164a] class 000200 header type 00 pci 0000:02:04.0: reg 10: [mem 0xf8000000-0xf9ffffff 64bit] pci 0000:02:04.0: reg 30: [mem 0x00000000-0x0001ffff pref] are indeed useful when remote debugging users' machines] Remove the noisy printks and save electrons everywhere. Cc: Bjorn Helgaas Cc: Yinghai Lu Cc: Andi Kleen Cc: Ingo Molnar Signed-off-by: Alex Chiang Signed-off-by: Jesse Barnes --- arch/x86/pci/early.c | 7 ------- 1 file changed, 7 deletions(-) diff --git a/arch/x86/pci/early.c b/arch/x86/pci/early.c index aaf26ae58cd5..d1067d539bee 100644 --- a/arch/x86/pci/early.c +++ b/arch/x86/pci/early.c @@ -12,8 +12,6 @@ u32 read_pci_config(u8 bus, u8 slot, u8 func, u8 offset) u32 v; outl(0x80000000 | (bus<<16) | (slot<<11) | (func<<8) | offset, 0xcf8); v = inl(0xcfc); - if (v != 0xffffffff) - pr_debug("%x reading 4 from %x: %x\n", slot, offset, v); return v; } @@ -22,7 +20,6 @@ u8 read_pci_config_byte(u8 bus, u8 slot, u8 func, u8 offset) u8 v; outl(0x80000000 | (bus<<16) | (slot<<11) | (func<<8) | offset, 0xcf8); v = inb(0xcfc + (offset&3)); - pr_debug("%x reading 1 from %x: %x\n", slot, offset, v); return v; } @@ -31,28 +28,24 @@ u16 read_pci_config_16(u8 bus, u8 slot, u8 func, u8 offset) u16 v; outl(0x80000000 | (bus<<16) | (slot<<11) | (func<<8) | offset, 0xcf8); v = inw(0xcfc + (offset&2)); - pr_debug("%x reading 2 from %x: %x\n", slot, offset, v); return v; } void write_pci_config(u8 bus, u8 slot, u8 func, u8 offset, u32 val) { - pr_debug("%x writing to %x: %x\n", slot, offset, val); outl(0x80000000 | (bus<<16) | (slot<<11) | (func<<8) | offset, 0xcf8); outl(val, 0xcfc); } void write_pci_config_byte(u8 bus, u8 slot, u8 func, u8 offset, u8 val) { - pr_debug("%x writing to %x: %x\n", slot, offset, val); outl(0x80000000 | (bus<<16) | (slot<<11) | (func<<8) | offset, 0xcf8); outb(val, 0xcfc + (offset&3)); } void write_pci_config_16(u8 bus, u8 slot, u8 func, u8 offset, u16 val) { - pr_debug("%x writing to %x: %x\n", slot, offset, val); outl(0x80000000 | (bus<<16) | (slot<<11) | (func<<8) | offset, 0xcf8); outw(val, 0xcfc + (offset&2)); } From 7b7a78594292d540720485544ad1043b71de14e0 Mon Sep 17 00:00:00 2001 From: Jiri Kosina Date: Tue, 17 Nov 2009 23:19:53 +0100 Subject: [PATCH 068/109] PCI: fix comment typo in bus_numa.h MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: André Goddard Rosa Signed-off-by: Jiri Kosina Signed-off-by: Jesse Barnes --- arch/x86/pci/bus_numa.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/x86/pci/bus_numa.h b/arch/x86/pci/bus_numa.h index 4ff126a3e887..730369f392a3 100644 --- a/arch/x86/pci/bus_numa.h +++ b/arch/x86/pci/bus_numa.h @@ -2,7 +2,7 @@ /* * sub bus (transparent) will use entres from 3 to store extra from - * root, so need to make sure we have enought slot there, Should we + * root, so need to make sure we have enough slot there, Should we * increase PCI_BUS_NUM_RESOURCES? */ #define RES_NUM 16 From 67f241f4579651ea4335b58967c8880c0a378249 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Wed, 11 Nov 2009 22:27:40 -0800 Subject: [PATCH 069/109] x86/pci: seperate x86_pci_rootbus_res_quirks from amd_bus.c Those functions are used by intel_bus.c so seperate them to another file. and make amd_bus a bit smaller. Signed-off-by: Yinghai Lu Signed-off-by: Jesse Barnes --- arch/x86/pci/Makefile | 2 +- arch/x86/pci/amd_bus.c | 99 --------------------------------------- arch/x86/pci/bus_numa.c | 101 ++++++++++++++++++++++++++++++++++++++++ arch/x86/pci/bus_numa.h | 1 + 4 files changed, 103 insertions(+), 100 deletions(-) create mode 100644 arch/x86/pci/bus_numa.c diff --git a/arch/x86/pci/Makefile b/arch/x86/pci/Makefile index d8a0a6279a4d..564b008a51c7 100644 --- a/arch/x86/pci/Makefile +++ b/arch/x86/pci/Makefile @@ -15,7 +15,7 @@ obj-$(CONFIG_X86_NUMAQ) += numaq_32.o obj-y += common.o early.o obj-y += amd_bus.o -obj-$(CONFIG_X86_64) += intel_bus.o +obj-$(CONFIG_X86_64) += bus_numa.o intel_bus.o ifeq ($(CONFIG_PCI_DEBUG),y) EXTRA_CFLAGS += -DDEBUG diff --git a/arch/x86/pci/amd_bus.c b/arch/x86/pci/amd_bus.c index 995f36096a42..95ecbd495955 100644 --- a/arch/x86/pci/amd_bus.c +++ b/arch/x86/pci/amd_bus.c @@ -6,8 +6,6 @@ #ifdef CONFIG_X86_64 #include -#include -#include #endif #include "bus_numa.h" @@ -19,54 +17,6 @@ #ifdef CONFIG_X86_64 -int pci_root_num; -struct pci_root_info pci_root_info[PCI_ROOT_NR]; -static int found_all_numa_early; - -void x86_pci_root_bus_res_quirks(struct pci_bus *b) -{ - int i; - int j; - struct pci_root_info *info; - - /* don't go for it if _CRS is used already */ - if (b->resource[0] != &ioport_resource || - b->resource[1] != &iomem_resource) - return; - - if (!pci_root_num) - return; - - /* for amd, if only one root bus, don't need to do anything */ - if (pci_root_num < 2 && found_all_numa_early) - return; - - for (i = 0; i < pci_root_num; i++) { - if (pci_root_info[i].bus_min == b->number) - break; - } - - if (i == pci_root_num) - return; - - printk(KERN_DEBUG "PCI: peer root bus %02x res updated from pci conf\n", - b->number); - - info = &pci_root_info[i]; - for (j = 0; j < info->res_num; j++) { - struct resource *res; - struct resource *root; - - res = &info->res[j]; - b->resource[j] = res; - if (res->flags & IORESOURCE_IO) - root = &ioport_resource; - else - root = &iomem_resource; - insert_resource(root, res); - } -} - #define RANGE_NUM 16 struct res_range { @@ -119,55 +69,6 @@ static void __init update_range(struct res_range *range, size_t start, } } -void __init update_res(struct pci_root_info *info, size_t start, - size_t end, unsigned long flags, int merge) -{ - int i; - struct resource *res; - - if (start > end) - return; - - if (!merge) - goto addit; - - /* try to merge it with old one */ - for (i = 0; i < info->res_num; i++) { - size_t final_start, final_end; - size_t common_start, common_end; - - res = &info->res[i]; - if (res->flags != flags) - continue; - - common_start = max((size_t)res->start, start); - common_end = min((size_t)res->end, end); - if (common_start > common_end + 1) - continue; - - final_start = min((size_t)res->start, start); - final_end = max((size_t)res->end, end); - - res->start = final_start; - res->end = final_end; - return; - } - -addit: - - /* need to add that */ - if (info->res_num >= RES_NUM) - return; - - res = &info->res[info->res_num]; - res->name = info->name; - res->flags = flags; - res->start = start; - res->end = end; - res->child = NULL; - info->res_num++; -} - struct pci_hostbridge_probe { u32 bus; u32 slot; diff --git a/arch/x86/pci/bus_numa.c b/arch/x86/pci/bus_numa.c new file mode 100644 index 000000000000..145df00e0387 --- /dev/null +++ b/arch/x86/pci/bus_numa.c @@ -0,0 +1,101 @@ +#include +#include + +#include "bus_numa.h" + +int pci_root_num; +struct pci_root_info pci_root_info[PCI_ROOT_NR]; +int found_all_numa_early; + +void x86_pci_root_bus_res_quirks(struct pci_bus *b) +{ + int i; + int j; + struct pci_root_info *info; + + /* don't go for it if _CRS is used already */ + if (b->resource[0] != &ioport_resource || + b->resource[1] != &iomem_resource) + return; + + if (!pci_root_num) + return; + + /* for amd, if only one root bus, don't need to do anything */ + if (pci_root_num < 2 && found_all_numa_early) + return; + + for (i = 0; i < pci_root_num; i++) { + if (pci_root_info[i].bus_min == b->number) + break; + } + + if (i == pci_root_num) + return; + + printk(KERN_DEBUG "PCI: peer root bus %02x res updated from pci conf\n", + b->number); + + info = &pci_root_info[i]; + for (j = 0; j < info->res_num; j++) { + struct resource *res; + struct resource *root; + + res = &info->res[j]; + b->resource[j] = res; + if (res->flags & IORESOURCE_IO) + root = &ioport_resource; + else + root = &iomem_resource; + insert_resource(root, res); + } +} + +void __init update_res(struct pci_root_info *info, size_t start, + size_t end, unsigned long flags, int merge) +{ + int i; + struct resource *res; + + if (start > end) + return; + + if (!merge) + goto addit; + + /* try to merge it with old one */ + for (i = 0; i < info->res_num; i++) { + size_t final_start, final_end; + size_t common_start, common_end; + + res = &info->res[i]; + if (res->flags != flags) + continue; + + common_start = max((size_t)res->start, start); + common_end = min((size_t)res->end, end); + if (common_start > common_end + 1) + continue; + + final_start = min((size_t)res->start, start); + final_end = max((size_t)res->end, end); + + res->start = final_start; + res->end = final_end; + return; + } + +addit: + + /* need to add that */ + if (info->res_num >= RES_NUM) + return; + + res = &info->res[info->res_num]; + res->name = info->name; + res->flags = flags; + res->start = start; + res->end = end; + res->child = NULL; + info->res_num++; +} diff --git a/arch/x86/pci/bus_numa.h b/arch/x86/pci/bus_numa.h index 730369f392a3..adbc23fe82ac 100644 --- a/arch/x86/pci/bus_numa.h +++ b/arch/x86/pci/bus_numa.h @@ -20,6 +20,7 @@ struct pci_root_info { #define PCI_ROOT_NR 4 extern int pci_root_num; extern struct pci_root_info pci_root_info[PCI_ROOT_NR]; +extern int found_all_numa_early; extern void update_res(struct pci_root_info *info, size_t start, size_t end, unsigned long flags, int merge); From 5663b1b963183e98ece3e77e471da833bb5ad2ff Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Fri, 13 Nov 2009 17:33:37 -0700 Subject: [PATCH 070/109] x86/PCI: MMCONFIG: remove unused definitions Reviewed-by: Yinghai Lu Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/pci/mmconfig-shared.c | 4 ---- 1 file changed, 4 deletions(-) diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index 02642773c29d..9bf04bcfb9c2 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -23,10 +23,6 @@ #define PREFIX "PCI: " -/* aperture is up to 256MB but BIOS may reserve less */ -#define MMCONFIG_APER_MIN (2 * 1024*1024) -#define MMCONFIG_APER_MAX (256 * 1024*1024) - /* Indicate if the mmcfg resources have been placed into the resource table. */ static int __initdata pci_mmcfg_resources_inserted; From e823d6ff581c5d1d76aa8c73a202d7d1419d34b8 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Fri, 13 Nov 2009 17:33:42 -0700 Subject: [PATCH 071/109] x86/PCI: MMCONFIG: count MCFG structures with local variable Use a local variable, not pci_mmcfg_config_num, to count MCFG entries. No functional change, but simplifies future changes. Reviewed-by: Yinghai Lu Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/pci/mmconfig-shared.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index 9bf04bcfb9c2..fbadb89c71eb 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -555,7 +555,7 @@ static int __init pci_parse_mcfg(struct acpi_table_header *header) { struct acpi_table_mcfg *mcfg; unsigned long i; - int config_size; + int entries, config_size; if (!header) return -EINVAL; @@ -564,17 +564,18 @@ static int __init pci_parse_mcfg(struct acpi_table_header *header) /* how many config structures do we have */ pci_mmcfg_config_num = 0; + entries = 0; i = header->length - sizeof(struct acpi_table_mcfg); while (i >= sizeof(struct acpi_mcfg_allocation)) { - ++pci_mmcfg_config_num; + entries++; i -= sizeof(struct acpi_mcfg_allocation); }; - if (pci_mmcfg_config_num == 0) { + if (entries == 0) { printk(KERN_ERR PREFIX "MMCONFIG has no entries\n"); return -ENODEV; } - config_size = pci_mmcfg_config_num * sizeof(*pci_mmcfg_config); + config_size = entries * sizeof(*pci_mmcfg_config); pci_mmcfg_config = kmalloc(config_size, GFP_KERNEL); if (!pci_mmcfg_config) { printk(KERN_WARNING PREFIX @@ -583,8 +584,9 @@ static int __init pci_parse_mcfg(struct acpi_table_header *header) } memcpy(pci_mmcfg_config, &mcfg[1], config_size); + pci_mmcfg_config_num = entries; - for (i = 0; i < pci_mmcfg_config_num; ++i) { + for (i = 0; i < entries; i++) { if (acpi_mcfg_check_entry(mcfg, &pci_mmcfg_config[i])) { kfree(pci_mmcfg_config); pci_mmcfg_config_num = 0; From d3578ef7aab5b9bb874d085609b3ed5d9abffc48 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Fri, 13 Nov 2009 17:33:47 -0700 Subject: [PATCH 072/109] x86/PCI: MMCONFIG: step through MCFG table, not pci_mmcfg_config[] Step through the ACPI MCFG table, not pci_mmcfg_config[]. No functional change, but simplifies future patches that encapsulate pci_mmcfg_config[]. Reviewed-by: Yinghai Lu Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/pci/mmconfig-shared.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index fbadb89c71eb..7a7b6ba3abbb 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -554,6 +554,7 @@ static int __init acpi_mcfg_check_entry(struct acpi_table_mcfg *mcfg, 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, config_size; @@ -586,8 +587,10 @@ static int __init pci_parse_mcfg(struct acpi_table_header *header) memcpy(pci_mmcfg_config, &mcfg[1], config_size); pci_mmcfg_config_num = entries; + cfg_table = (struct acpi_mcfg_allocation *) &mcfg[1]; for (i = 0; i < entries; i++) { - if (acpi_mcfg_check_entry(mcfg, &pci_mmcfg_config[i])) { + cfg = &cfg_table[i]; + if (acpi_mcfg_check_entry(mcfg, cfg)) { kfree(pci_mmcfg_config); pci_mmcfg_config_num = 0; return -ENODEV; From 7da7d360ae025158d09aab18d66f5d2fe3c02252 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Fri, 13 Nov 2009 17:33:53 -0700 Subject: [PATCH 073/109] x86/PCI: MMCONFIG: centralize MCFG structure management This patch encapsulate pci_mmcfg_config[] updates. All alloc/free is now done in pci_mmconfig_add() and free_all_mcfg(), so all updates to pci_mmcfg_config[] and pci_mmcfg_config_num are in those two functions. This replaces the previous sequence of extend_mmcfg(), fill_one_mmcfg() with the single pci_mmconfig_add() interface. This interface is currently static but will eventually be used in the host bridge hot-add path. Reviewed-by: Yinghai Lu Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/pci/mmconfig-shared.c | 85 ++++++++++++++++------------------ 1 file changed, 39 insertions(+), 46 deletions(-) diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index 7a7b6ba3abbb..62a8ecd96980 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -26,14 +26,24 @@ /* Indicate if the mmcfg resources have been placed into the resource table. */ static int __initdata pci_mmcfg_resources_inserted; -static __init int extend_mmcfg(int num) +static __init void free_all_mmcfg(void) +{ + pci_mmcfg_arch_free(); + pci_mmcfg_config_num = 0; + kfree(pci_mmcfg_config); + pci_mmcfg_config = NULL; +} + +static __init struct acpi_mcfg_allocation *pci_mmconfig_add(int segment, + int start, int end, u64 addr) { struct acpi_mcfg_allocation *new; - int new_num = pci_mmcfg_config_num + num; + int new_num = pci_mmcfg_config_num + 1; + int i = pci_mmcfg_config_num; new = kzalloc(sizeof(pci_mmcfg_config[0]) * new_num, GFP_KERNEL); if (!new) - return -1; + return NULL; if (pci_mmcfg_config) { memcpy(new, pci_mmcfg_config, @@ -42,18 +52,13 @@ static __init int extend_mmcfg(int num) } pci_mmcfg_config = new; - return 0; -} - -static __init void fill_one_mmcfg(u64 addr, int segment, int start, int end) -{ - int i = pci_mmcfg_config_num; - pci_mmcfg_config_num++; pci_mmcfg_config[i].address = addr; pci_mmcfg_config[i].pci_segment = segment; pci_mmcfg_config[i].start_bus_number = start; pci_mmcfg_config[i].end_bus_number = end; + + return &pci_mmcfg_config[i]; } static const char __init *pci_mmcfg_e7520(void) @@ -65,11 +70,9 @@ static const char __init *pci_mmcfg_e7520(void) if (win == 0x0000 || win == 0xf000) return NULL; - if (extend_mmcfg(1) == -1) + if (pci_mmconfig_add(0, 0, 255, win << 16) == NULL) return NULL; - fill_one_mmcfg(win << 16, 0, 0, 255); - return "Intel Corporation E7520 Memory Controller Hub"; } @@ -111,11 +114,9 @@ static const char __init *pci_mmcfg_intel_945(void) if ((pciexbar & mask) >= 0xf0000000U) return NULL; - if (extend_mmcfg(1) == -1) + if (pci_mmconfig_add(0, 0, (len >> 20) - 1, pciexbar & mask) == NULL) return NULL; - fill_one_mmcfg(pciexbar & mask, 0, 0, (len >> 20) - 1); - return "Intel Corporation 945G/GZ/P/PL Express Memory Controller Hub"; } @@ -124,7 +125,7 @@ static const char __init *pci_mmcfg_amd_fam10h(void) u32 low, high, address; u64 base, msr; int i; - unsigned segnbits = 0, busnbits; + unsigned segnbits = 0, busnbits, end_bus; if (!(pci_probe & PCI_CHECK_ENABLE_AMD_MMCONF)) return NULL; @@ -158,11 +159,13 @@ static const char __init *pci_mmcfg_amd_fam10h(void) busnbits = 8; } - if (extend_mmcfg(1 << segnbits) == -1) - return NULL; - + end_bus = (1 << busnbits) - 1; for (i = 0; i < (1 << segnbits); i++) - fill_one_mmcfg(base + (1<<28) * i, i, 0, (1 << busnbits) - 1); + if (pci_mmconfig_add(i, 0, end_bus, + base + (1<<28) * i) == NULL) { + free_all_mmcfg(); + return NULL; + } return "AMD Family 10h NB"; } @@ -210,16 +213,14 @@ static const char __init *pci_mmcfg_nvidia_mcp55(void) if (!(extcfg & extcfg_enable_mask)) continue; - if (extend_mmcfg(1) == -1) - continue; - size_index = (extcfg & extcfg_size_mask) >> extcfg_size_shift; base = extcfg & extcfg_base_mask[size_index]; /* base could > 4G */ base <<= extcfg_base_lshift; start = (extcfg & extcfg_start_mask) >> extcfg_start_shift; end = start + extcfg_sizebus[size_index] - 1; - fill_one_mmcfg(base, 0, start, end); + if (pci_mmconfig_add(0, start, end, base) == NULL) + continue; mcp55_mmconf_found++; } @@ -303,8 +304,7 @@ static int __init pci_mmcfg_check_hostbridge(void) if (!raw_pci_ops) return 0; - pci_mmcfg_config_num = 0; - pci_mmcfg_config = NULL; + free_all_mmcfg(); for (i = 0; i < ARRAY_SIZE(pci_mmcfg_probes); i++) { bus = pci_mmcfg_probes[i].bus; @@ -516,10 +516,7 @@ static void __init pci_mmcfg_reject_broken(int early) reject: printk(KERN_INFO "PCI: Not using MMCONFIG.\n"); - pci_mmcfg_arch_free(); - kfree(pci_mmcfg_config); - pci_mmcfg_config = NULL; - pci_mmcfg_config_num = 0; + free_all_mmcfg(); } static int __initdata known_bridge; @@ -556,7 +553,7 @@ 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, config_size; + int entries; if (!header) return -EINVAL; @@ -564,7 +561,7 @@ static int __init pci_parse_mcfg(struct acpi_table_header *header) mcfg = (struct acpi_table_mcfg *)header; /* how many config structures do we have */ - pci_mmcfg_config_num = 0; + free_all_mmcfg(); entries = 0; i = header->length - sizeof(struct acpi_table_mcfg); while (i >= sizeof(struct acpi_mcfg_allocation)) { @@ -576,25 +573,21 @@ static int __init pci_parse_mcfg(struct acpi_table_header *header) return -ENODEV; } - config_size = entries * sizeof(*pci_mmcfg_config); - pci_mmcfg_config = kmalloc(config_size, GFP_KERNEL); - if (!pci_mmcfg_config) { - printk(KERN_WARNING PREFIX - "No memory for MCFG config tables\n"); - return -ENOMEM; - } - - memcpy(pci_mmcfg_config, &mcfg[1], config_size); - pci_mmcfg_config_num = entries; - 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)) { - kfree(pci_mmcfg_config); - pci_mmcfg_config_num = 0; + free_all_mmcfg(); return -ENODEV; } + + if (pci_mmconfig_add(cfg->pci_segment, cfg->start_bus_number, + cfg->end_bus_number, cfg->address) == NULL) { + printk(KERN_WARNING PREFIX + "no memory for MCFG entries\n"); + free_all_mmcfg(); + return -ENOMEM; + } } return 0; From 463a5df175e3ceed684397ee2f8a3eb523d835a0 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Fri, 13 Nov 2009 17:33:58 -0700 Subject: [PATCH 074/109] x86/PCI: MMCONFIG: simplify tests for empty pci_mmcfg_config table We never set pci_mmcfg_config unless we increment pci_mmcfg_config_num, so there's no need to test both pci_mmcfg_config_num and pci_mmcfg_config. Reviewed-by: Yinghai Lu Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/pci/mmconfig-shared.c | 3 --- 1 file changed, 3 deletions(-) diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index 62a8ecd96980..a0cc4d2efb8a 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -472,7 +472,6 @@ static void __init pci_mmcfg_reject_broken(int early) int i; if ((pci_mmcfg_config_num == 0) || - (pci_mmcfg_config == NULL) || (pci_mmcfg_config[0].address == 0)) return; @@ -618,7 +617,6 @@ static void __init __pci_mmcfg_init(int early) pci_mmcfg_reject_broken(early); if ((pci_mmcfg_config_num == 0) || - (pci_mmcfg_config == NULL) || (pci_mmcfg_config[0].address == 0)) return; @@ -652,7 +650,6 @@ static int __init pci_mmcfg_late_insert_resources(void) if ((pci_mmcfg_resources_inserted == 1) || (pci_probe & PCI_PROBE_MMCONF) == 0 || (pci_mmcfg_config_num == 0) || - (pci_mmcfg_config == NULL) || (pci_mmcfg_config[0].address == 0)) return 1; From f7ca69848786bb99fdfafb511791b078c298438e Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Fri, 13 Nov 2009 17:34:03 -0700 Subject: [PATCH 075/109] x86/PCI: MMCONFIG: reject MMCONFIG apertures at address zero Since all MMCONFIG regions go through pci_mmconfig_add(), we can test the address once there. If the caller supplies an address of zero, we never insert it in the pci_mmcfg_config[] table, so no need to test it elsewhere. Reviewed-by: Yinghai Lu Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/pci/mmconfig-shared.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index a0cc4d2efb8a..067a2cfed15c 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -41,6 +41,9 @@ static __init struct acpi_mcfg_allocation *pci_mmconfig_add(int segment, int new_num = pci_mmcfg_config_num + 1; int i = pci_mmcfg_config_num; + if (addr == 0) + return NULL; + new = kzalloc(sizeof(pci_mmcfg_config[0]) * new_num, GFP_KERNEL); if (!new) return NULL; @@ -471,8 +474,7 @@ static void __init pci_mmcfg_reject_broken(int early) typeof(pci_mmcfg_config[0]) *cfg; int i; - if ((pci_mmcfg_config_num == 0) || - (pci_mmcfg_config[0].address == 0)) + if (pci_mmcfg_config_num == 0) return; for (i = 0; i < pci_mmcfg_config_num; i++) { @@ -616,8 +618,7 @@ static void __init __pci_mmcfg_init(int early) pci_mmcfg_reject_broken(early); - if ((pci_mmcfg_config_num == 0) || - (pci_mmcfg_config[0].address == 0)) + if (pci_mmcfg_config_num == 0) return; if (pci_mmcfg_arch_init()) @@ -649,8 +650,7 @@ static int __init pci_mmcfg_late_insert_resources(void) */ if ((pci_mmcfg_resources_inserted == 1) || (pci_probe & PCI_PROBE_MMCONF) == 0 || - (pci_mmcfg_config_num == 0) || - (pci_mmcfg_config[0].address == 0)) + (pci_mmcfg_config_num == 0)) return 1; /* From df5eb1d67e8074dfbc23cf396c556116728187b3 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Fri, 13 Nov 2009 17:34:08 -0700 Subject: [PATCH 076/109] x86/PCI: MMCONFIG: add PCI_MMCFG_BUS_OFFSET() to factor common expression This factors out the common "bus << 20" expression used when computing the MMCONFIG address. Reviewed-by: Yinghai Lu Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/include/asm/pci_x86.h | 2 ++ arch/x86/pci/mmconfig-shared.c | 16 ++++++++-------- arch/x86/pci/mmconfig_32.c | 2 +- arch/x86/pci/mmconfig_64.c | 15 +++++++-------- 4 files changed, 18 insertions(+), 17 deletions(-) diff --git a/arch/x86/include/asm/pci_x86.h b/arch/x86/include/asm/pci_x86.h index b399988eee3a..7d94a235ec82 100644 --- a/arch/x86/include/asm/pci_x86.h +++ b/arch/x86/include/asm/pci_x86.h @@ -124,6 +124,8 @@ extern void __init pci_mmcfg_arch_free(void); extern struct acpi_mcfg_allocation *pci_mmcfg_config; extern int pci_mmcfg_config_num; +#define PCI_MMCFG_BUS_OFFSET(bus) ((bus) << 20) + /* * 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 067a2cfed15c..4820f0e8c594 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -355,8 +355,9 @@ static void __init pci_mmcfg_insert_resources(void) snprintf(names, PCI_MMCFG_RESOURCE_NAME_LEN, "PCI MMCONFIG %u [%02x-%02x]", cfg->pci_segment, cfg->start_bus_number, cfg->end_bus_number); - res->start = cfg->address + (cfg->start_bus_number << 20); - res->end = res->start + (num_buses << 20) - 1; + res->start = cfg->address + + PCI_MMCFG_BUS_OFFSET(cfg->start_bus_number); + res->end = res->start + PCI_MMCFG_BUS_OFFSET(num_buses) - 1; res->flags = IORESOURCE_MEM | IORESOURCE_BUSY; insert_resource(&iomem_resource, res); names += PCI_MMCFG_RESOURCE_NAME_LEN; @@ -478,15 +479,14 @@ static void __init pci_mmcfg_reject_broken(int early) return; for (i = 0; i < pci_mmcfg_config_num; i++) { - int valid = 0; + int num_buses, valid = 0; u64 addr, size; cfg = &pci_mmcfg_config[i]; - addr = cfg->start_bus_number; - addr <<= 20; - addr += cfg->address; - size = cfg->end_bus_number + 1 - cfg->start_bus_number; - size <<= 20; + addr = cfg->address + + PCI_MMCFG_BUS_OFFSET(cfg->start_bus_number); + num_buses = cfg->end_bus_number - cfg->start_bus_number + 1; + size = PCI_MMCFG_BUS_OFFSET(num_buses); printk(KERN_NOTICE "PCI: MCFG configuration %d: base %lx " "segment %hu buses %u - %u\n", i, (unsigned long)cfg->address, cfg->pci_segment, diff --git a/arch/x86/pci/mmconfig_32.c b/arch/x86/pci/mmconfig_32.c index f10a7e94a84c..8c19df89ad75 100644 --- a/arch/x86/pci/mmconfig_32.c +++ b/arch/x86/pci/mmconfig_32.c @@ -47,7 +47,7 @@ static u32 get_base_addr(unsigned int seg, int bus, unsigned devfn) */ static void pci_exp_set_dev_base(unsigned int base, int bus, int devfn) { - u32 dev_base = base | (bus << 20) | (devfn << 12); + u32 dev_base = base | PCI_MMCFG_BUS_OFFSET(bus) | (devfn << 12); int cpu = smp_processor_id(); if (dev_base != mmcfg_last_accessed_device || cpu != mmcfg_last_accessed_cpu) { diff --git a/arch/x86/pci/mmconfig_64.c b/arch/x86/pci/mmconfig_64.c index 94349f8b2f96..8588711924cc 100644 --- a/arch/x86/pci/mmconfig_64.c +++ b/arch/x86/pci/mmconfig_64.c @@ -43,7 +43,7 @@ static char __iomem *pci_dev_base(unsigned int seg, unsigned int bus, unsigned i addr = get_virt(seg, bus); if (!addr) return NULL; - return addr + ((bus << 20) | (devfn << 12)); + return addr + (PCI_MMCFG_BUS_OFFSET(bus) | (devfn << 12)); } static int pci_mmcfg_read(unsigned int seg, unsigned int bus, @@ -113,17 +113,16 @@ static void __iomem * __init mcfg_ioremap(struct acpi_mcfg_allocation *cfg) { void __iomem *addr; u64 start, size; + int num_buses; - start = cfg->start_bus_number; - start <<= 20; - start += cfg->address; - size = cfg->end_bus_number + 1 - cfg->start_bus_number; - size <<= 20; + start = cfg->address + PCI_MMCFG_BUS_OFFSET(cfg->start_bus_number); + num_buses = cfg->end_bus_number - cfg->start_bus_number + 1; + size = PCI_MMCFG_BUS_OFFSET(num_buses); addr = ioremap_nocache(start, size); if (addr) { printk(KERN_INFO "PCI: Using MMCONFIG at %Lx - %Lx\n", start, start + size - 1); - addr -= cfg->start_bus_number << 20; + addr -= PCI_MMCFG_BUS_OFFSET(cfg->start_bus_number); } return addr; } @@ -162,7 +161,7 @@ void __init pci_mmcfg_arch_free(void) for (i = 0; i < pci_mmcfg_config_num; ++i) { if (pci_mmcfg_virt[i].virt) { - iounmap(pci_mmcfg_virt[i].virt + (pci_mmcfg_virt[i].cfg->start_bus_number << 20)); + iounmap(pci_mmcfg_virt[i].virt + PCI_MMCFG_BUS_OFFSET(pci_mmcfg_virt[i].cfg->start_bus_number)); pci_mmcfg_virt[i].virt = NULL; pci_mmcfg_virt[i].cfg = NULL; } From d215a9c8b46e55a1d3bc1cd907c943ef95938a0e Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Fri, 13 Nov 2009 17:34:13 -0700 Subject: [PATCH 077/109] x86/PCI: MMCONFIG: use a private structure rather than the ACPI MCFG one This adds a struct pci_mmcfg_region with a little more information than the struct acpi_mcfg_allocation used previously. The acpi_mcfg structure is defined by the spec, so we can't change it. To begin with, struct pci_mmcfg_region is basically the same as the ACPI MCFG version, but future patches will add more information. Reviewed-by: Yinghai Lu Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/include/asm/pci_x86.h | 9 ++++++++- arch/x86/pci/mmconfig-shared.c | 10 +++++----- arch/x86/pci/mmconfig_32.c | 2 +- arch/x86/pci/mmconfig_64.c | 6 +++--- 4 files changed, 17 insertions(+), 10 deletions(-) diff --git a/arch/x86/include/asm/pci_x86.h b/arch/x86/include/asm/pci_x86.h index 7d94a235ec82..3a2ca5f69521 100644 --- a/arch/x86/include/asm/pci_x86.h +++ b/arch/x86/include/asm/pci_x86.h @@ -118,10 +118,17 @@ extern int __init pcibios_init(void); /* pci-mmconfig.c */ +struct pci_mmcfg_region { + u64 address; + u16 pci_segment; + u8 start_bus_number; + u8 end_bus_number; +}; + extern int __init pci_mmcfg_arch_init(void); extern void __init pci_mmcfg_arch_free(void); -extern struct acpi_mcfg_allocation *pci_mmcfg_config; +extern struct pci_mmcfg_region *pci_mmcfg_config; extern int pci_mmcfg_config_num; #define PCI_MMCFG_BUS_OFFSET(bus) ((bus) << 20) diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index 4820f0e8c594..5f7afdd1e2d6 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -34,10 +34,10 @@ static __init void free_all_mmcfg(void) pci_mmcfg_config = NULL; } -static __init struct acpi_mcfg_allocation *pci_mmconfig_add(int segment, - int start, int end, u64 addr) +static __init struct pci_mmcfg_region *pci_mmconfig_add(int segment, int start, + int end, u64 addr) { - struct acpi_mcfg_allocation *new; + struct pci_mmcfg_region *new; int new_num = pci_mmcfg_config_num + 1; int i = pci_mmcfg_config_num; @@ -349,7 +349,7 @@ static void __init pci_mmcfg_insert_resources(void) names = (void *)&res[pci_mmcfg_config_num]; for (i = 0; i < pci_mmcfg_config_num; i++, res++) { - struct acpi_mcfg_allocation *cfg = &pci_mmcfg_config[i]; + struct pci_mmcfg_region *cfg = &pci_mmcfg_config[i]; num_buses = cfg->end_bus_number - cfg->start_bus_number + 1; res->name = names; snprintf(names, PCI_MMCFG_RESOURCE_NAME_LEN, @@ -523,7 +523,7 @@ reject: static int __initdata known_bridge; /* The physical address of the MMCONFIG aperture. Set from ACPI tables. */ -struct acpi_mcfg_allocation *pci_mmcfg_config; +struct pci_mmcfg_region *pci_mmcfg_config; int pci_mmcfg_config_num; static int __init acpi_mcfg_check_entry(struct acpi_table_mcfg *mcfg, diff --git a/arch/x86/pci/mmconfig_32.c b/arch/x86/pci/mmconfig_32.c index 8c19df89ad75..3936eced993c 100644 --- a/arch/x86/pci/mmconfig_32.c +++ b/arch/x86/pci/mmconfig_32.c @@ -27,7 +27,7 @@ static int mmcfg_last_accessed_cpu; */ static u32 get_base_addr(unsigned int seg, int bus, unsigned devfn) { - struct acpi_mcfg_allocation *cfg; + struct pci_mmcfg_region *cfg; int cfg_num; for (cfg_num = 0; cfg_num < pci_mmcfg_config_num; cfg_num++) { diff --git a/arch/x86/pci/mmconfig_64.c b/arch/x86/pci/mmconfig_64.c index 8588711924cc..7a6231c3335e 100644 --- a/arch/x86/pci/mmconfig_64.c +++ b/arch/x86/pci/mmconfig_64.c @@ -14,14 +14,14 @@ /* Static virtual mapping of the MMCONFIG aperture */ struct mmcfg_virt { - struct acpi_mcfg_allocation *cfg; + struct pci_mmcfg_region *cfg; char __iomem *virt; }; static struct mmcfg_virt *pci_mmcfg_virt; static char __iomem *get_virt(unsigned int seg, unsigned bus) { - struct acpi_mcfg_allocation *cfg; + struct pci_mmcfg_region *cfg; int cfg_num; for (cfg_num = 0; cfg_num < pci_mmcfg_config_num; cfg_num++) { @@ -109,7 +109,7 @@ static struct pci_raw_ops pci_mmcfg = { .write = pci_mmcfg_write, }; -static void __iomem * __init mcfg_ioremap(struct acpi_mcfg_allocation *cfg) +static void __iomem * __init mcfg_ioremap(struct pci_mmcfg_region *cfg) { void __iomem *addr; u64 start, size; From d7e6b66fe87c9f42480d73fc314aecaeae84ca6b Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Fri, 13 Nov 2009 17:34:18 -0700 Subject: [PATCH 078/109] x86/PCI: MMCONFIG: rename pci_mmcfg_region structure members This only renames the struct pci_mmcfg_region members; no functional change. Reviewed-by: Yinghai Lu Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/include/asm/pci_x86.h | 6 ++-- arch/x86/pci/mmconfig-shared.c | 50 +++++++++++++++++----------------- arch/x86/pci/mmconfig_32.c | 6 ++-- arch/x86/pci/mmconfig_64.c | 16 +++++------ 4 files changed, 39 insertions(+), 39 deletions(-) diff --git a/arch/x86/include/asm/pci_x86.h b/arch/x86/include/asm/pci_x86.h index 3a2ca5f69521..a752d618f196 100644 --- a/arch/x86/include/asm/pci_x86.h +++ b/arch/x86/include/asm/pci_x86.h @@ -120,9 +120,9 @@ extern int __init pcibios_init(void); struct pci_mmcfg_region { u64 address; - u16 pci_segment; - u8 start_bus_number; - u8 end_bus_number; + u16 segment; + u8 start_bus; + u8 end_bus; }; extern int __init pci_mmcfg_arch_init(void); diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index 5f7afdd1e2d6..5479fbb2d6ab 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -57,9 +57,9 @@ static __init struct pci_mmcfg_region *pci_mmconfig_add(int segment, int start, pci_mmcfg_config_num++; pci_mmcfg_config[i].address = addr; - pci_mmcfg_config[i].pci_segment = segment; - pci_mmcfg_config[i].start_bus_number = start; - pci_mmcfg_config[i].end_bus_number = end; + pci_mmcfg_config[i].segment = segment; + pci_mmcfg_config[i].start_bus = start; + pci_mmcfg_config[i].end_bus = end; return &pci_mmcfg_config[i]; } @@ -260,8 +260,8 @@ static int __init cmp_mmcfg(const void *x1, const void *x2) const typeof(pci_mmcfg_config[0]) *m2 = x2; int start1, start2; - start1 = m1->start_bus_number; - start2 = m2->start_bus_number; + start1 = m1->start_bus; + start2 = m2->start_bus; return start1 - start2; } @@ -279,8 +279,8 @@ static void __init pci_mmcfg_check_end_bus_number(void) if (pci_mmcfg_config_num > 0) { i = pci_mmcfg_config_num - 1; cfg = &pci_mmcfg_config[i]; - if (cfg->end_bus_number < cfg->start_bus_number) - cfg->end_bus_number = 255; + if (cfg->end_bus < cfg->start_bus) + cfg->end_bus = 255; } /* don't overlap please */ @@ -288,11 +288,11 @@ static void __init pci_mmcfg_check_end_bus_number(void) cfg = &pci_mmcfg_config[i]; cfgx = &pci_mmcfg_config[i+1]; - if (cfg->end_bus_number < cfg->start_bus_number) - cfg->end_bus_number = 255; + if (cfg->end_bus < cfg->start_bus) + cfg->end_bus = 255; - if (cfg->end_bus_number >= cfgx->start_bus_number) - cfg->end_bus_number = cfgx->start_bus_number - 1; + if (cfg->end_bus >= cfgx->start_bus) + cfg->end_bus = cfgx->start_bus - 1; } } @@ -350,13 +350,13 @@ static void __init pci_mmcfg_insert_resources(void) names = (void *)&res[pci_mmcfg_config_num]; for (i = 0; i < pci_mmcfg_config_num; i++, res++) { struct pci_mmcfg_region *cfg = &pci_mmcfg_config[i]; - num_buses = cfg->end_bus_number - cfg->start_bus_number + 1; + num_buses = cfg->end_bus - cfg->start_bus + 1; res->name = names; snprintf(names, PCI_MMCFG_RESOURCE_NAME_LEN, - "PCI MMCONFIG %u [%02x-%02x]", cfg->pci_segment, - cfg->start_bus_number, cfg->end_bus_number); + "PCI MMCONFIG %u [%02x-%02x]", cfg->segment, + cfg->start_bus, cfg->end_bus); res->start = cfg->address + - PCI_MMCFG_BUS_OFFSET(cfg->start_bus_number); + PCI_MMCFG_BUS_OFFSET(cfg->start_bus); res->end = res->start + PCI_MMCFG_BUS_OFFSET(num_buses) - 1; res->flags = IORESOURCE_MEM | IORESOURCE_BUSY; insert_resource(&iomem_resource, res); @@ -457,13 +457,13 @@ static int __init is_mmconf_reserved(check_reserved_t is_reserved, valid = 1; if (old_size != size) { - /* update end_bus_number */ - cfg->end_bus_number = cfg->start_bus_number + ((size>>20) - 1); + /* update end_bus */ + cfg->end_bus = cfg->start_bus + ((size>>20) - 1); printk(KERN_NOTICE "PCI: updated MCFG configuration %d: base %lx " "segment %hu buses %u - %u\n", - i, (unsigned long)cfg->address, cfg->pci_segment, - (unsigned int)cfg->start_bus_number, - (unsigned int)cfg->end_bus_number); + i, (unsigned long)cfg->address, cfg->segment, + (unsigned int)cfg->start_bus, + (unsigned int)cfg->end_bus); } } @@ -484,14 +484,14 @@ static void __init pci_mmcfg_reject_broken(int early) cfg = &pci_mmcfg_config[i]; addr = cfg->address + - PCI_MMCFG_BUS_OFFSET(cfg->start_bus_number); - num_buses = cfg->end_bus_number - cfg->start_bus_number + 1; + PCI_MMCFG_BUS_OFFSET(cfg->start_bus); + num_buses = cfg->end_bus - cfg->start_bus + 1; size = PCI_MMCFG_BUS_OFFSET(num_buses); printk(KERN_NOTICE "PCI: MCFG configuration %d: base %lx " "segment %hu buses %u - %u\n", - i, (unsigned long)cfg->address, cfg->pci_segment, - (unsigned int)cfg->start_bus_number, - (unsigned int)cfg->end_bus_number); + i, (unsigned long)cfg->address, cfg->segment, + (unsigned int)cfg->start_bus, + (unsigned int)cfg->end_bus); if (!early && !acpi_disabled) valid = is_mmconf_reserved(is_acpi_reserved, addr, size, i, cfg, 0); diff --git a/arch/x86/pci/mmconfig_32.c b/arch/x86/pci/mmconfig_32.c index 3936eced993c..a3cee532c935 100644 --- a/arch/x86/pci/mmconfig_32.c +++ b/arch/x86/pci/mmconfig_32.c @@ -32,9 +32,9 @@ static u32 get_base_addr(unsigned int seg, int bus, unsigned devfn) for (cfg_num = 0; cfg_num < pci_mmcfg_config_num; cfg_num++) { cfg = &pci_mmcfg_config[cfg_num]; - if (cfg->pci_segment == seg && - (cfg->start_bus_number <= bus) && - (cfg->end_bus_number >= bus)) + if (cfg->segment == seg && + (cfg->start_bus <= bus) && + (cfg->end_bus >= bus)) return cfg->address; } diff --git a/arch/x86/pci/mmconfig_64.c b/arch/x86/pci/mmconfig_64.c index 7a6231c3335e..fdf08f97131b 100644 --- a/arch/x86/pci/mmconfig_64.c +++ b/arch/x86/pci/mmconfig_64.c @@ -26,9 +26,9 @@ static char __iomem *get_virt(unsigned int seg, unsigned bus) for (cfg_num = 0; cfg_num < pci_mmcfg_config_num; cfg_num++) { cfg = pci_mmcfg_virt[cfg_num].cfg; - if (cfg->pci_segment == seg && - (cfg->start_bus_number <= bus) && - (cfg->end_bus_number >= bus)) + if (cfg->segment == seg && + (cfg->start_bus <= bus) && + (cfg->end_bus >= bus)) return pci_mmcfg_virt[cfg_num].virt; } @@ -115,14 +115,14 @@ static void __iomem * __init mcfg_ioremap(struct pci_mmcfg_region *cfg) u64 start, size; int num_buses; - start = cfg->address + PCI_MMCFG_BUS_OFFSET(cfg->start_bus_number); - num_buses = cfg->end_bus_number - cfg->start_bus_number + 1; + 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) { printk(KERN_INFO "PCI: Using MMCONFIG at %Lx - %Lx\n", start, start + size - 1); - addr -= PCI_MMCFG_BUS_OFFSET(cfg->start_bus_number); + addr -= PCI_MMCFG_BUS_OFFSET(cfg->start_bus); } return addr; } @@ -143,7 +143,7 @@ int __init pci_mmcfg_arch_init(void) if (!pci_mmcfg_virt[i].virt) { printk(KERN_ERR "PCI: Cannot map mmconfig aperture for " "segment %d\n", - pci_mmcfg_config[i].pci_segment); + pci_mmcfg_config[i].segment); pci_mmcfg_arch_free(); return 0; } @@ -161,7 +161,7 @@ void __init pci_mmcfg_arch_free(void) for (i = 0; i < pci_mmcfg_config_num; ++i) { if (pci_mmcfg_virt[i].virt) { - iounmap(pci_mmcfg_virt[i].virt + PCI_MMCFG_BUS_OFFSET(pci_mmcfg_virt[i].cfg->start_bus_number)); + iounmap(pci_mmcfg_virt[i].virt + PCI_MMCFG_BUS_OFFSET(pci_mmcfg_virt[i].cfg->start_bus)); pci_mmcfg_virt[i].virt = NULL; pci_mmcfg_virt[i].cfg = NULL; } From 95cf1cf0c5a767feb811dfed298b95b1df8824c7 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Fri, 13 Nov 2009 17:34:24 -0700 Subject: [PATCH 079/109] x86/PCI: MMCONFIG: use pointer to simplify pci_mmcfg_config[] structure access No functional change, but simplifies a future patch to convert the table to a list. Reviewed-by: Yinghai Lu Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/pci/mmconfig-shared.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index 5479fbb2d6ab..28ac9f58a986 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -54,12 +54,14 @@ static __init struct pci_mmcfg_region *pci_mmconfig_add(int segment, int start, kfree(pci_mmcfg_config); } pci_mmcfg_config = new; - pci_mmcfg_config_num++; - pci_mmcfg_config[i].address = addr; - pci_mmcfg_config[i].segment = segment; - pci_mmcfg_config[i].start_bus = start; - pci_mmcfg_config[i].end_bus = end; + + new = &pci_mmcfg_config[i]; + + new->address = addr; + new->segment = segment; + new->start_bus = start; + new->end_bus = end; return &pci_mmcfg_config[i]; } From 56ddf4d3cf04e80254d3d721c6bea2f8ec44c41a Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Fri, 13 Nov 2009 17:34:29 -0700 Subject: [PATCH 080/109] x86/PCI: MMCONFIG: add resource to struct pci_mmcfg_region This patch adds a resource and corresponding name to the MMCONFIG structure. This makes allocation simpler (we can allocate the resource and name at the same time we allocate the pci_mmcfg_region), and gives us a way to hang onto the resource after inserting it. This will be needed so we can release and free it when hot-removing a host bridge. Reviewed-by: Yinghai Lu Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/include/asm/pci_x86.h | 5 +++ arch/x86/pci/mmconfig-shared.c | 64 ++++++++++++++++++---------------- 2 files changed, 38 insertions(+), 31 deletions(-) diff --git a/arch/x86/include/asm/pci_x86.h b/arch/x86/include/asm/pci_x86.h index a752d618f196..a6d42c10b017 100644 --- a/arch/x86/include/asm/pci_x86.h +++ b/arch/x86/include/asm/pci_x86.h @@ -118,11 +118,16 @@ extern int __init pcibios_init(void); /* pci-mmconfig.c */ +/* "PCI MMCONFIG %04x [bus %02x-%02x]" */ +#define PCI_MMCFG_RESOURCE_NAME_LEN (22 + 4 + 2 + 2) + struct pci_mmcfg_region { + struct resource res; u64 address; u16 segment; u8 start_bus; u8 end_bus; + char name[PCI_MMCFG_RESOURCE_NAME_LEN]; }; extern int __init pci_mmcfg_arch_init(void); diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index 28ac9f58a986..ba3aa3697418 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -28,7 +28,15 @@ static int __initdata pci_mmcfg_resources_inserted; static __init void free_all_mmcfg(void) { + int i; + struct pci_mmcfg_region *cfg; + pci_mmcfg_arch_free(); + for (i = 0; i < pci_mmcfg_config_num; i++) { + cfg = &pci_mmcfg_config[i]; + if (cfg->res.parent) + release_resource(&cfg->res); + } pci_mmcfg_config_num = 0; kfree(pci_mmcfg_config); pci_mmcfg_config = NULL; @@ -40,6 +48,8 @@ static __init struct pci_mmcfg_region *pci_mmconfig_add(int segment, int start, struct pci_mmcfg_region *new; int new_num = pci_mmcfg_config_num + 1; int i = pci_mmcfg_config_num; + int num_buses; + struct resource *res; if (addr == 0) return NULL; @@ -63,6 +73,15 @@ static __init struct pci_mmcfg_region *pci_mmconfig_add(int segment, int start, new->start_bus = start; new->end_bus = end; + num_buses = end - start + 1; + res = &new->res; + res->start = addr + PCI_MMCFG_BUS_OFFSET(start); + res->end = addr + PCI_MMCFG_BUS_OFFSET(num_buses) - 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 &pci_mmcfg_config[i]; } @@ -336,33 +355,12 @@ static int __init pci_mmcfg_check_hostbridge(void) static void __init pci_mmcfg_insert_resources(void) { -#define PCI_MMCFG_RESOURCE_NAME_LEN 24 int i; - struct resource *res; - char *names; - unsigned num_buses; + struct pci_mmcfg_region *cfg; - res = kcalloc(PCI_MMCFG_RESOURCE_NAME_LEN + sizeof(*res), - pci_mmcfg_config_num, GFP_KERNEL); - if (!res) { - printk(KERN_ERR "PCI: Unable to allocate MMCONFIG resources\n"); - return; - } - - names = (void *)&res[pci_mmcfg_config_num]; - for (i = 0; i < pci_mmcfg_config_num; i++, res++) { - struct pci_mmcfg_region *cfg = &pci_mmcfg_config[i]; - num_buses = cfg->end_bus - cfg->start_bus + 1; - res->name = names; - snprintf(names, PCI_MMCFG_RESOURCE_NAME_LEN, - "PCI MMCONFIG %u [%02x-%02x]", cfg->segment, - cfg->start_bus, cfg->end_bus); - res->start = cfg->address + - PCI_MMCFG_BUS_OFFSET(cfg->start_bus); - res->end = res->start + PCI_MMCFG_BUS_OFFSET(num_buses) - 1; - res->flags = IORESOURCE_MEM | IORESOURCE_BUSY; - insert_resource(&iomem_resource, res); - names += PCI_MMCFG_RESOURCE_NAME_LEN; + for (i = 0; i < pci_mmcfg_config_num; i++) { + cfg = &pci_mmcfg_config[i]; + insert_resource(&iomem_resource, &cfg->res); } /* Mark that the resources have been inserted. */ @@ -444,7 +442,7 @@ static int __init is_mmconf_reserved(check_reserved_t is_reserved, typeof(pci_mmcfg_config[0]) *cfg, int with_e820) { u64 old_size = size; - int valid = 0; + int valid = 0, num_buses; while (!is_reserved(addr, addr + size, E820_RESERVED)) { size >>= 1; @@ -461,6 +459,12 @@ static int __init is_mmconf_reserved(check_reserved_t is_reserved, if (old_size != size) { /* update end_bus */ cfg->end_bus = cfg->start_bus + ((size>>20) - 1); + num_buses = cfg->end_bus - cfg->start_bus + 1; + cfg->res.end = cfg->res.start + + PCI_MMCFG_BUS_OFFSET(num_buses) - 1; + snprintf(cfg->name, PCI_MMCFG_RESOURCE_NAME_LEN, + "PCI MMCONFIG %04x [bus %02x-%02x]", + cfg->segment, cfg->start_bus, cfg->end_bus); printk(KERN_NOTICE "PCI: updated MCFG configuration %d: base %lx " "segment %hu buses %u - %u\n", i, (unsigned long)cfg->address, cfg->segment, @@ -481,14 +485,12 @@ static void __init pci_mmcfg_reject_broken(int early) return; for (i = 0; i < pci_mmcfg_config_num; i++) { - int num_buses, valid = 0; + int valid = 0; u64 addr, size; cfg = &pci_mmcfg_config[i]; - addr = 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 = cfg->res.start; + size = resource_size(&cfg->res); printk(KERN_NOTICE "PCI: MCFG configuration %d: base %lx " "segment %hu buses %u - %u\n", i, (unsigned long)cfg->address, cfg->segment, From 2f2a8b9c90279e75f87aaf322a948bdced27e89f Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Fri, 13 Nov 2009 17:34:34 -0700 Subject: [PATCH 081/109] x86/PCI: MMCONFIG: trivial is_mmconf_reserved() interface simplification Since pci_mmcfg_region contains the struct resource, no need to pass the pci_mmcfg_region *and* the resource start/size. Reviewed-by: Yinghai Lu Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/pci/mmconfig-shared.c | 12 +++++------- 1 file changed, 5 insertions(+), 7 deletions(-) diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index ba3aa3697418..90422b4a7c91 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -438,9 +438,10 @@ static int __init is_acpi_reserved(u64 start, u64 end, unsigned not_used) typedef int (*check_reserved_t)(u64 start, u64 end, unsigned type); static int __init is_mmconf_reserved(check_reserved_t is_reserved, - u64 addr, u64 size, int i, - typeof(pci_mmcfg_config[0]) *cfg, int with_e820) + int i, typeof(pci_mmcfg_config[0]) *cfg, int with_e820) { + u64 addr = cfg->res.start; + u64 size = resource_size(&cfg->res); u64 old_size = size; int valid = 0, num_buses; @@ -486,11 +487,8 @@ static void __init pci_mmcfg_reject_broken(int early) for (i = 0; i < pci_mmcfg_config_num; i++) { int valid = 0; - u64 addr, size; cfg = &pci_mmcfg_config[i]; - addr = cfg->res.start; - size = resource_size(&cfg->res); printk(KERN_NOTICE "PCI: MCFG configuration %d: base %lx " "segment %hu buses %u - %u\n", i, (unsigned long)cfg->address, cfg->segment, @@ -498,7 +496,7 @@ static void __init pci_mmcfg_reject_broken(int early) (unsigned int)cfg->end_bus); if (!early && !acpi_disabled) - valid = is_mmconf_reserved(is_acpi_reserved, addr, size, i, cfg, 0); + valid = is_mmconf_reserved(is_acpi_reserved, i, cfg, 0); if (valid) continue; @@ -511,7 +509,7 @@ static void __init pci_mmcfg_reject_broken(int early) /* Don't try to do this check unless configuration type 1 is available. how about type 2 ?*/ if (raw_pci_ops) - valid = is_mmconf_reserved(e820_all_mapped, addr, size, i, cfg, 1); + valid = is_mmconf_reserved(e820_all_mapped, i, cfg, 1); if (!valid) goto reject; From 3f0f5503926f7447615f083c2d57545a83b6357c Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Fri, 13 Nov 2009 17:34:39 -0700 Subject: [PATCH 082/109] x86/PCI: MMCONFIG: add virtual address to struct pci_mmcfg_region The virtual address is only used for x86_64, but it's so much simpler to manage it as part of the pci_mmcfg_region that I think it's worth wasting a pointer per MMCONFIG region on x86_32. Reviewed-by: Yinghai Lu Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/include/asm/pci_x86.h | 1 + arch/x86/pci/mmconfig_64.c | 45 +++++++++++----------------------- 2 files changed, 15 insertions(+), 31 deletions(-) diff --git a/arch/x86/include/asm/pci_x86.h b/arch/x86/include/asm/pci_x86.h index a6d42c10b017..7aa2ed8f25ab 100644 --- a/arch/x86/include/asm/pci_x86.h +++ b/arch/x86/include/asm/pci_x86.h @@ -124,6 +124,7 @@ extern int __init pcibios_init(void); struct pci_mmcfg_region { struct resource res; u64 address; + char __iomem *virt; u16 segment; u8 start_bus; u8 end_bus; diff --git a/arch/x86/pci/mmconfig_64.c b/arch/x86/pci/mmconfig_64.c index fdf08f97131b..78fa05c6c04d 100644 --- a/arch/x86/pci/mmconfig_64.c +++ b/arch/x86/pci/mmconfig_64.c @@ -12,24 +12,17 @@ #include #include -/* Static virtual mapping of the MMCONFIG aperture */ -struct mmcfg_virt { - struct pci_mmcfg_region *cfg; - char __iomem *virt; -}; -static struct mmcfg_virt *pci_mmcfg_virt; - static char __iomem *get_virt(unsigned int seg, unsigned bus) { + int i; struct pci_mmcfg_region *cfg; - int cfg_num; - for (cfg_num = 0; cfg_num < pci_mmcfg_config_num; cfg_num++) { - cfg = pci_mmcfg_virt[cfg_num].cfg; + for (i = 0; i < pci_mmcfg_config_num; ++i) { + cfg = &pci_mmcfg_config[i]; if (cfg->segment == seg && (cfg->start_bus <= bus) && (cfg->end_bus >= bus)) - return pci_mmcfg_virt[cfg_num].virt; + return cfg->virt; } /* Fall back to type 0 */ @@ -130,20 +123,15 @@ static void __iomem * __init mcfg_ioremap(struct pci_mmcfg_region *cfg) int __init pci_mmcfg_arch_init(void) { int i; - pci_mmcfg_virt = kzalloc(sizeof(*pci_mmcfg_virt) * - pci_mmcfg_config_num, GFP_KERNEL); - if (pci_mmcfg_virt == NULL) { - printk(KERN_ERR "PCI: Can not allocate memory for mmconfig structures\n"); - return 0; - } + struct pci_mmcfg_region *cfg; for (i = 0; i < pci_mmcfg_config_num; ++i) { - pci_mmcfg_virt[i].cfg = &pci_mmcfg_config[i]; - pci_mmcfg_virt[i].virt = mcfg_ioremap(&pci_mmcfg_config[i]); - if (!pci_mmcfg_virt[i].virt) { + cfg = &pci_mmcfg_config[i]; + cfg->virt = mcfg_ioremap(cfg); + if (!cfg->virt) { printk(KERN_ERR "PCI: Cannot map mmconfig aperture for " "segment %d\n", - pci_mmcfg_config[i].segment); + cfg->segment); pci_mmcfg_arch_free(); return 0; } @@ -155,18 +143,13 @@ int __init pci_mmcfg_arch_init(void) void __init pci_mmcfg_arch_free(void) { int i; - - if (pci_mmcfg_virt == NULL) - return; + struct pci_mmcfg_region *cfg; for (i = 0; i < pci_mmcfg_config_num; ++i) { - if (pci_mmcfg_virt[i].virt) { - iounmap(pci_mmcfg_virt[i].virt + PCI_MMCFG_BUS_OFFSET(pci_mmcfg_virt[i].cfg->start_bus)); - pci_mmcfg_virt[i].virt = NULL; - pci_mmcfg_virt[i].cfg = NULL; + cfg = &pci_mmcfg_config[i]; + if (cfg->virt) { + iounmap(cfg->virt + PCI_MMCFG_BUS_OFFSET(cfg->start_bus)); + cfg->virt = NULL; } } - - kfree(pci_mmcfg_virt); - pci_mmcfg_virt = NULL; } From 987c367b4e93be6826394e7c9cc14d28bb5c8810 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Fri, 13 Nov 2009 17:34:44 -0700 Subject: [PATCH 083/109] x86/PCI: MMCONFIG: remove typeof so we can use a list This replaces "typeof(pci_mmcfg_config[0])" with the actual type because I plan to convert pci_mmcfg_config to a list, and then "pci_mmcfg_config[0]" won't mean anything. Reviewed-by: Yinghai Lu Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/pci/mmconfig-shared.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index 90422b4a7c91..6eeeac0d25f4 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -277,8 +277,8 @@ static struct pci_mmcfg_hostbridge_probe pci_mmcfg_probes[] __initdata = { static int __init cmp_mmcfg(const void *x1, const void *x2) { - const typeof(pci_mmcfg_config[0]) *m1 = x1; - const typeof(pci_mmcfg_config[0]) *m2 = x2; + const struct pci_mmcfg_region *m1 = x1; + const struct pci_mmcfg_region *m2 = x2; int start1, start2; start1 = m1->start_bus; @@ -290,7 +290,7 @@ static int __init cmp_mmcfg(const void *x1, const void *x2) static void __init pci_mmcfg_check_end_bus_number(void) { int i; - typeof(pci_mmcfg_config[0]) *cfg, *cfgx; + struct pci_mmcfg_region *cfg, *cfgx; /* sort them at first */ sort(pci_mmcfg_config, pci_mmcfg_config_num, @@ -438,7 +438,7 @@ static int __init is_acpi_reserved(u64 start, u64 end, unsigned not_used) typedef int (*check_reserved_t)(u64 start, u64 end, unsigned type); static int __init is_mmconf_reserved(check_reserved_t is_reserved, - int i, typeof(pci_mmcfg_config[0]) *cfg, int with_e820) + int i, struct pci_mmcfg_region *cfg, int with_e820) { u64 addr = cfg->res.start; u64 size = resource_size(&cfg->res); @@ -479,7 +479,7 @@ static int __init is_mmconf_reserved(check_reserved_t is_reserved, static void __init pci_mmcfg_reject_broken(int early) { - typeof(pci_mmcfg_config[0]) *cfg; + struct pci_mmcfg_region *cfg; int i; if (pci_mmcfg_config_num == 0) From ff097ddd4aeac790fd51d013c79c2f18ec9a7117 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Fri, 13 Nov 2009 17:34:49 -0700 Subject: [PATCH 084/109] x86/PCI: MMCONFIG: manage pci_mmcfg_region as a list, not a table This changes pci_mmcfg_region from a table to a list, to make it easier to add and remove MMCONFIG regions for PCI host bridge hotplug. Reviewed-by: Yinghai Lu Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/include/asm/pci_x86.h | 4 +- arch/x86/pci/mmconfig-shared.c | 106 +++++++++++++-------------------- arch/x86/pci/mmconfig_32.c | 5 +- arch/x86/pci/mmconfig_64.c | 13 +--- 4 files changed, 47 insertions(+), 81 deletions(-) diff --git a/arch/x86/include/asm/pci_x86.h b/arch/x86/include/asm/pci_x86.h index 7aa2ed8f25ab..0b7c316a70c3 100644 --- a/arch/x86/include/asm/pci_x86.h +++ b/arch/x86/include/asm/pci_x86.h @@ -122,6 +122,7 @@ extern int __init pcibios_init(void); #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; @@ -134,8 +135,7 @@ struct pci_mmcfg_region { extern int __init pci_mmcfg_arch_init(void); extern void __init pci_mmcfg_arch_free(void); -extern struct pci_mmcfg_region *pci_mmcfg_config; -extern int pci_mmcfg_config_num; +extern struct list_head pci_mmcfg_list; #define PCI_MMCFG_BUS_OFFSET(bus) ((bus) << 20) diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index 6eeeac0d25f4..2709aa81801d 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -16,7 +16,6 @@ #include #include #include -#include #include #include #include @@ -26,53 +25,58 @@ /* Indicate if the mmcfg resources have been placed into the resource table. */ static int __initdata pci_mmcfg_resources_inserted; +LIST_HEAD(pci_mmcfg_list); + static __init void free_all_mmcfg(void) { - int i; - struct pci_mmcfg_region *cfg; + struct pci_mmcfg_region *cfg, *tmp; pci_mmcfg_arch_free(); - for (i = 0; i < pci_mmcfg_config_num; i++) { - cfg = &pci_mmcfg_config[i]; + list_for_each_entry_safe(cfg, tmp, &pci_mmcfg_list, list) { if (cfg->res.parent) release_resource(&cfg->res); + list_del(&cfg->list); + kfree(cfg); } - pci_mmcfg_config_num = 0; - kfree(pci_mmcfg_config); - pci_mmcfg_config = NULL; +} + +static __init 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(cfg, &pci_mmcfg_list, list) { + if (cfg->segment > new->segment || + (cfg->segment == new->segment && + cfg->start_bus >= new->start_bus)) { + list_add_tail(&new->list, &cfg->list); + return; + } + } + list_add_tail(&new->list, &pci_mmcfg_list); } static __init struct pci_mmcfg_region *pci_mmconfig_add(int segment, int start, int end, u64 addr) { struct pci_mmcfg_region *new; - int new_num = pci_mmcfg_config_num + 1; - int i = pci_mmcfg_config_num; int num_buses; struct resource *res; if (addr == 0) return NULL; - new = kzalloc(sizeof(pci_mmcfg_config[0]) * new_num, GFP_KERNEL); + new = kzalloc(sizeof(*new), GFP_KERNEL); if (!new) return NULL; - if (pci_mmcfg_config) { - memcpy(new, pci_mmcfg_config, - sizeof(pci_mmcfg_config[0]) * new_num); - kfree(pci_mmcfg_config); - } - pci_mmcfg_config = new; - pci_mmcfg_config_num++; - - new = &pci_mmcfg_config[i]; - new->address = addr; new->segment = segment; new->start_bus = start; new->end_bus = end; + list_add_sorted(new); + num_buses = end - start + 1; res = &new->res; res->start = addr + PCI_MMCFG_BUS_OFFSET(start); @@ -82,7 +86,7 @@ static __init struct pci_mmcfg_region *pci_mmconfig_add(int segment, int start, "PCI MMCONFIG %04x [bus %02x-%02x]", segment, start, end); res->name = new->name; - return &pci_mmcfg_config[i]; + return new; } static const char __init *pci_mmcfg_e7520(void) @@ -214,7 +218,7 @@ static const char __init *pci_mmcfg_nvidia_mcp55(void) /* * do check if amd fam10h already took over */ - if (!acpi_disabled || pci_mmcfg_config_num || mcp55_checked) + if (!acpi_disabled || !list_empty(&pci_mmcfg_list) || mcp55_checked) return NULL; mcp55_checked = true; @@ -275,44 +279,26 @@ static struct pci_mmcfg_hostbridge_probe pci_mmcfg_probes[] __initdata = { 0x0369, pci_mmcfg_nvidia_mcp55 }, }; -static int __init cmp_mmcfg(const void *x1, const void *x2) -{ - const struct pci_mmcfg_region *m1 = x1; - const struct pci_mmcfg_region *m2 = x2; - int start1, start2; - - start1 = m1->start_bus; - start2 = m2->start_bus; - - return start1 - start2; -} - static void __init pci_mmcfg_check_end_bus_number(void) { - int i; struct pci_mmcfg_region *cfg, *cfgx; - /* sort them at first */ - sort(pci_mmcfg_config, pci_mmcfg_config_num, - sizeof(pci_mmcfg_config[0]), cmp_mmcfg, NULL); - /* last one*/ - if (pci_mmcfg_config_num > 0) { - i = pci_mmcfg_config_num - 1; - cfg = &pci_mmcfg_config[i]; + cfg = list_entry(pci_mmcfg_list.prev, typeof(*cfg), list); + if (cfg) if (cfg->end_bus < cfg->start_bus) cfg->end_bus = 255; - } + + if (list_is_singular(&pci_mmcfg_list)) + return; /* don't overlap please */ - for (i = 0; i < pci_mmcfg_config_num - 1; i++) { - cfg = &pci_mmcfg_config[i]; - cfgx = &pci_mmcfg_config[i+1]; - + list_for_each_entry(cfg, &pci_mmcfg_list, list) { if (cfg->end_bus < cfg->start_bus) cfg->end_bus = 255; - if (cfg->end_bus >= cfgx->start_bus) + cfgx = list_entry(cfg->list.next, typeof(*cfg), list); + if (cfg != cfgx && cfg->end_bus >= cfgx->start_bus) cfg->end_bus = cfgx->start_bus - 1; } } @@ -350,18 +336,15 @@ static int __init pci_mmcfg_check_hostbridge(void) /* some end_bus_number is crazy, fix it */ pci_mmcfg_check_end_bus_number(); - return pci_mmcfg_config_num != 0; + return !list_empty(&pci_mmcfg_list); } static void __init pci_mmcfg_insert_resources(void) { - int i; struct pci_mmcfg_region *cfg; - for (i = 0; i < pci_mmcfg_config_num; i++) { - cfg = &pci_mmcfg_config[i]; + list_for_each_entry(cfg, &pci_mmcfg_list, list) insert_resource(&iomem_resource, &cfg->res); - } /* Mark that the resources have been inserted. */ pci_mmcfg_resources_inserted = 1; @@ -482,18 +465,15 @@ static void __init pci_mmcfg_reject_broken(int early) struct pci_mmcfg_region *cfg; int i; - if (pci_mmcfg_config_num == 0) - return; - - for (i = 0; i < pci_mmcfg_config_num; i++) { + list_for_each_entry(cfg, &pci_mmcfg_list, list) { int valid = 0; - cfg = &pci_mmcfg_config[i]; printk(KERN_NOTICE "PCI: MCFG configuration %d: base %lx " "segment %hu buses %u - %u\n", i, (unsigned long)cfg->address, cfg->segment, (unsigned int)cfg->start_bus, (unsigned int)cfg->end_bus); + i++; if (!early && !acpi_disabled) valid = is_mmconf_reserved(is_acpi_reserved, i, cfg, 0); @@ -524,10 +504,6 @@ reject: static int __initdata known_bridge; -/* The physical address of the MMCONFIG aperture. Set from ACPI tables. */ -struct pci_mmcfg_region *pci_mmcfg_config; -int pci_mmcfg_config_num; - static int __init acpi_mcfg_check_entry(struct acpi_table_mcfg *mcfg, struct acpi_mcfg_allocation *cfg) { @@ -620,7 +596,7 @@ static void __init __pci_mmcfg_init(int early) pci_mmcfg_reject_broken(early); - if (pci_mmcfg_config_num == 0) + if (list_empty(&pci_mmcfg_list)) return; if (pci_mmcfg_arch_init()) @@ -652,7 +628,7 @@ static int __init pci_mmcfg_late_insert_resources(void) */ if ((pci_mmcfg_resources_inserted == 1) || (pci_probe & PCI_PROBE_MMCONF) == 0 || - (pci_mmcfg_config_num == 0)) + list_empty(&pci_mmcfg_list)) return 1; /* diff --git a/arch/x86/pci/mmconfig_32.c b/arch/x86/pci/mmconfig_32.c index a3cee532c935..c04523e09649 100644 --- a/arch/x86/pci/mmconfig_32.c +++ b/arch/x86/pci/mmconfig_32.c @@ -28,15 +28,12 @@ static int mmcfg_last_accessed_cpu; static u32 get_base_addr(unsigned int seg, int bus, unsigned devfn) { struct pci_mmcfg_region *cfg; - int cfg_num; - for (cfg_num = 0; cfg_num < pci_mmcfg_config_num; cfg_num++) { - cfg = &pci_mmcfg_config[cfg_num]; + list_for_each_entry(cfg, &pci_mmcfg_list, list) if (cfg->segment == seg && (cfg->start_bus <= bus) && (cfg->end_bus >= bus)) return cfg->address; - } /* Fall back to type 0 */ return 0; diff --git a/arch/x86/pci/mmconfig_64.c b/arch/x86/pci/mmconfig_64.c index 78fa05c6c04d..ed1f479b4d0e 100644 --- a/arch/x86/pci/mmconfig_64.c +++ b/arch/x86/pci/mmconfig_64.c @@ -14,16 +14,13 @@ static char __iomem *get_virt(unsigned int seg, unsigned bus) { - int i; struct pci_mmcfg_region *cfg; - for (i = 0; i < pci_mmcfg_config_num; ++i) { - cfg = &pci_mmcfg_config[i]; + list_for_each_entry(cfg, &pci_mmcfg_list, list) if (cfg->segment == seg && (cfg->start_bus <= bus) && (cfg->end_bus >= bus)) return cfg->virt; - } /* Fall back to type 0 */ return NULL; @@ -122,11 +119,9 @@ static void __iomem * __init mcfg_ioremap(struct pci_mmcfg_region *cfg) int __init pci_mmcfg_arch_init(void) { - int i; struct pci_mmcfg_region *cfg; - for (i = 0; i < pci_mmcfg_config_num; ++i) { - cfg = &pci_mmcfg_config[i]; + list_for_each_entry(cfg, &pci_mmcfg_list, list) { cfg->virt = mcfg_ioremap(cfg); if (!cfg->virt) { printk(KERN_ERR "PCI: Cannot map mmconfig aperture for " @@ -142,11 +137,9 @@ int __init pci_mmcfg_arch_init(void) void __init pci_mmcfg_arch_free(void) { - int i; struct pci_mmcfg_region *cfg; - for (i = 0; i < pci_mmcfg_config_num; ++i) { - cfg = &pci_mmcfg_config[i]; + list_for_each_entry(cfg, &pci_mmcfg_list, list) { if (cfg->virt) { iounmap(cfg->virt + PCI_MMCFG_BUS_OFFSET(cfg->start_bus)); cfg->virt = NULL; From ba2afbabfc44d6322e8607c004f37868ff786cf8 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Fri, 13 Nov 2009 17:34:54 -0700 Subject: [PATCH 085/109] x86/PCI: MMCONFIG: add pci_mmconfig_remove() to remove MMCONFIG region This is only used internally now, but eventually will be used in the hot-remove path to remove the MMCONFIG region associated with a host bridge. Reviewed-by: Yinghai Lu Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/pci/mmconfig-shared.c | 16 ++++++++++------ 1 file changed, 10 insertions(+), 6 deletions(-) diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index 2709aa81801d..392f8fe16955 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -27,17 +27,21 @@ static int __initdata pci_mmcfg_resources_inserted; LIST_HEAD(pci_mmcfg_list); +static __init void pci_mmconfig_remove(struct pci_mmcfg_region *cfg) +{ + if (cfg->res.parent) + release_resource(&cfg->res); + list_del(&cfg->list); + kfree(cfg); +} + static __init void 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) { - if (cfg->res.parent) - release_resource(&cfg->res); - list_del(&cfg->list); - kfree(cfg); - } + list_for_each_entry_safe(cfg, tmp, &pci_mmcfg_list, list) + pci_mmconfig_remove(cfg); } static __init void list_add_sorted(struct pci_mmcfg_region *new) From 8c57786ad3d921713c7ad8e44132aa537a1d0fec Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Fri, 13 Nov 2009 17:34:59 -0700 Subject: [PATCH 086/109] x86/PCI: MMCONFIG: clean up printks No functional change; just tidy up printks and make them more consistent with the rest of PCI. Reviewed-by: Yinghai Lu Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/pci/mmconfig-shared.c | 46 ++++++++++++++++------------------ arch/x86/pci/mmconfig_64.c | 12 ++++----- 2 files changed, 26 insertions(+), 32 deletions(-) diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index 392f8fe16955..71d69b88fa33 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -90,6 +90,10 @@ static __init struct pci_mmcfg_region *pci_mmconfig_add(int segment, int start, "PCI MMCONFIG %04x [bus %02x-%02x]", segment, start, end); res->name = new->name; + printk(KERN_INFO PREFIX "MMCONFIG for domain %04x [bus %02x-%02x] at " + "%pR (base %#lx)\n", segment, start, end, &new->res, + (unsigned long) addr); + return new; } @@ -333,7 +337,7 @@ static int __init pci_mmcfg_check_hostbridge(void) name = pci_mmcfg_probes[i].probe(); if (name) - printk(KERN_INFO "PCI: Found %s with MMCONFIG support.\n", + printk(KERN_INFO PREFIX "%s with MMCONFIG support\n", name); } @@ -425,7 +429,7 @@ static int __init is_acpi_reserved(u64 start, u64 end, unsigned not_used) typedef int (*check_reserved_t)(u64 start, u64 end, unsigned type); static int __init is_mmconf_reserved(check_reserved_t is_reserved, - int i, struct pci_mmcfg_region *cfg, int with_e820) + struct pci_mmcfg_region *cfg, int with_e820) { u64 addr = cfg->res.start; u64 size = resource_size(&cfg->res); @@ -439,9 +443,9 @@ static int __init is_mmconf_reserved(check_reserved_t is_reserved, } if (size >= (16UL<<20) || size == old_size) { - printk(KERN_NOTICE - "PCI: MCFG area at %Lx reserved in %s\n", - addr, with_e820?"E820":"ACPI motherboard resources"); + printk(KERN_INFO PREFIX "MMCONFIG at %pR reserved in %s\n", + &cfg->res, + with_e820 ? "E820" : "ACPI motherboard resources"); valid = 1; if (old_size != size) { @@ -453,11 +457,11 @@ static int __init is_mmconf_reserved(check_reserved_t is_reserved, snprintf(cfg->name, PCI_MMCFG_RESOURCE_NAME_LEN, "PCI MMCONFIG %04x [bus %02x-%02x]", cfg->segment, cfg->start_bus, cfg->end_bus); - printk(KERN_NOTICE "PCI: updated MCFG configuration %d: base %lx " - "segment %hu buses %u - %u\n", - i, (unsigned long)cfg->address, cfg->segment, - (unsigned int)cfg->start_bus, - (unsigned int)cfg->end_bus); + printk(KERN_INFO PREFIX + "MMCONFIG for %04x [bus%02x-%02x] " + "at %pR (base %#lx) (size reduced!)\n", + cfg->segment, cfg->start_bus, cfg->end_bus, + &cfg->res, (unsigned long) cfg->address); } } @@ -467,33 +471,25 @@ static int __init is_mmconf_reserved(check_reserved_t is_reserved, static void __init pci_mmcfg_reject_broken(int early) { struct pci_mmcfg_region *cfg; - int i; list_for_each_entry(cfg, &pci_mmcfg_list, list) { int valid = 0; - printk(KERN_NOTICE "PCI: MCFG configuration %d: base %lx " - "segment %hu buses %u - %u\n", - i, (unsigned long)cfg->address, cfg->segment, - (unsigned int)cfg->start_bus, - (unsigned int)cfg->end_bus); - i++; - if (!early && !acpi_disabled) - valid = is_mmconf_reserved(is_acpi_reserved, i, cfg, 0); + valid = is_mmconf_reserved(is_acpi_reserved, cfg, 0); if (valid) continue; if (!early) - printk(KERN_ERR "PCI: BIOS Bug: MCFG area at %Lx is not" - " reserved in ACPI motherboard resources\n", - cfg->address); + printk(KERN_ERR FW_BUG PREFIX + "MMCONFIG at %pR not reserved in " + "ACPI motherboard resources\n", &cfg->res); /* Don't try to do this check unless configuration type 1 is available. how about type 2 ?*/ if (raw_pci_ops) - valid = is_mmconf_reserved(e820_all_mapped, i, cfg, 1); + valid = is_mmconf_reserved(e820_all_mapped, cfg, 1); if (!valid) goto reject; @@ -502,7 +498,7 @@ static void __init pci_mmcfg_reject_broken(int early) return; reject: - printk(KERN_INFO "PCI: Not using MMCONFIG.\n"); + printk(KERN_INFO PREFIX "not using MMCONFIG\n"); free_all_mmcfg(); } @@ -525,7 +521,7 @@ static int __init acpi_mcfg_check_entry(struct acpi_table_mcfg *mcfg, return 0; } - printk(KERN_ERR PREFIX "MCFG region for %04x:%02x-%02x at %#llx " + printk(KERN_ERR PREFIX "MCFG region for %04x [bus %02x-%02x] at %#llx " "is above 4GB, ignored\n", cfg->pci_segment, cfg->start_bus_number, cfg->end_bus_number, cfg->address); return -EINVAL; diff --git a/arch/x86/pci/mmconfig_64.c b/arch/x86/pci/mmconfig_64.c index ed1f479b4d0e..cfa6cdb6d262 100644 --- a/arch/x86/pci/mmconfig_64.c +++ b/arch/x86/pci/mmconfig_64.c @@ -12,6 +12,8 @@ #include #include +#define PREFIX "PCI: " + static char __iomem *get_virt(unsigned int seg, unsigned bus) { struct pci_mmcfg_region *cfg; @@ -109,11 +111,8 @@ static void __iomem * __init mcfg_ioremap(struct pci_mmcfg_region *cfg) num_buses = cfg->end_bus - cfg->start_bus + 1; size = PCI_MMCFG_BUS_OFFSET(num_buses); addr = ioremap_nocache(start, size); - if (addr) { - printk(KERN_INFO "PCI: Using MMCONFIG at %Lx - %Lx\n", - start, start + size - 1); + if (addr) addr -= PCI_MMCFG_BUS_OFFSET(cfg->start_bus); - } return addr; } @@ -124,9 +123,8 @@ int __init pci_mmcfg_arch_init(void) list_for_each_entry(cfg, &pci_mmcfg_list, list) { cfg->virt = mcfg_ioremap(cfg); if (!cfg->virt) { - printk(KERN_ERR "PCI: Cannot map mmconfig aperture for " - "segment %d\n", - cfg->segment); + printk(KERN_ERR PREFIX "can't map MMCONFIG at %pR\n", + &cfg->res); pci_mmcfg_arch_free(); return 0; } From f6e1d8cc38b3776038fb15d3acc82ed8bb552f82 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Fri, 13 Nov 2009 17:35:04 -0700 Subject: [PATCH 087/109] x86/PCI: MMCONFIG: add lookup function This patch factors out the search for an MMCONFIG region, which was previously implemented in both mmconfig_32 and mmconfig_64. No functional change. Reviewed-by: Yinghai Lu Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- arch/x86/include/asm/pci_x86.h | 1 + arch/x86/pci/mmconfig-shared.c | 12 ++++++++++++ arch/x86/pci/mmconfig_32.c | 11 +++-------- arch/x86/pci/mmconfig_64.c | 23 ++++------------------- 4 files changed, 20 insertions(+), 27 deletions(-) diff --git a/arch/x86/include/asm/pci_x86.h b/arch/x86/include/asm/pci_x86.h index 0b7c316a70c3..b4bf9a942ed0 100644 --- a/arch/x86/include/asm/pci_x86.h +++ b/arch/x86/include/asm/pci_x86.h @@ -134,6 +134,7 @@ struct pci_mmcfg_region { extern int __init pci_mmcfg_arch_init(void); extern void __init pci_mmcfg_arch_free(void); +extern struct pci_mmcfg_region *pci_mmconfig_lookup(int segment, int bus); extern struct list_head pci_mmcfg_list; diff --git a/arch/x86/pci/mmconfig-shared.c b/arch/x86/pci/mmconfig-shared.c index 71d69b88fa33..b19d1e54201e 100644 --- a/arch/x86/pci/mmconfig-shared.c +++ b/arch/x86/pci/mmconfig-shared.c @@ -97,6 +97,18 @@ static __init struct pci_mmcfg_region *pci_mmconfig_add(int segment, int start, return new; } +struct pci_mmcfg_region *pci_mmconfig_lookup(int segment, int bus) +{ + struct pci_mmcfg_region *cfg; + + list_for_each_entry(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) { u32 win; diff --git a/arch/x86/pci/mmconfig_32.c b/arch/x86/pci/mmconfig_32.c index c04523e09649..90d5fd476ed4 100644 --- a/arch/x86/pci/mmconfig_32.c +++ b/arch/x86/pci/mmconfig_32.c @@ -27,15 +27,10 @@ static int mmcfg_last_accessed_cpu; */ static u32 get_base_addr(unsigned int seg, int bus, unsigned devfn) { - struct pci_mmcfg_region *cfg; + struct pci_mmcfg_region *cfg = pci_mmconfig_lookup(seg, bus); - list_for_each_entry(cfg, &pci_mmcfg_list, list) - if (cfg->segment == seg && - (cfg->start_bus <= bus) && - (cfg->end_bus >= bus)) - return cfg->address; - - /* Fall back to type 0 */ + if (cfg) + return cfg->address; return 0; } diff --git a/arch/x86/pci/mmconfig_64.c b/arch/x86/pci/mmconfig_64.c index cfa6cdb6d262..e783841bd1d7 100644 --- a/arch/x86/pci/mmconfig_64.c +++ b/arch/x86/pci/mmconfig_64.c @@ -14,28 +14,13 @@ #define PREFIX "PCI: " -static char __iomem *get_virt(unsigned int seg, unsigned bus) -{ - struct pci_mmcfg_region *cfg; - - list_for_each_entry(cfg, &pci_mmcfg_list, list) - if (cfg->segment == seg && - (cfg->start_bus <= bus) && - (cfg->end_bus >= bus)) - return cfg->virt; - - /* Fall back to type 0 */ - return NULL; -} - static char __iomem *pci_dev_base(unsigned int seg, unsigned int bus, unsigned int devfn) { - char __iomem *addr; + struct pci_mmcfg_region *cfg = pci_mmconfig_lookup(seg, bus); - addr = get_virt(seg, bus); - if (!addr) - return NULL; - return addr + (PCI_MMCFG_BUS_OFFSET(bus) | (devfn << 12)); + if (cfg && cfg->virt) + return cfg->virt + (PCI_MMCFG_BUS_OFFSET(bus) | (devfn << 12)); + return NULL; } static int pci_mmcfg_read(unsigned int seg, unsigned int bus, From 5c788a695ab5740413d9f9c0035d0d7aeef1c708 Mon Sep 17 00:00:00 2001 From: Jesse Barnes Date: Fri, 4 Dec 2009 15:18:01 -0800 Subject: [PATCH 088/109] PCI: ibmphp_hpc: don't release hw sem twice if kthread stops If we stop the kthread, we may end up up'ing the sem twice, which seems unintended. Reported-by: Dan Carpenter Signed-off-by: Jesse Barnes --- drivers/pci/hotplug/ibmphp_hpc.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/pci/hotplug/ibmphp_hpc.c b/drivers/pci/hotplug/ibmphp_hpc.c index 83f337c891a9..c7084f0eca5a 100644 --- a/drivers/pci/hotplug/ibmphp_hpc.c +++ b/drivers/pci/hotplug/ibmphp_hpc.c @@ -890,7 +890,7 @@ static int poll_hpc(void *data) msleep(POLL_INTERVAL_SEC * 1000); if (kthread_should_stop()) - break; + goto out_sleep; down (&semOperations); @@ -904,6 +904,7 @@ static int poll_hpc(void *data) /* give up the hardware semaphore */ up (&semOperations); /* sleep for a short time just for good measure */ +out_sleep: msleep(100); } up (&sem_exit); From c6a415761c59adabb53699c84e5cb42868d97c67 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Wed, 25 Nov 2009 16:28:50 -0800 Subject: [PATCH 089/109] PCI: add debug output for DMA mask info This allows us to find out what DMA mask is used for each PCI device at boot time; useful for debugging. After the patch: ehci_hcd 0000:00:02.1: using 31bit consistent DMA mask e1000 0000:0b:01.0: using 64bit DMA mask e1000 0000:0b:01.0: using 64bit consistent DMA mask e1000e 0000:04:00.0: using 64bit DMA mask e1000e 0000:04:00.0: using 64bit consistent DMA mask ixgb 0000:0c:01.0: using 64bit DMA mask ixgb 0000:0c:01.0: using 64bit consistent DMA mask aacraid 0000:86:00.0: using 32bit DMA mask aacraid 0000:86:00.0: using 32bit consistent DMA mask aacraid 0000:86:00.0: using 64bit DMA mask aacraid 0000:86:00.0: using 64bit consistent DMA mask qla2xxx 0000:0c:02.0: using 64bit consistent DMA mask qla2xxx 0000:0c:02.1: using 64bit consistent DMA mask lpfc 0000:06:00.0: using 64bit DMA mask lpfc 0000:06:00.1: using 64bit DMA mask pata_amd 0000:00:06.0: using 32bit DMA mask pata_amd 0000:00:06.0: using 32bit consistent DMA mask mptsas 0000:0c:04.0: using 64bit DMA mask mptsas 0000:0c:04.0: using 64bit consistent DMA mask forcedeth 0000:00:08.0: using 39bit DMA mask forcedeth 0000:00:08.0: using 39bit consistent DMA mask niu 0000:02:00.0: using 44bit DMA mask niu 0000:02:00.0: using 44bit consistent DMA mask sata_nv 0000:00:05.0: using 32bit DMA mask sata_nv 0000:00:05.0: using 32bit consistent DMA mask ib_mthca 0000:03:00.0: using 64bit DMA mask ib_mthca 0000:03:00.0: using 64bit consistent DMA mask Reviewed-by: Grant Grundler Signed-off-by: Yinghai Lu Signed-off-by: Jesse Barnes --- drivers/pci/pci.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index 453d6ba6811e..be91a09c74a5 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -2101,6 +2101,7 @@ pci_set_dma_mask(struct pci_dev *dev, u64 mask) return -EIO; dev->dma_mask = mask; + dev_dbg(&dev->dev, "using %dbit DMA mask\n", fls64(mask)); return 0; } @@ -2112,6 +2113,7 @@ pci_set_consistent_dma_mask(struct pci_dev *dev, u64 mask) return -EIO; dev->dev.coherent_dma_mask = mask; + dev_dbg(&dev->dev, "using %dbit consistent DMA mask\n", fls64(mask)); return 0; } From bb965401fd2afa26629b244e7bb2e48a117dc238 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Tue, 24 Nov 2009 18:21:21 -0800 Subject: [PATCH 090/109] PCI: show dma_mask bits in /sys So we can catch if the driver sets an incorrect dma_mask. Reviewed-by: Grant Grundler Signed-off-by: Yinghai Lu Signed-off-by: Jesse Barnes --- drivers/pci/pci-sysfs.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+) diff --git a/drivers/pci/pci-sysfs.c b/drivers/pci/pci-sysfs.c index 0fa707e2a0f8..c5df94e86678 100644 --- a/drivers/pci/pci-sysfs.c +++ b/drivers/pci/pci-sysfs.c @@ -183,6 +183,21 @@ numa_node_show(struct device *dev, struct device_attribute *attr, char *buf) } #endif +static ssize_t +dma_mask_bits_show(struct device *dev, struct device_attribute *attr, char *buf) +{ + struct pci_dev *pdev = to_pci_dev(dev); + + return sprintf (buf, "%d\n", fls64(pdev->dma_mask)); +} + +static ssize_t +consistent_dma_mask_bits_show(struct device *dev, struct device_attribute *attr, + char *buf) +{ + return sprintf (buf, "%d\n", fls64(dev->coherent_dma_mask)); +} + static ssize_t msi_bus_show(struct device *dev, struct device_attribute *attr, char *buf) { @@ -314,6 +329,8 @@ struct device_attribute pci_dev_attrs[] = { #ifdef CONFIG_NUMA __ATTR_RO(numa_node), #endif + __ATTR_RO(dma_mask_bits), + __ATTR_RO(consistent_dma_mask_bits), __ATTR(enable, 0600, is_enabled_show, is_enabled_store), __ATTR(broken_parity_status,(S_IRUGO|S_IWUSR), broken_parity_status_show,broken_parity_status_store), From 04b55c4732780381410e52db0e9bfb7661f2b4b3 Mon Sep 17 00:00:00 2001 From: Shmulik Ravid Date: Thu, 3 Dec 2009 22:27:51 +0200 Subject: [PATCH 091/109] PCI: read-modify-write the pcie device control register when initiating pcie flr The pcie_flr routine writes the device control register with the FLR bit set clearing all other fields for the FLR duration. Among other fields, the Max_Payload_Size is also cleared which can cause errors if there are transactions lurking in the HW pipeline. The patch replaces the blank write with read-modify-write of the control register keeping the other fields intact. Signed-off-by: Shmulik Ravid Signed-off-by: Jesse Barnes --- drivers/pci/pci.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index be91a09c74a5..6af212c509c5 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -2140,7 +2140,7 @@ static int pcie_flr(struct pci_dev *dev, int probe) int i; int pos; u32 cap; - u16 status; + u16 status, control; pos = pci_pcie_cap(dev); if (!pos) @@ -2167,8 +2167,10 @@ static int pcie_flr(struct pci_dev *dev, int probe) "proceeding with reset anyway\n"); clear: - pci_write_config_word(dev, pos + PCI_EXP_DEVCTL, - PCI_EXP_DEVCTL_BCR_FLR); + pci_read_config_word(dev, pos + PCI_EXP_DEVCTL, &control); + control |= PCI_EXP_DEVCTL_BCR_FLR; + pci_write_config_word(dev, pos + PCI_EXP_DEVCTL, control); + msleep(100); return 0; From 59353ea30e65ab3ae181d6175e3212e1361c3787 Mon Sep 17 00:00:00 2001 From: Alex Williamson Date: Mon, 30 Nov 2009 14:51:44 -0700 Subject: [PATCH 092/109] PCI: Always set prefetchable base/limit upper32 registers Prior to 1f82de10 we always initialized the upper 32bits of the prefetchable memory window, regardless of the address range used. Now we only touch it for a >32bit address, which means the upper32 registers remain whatever the BIOS initialized them too. It's valid for the BIOS to set the upper32 base/limit to 0xffffffff/0x00000000, which makes us program prefetchable ranges like 0xffffffffabc00000 - 0x00000000abc00000 Revert the chunk of 1f82de10 that made this conditional so we always write the upper32 registers and remove now unused pref_mem64 variable. Signed-off-by: Alex Williamson Signed-off-by: Jesse Barnes --- drivers/pci/setup-bus.c | 11 +++-------- 1 file changed, 3 insertions(+), 8 deletions(-) diff --git a/drivers/pci/setup-bus.c b/drivers/pci/setup-bus.c index 502d1704c533..c48cd377b3f5 100644 --- a/drivers/pci/setup-bus.c +++ b/drivers/pci/setup-bus.c @@ -140,7 +140,6 @@ static void pci_setup_bridge(struct pci_bus *bus) struct resource *res; struct pci_bus_region region; u32 l, bu, lu, io_upper16; - int pref_mem64; if (pci_is_enabled(bridge)) return; @@ -194,7 +193,6 @@ static void pci_setup_bridge(struct pci_bus *bus) pci_write_config_dword(bridge, PCI_PREF_LIMIT_UPPER32, 0); /* Set up PREF base/limit. */ - pref_mem64 = 0; bu = lu = 0; res = bus->resource[2]; pcibios_resource_to_bus(bridge, ®ion, res); @@ -202,7 +200,6 @@ static void pci_setup_bridge(struct pci_bus *bus) l = (region.start >> 16) & 0xfff0; l |= region.end & 0xfff00000; if (res->flags & IORESOURCE_MEM_64) { - pref_mem64 = 1; bu = upper_32_bits(region.start); lu = upper_32_bits(region.end); } @@ -214,11 +211,9 @@ static void pci_setup_bridge(struct pci_bus *bus) } pci_write_config_dword(bridge, PCI_PREF_MEMORY_BASE, l); - if (pref_mem64) { - /* Set the upper 32 bits of PREF base & limit. */ - pci_write_config_dword(bridge, PCI_PREF_BASE_UPPER32, bu); - pci_write_config_dword(bridge, PCI_PREF_LIMIT_UPPER32, lu); - } + /* Set the upper 32 bits of PREF base & limit. */ + pci_write_config_dword(bridge, PCI_PREF_BASE_UPPER32, bu); + pci_write_config_dword(bridge, PCI_PREF_LIMIT_UPPER32, lu); pci_write_config_word(bridge, PCI_BRIDGE_CONTROL, bus->bridge_ctl); } From 898294c97500b1cdff6edce52fd34e024eb070ec Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 25 Nov 2009 21:00:53 +0900 Subject: [PATCH 093/109] PCI: portdrv: remove redundant pcie_port_device_probe We don't need pcie_port_device_probe() because we can get pci device/port type using pci_is_pcie() and 'pcie_type' fields in struct pci_dev. Remove pcie_port_device_probe(). Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/pcie/portdrv.h | 1 - drivers/pci/pcie/portdrv_core.c | 22 ---------------------- drivers/pci/pcie/portdrv_pci.c | 14 ++++++++------ 3 files changed, 8 insertions(+), 29 deletions(-) diff --git a/drivers/pci/pcie/portdrv.h b/drivers/pci/pcie/portdrv.h index 17ad53868f9f..cc1f8435ee20 100644 --- a/drivers/pci/pcie/portdrv.h +++ b/drivers/pci/pcie/portdrv.h @@ -35,7 +35,6 @@ #define get_descriptor_id(type, service) (((type - 4) << 4) | service) extern struct bus_type pcie_port_bus_type; -extern int pcie_port_device_probe(struct pci_dev *dev); extern int pcie_port_device_register(struct pci_dev *dev); #ifdef CONFIG_PM extern int pcie_port_device_suspend(struct device *dev); diff --git a/drivers/pci/pcie/portdrv_core.c b/drivers/pci/pcie/portdrv_core.c index ce99c7121372..a0376f80bc5e 100644 --- a/drivers/pci/pcie/portdrv_core.c +++ b/drivers/pci/pcie/portdrv_core.c @@ -296,28 +296,6 @@ static struct pcie_device* alloc_pcie_device(struct pci_dev *parent, return device; } -/** - * pcie_port_device_probe - check if device is a PCI Express port - * @dev: Device to check - */ -int pcie_port_device_probe(struct pci_dev *dev) -{ - int pos, type; - u16 reg; - - pos = pci_pcie_cap(dev); - if (!pos) - return -ENODEV; - - pci_read_config_word(dev, pos + PCIE_CAPABILITIES_REG, ®); - type = (reg >> 4) & PORT_TYPE_MASK; - if ( type == PCIE_RC_PORT || type == PCIE_SW_UPSTREAM_PORT || - type == PCIE_SW_DOWNSTREAM_PORT ) - return 0; - - return -ENODEV; -} - /** * pcie_port_device_register - register PCI Express port * @dev: PCI Express port to register diff --git a/drivers/pci/pcie/portdrv_pci.c b/drivers/pci/pcie/portdrv_pci.c index f635e476d632..ce52ea34fee5 100644 --- a/drivers/pci/pcie/portdrv_pci.c +++ b/drivers/pci/pcie/portdrv_pci.c @@ -67,14 +67,16 @@ static struct dev_pm_ops pcie_portdrv_pm_ops = { * this port device. * */ -static int __devinit pcie_portdrv_probe (struct pci_dev *dev, - const struct pci_device_id *id ) +static int __devinit pcie_portdrv_probe(struct pci_dev *dev, + const struct pci_device_id *id) { - int status; + int status; - status = pcie_port_device_probe(dev); - if (status) - return status; + if (!pci_is_pcie(dev) || + ((dev->pcie_type != PCI_EXP_TYPE_ROOT_PORT) && + (dev->pcie_type != PCI_EXP_TYPE_UPSTREAM) && + (dev->pcie_type != PCI_EXP_TYPE_DOWNSTREAM))) + return -ENODEV; if (!dev->irq && dev->pin) { dev_warn(&dev->dev, "device [%04x:%04x] has invalid IRQ; " From 52a0f24beabe9e89223e367c65a0156dff17265c Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 25 Nov 2009 21:01:28 +0900 Subject: [PATCH 094/109] PCI: portdrv: cleanup pcie_device registration In the current port bus driver implementation, pcie_device allocation, initialization and registration are done in separated functions. Doing those in one function make the code simple and easier to read. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/pcie/portdrv_core.c | 75 +++++++++++---------------------- 1 file changed, 25 insertions(+), 50 deletions(-) diff --git a/drivers/pci/pcie/portdrv_core.c b/drivers/pci/pcie/portdrv_core.c index a0376f80bc5e..7ea37c075d7e 100644 --- a/drivers/pci/pcie/portdrv_core.c +++ b/drivers/pci/pcie/portdrv_core.c @@ -246,54 +246,39 @@ static int get_port_device_capability(struct pci_dev *dev) } /** - * pcie_device_init - initialize PCI Express port service device - * @dev: Port service device to initialize - * @parent: PCI Express port to associate the service device with - * @port_type: Type of the port - * @service_type: Type of service to associate with the service device + * pcie_device_init - allocate and initialize PCI Express port service device + * @pdev: PCI Express port to associate the service device with + * @service: Type of service to associate with the service device * @irq: Interrupt vector to associate with the service device */ -static void pcie_device_init(struct pci_dev *parent, struct pcie_device *dev, - int service_type, int irq) +static int pcie_device_init(struct pci_dev *pdev, int service, int irq) { - struct pcie_port_data *port_data = pci_get_drvdata(parent); + int retval; + struct pcie_device *pcie; struct device *device; - int port_type = port_data->port_type; - dev->port = parent; - dev->irq = irq; - dev->service = service_type; + pcie = kzalloc(sizeof(*pcie), GFP_KERNEL); + if (!pcie) + return -ENOMEM; + pcie->port = pdev; + pcie->irq = irq; + pcie->service = service; /* Initialize generic device interface */ - device = &dev->device; - memset(device, 0, sizeof(struct device)); + device = &pcie->device; device->bus = &pcie_port_bus_type; - device->driver = NULL; - dev_set_drvdata(device, NULL); device->release = release_pcie_device; /* callback to free pcie dev */ dev_set_name(device, "%s:pcie%02x", - pci_name(parent), get_descriptor_id(port_type, service_type)); - device->parent = &parent->dev; -} + pci_name(pdev), + get_descriptor_id(pdev->pcie_type, service)); + device->parent = &pdev->dev; -/** - * alloc_pcie_device - allocate PCI Express port service device structure - * @parent: PCI Express port to associate the service device with - * @port_type: Type of the port - * @service_type: Type of service to associate with the service device - * @irq: Interrupt vector to associate with the service device - */ -static struct pcie_device* alloc_pcie_device(struct pci_dev *parent, - int service_type, int irq) -{ - struct pcie_device *device; - - device = kzalloc(sizeof(struct pcie_device), GFP_KERNEL); - if (!device) - return NULL; - - pcie_device_init(parent, device, service_type, irq); - return device; + retval = device_register(device); + if (retval) + kfree(pcie); + else + get_device(device); + return retval; } /** @@ -346,24 +331,14 @@ int pcie_port_device_register(struct pci_dev *dev) /* Allocate child services if any */ for (i = 0, nr_serv = 0; i < PCIE_PORT_DEVICE_MAXSERVICES; i++) { - struct pcie_device *child; int service = 1 << i; if (!(capabilities & service)) continue; - child = alloc_pcie_device(dev, service, vectors[i]); - if (!child) - continue; - - status = device_register(&child->device); - if (status) { - kfree(child); - continue; - } - - get_device(&child->device); - nr_serv++; + status = pcie_device_init(dev, service, vectors[i]); + if (!status) + nr_serv++; } if (!nr_serv) { pci_disable_device(dev); From 2dd60e96b4d52bccd2dd585e776a3449d7b34b8f Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 25 Nov 2009 21:02:13 +0900 Subject: [PATCH 095/109] PCI: portdrv: remove redundant pcie type calculation PCIe port type is already stored in 'pcie_type' field of struct pci_dev. So we don't need to get it from pci configuration space. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/pcie/portdrv_core.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-) diff --git a/drivers/pci/pcie/portdrv_core.c b/drivers/pci/pcie/portdrv_core.c index 7ea37c075d7e..42b21eec15f0 100644 --- a/drivers/pci/pcie/portdrv_core.c +++ b/drivers/pci/pcie/portdrv_core.c @@ -291,19 +291,15 @@ static int pcie_device_init(struct pci_dev *pdev, int service, int irq) int pcie_port_device_register(struct pci_dev *dev) { struct pcie_port_data *port_data; - int status, capabilities, irq_mode, i, nr_serv, pos; + int status, capabilities, irq_mode, i, nr_serv; int vectors[PCIE_PORT_DEVICE_MAXSERVICES]; - u16 reg16; port_data = kzalloc(sizeof(*port_data), GFP_KERNEL); if (!port_data) return -ENOMEM; pci_set_drvdata(dev, port_data); - /* Get port type */ - pos = pci_pcie_cap(dev); - pci_read_config_word(dev, pos + PCIE_CAPABILITIES_REG, ®16); - port_data->port_type = (reg16 >> 4) & PORT_TYPE_MASK; + port_data->port_type = dev->pcie_type; capabilities = get_port_device_capability(dev); /* Root ports are capable of generating PME too */ From 9e5d0b16dada536dfe2f1e893b6ad0225ff8a2c9 Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 25 Nov 2009 21:02:51 +0900 Subject: [PATCH 096/109] PCI: portdrv: move PME capability check No reason to check PME capability outside get_port_device_capability(). Do it in get_port_device_capability(). Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/pcie/portdrv_core.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/pci/pcie/portdrv_core.c b/drivers/pci/pcie/portdrv_core.c index 42b21eec15f0..079bbc3ed4f8 100644 --- a/drivers/pci/pcie/portdrv_core.c +++ b/drivers/pci/pcie/portdrv_core.c @@ -241,6 +241,9 @@ static int get_port_device_capability(struct pci_dev *dev) /* VC support */ if (pci_find_ext_capability(dev, PCI_EXT_CAP_ID_VC)) services |= PCIE_PORT_SERVICE_VC; + /* Root ports are capable of generating PME too */ + if (dev->pcie_type == PCI_EXP_TYPE_ROOT_PORT) + services |= PCIE_PORT_SERVICE_PME; return services; } @@ -302,9 +305,6 @@ int pcie_port_device_register(struct pci_dev *dev) port_data->port_type = dev->pcie_type; capabilities = get_port_device_capability(dev); - /* Root ports are capable of generating PME too */ - if (port_data->port_type == PCIE_RC_PORT) - capabilities |= PCIE_PORT_SERVICE_PME; irq_mode = assign_interrupt_mode(dev, vectors, capabilities); if (irq_mode == PCIE_PORT_NO_IRQ) { From d013598d9a46befebdfd37195829ce411e4878ea Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 25 Nov 2009 21:03:27 +0900 Subject: [PATCH 097/109] PCI: portdrv: check capabilities first Move capability check capability to the beginning of pcie_port_device_register() prevents redundant execution path. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/pcie/portdrv_core.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/drivers/pci/pcie/portdrv_core.c b/drivers/pci/pcie/portdrv_core.c index 079bbc3ed4f8..52c4e6fd6fd4 100644 --- a/drivers/pci/pcie/portdrv_core.c +++ b/drivers/pci/pcie/portdrv_core.c @@ -297,6 +297,10 @@ int pcie_port_device_register(struct pci_dev *dev) int status, capabilities, irq_mode, i, nr_serv; int vectors[PCIE_PORT_DEVICE_MAXSERVICES]; + capabilities = get_port_device_capability(dev); + if (!capabilities) + return -ENODEV; + port_data = kzalloc(sizeof(*port_data), GFP_KERNEL); if (!port_data) return -ENOMEM; @@ -304,8 +308,6 @@ int pcie_port_device_register(struct pci_dev *dev) port_data->port_type = dev->pcie_type; - capabilities = get_port_device_capability(dev); - irq_mode = assign_interrupt_mode(dev, vectors, capabilities); if (irq_mode == PCIE_PORT_NO_IRQ) { /* From dc5351784eb36f1fec4efa88e01581be72c0b711 Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 25 Nov 2009 21:04:00 +0900 Subject: [PATCH 098/109] PCI: portdrv: cleanup service irqs initialization This patch cleans up the service irqs initialization as follows: - Remove 'irq_mode' field in pcie_port_data and related definitions, which is not needed because we can get the same information from 'is_msix', 'is_msi' and 'pin' fields in struct pci_dev. - Change the name of 'vectors' argument of assign_interrupt_mode() to 'irqs' because it holds irq numbers actually. People might confuse it with CPU vector or MSI/MSI-X vector. - Change function name assign_interrupt_mode() to init_service_irqs() becasuse we no longer have 'irq_mode' data structure, and new name is more straightforward (IMO). Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/pcie/portdrv_core.c | 71 ++++++++++++++------------------- include/linux/pcieport_if.h | 7 ---- 2 files changed, 29 insertions(+), 49 deletions(-) diff --git a/drivers/pci/pcie/portdrv_core.c b/drivers/pci/pcie/portdrv_core.c index 52c4e6fd6fd4..d208dc2d62fd 100644 --- a/drivers/pci/pcie/portdrv_core.c +++ b/drivers/pci/pcie/portdrv_core.c @@ -177,37 +177,32 @@ static int pcie_port_enable_msix(struct pci_dev *dev, int *vectors, int mask) } /** - * assign_interrupt_mode - choose interrupt mode for PCI Express port services - * (INTx, MSI-X, MSI) and set up vectors + * init_service_irqs - initialize irqs for PCI Express port services * @dev: PCI Express port to handle - * @vectors: Array of interrupt vectors to populate + * @irqs: Array of irqs to populate * @mask: Bitmask of port capabilities returned by get_port_device_capability() * * Return value: Interrupt mode associated with the port */ -static int assign_interrupt_mode(struct pci_dev *dev, int *vectors, int mask) +static int init_service_irqs(struct pci_dev *dev, int *irqs, int mask) { - int irq, interrupt_mode = PCIE_PORT_NO_IRQ; - int i; + int i, irq; /* Try to use MSI-X if supported */ - if (!pcie_port_enable_msix(dev, vectors, mask)) - return PCIE_PORT_MSIX_MODE; - + if (!pcie_port_enable_msix(dev, irqs, mask)) + return 0; /* We're not going to use MSI-X, so try MSI and fall back to INTx */ - if (!pci_enable_msi(dev)) - interrupt_mode = PCIE_PORT_MSI_MODE; + irq = -1; + if (!pci_enable_msi(dev) || dev->pin) + irq = dev->irq; - if (interrupt_mode == PCIE_PORT_NO_IRQ && dev->pin) - interrupt_mode = PCIE_PORT_INTx_MODE; - - irq = interrupt_mode != PCIE_PORT_NO_IRQ ? dev->irq : -1; for (i = 0; i < PCIE_PORT_DEVICE_MAXSERVICES; i++) - vectors[i] = irq; + irqs[i] = irq; + irqs[PCIE_PORT_SERVICE_VC_SHIFT] = -1; - vectors[PCIE_PORT_SERVICE_VC_SHIFT] = -1; - - return interrupt_mode; + if (irq < 0) + return -ENODEV; + return 0; } /** @@ -294,8 +289,8 @@ static int pcie_device_init(struct pci_dev *pdev, int service, int irq) int pcie_port_device_register(struct pci_dev *dev) { struct pcie_port_data *port_data; - int status, capabilities, irq_mode, i, nr_serv; - int vectors[PCIE_PORT_DEVICE_MAXSERVICES]; + int status, capabilities, i, nr_serv; + int irqs[PCIE_PORT_DEVICE_MAXSERVICES]; capabilities = get_port_device_capability(dev); if (!capabilities) @@ -304,23 +299,19 @@ int pcie_port_device_register(struct pci_dev *dev) port_data = kzalloc(sizeof(*port_data), GFP_KERNEL); if (!port_data) return -ENOMEM; + port_data->port_type = dev->pcie_type; pci_set_drvdata(dev, port_data); - port_data->port_type = dev->pcie_type; - - irq_mode = assign_interrupt_mode(dev, vectors, capabilities); - if (irq_mode == PCIE_PORT_NO_IRQ) { - /* - * Don't use service devices that require interrupts if there is - * no way to generate them. - */ - if (!(capabilities & PCIE_PORT_SERVICE_VC)) { - status = -ENODEV; + /* + * Initialize service irqs. Don't use service devices that + * require interrupts if there is no way to generate them. + */ + status = init_service_irqs(dev, irqs, capabilities); + if (status) { + capabilities &= PCIE_PORT_SERVICE_VC; + if (!capabilities) goto Error; - } - capabilities = PCIE_PORT_SERVICE_VC; } - port_data->port_irq_mode = irq_mode; status = pci_enable_device(dev); if (status) @@ -334,7 +325,7 @@ int pcie_port_device_register(struct pci_dev *dev) if (!(capabilities & service)) continue; - status = pcie_device_init(dev, service, vectors[i]); + status = pcie_device_init(dev, service, irqs[i]); if (!status) nr_serv++; } @@ -418,17 +409,13 @@ void pcie_port_device_remove(struct pci_dev *dev) struct pcie_port_data *port_data = pci_get_drvdata(dev); device_for_each_child(&dev->dev, NULL, remove_iter); - pci_disable_device(dev); - switch (port_data->port_irq_mode) { - case PCIE_PORT_MSIX_MODE: + if (dev->msix_enabled) pci_disable_msix(dev); - break; - case PCIE_PORT_MSI_MODE: + else if (dev->msi_enabled) pci_disable_msi(dev); - break; - } + pci_disable_device(dev); kfree(port_data); } diff --git a/include/linux/pcieport_if.h b/include/linux/pcieport_if.h index b4c79545330b..ec2e1eb1bacc 100644 --- a/include/linux/pcieport_if.h +++ b/include/linux/pcieport_if.h @@ -25,15 +25,8 @@ #define PCIE_PORT_SERVICE_VC_SHIFT 3 /* Virtual Channel */ #define PCIE_PORT_SERVICE_VC (1 << PCIE_PORT_SERVICE_VC_SHIFT) -/* Root/Upstream/Downstream Port's Interrupt Mode */ -#define PCIE_PORT_NO_IRQ (-1) -#define PCIE_PORT_INTx_MODE 0 -#define PCIE_PORT_MSI_MODE 1 -#define PCIE_PORT_MSIX_MODE 2 - struct pcie_port_data { int port_type; /* Type of the port */ - int port_irq_mode; /* [0:INTx | 1:MSI | 2:MSI-X] */ }; struct pcie_device { From 1ce5e83063bf388a2c9fa1e3d4d3122146ad305d Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 25 Nov 2009 21:04:30 +0900 Subject: [PATCH 099/109] PCI: portdrv: enable device before irq initialization Call pci_enable_device() before initializing service irqs, because legacy interrupt is initialized in pci_enable_device() on some architectures. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/pcie/portdrv_core.c | 21 +++++++++++---------- 1 file changed, 11 insertions(+), 10 deletions(-) diff --git a/drivers/pci/pcie/portdrv_core.c b/drivers/pci/pcie/portdrv_core.c index d208dc2d62fd..a2ac618a95be 100644 --- a/drivers/pci/pcie/portdrv_core.c +++ b/drivers/pci/pcie/portdrv_core.c @@ -302,6 +302,12 @@ int pcie_port_device_register(struct pci_dev *dev) port_data->port_type = dev->pcie_type; pci_set_drvdata(dev, port_data); + /* Enable PCI Express port device */ + status = pci_enable_device(dev); + if (status) + goto error_kfree; + pci_set_master(dev); + /* * Initialize service irqs. Don't use service devices that * require interrupts if there is no way to generate them. @@ -310,14 +316,9 @@ int pcie_port_device_register(struct pci_dev *dev) if (status) { capabilities &= PCIE_PORT_SERVICE_VC; if (!capabilities) - goto Error; + goto error_disable; } - status = pci_enable_device(dev); - if (status) - goto Error; - pci_set_master(dev); - /* Allocate child services if any */ for (i = 0, nr_serv = 0; i < PCIE_PORT_DEVICE_MAXSERVICES; i++) { int service = 1 << i; @@ -330,14 +331,14 @@ int pcie_port_device_register(struct pci_dev *dev) nr_serv++; } if (!nr_serv) { - pci_disable_device(dev); status = -ENODEV; - goto Error; + goto error_disable; } - return 0; - Error: +error_disable: + pci_disable_device(dev); +error_kfree: kfree(port_data); return status; } From fbb5de70bbe13ecbebb04226dd6d52b1258dc247 Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 25 Nov 2009 21:05:01 +0900 Subject: [PATCH 100/109] PCI: portdrv: add missing irq cleanup Add missing service irqs cleanup in the error code path of pcie_port_device_register(). Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/pcie/portdrv_core.c | 19 ++++++++++++------- 1 file changed, 12 insertions(+), 7 deletions(-) diff --git a/drivers/pci/pcie/portdrv_core.c b/drivers/pci/pcie/portdrv_core.c index a2ac618a95be..82a27f93d38b 100644 --- a/drivers/pci/pcie/portdrv_core.c +++ b/drivers/pci/pcie/portdrv_core.c @@ -205,6 +205,14 @@ static int init_service_irqs(struct pci_dev *dev, int *irqs, int mask) return 0; } +static void cleanup_service_irqs(struct pci_dev *dev) +{ + if (dev->msix_enabled) + pci_disable_msix(dev); + else if (dev->msi_enabled) + pci_disable_msi(dev); +} + /** * get_port_device_capability - discover capabilities of a PCI Express port * @dev: PCI Express port to examine @@ -332,10 +340,12 @@ int pcie_port_device_register(struct pci_dev *dev) } if (!nr_serv) { status = -ENODEV; - goto error_disable; + goto error_cleanup_irqs; } return 0; +error_cleanup_irqs: + cleanup_service_irqs(dev); error_disable: pci_disable_device(dev); error_kfree: @@ -410,12 +420,7 @@ void pcie_port_device_remove(struct pci_dev *dev) struct pcie_port_data *port_data = pci_get_drvdata(dev); device_for_each_child(&dev->dev, NULL, remove_iter); - - if (dev->msix_enabled) - pci_disable_msix(dev); - else if (dev->msi_enabled) - pci_disable_msi(dev); - + cleanup_service_irqs(dev); pci_disable_device(dev); kfree(port_data); } From 40717c39b1e6c064f48a263a27e58642221e8661 Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 25 Nov 2009 21:05:35 +0900 Subject: [PATCH 101/109] PCI: portdrv: minor cleanup for pcie_port_device_register Minor cleanups for pcie_port_device_register(). Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/pcie/portdrv_core.c | 21 ++++++++++----------- 1 file changed, 10 insertions(+), 11 deletions(-) diff --git a/drivers/pci/pcie/portdrv_core.c b/drivers/pci/pcie/portdrv_core.c index 82a27f93d38b..758e3d339287 100644 --- a/drivers/pci/pcie/portdrv_core.c +++ b/drivers/pci/pcie/portdrv_core.c @@ -297,13 +297,15 @@ static int pcie_device_init(struct pci_dev *pdev, int service, int irq) int pcie_port_device_register(struct pci_dev *dev) { struct pcie_port_data *port_data; - int status, capabilities, i, nr_serv; + int status, capabilities, i, nr_service; int irqs[PCIE_PORT_DEVICE_MAXSERVICES]; + /* Get and check PCI Express port services */ capabilities = get_port_device_capability(dev); if (!capabilities) return -ENODEV; + /* Allocate driver data for port device */ port_data = kzalloc(sizeof(*port_data), GFP_KERNEL); if (!port_data) return -ENOMEM; @@ -315,7 +317,6 @@ int pcie_port_device_register(struct pci_dev *dev) if (status) goto error_kfree; pci_set_master(dev); - /* * Initialize service irqs. Don't use service devices that * require interrupts if there is no way to generate them. @@ -328,20 +329,18 @@ int pcie_port_device_register(struct pci_dev *dev) } /* Allocate child services if any */ - for (i = 0, nr_serv = 0; i < PCIE_PORT_DEVICE_MAXSERVICES; i++) { + status = -ENODEV; + nr_service = 0; + for (i = 0; i < PCIE_PORT_DEVICE_MAXSERVICES; i++) { int service = 1 << i; - if (!(capabilities & service)) continue; - - status = pcie_device_init(dev, service, irqs[i]); - if (!status) - nr_serv++; + if (!pcie_device_init(dev, service, irqs[i])) + nr_service++; } - if (!nr_serv) { - status = -ENODEV; + if (!nr_service) goto error_cleanup_irqs; - } + return 0; error_cleanup_irqs: From 694f88ef7ada0d99e304f687ba92e268a594358b Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 25 Nov 2009 21:06:15 +0900 Subject: [PATCH 102/109] PCI: portdrv: remove unnecessary struct pcie_port_data Remove 'port_type' field in struct pcie_port_data(), because we can get port type information from struct pci_dev. With this change, this patch also does followings: - Remove struct pcie_port_data because it no longer has any field. - Remove portdrv private definitions about port type (PCIE_RC_PORT, PCIE_SW_UPSTREAM_PORT and PCIE_SW_DOWNSTREAM_PORT), and use generic definitions instead. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/pcie/aer/aerdrv.c | 2 +- drivers/pci/pcie/aer/aerdrv_core.c | 11 +++++------ drivers/pci/pcie/portdrv_bus.c | 7 ++----- drivers/pci/pcie/portdrv_core.c | 15 +-------------- include/linux/pcieport_if.h | 9 +-------- 5 files changed, 10 insertions(+), 34 deletions(-) diff --git a/drivers/pci/pcie/aer/aerdrv.c b/drivers/pci/pcie/aer/aerdrv.c index 6d30e795a10d..97a345927b55 100644 --- a/drivers/pci/pcie/aer/aerdrv.c +++ b/drivers/pci/pcie/aer/aerdrv.c @@ -53,7 +53,7 @@ static struct pci_error_handlers aer_error_handlers = { static struct pcie_port_service_driver aerdriver = { .name = "aer", - .port_type = PCIE_RC_PORT, + .port_type = PCI_EXP_TYPE_ROOT_PORT, .service = PCIE_PORT_SERVICE_AER, .probe = aer_probe, diff --git a/drivers/pci/pcie/aer/aerdrv_core.c b/drivers/pci/pcie/aer/aerdrv_core.c index 5391a9b412e5..2fbbcee033a7 100644 --- a/drivers/pci/pcie/aer/aerdrv_core.c +++ b/drivers/pci/pcie/aer/aerdrv_core.c @@ -123,9 +123,9 @@ static int set_device_error_reporting(struct pci_dev *dev, void *data) { bool enable = *((bool *)data); - if (dev->pcie_type == PCIE_RC_PORT || - dev->pcie_type == PCIE_SW_UPSTREAM_PORT || - dev->pcie_type == PCIE_SW_DOWNSTREAM_PORT) { + if ((dev->pcie_type == PCI_EXP_TYPE_ROOT_PORT) || + (dev->pcie_type == PCI_EXP_TYPE_UPSTREAM) || + (dev->pcie_type == PCI_EXP_TYPE_DOWNSTREAM)) { if (enable) pci_enable_pcie_error_reporting(dev); else @@ -437,10 +437,9 @@ static int find_aer_service_iter(struct device *device, void *data) result = (struct find_aer_service_data *) data; if (device->bus == &pcie_port_bus_type) { - struct pcie_port_data *port_data; + struct pcie_device *pcie = to_pcie_device(device); - port_data = pci_get_drvdata(to_pcie_device(device)->port); - if (port_data->port_type == PCIE_SW_DOWNSTREAM_PORT) + if (pcie->port->pcie_type == PCI_EXP_TYPE_DOWNSTREAM) result->is_downstream = 1; driver = device->driver; diff --git a/drivers/pci/pcie/portdrv_bus.c b/drivers/pci/pcie/portdrv_bus.c index ef3a4eeaebb4..18bf90f748f6 100644 --- a/drivers/pci/pcie/portdrv_bus.c +++ b/drivers/pci/pcie/portdrv_bus.c @@ -26,7 +26,6 @@ EXPORT_SYMBOL_GPL(pcie_port_bus_type); static int pcie_port_bus_match(struct device *dev, struct device_driver *drv) { struct pcie_device *pciedev; - struct pcie_port_data *port_data; struct pcie_port_service_driver *driver; if (drv->bus != &pcie_port_bus_type || dev->bus != &pcie_port_bus_type) @@ -38,10 +37,8 @@ static int pcie_port_bus_match(struct device *dev, struct device_driver *drv) if (driver->service != pciedev->service) return 0; - port_data = pci_get_drvdata(pciedev->port); - - if (driver->port_type != PCIE_ANY_PORT - && driver->port_type != port_data->port_type) + if ((driver->port_type != PCIE_ANY_PORT) && + (driver->port_type != pciedev->port->pcie_type)) return 0; return 1; diff --git a/drivers/pci/pcie/portdrv_core.c b/drivers/pci/pcie/portdrv_core.c index 758e3d339287..9318b96bb255 100644 --- a/drivers/pci/pcie/portdrv_core.c +++ b/drivers/pci/pcie/portdrv_core.c @@ -296,7 +296,6 @@ static int pcie_device_init(struct pci_dev *pdev, int service, int irq) */ int pcie_port_device_register(struct pci_dev *dev) { - struct pcie_port_data *port_data; int status, capabilities, i, nr_service; int irqs[PCIE_PORT_DEVICE_MAXSERVICES]; @@ -305,17 +304,10 @@ int pcie_port_device_register(struct pci_dev *dev) if (!capabilities) return -ENODEV; - /* Allocate driver data for port device */ - port_data = kzalloc(sizeof(*port_data), GFP_KERNEL); - if (!port_data) - return -ENOMEM; - port_data->port_type = dev->pcie_type; - pci_set_drvdata(dev, port_data); - /* Enable PCI Express port device */ status = pci_enable_device(dev); if (status) - goto error_kfree; + return status; pci_set_master(dev); /* * Initialize service irqs. Don't use service devices that @@ -347,8 +339,6 @@ error_cleanup_irqs: cleanup_service_irqs(dev); error_disable: pci_disable_device(dev); -error_kfree: - kfree(port_data); return status; } @@ -416,12 +406,9 @@ static int remove_iter(struct device *dev, void *data) */ void pcie_port_device_remove(struct pci_dev *dev) { - struct pcie_port_data *port_data = pci_get_drvdata(dev); - device_for_each_child(&dev->dev, NULL, remove_iter); cleanup_service_irqs(dev); pci_disable_device(dev); - kfree(port_data); } /** diff --git a/include/linux/pcieport_if.h b/include/linux/pcieport_if.h index ec2e1eb1bacc..6775532b92a9 100644 --- a/include/linux/pcieport_if.h +++ b/include/linux/pcieport_if.h @@ -10,10 +10,7 @@ #define _PCIEPORT_IF_H_ /* Port Type */ -#define PCIE_RC_PORT 4 /* Root port of RC */ -#define PCIE_SW_UPSTREAM_PORT 5 /* Upstream port of Switch */ -#define PCIE_SW_DOWNSTREAM_PORT 6 /* Downstream port of Switch */ -#define PCIE_ANY_PORT 7 +#define PCIE_ANY_PORT (~0) /* Service Type */ #define PCIE_PORT_SERVICE_PME_SHIFT 0 /* Power Management Event */ @@ -25,10 +22,6 @@ #define PCIE_PORT_SERVICE_VC_SHIFT 3 /* Virtual Channel */ #define PCIE_PORT_SERVICE_VC (1 << PCIE_PORT_SERVICE_VC_SHIFT) -struct pcie_port_data { - int port_type; /* Type of the port */ -}; - struct pcie_device { int irq; /* Service IRQ/MSI/MSI-X Vector */ struct pci_dev *port; /* Root/Upstream/Downstream Port */ From f9f45604edcf87ac86a9d68ca54106c5fb743719 Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Wed, 25 Nov 2009 21:06:51 +0900 Subject: [PATCH 103/109] PCI: portdrv: remove redundant definitions Remove unnecessary definitions from portdrv.h and use generic definitions instead. Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/pcie/portdrv.h | 20 +++----------------- drivers/pci/pcie/portdrv_core.c | 13 ++++++------- 2 files changed, 9 insertions(+), 24 deletions(-) diff --git a/drivers/pci/pcie/portdrv.h b/drivers/pci/pcie/portdrv.h index cc1f8435ee20..aaeb9d21cba5 100644 --- a/drivers/pci/pcie/portdrv.h +++ b/drivers/pci/pcie/portdrv.h @@ -11,24 +11,10 @@ #include -#if !defined(PCI_CAP_ID_PME) -#define PCI_CAP_ID_PME 1 -#endif - -#if !defined(PCI_CAP_ID_EXP) -#define PCI_CAP_ID_EXP 0x10 -#endif - -#define PORT_TYPE_MASK 0xf -#define PORT_TO_SLOT_MASK 0x100 -#define SLOT_HP_CAPABLE_MASK 0x40 -#define PCIE_CAPABILITIES_REG 0x2 -#define PCIE_SLOT_CAPABILITIES_REG 0x14 -#define PCIE_PORT_DEVICE_MAXSERVICES 4 -#define PCIE_PORT_MSI_VECTOR_MASK 0x1f +#define PCIE_PORT_DEVICE_MAXSERVICES 4 /* - * According to the PCI Express Base Specification 2.0, the indices of the MSI-X - * table entires used by port services must not exceed 31 + * According to the PCI Express Base Specification 2.0, the indices of + * the MSI-X table entires used by port services must not exceed 31 */ #define PCIE_PORT_MAX_MSIX_ENTRIES 32 diff --git a/drivers/pci/pcie/portdrv_core.c b/drivers/pci/pcie/portdrv_core.c index 9318b96bb255..413262eb95b7 100644 --- a/drivers/pci/pcie/portdrv_core.c +++ b/drivers/pci/pcie/portdrv_core.c @@ -109,8 +109,8 @@ static int pcie_port_enable_msix(struct pci_dev *dev, int *vectors, int mask) * used to generate the interrupt message." */ pos = pci_pcie_cap(dev); - pci_read_config_word(dev, pos + PCIE_CAPABILITIES_REG, ®16); - entry = (reg16 >> 9) & PCIE_PORT_MSI_VECTOR_MASK; + pci_read_config_word(dev, pos + PCI_EXP_FLAGS, ®16); + entry = (reg16 & PCI_EXP_FLAGS_IRQ) >> 9; if (entry >= nr_entries) goto Error; @@ -230,12 +230,11 @@ static int get_port_device_capability(struct pci_dev *dev) u32 reg32; pos = pci_pcie_cap(dev); - pci_read_config_word(dev, pos + PCIE_CAPABILITIES_REG, ®16); + pci_read_config_word(dev, pos + PCI_EXP_FLAGS, ®16); /* Hot-Plug Capable */ - if (reg16 & PORT_TO_SLOT_MASK) { - pci_read_config_dword(dev, - pos + PCIE_SLOT_CAPABILITIES_REG, ®32); - if (reg32 & SLOT_HP_CAPABLE_MASK) + if (reg16 & PCI_EXP_FLAGS_SLOT) { + pci_read_config_dword(dev, pos + PCI_EXP_SLTCAP, ®32); + if (reg32 & PCI_EXP_SLTCAP_HPC) services |= PCIE_PORT_SERVICE_HP; } /* AER capable */ From 575939cf548951dde8df0786899ea5a91bb669b2 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Tue, 24 Nov 2009 18:05:12 -0800 Subject: [PATCH 104/109] x86/PCI: claim SR-IOV BARs in pcibios_allocate_resource This allows us to use the BIOS SR-IOV allocations rather than assigning our own later on. Signed-off-by: Yinghai Lu Signed-off-by: Jesse Barnes --- arch/x86/pci/i386.c | 17 +++++++++++++++-- 1 file changed, 15 insertions(+), 2 deletions(-) diff --git a/arch/x86/pci/i386.c b/arch/x86/pci/i386.c index b73c09f45210..5dc9e8c63fcd 100644 --- a/arch/x86/pci/i386.c +++ b/arch/x86/pci/i386.c @@ -146,16 +146,29 @@ static void __init pcibios_allocate_bus_resources(struct list_head *bus_list) } } +struct pci_check_idx_range { + int start; + int end; +}; + static void __init pcibios_allocate_resources(int pass) { struct pci_dev *dev = NULL; - int idx, disabled; + int idx, disabled, i; u16 command; struct resource *r; + struct pci_check_idx_range idx_range[] = { + { PCI_STD_RESOURCES, PCI_STD_RESOURCE_END }, +#ifdef CONFIG_PCI_IOV + { PCI_IOV_RESOURCES, PCI_IOV_RESOURCE_END }, +#endif + }; + for_each_pci_dev(dev) { pci_read_config_word(dev, PCI_COMMAND, &command); - for (idx = 0; idx < PCI_ROM_RESOURCE; idx++) { + for (i = 0; i < ARRAY_SIZE(idx_range); i++) + for (idx = idx_range[i].start; idx <= idx_range[i].end; idx++) { r = &dev->resource[idx]; if (r->parent) /* Already allocated */ continue; From 6cdfd995a65a52e05b99e3a72a9b979abe73b312 Mon Sep 17 00:00:00 2001 From: Andrew Patterson Date: Thu, 3 Dec 2009 10:28:20 -0700 Subject: [PATCH 105/109] PCI: unconditionally clear AER uncorr status register during cleanup The current implementation of pci_cleanup_aer_uncorrect_error_status only clears either fatal or non-fatal error status bits depending on the state of the I/O channel. This implementation will then often leave some bits set after PCI error recovery completes. The uncleared bit settings will then be falsely reported the next time an AER interrupt is generated for that hierarchy. An easy way to illustrate this issue is to use the aer-inject module to simultaneously inject both an uncorrectable non-fatal and uncorrectable fatal error. One of the errors will not be cleared. This patch resolves this issue by unconditionally clearing all bits in the AER uncorrectable status register. All settings and corrective action strategies are saved and determined before pci_cleanup_aer_uncorrect_error_status is called, so this change should not affect errory handling functionality. Signed-off-by: Andrew Patterson Signed-off-by: Jesse Barnes --- drivers/pci/pcie/aer/aerdrv_core.c | 10 +++------- 1 file changed, 3 insertions(+), 7 deletions(-) diff --git a/drivers/pci/pcie/aer/aerdrv_core.c b/drivers/pci/pcie/aer/aerdrv_core.c index 2fbbcee033a7..ff4720583aa7 100644 --- a/drivers/pci/pcie/aer/aerdrv_core.c +++ b/drivers/pci/pcie/aer/aerdrv_core.c @@ -84,19 +84,15 @@ EXPORT_SYMBOL_GPL(pci_disable_pcie_error_reporting); int pci_cleanup_aer_uncorrect_error_status(struct pci_dev *dev) { int pos; - u32 status, mask; + u32 status; pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ERR); if (!pos) return -EIO; pci_read_config_dword(dev, pos + PCI_ERR_UNCOR_STATUS, &status); - pci_read_config_dword(dev, pos + PCI_ERR_UNCOR_SEVER, &mask); - if (dev->error_state == pci_channel_io_normal) - status &= ~mask; /* Clear corresponding nonfatal bits */ - else - status &= mask; /* Clear corresponding fatal bits */ - pci_write_config_dword(dev, pos + PCI_ERR_UNCOR_STATUS, status); + if (status) + pci_write_config_dword(dev, pos + PCI_ERR_UNCOR_STATUS, status); return 0; } From 638bba08282fb50ba4ebde073ad70551b929e0f2 Mon Sep 17 00:00:00 2001 From: Andrew Patterson Date: Thu, 3 Dec 2009 10:28:25 -0700 Subject: [PATCH 106/109] PCI: remove ifdefed pci_cleanup_aer_correct_error_status The pci_cleanup_aer_correct_error_status() function has been #if 0'd out since 2.6.25. Time to remove the dead code. Signed-off-by: Andrew Patterson Signed-off-by: Jesse Barnes --- drivers/pci/pcie/aer/aerdrv_core.c | 17 ----------------- 1 file changed, 17 deletions(-) diff --git a/drivers/pci/pcie/aer/aerdrv_core.c b/drivers/pci/pcie/aer/aerdrv_core.c index ff4720583aa7..ae672ca80333 100644 --- a/drivers/pci/pcie/aer/aerdrv_core.c +++ b/drivers/pci/pcie/aer/aerdrv_core.c @@ -98,23 +98,6 @@ int pci_cleanup_aer_uncorrect_error_status(struct pci_dev *dev) } EXPORT_SYMBOL_GPL(pci_cleanup_aer_uncorrect_error_status); -#if 0 -int pci_cleanup_aer_correct_error_status(struct pci_dev *dev) -{ - int pos; - u32 status; - - pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ERR); - if (!pos) - return -EIO; - - pci_read_config_dword(dev, pos + PCI_ERR_COR_STATUS, &status); - pci_write_config_dword(dev, pos + PCI_ERR_COR_STATUS, status); - - return 0; -} -#endif /* 0 */ - static int set_device_error_reporting(struct pci_dev *dev, void *data) { bool enable = *((bool *)data); From b26a34aa4792b3db2500b8a98cb7702765c1a92e Mon Sep 17 00:00:00 2001 From: Kenji Kaneshige Date: Fri, 6 Nov 2009 11:25:13 +0900 Subject: [PATCH 107/109] PCI: fix BUG_ON triggered by logical PCIe root port removal This problem happened when removing PCIe root port using PCI logical hotplug operation. The immediate cause of this problem is that the pointer to invalid data structure is passed to pcie_update_aspm_capable() by pcie_aspm_exit_link_state(). When pcie_aspm_exit_link_state() received a pointer to root port link, it unconfigures the root port link and frees its data structure at first. At this point, there are not links to configure under the root port and the data structure for root port link is already freed. So pcie_aspm_exit_link_state() must not call pcie_update_aspm_capable() and pcie_config_aspm_path(). This patch fixes the problem by changing pcie_aspm_exit_link_state() not to call pcie_update_aspm_capable() and pcie_config_aspm_path() if the specified link is root port link. ------------[ cut here ]------------ kernel BUG at drivers/pci/pcie/aspm.c:606! invalid opcode: 0000 [#1] SMP DEBUG_PAGEALLOC last sysfs file: /sys/devices/pci0000:40/0000:40:13.0/remove CPU 1 Modules linked in: shpchp Pid: 9345, comm: sysfsd Not tainted 2.6.32-rc5 #98 ProLiant DL785 G6 RIP: 0010:[] [] pcie_update_aspm_capable+0x15/0xbe RSP: 0018:ffff88082a2f5ca0 EFLAGS: 00010202 RAX: 0000000000000e77 RBX: ffff88182cc3e000 RCX: ffff88082a33d006 RDX: 0000000000000001 RSI: ffffffff811dff4a RDI: ffff88182cc3e000 RBP: ffff88082a2f5cc0 R08: ffff88182cc3e000 R09: 0000000000000000 R10: ffff88182fc00180 R11: ffff88182fc00198 R12: ffff88182cc3e000 R13: 0000000000000000 R14: ffff88182cc3e000 R15: ffff88082a2f5e20 FS: 00007f259a64b6f0(0000) GS:ffff880864600000(0000) knlGS:0000000000000000 CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b CR2: 00007feb53f73da0 CR3: 000000102cc94000 CR4: 00000000000006e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process sysfsd (pid: 9345, threadinfo ffff88082a2f4000, task ffff88082a33cf00) Stack: ffff88182cc3e000 ffff88182cc3e000 0000000000000000 ffff88082a33cf00 <0> ffff88082a2f5cf0 ffffffff811dff52 ffff88082a2f5cf0 ffff88082c525168 <0> ffff88402c9fd2f8 ffff88402c9fd2f8 ffff88082a2f5d20 ffffffff811d7db2 Call Trace: [] pcie_aspm_exit_link_state+0xf5/0x11e [] pci_stop_bus_device+0x76/0x7e [] pci_stop_bus_device+0x2b/0x7e [] pci_remove_bus_device+0x15/0xb9 [] remove_callback+0x29/0x3a [] sysfs_schedule_callback_work+0x15/0x6d [] worker_thread+0x19d/0x298 [] ? worker_thread+0x148/0x298 [] ? sysfs_schedule_callback_work+0x0/0x6d [] ? autoremove_wake_function+0x0/0x38 [] ? worker_thread+0x0/0x298 [] kthread+0x7d/0x85 [] child_rip+0xa/0x20 [] ? restore_args+0x0/0x30 [] ? kthread+0x0/0x85 [] ? child_rip+0x0/0x20 Code: 89 e5 8a 50 48 31 c0 c0 ea 03 83 e2 07 e8 b2 de fe ff c9 48 98 c3 55 48 89 e5 41 56 49 89 fe 41 55 41 54 53 48 83 7f 10 00 74 04 <0f> 0b eb fe 48 8b 05 da 7d 63 00 4c 8d 60 e8 4c 89 e1 eb 24 4c RIP [] pcie_update_aspm_capable+0x15/0xbe RSP ---[ end trace 6ae0f65bdeab8555 ]--- Reported-by: Alex Chiang Tested-by: Alex Chiang Signed-off-by: Kenji Kaneshige Signed-off-by: Jesse Barnes --- drivers/pci/pcie/aspm.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/drivers/pci/pcie/aspm.c b/drivers/pci/pcie/aspm.c index 05b8e25b551b..5a01fc7fbf05 100644 --- a/drivers/pci/pcie/aspm.c +++ b/drivers/pci/pcie/aspm.c @@ -657,8 +657,10 @@ void pcie_aspm_exit_link_state(struct pci_dev *pdev) free_link_state(link); /* Recheck latencies and configure upstream links */ - pcie_update_aspm_capable(root); - pcie_config_aspm_path(parent_link); + if (parent_link) { + pcie_update_aspm_capable(root); + pcie_config_aspm_path(parent_link); + } out: mutex_unlock(&aspm_lock); up_read(&pci_bus_sem); From 5d990b627537e59a3a2f039ff588a4750e9c1a6a Mon Sep 17 00:00:00 2001 From: Chris Wright Date: Fri, 4 Dec 2009 12:15:21 -0800 Subject: [PATCH 108/109] PCI: add pci_request_acs Commit ae21ee65e8bc228416bbcc8a1da01c56a847a60c "PCI: acs p2p upsteram forwarding enabling" doesn't actually enable ACS. Add a function to pci core to allow an IOMMU to request that ACS be enabled. The existing mechanism of using iommu_found() in the pci core to know when ACS should be enabled doesn't actually work due to initialization order; iommu has only been detected not initialized. Have Intel and AMD IOMMUs request ACS, and Xen does as well during early init of dom0. Cc: Allen Kay Cc: David Woodhouse Cc: Jeremy Fitzhardinge Cc: Joerg Roedel Signed-off-by: Chris Wright Signed-off-by: Jesse Barnes --- arch/x86/kernel/amd_iommu_init.c | 2 ++ arch/x86/xen/enlighten.c | 5 +++++ drivers/pci/dmar.c | 5 ++++- drivers/pci/pci.c | 13 +++++++++++++ drivers/pci/probe.c | 5 +---- include/linux/pci.h | 2 ++ 6 files changed, 27 insertions(+), 5 deletions(-) diff --git a/arch/x86/kernel/amd_iommu_init.c b/arch/x86/kernel/amd_iommu_init.c index b4b61d462dcc..e60530a5f524 100644 --- a/arch/x86/kernel/amd_iommu_init.c +++ b/arch/x86/kernel/amd_iommu_init.c @@ -1330,6 +1330,8 @@ void __init amd_iommu_detect(void) gart_iommu_aperture_disabled = 1; gart_iommu_aperture = 0; #endif + /* Make sure ACS will be enabled */ + pci_request_acs(); } } diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index 5bccd706232c..e2511bccbc8d 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -27,6 +27,7 @@ #include #include #include +#include #include #include @@ -1170,7 +1171,11 @@ asmlinkage void __init xen_start_kernel(void) add_preferred_console("xenboot", 0, NULL); add_preferred_console("tty", 0, NULL); add_preferred_console("hvc", 0, NULL); + } else { + /* Make sure ACS will be enabled */ + pci_request_acs(); } + xen_raw_console_write("about to get started...\n"); diff --git a/drivers/pci/dmar.c b/drivers/pci/dmar.c index e01ca4d6b3e6..0e98f6b6f515 100644 --- a/drivers/pci/dmar.c +++ b/drivers/pci/dmar.c @@ -614,8 +614,11 @@ void __init detect_intel_iommu(void) #endif #ifdef CONFIG_DMAR if (ret && !no_iommu && !iommu_detected && !swiotlb && - !dmar_disabled) + !dmar_disabled) { iommu_detected = 1; + /* Make sure ACS will be enabled */ + pci_request_acs(); + } #endif } early_acpi_os_unmap_memory(dmar_tbl, dmar_tbl_size); diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index 6af212c509c5..cd9b375f49d5 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -1550,6 +1550,16 @@ void pci_enable_ari(struct pci_dev *dev) bridge->ari_enabled = 1; } +static int pci_acs_enable; + +/** + * pci_request_acs - ask for ACS to be enabled if supported + */ +void pci_request_acs(void) +{ + pci_acs_enable = 1; +} + /** * pci_enable_acs - enable ACS if hardware support it * @dev: the PCI device @@ -1560,6 +1570,9 @@ void pci_enable_acs(struct pci_dev *dev) u16 cap; u16 ctrl; + if (!pci_acs_enable) + return; + if (!pci_is_pcie(dev)) return; diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index 2fdffc02a308..98ffb2de22e9 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -10,9 +10,7 @@ #include #include #include -#include #include -#include #include "pci.h" #define CARDBUS_LATENCY_TIMER 176 /* secondary latency timer */ @@ -1029,8 +1027,7 @@ static void pci_init_capabilities(struct pci_dev *dev) pci_iov_init(dev); /* Enable ACS P2P upstream forwarding */ - if (iommu_found() || xen_initial_domain()) - pci_enable_acs(dev); + pci_enable_acs(dev); } void pci_device_add(struct pci_dev *dev, struct pci_bus *bus) diff --git a/include/linux/pci.h b/include/linux/pci.h index 2891c3d3e51a..04771b9c3316 100644 --- a/include/linux/pci.h +++ b/include/linux/pci.h @@ -1328,5 +1328,7 @@ static inline bool pci_is_pcie(struct pci_dev *dev) return !!pci_pcie_cap(dev); } +void pci_request_acs(void); + #endif /* __KERNEL__ */ #endif /* LINUX_PCI_H */ From 9e0b5b2c447ad0caa075a5cfef86def62e1782ff Mon Sep 17 00:00:00 2001 From: Kleber Sacilotto de Souza Date: Wed, 25 Nov 2009 00:55:51 -0200 Subject: [PATCH 109/109] PCI: fix coding style issue in pci_save_state() Remove a stray space in pci_save_state(). Signed-off-by: Kleber Sacilotto de Souza Signed-off-by: Jesse Barnes --- drivers/pci/pci.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index cd9b375f49d5..0bc27e059019 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -850,7 +850,7 @@ pci_save_state(struct pci_dev *dev) int i; /* XXX: 100% dword access ok here? */ for (i = 0; i < 16; i++) - pci_read_config_dword(dev, i * 4,&dev->saved_config_space[i]); + pci_read_config_dword(dev, i * 4, &dev->saved_config_space[i]); dev->state_saved = true; if ((i = pci_save_pcie_state(dev)) != 0) return i;